diff --git a/.clang-tidy b/.clang-tidy index 3903911a277..0dacf813c7e 100644 --- a/.clang-tidy +++ b/.clang-tidy @@ -10,7 +10,7 @@ # TODO Let clang-tidy check headers in further directories # --> HeaderFilterRegex: '^.*/(src|base|programs|utils)/.*(h|hpp)$' -HeaderFilterRegex: '^.*/(base)/.*(h|hpp)$' +HeaderFilterRegex: '^.*/(base|programs|utils)/.*(h|hpp)$' Checks: '*, -abseil-*, diff --git a/.github/workflows/master.yml b/.github/workflows/master.yml index f18a83e1b97..2853adff48a 100644 --- a/.github/workflows/master.yml +++ b/.github/workflows/master.yml @@ -305,7 +305,7 @@ jobs: runner_type: style-checker-aarch64 data: ${{ needs.RunConfig.outputs.data }} MarkReleaseReady: - if: ${{ ! (contains(needs.*.result, 'skipped') || contains(needs.*.result, 'failure')) }} + if: ${{ !failure() && !cancelled() }} needs: - BuilderBinDarwin - BuilderBinDarwinAarch64 @@ -313,9 +313,25 @@ jobs: - BuilderDebAarch64 runs-on: [self-hosted, style-checker] steps: + - name: Debug + run: | + echo need with different filters + cat << 'EOF' + ${{ toJSON(needs) }} + ${{ toJSON(needs.*.result) }} + no failures ${{ !contains(needs.*.result, 'failure') }} + no skips ${{ !contains(needs.*.result, 'skipped') }} + no both ${{ !(contains(needs.*.result, 'skipped') || contains(needs.*.result, 'failure')) }} + EOF + - name: Not ready + # fail the job to be able restart it + if: ${{ contains(needs.*.result, 'skipped') || contains(needs.*.result, 'failure') }} + run: exit 1 - name: Check out repository code + if: ${{ ! (contains(needs.*.result, 'skipped') || contains(needs.*.result, 'failure')) }} uses: ClickHouse/checkout@v1 - name: Mark Commit Release Ready + if: ${{ ! (contains(needs.*.result, 'skipped') || contains(needs.*.result, 'failure')) }} run: | cd "$GITHUB_WORKSPACE/tests/ci" python3 mark_release_ready.py diff --git a/.github/workflows/pull_request.yml b/.github/workflows/pull_request.yml index 8783f959ec6..c065219f980 100644 --- a/.github/workflows/pull_request.yml +++ b/.github/workflows/pull_request.yml @@ -13,9 +13,7 @@ on: # yamllint disable-line rule:truthy - opened branches: - master -########################################################################################## -##################################### SMALL CHECKS ####################################### -########################################################################################## + jobs: RunConfig: runs-on: [self-hosted, style-checker-aarch64] @@ -70,13 +68,13 @@ jobs: python3 "$GITHUB_WORKSPACE/tests/ci/ci.py" --infile ${{ runner.temp }}/ci_run_data.json --post --job-name 'Style check' BuildDockers: needs: [RunConfig] - if: ${{ !failure() && !cancelled() }} + if: ${{ !failure() && !cancelled() && toJson(fromJson(needs.RunConfig.outputs.data).docker_data.missing_multi) != '[]' }} uses: ./.github/workflows/reusable_docker.yml with: data: ${{ needs.RunConfig.outputs.data }} StyleCheck: needs: [RunConfig, BuildDockers] - if: ${{ !failure() && !cancelled() }} + if: ${{ !failure() && !cancelled() && contains(fromJson(needs.RunConfig.outputs.data).jobs_data.jobs_to_do, 'Style check')}} uses: ./.github/workflows/reusable_test.yml with: test_name: Style check @@ -89,19 +87,9 @@ jobs: ROBOT_CLICKHOUSE_SSH_KEY<` clickhouse `` com. +Keep an eye out for upcoming meetups and eventsaround the world. Somewhere else you want us to be? Please feel free to reach out to tyler `` clickhouse `` com. You can also peruse [ClickHouse Events](https://clickhouse.com/company/news-events) for a list of all upcoming trainings, meetups, speaking engagements, etc. + +* [ClickHouse Meetup in Bellevue](https://www.meetup.com/clickhouse-seattle-user-group/events/298650371/) - Mar 11 +* [ClickHouse Meetup at Ramp's Offices in NYC](https://www.meetup.com/clickhouse-new-york-user-group/events/298640542/) - Mar 19 +* [ClickHouse Melbourne Meetup](https://www.meetup.com/clickhouse-australia-user-group/events/299479750/) - Mar 20 +* [ClickHouse Meetup in Paris](https://www.meetup.com/clickhouse-france-user-group/events/298997115/) - Mar 21 +* [ClickHouse Meetup in Bengaluru](https://www.meetup.com/clickhouse-bangalore-user-group/events/299479850/) - Mar 23 +* [ClickHouse Meetup in Zurich](https://www.meetup.com/clickhouse-switzerland-meetup-group/events/299628922/) - Apr 16 +* [ClickHouse Meetup in Copenhagen](https://www.meetup.com/clickhouse-denmark-meetup-group/events/299629133/) - Apr 23 +* [ClickHouse Meetup in Dubai](https://www.meetup.com/clickhouse-dubai-meetup-group/events/299629189/) - May 28 + ## Recent Recordings * **Recent Meetup Videos**: [Meetup Playlist](https://www.youtube.com/playlist?list=PL0Z2YDlm0b3iNDUzpY1S3L_iV4nARda_U) Whenever possible recordings of the ClickHouse Community Meetups are edited and presented as individual talks. Current featuring "Modern SQL in 2023", "Fast, Concurrent, and Consistent Asynchronous INSERTS in ClickHouse", and "Full-Text Indices: Design and Experiments" -* **Recording available**: [**v24.1 Release Webinar**](https://www.youtube.com/watch?v=pBF9g0wGAGs) All the features of 24.1, one convenient video! Watch it now! -* **All release webinar recordings**: [YouTube playlist](https://www.youtube.com/playlist?list=PL0Z2YDlm0b3jAlSy1JxyP8zluvXaN3nxU) - +* **Recording available**: [**v24.2 Release Call**](https://www.youtube.com/watch?v=iN2y-TK8f3A) All the features of 24.2, one convenient video! Watch it now! ## Interested in joining ClickHouse and making it your full-time job? diff --git a/base/base/CMakeLists.txt b/base/base/CMakeLists.txt index 548ba01d86a..610877eae73 100644 --- a/base/base/CMakeLists.txt +++ b/base/base/CMakeLists.txt @@ -13,6 +13,7 @@ set (SRCS cgroupsv2.cpp coverage.cpp demangle.cpp + Decimal.cpp getAvailableMemoryAmount.cpp getFQDNOrHostName.cpp getMemoryAmount.cpp diff --git a/base/base/Decimal.cpp b/base/base/Decimal.cpp new file mode 100644 index 00000000000..7e65c0eb8d1 --- /dev/null +++ b/base/base/Decimal.cpp @@ -0,0 +1,87 @@ +#include +#include + +namespace DB +{ + +/// Explicit template instantiations. + +#define FOR_EACH_UNDERLYING_DECIMAL_TYPE(M) \ + M(Int32) \ + M(Int64) \ + M(Int128) \ + M(Int256) + +#define FOR_EACH_UNDERLYING_DECIMAL_TYPE_PASS(M, X) \ + M(Int32, X) \ + M(Int64, X) \ + M(Int128, X) \ + M(Int256, X) + +template const Decimal & Decimal::operator += (const T & x) { value += x; return *this; } +template const Decimal & Decimal::operator -= (const T & x) { value -= x; return *this; } +template const Decimal & Decimal::operator *= (const T & x) { value *= x; return *this; } +template const Decimal & Decimal::operator /= (const T & x) { value /= x; return *this; } +template const Decimal & Decimal::operator %= (const T & x) { value %= x; return *this; } + +template void NO_SANITIZE_UNDEFINED Decimal::addOverflow(const T & x) { value += x; } + +/// Maybe this explicit instantiation affects performance since operators cannot be inlined. + +template template const Decimal & Decimal::operator += (const Decimal & x) { value += static_cast(x.value); return *this; } +template template const Decimal & Decimal::operator -= (const Decimal & x) { value -= static_cast(x.value); return *this; } +template template const Decimal & Decimal::operator *= (const Decimal & x) { value *= static_cast(x.value); return *this; } +template template const Decimal & Decimal::operator /= (const Decimal & x) { value /= static_cast(x.value); return *this; } +template template const Decimal & Decimal::operator %= (const Decimal & x) { value %= static_cast(x.value); return *this; } + +#define DISPATCH(TYPE_T, TYPE_U) \ + template const Decimal & Decimal::operator += (const Decimal & x); \ + template const Decimal & Decimal::operator -= (const Decimal & x); \ + template const Decimal & Decimal::operator *= (const Decimal & x); \ + template const Decimal & Decimal::operator /= (const Decimal & x); \ + template const Decimal & Decimal::operator %= (const Decimal & x); +#define INVOKE(X) FOR_EACH_UNDERLYING_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_UNDERLYING_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + +#define DISPATCH(TYPE) template struct Decimal; +FOR_EACH_UNDERLYING_DECIMAL_TYPE(DISPATCH) +#undef DISPATCH + +template bool operator< (const Decimal & x, const Decimal & y) { return x.value < y.value; } +template bool operator> (const Decimal & x, const Decimal & y) { return x.value > y.value; } +template bool operator<= (const Decimal & x, const Decimal & y) { return x.value <= y.value; } +template bool operator>= (const Decimal & x, const Decimal & y) { return x.value >= y.value; } +template bool operator== (const Decimal & x, const Decimal & y) { return x.value == y.value; } +template bool operator!= (const Decimal & x, const Decimal & y) { return x.value != y.value; } + +#define DISPATCH(TYPE) \ +template bool operator< (const Decimal & x, const Decimal & y); \ +template bool operator> (const Decimal & x, const Decimal & y); \ +template bool operator<= (const Decimal & x, const Decimal & y); \ +template bool operator>= (const Decimal & x, const Decimal & y); \ +template bool operator== (const Decimal & x, const Decimal & y); \ +template bool operator!= (const Decimal & x, const Decimal & y); +FOR_EACH_UNDERLYING_DECIMAL_TYPE(DISPATCH) +#undef DISPATCH + + +template Decimal operator+ (const Decimal & x, const Decimal & y) { return x.value + y.value; } +template Decimal operator- (const Decimal & x, const Decimal & y) { return x.value - y.value; } +template Decimal operator* (const Decimal & x, const Decimal & y) { return x.value * y.value; } +template Decimal operator/ (const Decimal & x, const Decimal & y) { return x.value / y.value; } +template Decimal operator- (const Decimal & x) { return -x.value; } + +#define DISPATCH(TYPE) \ +template Decimal operator+ (const Decimal & x, const Decimal & y); \ +template Decimal operator- (const Decimal & x, const Decimal & y); \ +template Decimal operator* (const Decimal & x, const Decimal & y); \ +template Decimal operator/ (const Decimal & x, const Decimal & y); \ +template Decimal operator- (const Decimal & x); +FOR_EACH_UNDERLYING_DECIMAL_TYPE(DISPATCH) +#undef DISPATCH + +#undef FOR_EACH_UNDERLYING_DECIMAL_TYPE_PASS +#undef FOR_EACH_UNDERLYING_DECIMAL_TYPE +} diff --git a/base/base/Decimal.h b/base/base/Decimal.h index 66ff623217c..42f9e67c49d 100644 --- a/base/base/Decimal.h +++ b/base/base/Decimal.h @@ -2,6 +2,7 @@ #include #include +#include #include @@ -10,6 +11,18 @@ namespace DB template struct Decimal; class DateTime64; +#define FOR_EACH_UNDERLYING_DECIMAL_TYPE(M) \ + M(Int32) \ + M(Int64) \ + M(Int128) \ + M(Int256) + +#define FOR_EACH_UNDERLYING_DECIMAL_TYPE_PASS(M, X) \ + M(Int32, X) \ + M(Int64, X) \ + M(Int128, X) \ + M(Int256, X) + using Decimal32 = Decimal; using Decimal64 = Decimal; using Decimal128 = Decimal; @@ -50,36 +63,73 @@ struct Decimal return static_cast(value); } - const Decimal & operator += (const T & x) { value += x; return *this; } - const Decimal & operator -= (const T & x) { value -= x; return *this; } - const Decimal & operator *= (const T & x) { value *= x; return *this; } - const Decimal & operator /= (const T & x) { value /= x; return *this; } - const Decimal & operator %= (const T & x) { value %= x; return *this; } + const Decimal & operator += (const T & x); + const Decimal & operator -= (const T & x); + const Decimal & operator *= (const T & x); + const Decimal & operator /= (const T & x); + const Decimal & operator %= (const T & x); - template const Decimal & operator += (const Decimal & x) { value += x.value; return *this; } - template const Decimal & operator -= (const Decimal & x) { value -= x.value; return *this; } - template const Decimal & operator *= (const Decimal & x) { value *= x.value; return *this; } - template const Decimal & operator /= (const Decimal & x) { value /= x.value; return *this; } - template const Decimal & operator %= (const Decimal & x) { value %= x.value; return *this; } + template const Decimal & operator += (const Decimal & x); + template const Decimal & operator -= (const Decimal & x); + template const Decimal & operator *= (const Decimal & x); + template const Decimal & operator /= (const Decimal & x); + template const Decimal & operator %= (const Decimal & x); /// This is to avoid UB for sumWithOverflow() - void NO_SANITIZE_UNDEFINED addOverflow(const T & x) { value += x; } + void NO_SANITIZE_UNDEFINED addOverflow(const T & x); T value; }; -template inline bool operator< (const Decimal & x, const Decimal & y) { return x.value < y.value; } -template inline bool operator> (const Decimal & x, const Decimal & y) { return x.value > y.value; } -template inline bool operator<= (const Decimal & x, const Decimal & y) { return x.value <= y.value; } -template inline bool operator>= (const Decimal & x, const Decimal & y) { return x.value >= y.value; } -template inline bool operator== (const Decimal & x, const Decimal & y) { return x.value == y.value; } -template inline bool operator!= (const Decimal & x, const Decimal & y) { return x.value != y.value; } +#define DISPATCH(TYPE) extern template struct Decimal; +FOR_EACH_UNDERLYING_DECIMAL_TYPE(DISPATCH) +#undef DISPATCH -template inline Decimal operator+ (const Decimal & x, const Decimal & y) { return x.value + y.value; } -template inline Decimal operator- (const Decimal & x, const Decimal & y) { return x.value - y.value; } -template inline Decimal operator* (const Decimal & x, const Decimal & y) { return x.value * y.value; } -template inline Decimal operator/ (const Decimal & x, const Decimal & y) { return x.value / y.value; } -template inline Decimal operator- (const Decimal & x) { return -x.value; } +#define DISPATCH(TYPE_T, TYPE_U) \ + extern template const Decimal & Decimal::operator += (const Decimal & x); \ + extern template const Decimal & Decimal::operator -= (const Decimal & x); \ + extern template const Decimal & Decimal::operator *= (const Decimal & x); \ + extern template const Decimal & Decimal::operator /= (const Decimal & x); \ + extern template const Decimal & Decimal::operator %= (const Decimal & x); +#define INVOKE(X) FOR_EACH_UNDERLYING_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_UNDERLYING_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + +template bool operator< (const Decimal & x, const Decimal & y); +template bool operator> (const Decimal & x, const Decimal & y); +template bool operator<= (const Decimal & x, const Decimal & y); +template bool operator>= (const Decimal & x, const Decimal & y); +template bool operator== (const Decimal & x, const Decimal & y); +template bool operator!= (const Decimal & x, const Decimal & y); + +#define DISPATCH(TYPE) \ +extern template bool operator< (const Decimal & x, const Decimal & y); \ +extern template bool operator> (const Decimal & x, const Decimal & y); \ +extern template bool operator<= (const Decimal & x, const Decimal & y); \ +extern template bool operator>= (const Decimal & x, const Decimal & y); \ +extern template bool operator== (const Decimal & x, const Decimal & y); \ +extern template bool operator!= (const Decimal & x, const Decimal & y); +FOR_EACH_UNDERLYING_DECIMAL_TYPE(DISPATCH) +#undef DISPATCH + +template Decimal operator+ (const Decimal & x, const Decimal & y); +template Decimal operator- (const Decimal & x, const Decimal & y); +template Decimal operator* (const Decimal & x, const Decimal & y); +template Decimal operator/ (const Decimal & x, const Decimal & y); +template Decimal operator- (const Decimal & x); + +#define DISPATCH(TYPE) \ +extern template Decimal operator+ (const Decimal & x, const Decimal & y); \ +extern template Decimal operator- (const Decimal & x, const Decimal & y); \ +extern template Decimal operator* (const Decimal & x, const Decimal & y); \ +extern template Decimal operator/ (const Decimal & x, const Decimal & y); \ +extern template Decimal operator- (const Decimal & x); +FOR_EACH_UNDERLYING_DECIMAL_TYPE(DISPATCH) +#undef DISPATCH + +#undef FOR_EACH_UNDERLYING_DECIMAL_TYPE_PASS +#undef FOR_EACH_UNDERLYING_DECIMAL_TYPE /// Distinguishable type to allow function resolution/deduction based on value type, /// but also relatively easy to convert to/from Decimal64. diff --git a/base/base/extended_types.h b/base/base/extended_types.h index b58df45a97e..796167ab45d 100644 --- a/base/base/extended_types.h +++ b/base/base/extended_types.h @@ -64,6 +64,44 @@ template <> struct is_arithmetic { static constexpr bool value = true; template inline constexpr bool is_arithmetic_v = is_arithmetic::value; +#define FOR_EACH_ARITHMETIC_TYPE(M) \ + M(DataTypeDate) \ + M(DataTypeDate32) \ + M(DataTypeDateTime) \ + M(DataTypeInt8) \ + M(DataTypeUInt8) \ + M(DataTypeInt16) \ + M(DataTypeUInt16) \ + M(DataTypeInt32) \ + M(DataTypeUInt32) \ + M(DataTypeInt64) \ + M(DataTypeUInt64) \ + M(DataTypeInt128) \ + M(DataTypeUInt128) \ + M(DataTypeInt256) \ + M(DataTypeUInt256) \ + M(DataTypeFloat32) \ + M(DataTypeFloat64) + +#define FOR_EACH_ARITHMETIC_TYPE_PASS(M, X) \ + M(DataTypeDate, X) \ + M(DataTypeDate32, X) \ + M(DataTypeDateTime, X) \ + M(DataTypeInt8, X) \ + M(DataTypeUInt8, X) \ + M(DataTypeInt16, X) \ + M(DataTypeUInt16, X) \ + M(DataTypeInt32, X) \ + M(DataTypeUInt32, X) \ + M(DataTypeInt64, X) \ + M(DataTypeUInt64, X) \ + M(DataTypeInt128, X) \ + M(DataTypeUInt128, X) \ + M(DataTypeInt256, X) \ + M(DataTypeUInt256, X) \ + M(DataTypeFloat32, X) \ + M(DataTypeFloat64, X) + template struct make_unsigned // NOLINT(readability-identifier-naming) { diff --git a/base/poco/Net/include/Poco/Net/HTTPChunkedStream.h b/base/poco/Net/include/Poco/Net/HTTPChunkedStream.h index 5f4729c9278..a6576aa561d 100644 --- a/base/poco/Net/include/Poco/Net/HTTPChunkedStream.h +++ b/base/poco/Net/include/Poco/Net/HTTPChunkedStream.h @@ -45,6 +45,8 @@ namespace Net ~HTTPChunkedStreamBuf(); void close(); + bool isComplete() const { return _chunk == std::char_traits::eof(); } + protected: int readFromDevice(char * buffer, std::streamsize length); int writeToDevice(const char * buffer, std::streamsize length); @@ -68,6 +70,8 @@ namespace Net ~HTTPChunkedIOS(); HTTPChunkedStreamBuf * rdbuf(); + bool isComplete() const { return _buf.isComplete(); } + protected: HTTPChunkedStreamBuf _buf; }; diff --git a/base/poco/Net/include/Poco/Net/HTTPClientSession.h b/base/poco/Net/include/Poco/Net/HTTPClientSession.h index 7c0caa1c18b..1cef988566c 100644 --- a/base/poco/Net/include/Poco/Net/HTTPClientSession.h +++ b/base/poco/Net/include/Poco/Net/HTTPClientSession.h @@ -210,7 +210,7 @@ namespace Net void setKeepAliveTimeout(const Poco::Timespan & timeout); /// Sets the connection timeout for HTTP connections. - const Poco::Timespan & getKeepAliveTimeout() const; + Poco::Timespan getKeepAliveTimeout() const; /// Returns the connection timeout for HTTP connections. virtual std::ostream & sendRequest(HTTPRequest & request); @@ -275,7 +275,7 @@ namespace Net /// This method should only be called if the request contains /// a "Expect: 100-continue" header. - void flushRequest(); + virtual void flushRequest(); /// Flushes the request stream. /// /// Normally this method does not need to be called. @@ -283,7 +283,7 @@ namespace Net /// fully sent if receiveResponse() is not called, e.g., /// because the underlying socket will be detached. - void reset(); + virtual void reset(); /// Resets the session and closes the socket. /// /// The next request will initiate a new connection, @@ -303,6 +303,9 @@ namespace Net /// Returns true if the proxy should be bypassed /// for the current host. + const Poco::Timestamp & getLastRequest() const; + /// Returns time when connection has been used last time + protected: enum { @@ -338,6 +341,10 @@ namespace Net /// Calls proxyConnect() and attaches the resulting StreamSocket /// to the HTTPClientSession. + void setLastRequest(Poco::Timestamp time); + + void assign(HTTPClientSession & session); + HTTPSessionFactory _proxySessionFactory; /// Factory to create HTTPClientSession to proxy. private: @@ -433,11 +440,20 @@ namespace Net } - inline const Poco::Timespan & HTTPClientSession::getKeepAliveTimeout() const + inline Poco::Timespan HTTPClientSession::getKeepAliveTimeout() const { return _keepAliveTimeout; } + inline const Poco::Timestamp & HTTPClientSession::getLastRequest() const + { + return _lastRequest; + } + + inline void HTTPClientSession::setLastRequest(Poco::Timestamp time) + { + _lastRequest = time; + } } } // namespace Poco::Net diff --git a/base/poco/Net/include/Poco/Net/HTTPFixedLengthStream.h b/base/poco/Net/include/Poco/Net/HTTPFixedLengthStream.h index 2f4df102605..17fa47cfa9b 100644 --- a/base/poco/Net/include/Poco/Net/HTTPFixedLengthStream.h +++ b/base/poco/Net/include/Poco/Net/HTTPFixedLengthStream.h @@ -48,6 +48,8 @@ namespace Net HTTPFixedLengthStreamBuf(HTTPSession & session, ContentLength length, openmode mode); ~HTTPFixedLengthStreamBuf(); + bool isComplete() const; + protected: int readFromDevice(char * buffer, std::streamsize length); int writeToDevice(const char * buffer, std::streamsize length); @@ -67,6 +69,8 @@ namespace Net ~HTTPFixedLengthIOS(); HTTPFixedLengthStreamBuf * rdbuf(); + bool isComplete() const { return _buf.isComplete(); } + protected: HTTPFixedLengthStreamBuf _buf; }; diff --git a/base/poco/Net/include/Poco/Net/HTTPSession.h b/base/poco/Net/include/Poco/Net/HTTPSession.h index 934b34be5d5..cac14f479db 100644 --- a/base/poco/Net/include/Poco/Net/HTTPSession.h +++ b/base/poco/Net/include/Poco/Net/HTTPSession.h @@ -64,6 +64,15 @@ namespace Net Poco::Timespan getTimeout() const; /// Returns the timeout for the HTTP session. + Poco::Timespan getConnectionTimeout() const; + /// Returns connection timeout for the HTTP session. + + Poco::Timespan getSendTimeout() const; + /// Returns send timeout for the HTTP session. + + Poco::Timespan getReceiveTimeout() const; + /// Returns receive timeout for the HTTP session. + bool connected() const; /// Returns true if the underlying socket is connected. @@ -217,12 +226,25 @@ namespace Net return _keepAlive; } - inline Poco::Timespan HTTPSession::getTimeout() const { return _receiveTimeout; } + inline Poco::Timespan HTTPSession::getConnectionTimeout() const + { + return _connectionTimeout; + } + + inline Poco::Timespan HTTPSession::getSendTimeout() const + { + return _sendTimeout; + } + + inline Poco::Timespan HTTPSession::getReceiveTimeout() const + { + return _receiveTimeout; + } inline StreamSocket & HTTPSession::socket() { diff --git a/base/poco/Net/include/Poco/Net/HTTPStream.h b/base/poco/Net/include/Poco/Net/HTTPStream.h index 48502347b2c..a00a861880f 100644 --- a/base/poco/Net/include/Poco/Net/HTTPStream.h +++ b/base/poco/Net/include/Poco/Net/HTTPStream.h @@ -63,6 +63,8 @@ namespace Net ~HTTPIOS(); HTTPStreamBuf * rdbuf(); + bool isComplete() const { return false; } + protected: HTTPStreamBuf _buf; }; diff --git a/base/poco/Net/src/HTTPChunkedStream.cpp b/base/poco/Net/src/HTTPChunkedStream.cpp index 376e3f55492..16ed1e71c31 100644 --- a/base/poco/Net/src/HTTPChunkedStream.cpp +++ b/base/poco/Net/src/HTTPChunkedStream.cpp @@ -49,10 +49,12 @@ HTTPChunkedStreamBuf::~HTTPChunkedStreamBuf() void HTTPChunkedStreamBuf::close() { - if (_mode & std::ios::out) + if (_mode & std::ios::out && _chunk != std::char_traits::eof()) { sync(); _session.write("0\r\n\r\n", 5); + + _chunk = std::char_traits::eof(); } } diff --git a/base/poco/Net/src/HTTPClientSession.cpp b/base/poco/Net/src/HTTPClientSession.cpp index 2282cca682b..33a3dcc4901 100644 --- a/base/poco/Net/src/HTTPClientSession.cpp +++ b/base/poco/Net/src/HTTPClientSession.cpp @@ -227,7 +227,7 @@ void HTTPClientSession::setKeepAliveTimeout(const Poco::Timespan& timeout) std::ostream& HTTPClientSession::sendRequest(HTTPRequest& request) { _pRequestStream = 0; - _pResponseStream = 0; + _pResponseStream = 0; clearException(); _responseReceived = false; @@ -501,5 +501,26 @@ bool HTTPClientSession::bypassProxy() const else return false; } +void HTTPClientSession::assign(Poco::Net::HTTPClientSession & session) +{ + poco_assert (this != &session); + + if (session.buffered()) + throw Poco::LogicException("assign a session with not empty buffered data"); + + if (buffered()) + throw Poco::LogicException("assign to a session with not empty buffered data"); + + attachSocket(session.detachSocket()); + setLastRequest(session.getLastRequest()); + setResolvedHost(session.getResolvedHost()); + setKeepAlive(session.getKeepAlive()); + + setTimeout(session.getConnectionTimeout(), session.getSendTimeout(), session.getReceiveTimeout()); + setKeepAliveTimeout(session.getKeepAliveTimeout()); + setProxyConfig(session.getProxyConfig()); + + session.reset(); +} } } // namespace Poco::Net diff --git a/base/poco/Net/src/HTTPFixedLengthStream.cpp b/base/poco/Net/src/HTTPFixedLengthStream.cpp index fd77ff71cd9..837e5723c57 100644 --- a/base/poco/Net/src/HTTPFixedLengthStream.cpp +++ b/base/poco/Net/src/HTTPFixedLengthStream.cpp @@ -43,6 +43,12 @@ HTTPFixedLengthStreamBuf::~HTTPFixedLengthStreamBuf() } +bool HTTPFixedLengthStreamBuf::isComplete() const +{ + return _count == _length; +} + + int HTTPFixedLengthStreamBuf::readFromDevice(char* buffer, std::streamsize length) { int n = 0; diff --git a/contrib/aws b/contrib/aws index 5f0542b3ad7..32870e234ca 160000 --- a/contrib/aws +++ b/contrib/aws @@ -1 +1 @@ -Subproject commit 5f0542b3ad7eef25b0540d37d778207e0345ea8f +Subproject commit 32870e234cac03e0ac46370c26858b0ffdf14200 diff --git a/docker/images.json b/docker/images.json index 2bf1efe005f..7439517379b 100644 --- a/docker/images.json +++ b/docker/images.json @@ -1,8 +1,12 @@ { - "docker/packager/binary": { + "docker/packager/binary-builder": { "name": "clickhouse/binary-builder", "dependent": [] }, + "docker/packager/cctools": { + "name": "clickhouse/cctools", + "dependent": [] + }, "docker/test/compatibility/centos": { "name": "clickhouse/test-old-centos", "dependent": [] @@ -30,7 +34,6 @@ "docker/test/util": { "name": "clickhouse/test-util", "dependent": [ - "docker/packager/binary", "docker/test/base", "docker/test/fasttest" ] @@ -67,7 +70,9 @@ }, "docker/test/fasttest": { "name": "clickhouse/fasttest", - "dependent": [] + "dependent": [ + "docker/packager/binary-builder" + ] }, "docker/test/style": { "name": "clickhouse/style-test", diff --git a/docker/packager/binary/Dockerfile b/docker/packager/binary-builder/Dockerfile similarity index 64% rename from docker/packager/binary/Dockerfile rename to docker/packager/binary-builder/Dockerfile index e20cbe9781c..96c90403187 100644 --- a/docker/packager/binary/Dockerfile +++ b/docker/packager/binary-builder/Dockerfile @@ -1,43 +1,6 @@ # docker build -t clickhouse/binary-builder . ARG FROM_TAG=latest -FROM clickhouse/test-util:latest AS cctools -# The cctools are built always from the clickhouse/test-util:latest and cached inline -# Theoretically, it should improve rebuild speed significantly -ENV CC=clang-${LLVM_VERSION} -ENV CXX=clang++-${LLVM_VERSION} -# !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! -# DO NOT PUT ANYTHING BEFORE THE NEXT TWO `RUN` DIRECTIVES -# THE MOST HEAVY OPERATION MUST BE THE FIRST IN THE CACHE -# !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! -# libtapi is required to support .tbh format from recent MacOS SDKs -RUN git clone https://github.com/tpoechtrager/apple-libtapi.git \ - && cd apple-libtapi \ - && git checkout 15dfc2a8c9a2a89d06ff227560a69f5265b692f9 \ - && INSTALLPREFIX=/cctools ./build.sh \ - && ./install.sh \ - && cd .. \ - && rm -rf apple-libtapi - -# Build and install tools for cross-linking to Darwin (x86-64) -# Build and install tools for cross-linking to Darwin (aarch64) -RUN git clone https://github.com/tpoechtrager/cctools-port.git \ - && cd cctools-port/cctools \ - && git checkout 2a3e1c2a6ff54a30f898b70cfb9ba1692a55fad7 \ - && ./configure --prefix=/cctools --with-libtapi=/cctools \ - --target=x86_64-apple-darwin \ - && make install -j$(nproc) \ - && make clean \ - && ./configure --prefix=/cctools --with-libtapi=/cctools \ - --target=aarch64-apple-darwin \ - && make install -j$(nproc) \ - && cd ../.. \ - && rm -rf cctools-port - -# !!!!!!!!!!! -# END COMPILE -# !!!!!!!!!!! - -FROM clickhouse/test-util:$FROM_TAG +FROM clickhouse/fasttest:$FROM_TAG ENV CC=clang-${LLVM_VERSION} ENV CXX=clang++-${LLVM_VERSION} @@ -110,7 +73,8 @@ RUN curl -Lo /usr/bin/clang-tidy-cache \ "https://raw.githubusercontent.com/matus-chochlik/ctcache/$CLANG_TIDY_SHA1/clang-tidy-cache" \ && chmod +x /usr/bin/clang-tidy-cache -COPY --from=cctools /cctools /cctools +# If the cctools is updated, then first build it in the CI, then update here in a different commit +COPY --from=clickhouse/cctools:5a908f73878a /cctools /cctools RUN mkdir /workdir && chmod 777 /workdir WORKDIR /workdir diff --git a/docker/packager/binary/build.sh b/docker/packager/binary-builder/build.sh similarity index 100% rename from docker/packager/binary/build.sh rename to docker/packager/binary-builder/build.sh diff --git a/docker/packager/cctools/Dockerfile b/docker/packager/cctools/Dockerfile new file mode 100644 index 00000000000..1b8c675a5c5 --- /dev/null +++ b/docker/packager/cctools/Dockerfile @@ -0,0 +1,31 @@ +# This is a hack to significantly reduce the build time of the clickhouse/binary-builder +# It's based on the assumption that we don't care of the cctools version so much +# It event does not depend on the clickhouse/fasttest in the `docker/images.json` +ARG FROM_TAG=latest +FROM clickhouse/fasttest:$FROM_TAG + +ENV CC=clang-${LLVM_VERSION} +ENV CXX=clang++-${LLVM_VERSION} + +RUN git clone https://github.com/tpoechtrager/apple-libtapi.git \ + && cd apple-libtapi \ + && git checkout 15dfc2a8c9a2a89d06ff227560a69f5265b692f9 \ + && INSTALLPREFIX=/cctools ./build.sh \ + && ./install.sh \ + && cd .. \ + && rm -rf apple-libtapi + +# Build and install tools for cross-linking to Darwin (x86-64) +# Build and install tools for cross-linking to Darwin (aarch64) +RUN git clone https://github.com/tpoechtrager/cctools-port.git \ + && cd cctools-port/cctools \ + && git checkout 2a3e1c2a6ff54a30f898b70cfb9ba1692a55fad7 \ + && ./configure --prefix=/cctools --with-libtapi=/cctools \ + --target=x86_64-apple-darwin \ + && make install -j$(nproc) \ + && make clean \ + && ./configure --prefix=/cctools --with-libtapi=/cctools \ + --target=aarch64-apple-darwin \ + && make install -j$(nproc) \ + && cd ../.. \ + && rm -rf cctools-port diff --git a/docker/packager/packager b/docker/packager/packager index ca0ae8358f3..23fc26bc1a4 100755 --- a/docker/packager/packager +++ b/docker/packager/packager @@ -1,16 +1,16 @@ #!/usr/bin/env python3 # -*- coding: utf-8 -*- -import subprocess -import os import argparse import logging +import os +import subprocess import sys from pathlib import Path from typing import List, Optional SCRIPT_PATH = Path(__file__).absolute() -IMAGE_TYPE = "binary" -IMAGE_NAME = f"clickhouse/{IMAGE_TYPE}-builder" +IMAGE_TYPE = "binary-builder" +IMAGE_NAME = f"clickhouse/{IMAGE_TYPE}" class BuildException(Exception): diff --git a/docker/test/fasttest/Dockerfile b/docker/test/fasttest/Dockerfile index e10555d4d4a..62cdcc3f830 100644 --- a/docker/test/fasttest/Dockerfile +++ b/docker/test/fasttest/Dockerfile @@ -6,9 +6,18 @@ FROM clickhouse/test-util:$FROM_TAG RUN apt-get update \ && apt-get install \ brotli \ + clang-${LLVM_VERSION} \ + clang-tidy-${LLVM_VERSION} \ + cmake \ expect \ file \ + libclang-${LLVM_VERSION}-dev \ + libclang-rt-${LLVM_VERSION}-dev \ + lld-${LLVM_VERSION} \ + llvm-${LLVM_VERSION} \ + llvm-${LLVM_VERSION}-dev \ lsof \ + ninja-build \ odbcinst \ psmisc \ python3 \ @@ -26,14 +35,50 @@ RUN apt-get update \ RUN pip3 install numpy==1.26.3 scipy==1.12.0 pandas==1.5.3 Jinja2==3.1.3 -ARG odbc_driver_url="https://github.com/ClickHouse/clickhouse-odbc/releases/download/v1.1.4.20200302/clickhouse-odbc-1.1.4-Linux.tar.gz" +# This symlink is required by gcc to find the lld linker +RUN ln -s /usr/bin/lld-${LLVM_VERSION} /usr/bin/ld.lld +# for external_symbolizer_path +RUN ln -s /usr/bin/llvm-symbolizer-${LLVM_VERSION} /usr/bin/llvm-symbolizer +# FIXME: workaround for "The imported target "merge-fdata" references the file" error +# https://salsa.debian.org/pkg-llvm-team/llvm-toolchain/-/commit/992e52c0b156a5ba9c6a8a54f8c4857ddd3d371d +RUN sed -i '/_IMPORT_CHECK_FILES_FOR_\(mlir-\|llvm-bolt\|merge-fdata\|MLIR\)/ {s|^|#|}' /usr/lib/llvm-${LLVM_VERSION}/lib/cmake/llvm/LLVMExports-*.cmake -RUN mkdir -p /tmp/clickhouse-odbc-tmp \ - && wget -nv -O - ${odbc_driver_url} | tar --strip-components=1 -xz -C /tmp/clickhouse-odbc-tmp \ - && cp /tmp/clickhouse-odbc-tmp/lib64/*.so /usr/local/lib/ \ - && odbcinst -i -d -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbcinst.ini.sample \ - && odbcinst -i -s -l -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbc.ini.sample \ - && rm -rf /tmp/clickhouse-odbc-tmp +ARG CCACHE_VERSION=4.6.1 +RUN mkdir /tmp/ccache \ + && cd /tmp/ccache \ + && curl -L \ + -O https://github.com/ccache/ccache/releases/download/v$CCACHE_VERSION/ccache-$CCACHE_VERSION.tar.xz \ + -O https://github.com/ccache/ccache/releases/download/v$CCACHE_VERSION/ccache-$CCACHE_VERSION.tar.xz.asc \ + && gpg --recv-keys --keyserver hkps://keyserver.ubuntu.com 5A939A71A46792CF57866A51996DDA075594ADB8 \ + && gpg --verify ccache-4.6.1.tar.xz.asc \ + && tar xf ccache-$CCACHE_VERSION.tar.xz \ + && cd /tmp/ccache/ccache-$CCACHE_VERSION \ + && cmake -DCMAKE_INSTALL_PREFIX=/usr \ + -DCMAKE_BUILD_TYPE=None \ + -DZSTD_FROM_INTERNET=ON \ + -DREDIS_STORAGE_BACKEND=OFF \ + -Wno-dev \ + -B build \ + -S . \ + && make VERBOSE=1 -C build \ + && make install -C build \ + && cd / \ + && rm -rf /tmp/ccache + +ARG TARGETARCH +ARG SCCACHE_VERSION=v0.7.7 +ENV SCCACHE_IGNORE_SERVER_IO_ERROR=1 +# sccache requires a value for the region. So by default we use The Default Region +ENV SCCACHE_REGION=us-east-1 +RUN arch=${TARGETARCH:-amd64} \ + && case $arch in \ + amd64) rarch=x86_64 ;; \ + arm64) rarch=aarch64 ;; \ + esac \ + && curl -Ls "https://github.com/mozilla/sccache/releases/download/$SCCACHE_VERSION/sccache-$SCCACHE_VERSION-$rarch-unknown-linux-musl.tar.gz" | \ + tar xz -C /tmp \ + && mv "/tmp/sccache-$SCCACHE_VERSION-$rarch-unknown-linux-musl/sccache" /usr/bin \ + && rm "/tmp/sccache-$SCCACHE_VERSION-$rarch-unknown-linux-musl" -r # Give suid to gdb to grant it attach permissions # chmod 777 to make the container user independent diff --git a/docker/test/fuzzer/run-fuzzer.sh b/docker/test/fuzzer/run-fuzzer.sh index 9358e88e1e8..dc1b35b3a21 100755 --- a/docker/test/fuzzer/run-fuzzer.sh +++ b/docker/test/fuzzer/run-fuzzer.sh @@ -247,6 +247,12 @@ quit fuzzer_pid=$! echo "Fuzzer pid is $fuzzer_pid" + # The fuzzer_pid belongs to the timeout process. + actual_fuzzer_pid=$(ps -o pid= --ppid "$fuzzer_pid") + + echo "Attaching gdb to the fuzzer itself" + gdb -batch -command script.gdb -p $actual_fuzzer_pid & + # Wait for the fuzzer to complete. # Note that the 'wait || ...' thing is required so that the script doesn't # exit because of 'set -e' when 'wait' returns nonzero code. @@ -337,10 +343,9 @@ quit # which is confusing. task_exit_code=$fuzzer_exit_code echo "failure" > status.txt - { rg -ao "Found error:.*" fuzzer.log \ - || rg -ao "Exception:.*" fuzzer.log \ - || echo "Fuzzer failed ($fuzzer_exit_code). See the logs." ; } \ - | tail -1 > description.txt + echo "Achtung!" > description.txt + echo "Fuzzer went wrong with error code: ($fuzzer_exit_code). Its process died somehow when the server stayed alive. The server log probably won't tell you much so try to find information in other files." >>description.txt + { rg -ao "Found error:.*" fuzzer.log || rg -ao "Exception:.*" fuzzer.log; } | tail -1 >>description.txt fi if test -f core.*; then @@ -387,7 +392,7 @@ if [ -f core.zst ]; then fi # Keep all the lines in the paragraphs containing that either contain or don't start with 20... (year) -sed -n '//,/^$/p' s.log | awk '// || !/^20/' server.log > fatal.log ||: +sed -n '//,/^$/p' server.log | awk '// || !/^20/' > fatal.log ||: FATAL_LINK='' if [ -s fatal.log ]; then FATAL_LINK='fatal.log' diff --git a/docker/test/sqllogic/Dockerfile b/docker/test/sqllogic/Dockerfile index 05130044c45..1ea1e52e6fa 100644 --- a/docker/test/sqllogic/Dockerfile +++ b/docker/test/sqllogic/Dockerfile @@ -24,17 +24,18 @@ RUN pip3 install \ deepdiff \ sqlglot -ARG odbc_repo="https://github.com/ClickHouse/clickhouse-odbc.git" +ARG odbc_driver_url="https://github.com/ClickHouse/clickhouse-odbc/releases/download/v1.1.6.20200320/clickhouse-odbc-1.1.6-Linux.tar.gz" + +RUN mkdir -p /tmp/clickhouse-odbc-tmp \ + && cd /tmp/clickhouse-odbc-tmp \ + && curl -L ${odbc_driver_url} | tar --strip-components=1 -xz clickhouse-odbc-1.1.6-Linux \ + && mkdir /usr/local/lib64 -p \ + && cp /tmp/clickhouse-odbc-tmp/lib64/*.so /usr/local/lib64/ \ + && odbcinst -i -d -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbcinst.ini.sample \ + && odbcinst -i -s -l -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbc.ini.sample \ + && sed -i 's"=libclickhouseodbc"=/usr/local/lib64/libclickhouseodbc"' /etc/odbcinst.ini \ + && rm -rf /tmp/clickhouse-odbc-tmp -RUN git clone --recursive ${odbc_repo} \ - && mkdir -p /clickhouse-odbc/build \ - && cmake -S /clickhouse-odbc -B /clickhouse-odbc/build \ - && ls /clickhouse-odbc/build/driver \ - && make -j 10 -C /clickhouse-odbc/build \ - && ls /clickhouse-odbc/build/driver \ - && mkdir -p /usr/local/lib64/ && cp /clickhouse-odbc/build/driver/lib*.so /usr/local/lib64/ \ - && odbcinst -i -d -f /clickhouse-odbc/packaging/odbcinst.ini.sample \ - && odbcinst -i -s -l -f /clickhouse-odbc/packaging/odbc.ini.sample ENV TZ=Europe/Amsterdam ENV MAX_RUN_TIME=9000 diff --git a/docker/test/stateless/.gitignore b/docker/test/stateless/.gitignore new file mode 100644 index 00000000000..928fed26d6d --- /dev/null +++ b/docker/test/stateless/.gitignore @@ -0,0 +1 @@ +/minio_data diff --git a/docker/test/stateless/Dockerfile b/docker/test/stateless/Dockerfile index 7f4bad3d4e6..cd8864c6299 100644 --- a/docker/test/stateless/Dockerfile +++ b/docker/test/stateless/Dockerfile @@ -3,7 +3,7 @@ ARG FROM_TAG=latest FROM clickhouse/test-base:$FROM_TAG -ARG odbc_driver_url="https://github.com/ClickHouse/clickhouse-odbc/releases/download/v1.1.4.20200302/clickhouse-odbc-1.1.4-Linux.tar.gz" +ARG odbc_driver_url="https://github.com/ClickHouse/clickhouse-odbc/releases/download/v1.1.6.20200320/clickhouse-odbc-1.1.6-Linux.tar.gz" # golang version 1.13 on Ubuntu 20 is enough for tests RUN apt-get update -y \ @@ -35,7 +35,6 @@ RUN apt-get update -y \ sudo \ tree \ unixodbc \ - wget \ rustc \ cargo \ zstd \ @@ -50,11 +49,14 @@ RUN apt-get update -y \ RUN pip3 install numpy==1.26.3 scipy==1.12.0 pandas==1.5.3 Jinja2==3.1.3 pyarrow==15.0.0 RUN mkdir -p /tmp/clickhouse-odbc-tmp \ - && wget -nv -O - ${odbc_driver_url} | tar --strip-components=1 -xz -C /tmp/clickhouse-odbc-tmp \ - && cp /tmp/clickhouse-odbc-tmp/lib64/*.so /usr/local/lib/ \ - && odbcinst -i -d -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbcinst.ini.sample \ - && odbcinst -i -s -l -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbc.ini.sample \ - && rm -rf /tmp/clickhouse-odbc-tmp + && cd /tmp/clickhouse-odbc-tmp \ + && curl -L ${odbc_driver_url} | tar --strip-components=1 -xz clickhouse-odbc-1.1.6-Linux \ + && mkdir /usr/local/lib64 -p \ + && cp /tmp/clickhouse-odbc-tmp/lib64/*.so /usr/local/lib64/ \ + && odbcinst -i -d -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbcinst.ini.sample \ + && odbcinst -i -s -l -f /tmp/clickhouse-odbc-tmp/share/doc/clickhouse-odbc/config/odbc.ini.sample \ + && sed -i 's"=libclickhouseodbc"=/usr/local/lib64/libclickhouseodbc"' /etc/odbcinst.ini \ + && rm -rf /tmp/clickhouse-odbc-tmp ENV TZ=Europe/Amsterdam RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone @@ -70,11 +72,11 @@ ARG TARGETARCH # Download Minio-related binaries RUN arch=${TARGETARCH:-amd64} \ - && wget "https://dl.min.io/server/minio/release/linux-${arch}/archive/minio.RELEASE.${MINIO_SERVER_VERSION}" -O ./minio \ - && wget "https://dl.min.io/client/mc/release/linux-${arch}/archive/mc.RELEASE.${MINIO_CLIENT_VERSION}" -O ./mc \ + && curl -L "https://dl.min.io/server/minio/release/linux-${arch}/archive/minio.RELEASE.${MINIO_SERVER_VERSION}" -o ./minio \ + && curl -L "https://dl.min.io/client/mc/release/linux-${arch}/archive/mc.RELEASE.${MINIO_CLIENT_VERSION}" -o ./mc \ && chmod +x ./mc ./minio -RUN wget --no-verbose 'https://archive.apache.org/dist/hadoop/common/hadoop-3.3.1/hadoop-3.3.1.tar.gz' \ +RUN curl -L --no-verbose -O 'https://archive.apache.org/dist/hadoop/common/hadoop-3.3.1/hadoop-3.3.1.tar.gz' \ && tar -xvf hadoop-3.3.1.tar.gz \ && rm -rf hadoop-3.3.1.tar.gz diff --git a/docker/test/style/Dockerfile b/docker/test/style/Dockerfile index abc2dba0e9d..b4ffcfb597c 100644 --- a/docker/test/style/Dockerfile +++ b/docker/test/style/Dockerfile @@ -18,7 +18,8 @@ RUN apt-get update && env DEBIAN_FRONTEND=noninteractive apt-get install --yes \ python3-pip \ yamllint \ locales \ - && pip3 install black==23.1.0 boto3 codespell==2.2.1 mypy==1.3.0 PyGithub unidiff pylint==2.6.2 \ + && pip3 install black==23.12.0 boto3 codespell==2.2.1 mypy==1.8.0 PyGithub unidiff pylint==3.1.0 \ + requests types-requests \ && apt-get clean \ && rm -rf /var/lib/apt/lists/* /var/cache/debconf /tmp/* \ && rm -rf /root/.cache/pip @@ -59,5 +60,4 @@ RUN arch=${TARGETARCH:-amd64} \ COPY run.sh / -COPY process_style_check_result.py / CMD ["/bin/bash", "/run.sh"] diff --git a/docker/test/upgrade/run.sh b/docker/test/upgrade/run.sh index bf08d537260..69228b1bfc9 100644 --- a/docker/test/upgrade/run.sh +++ b/docker/test/upgrade/run.sh @@ -79,6 +79,18 @@ remove_keeper_config "async_replication" "1" # create_if_not_exists feature flag doesn't exist on some older versions remove_keeper_config "create_if_not_exists" "[01]" +#todo: remove these after 24.3 released. +sudo cat /etc/clickhouse-server/config.d/azure_storage_conf.xml \ + | sed "s|azure|azure_blob_storage|" \ + > /etc/clickhouse-server/config.d/azure_storage_conf.xml.tmp +sudo mv /etc/clickhouse-server/config.d/azure_storage_conf.xml.tmp /etc/clickhouse-server/config.d/azure_storage_conf.xml + +#todo: remove these after 24.3 released. +sudo cat /etc/clickhouse-server/config.d/storage_conf.xml \ + | sed "s|local|local_blob_storage|" \ + > /etc/clickhouse-server/config.d/storage_conf.xml.tmp +sudo mv /etc/clickhouse-server/config.d/storage_conf.xml.tmp /etc/clickhouse-server/config.d/storage_conf.xml + # latest_logs_cache_size_threshold setting doesn't exist on some older versions remove_keeper_config "latest_logs_cache_size_threshold" "[[:digit:]]\+" @@ -113,6 +125,18 @@ sudo cat /etc/clickhouse-server/config.d/keeper_port.xml \ > /etc/clickhouse-server/config.d/keeper_port.xml.tmp sudo mv /etc/clickhouse-server/config.d/keeper_port.xml.tmp /etc/clickhouse-server/config.d/keeper_port.xml +#todo: remove these after 24.3 released. +sudo cat /etc/clickhouse-server/config.d/azure_storage_conf.xml \ + | sed "s|azure|azure_blob_storage|" \ + > /etc/clickhouse-server/config.d/azure_storage_conf.xml.tmp +sudo mv /etc/clickhouse-server/config.d/azure_storage_conf.xml.tmp /etc/clickhouse-server/config.d/azure_storage_conf.xml + +#todo: remove these after 24.3 released. +sudo cat /etc/clickhouse-server/config.d/storage_conf.xml \ + | sed "s|local|local_blob_storage|" \ + > /etc/clickhouse-server/config.d/storage_conf.xml.tmp +sudo mv /etc/clickhouse-server/config.d/storage_conf.xml.tmp /etc/clickhouse-server/config.d/storage_conf.xml + # async_replication setting doesn't exist on some older versions remove_keeper_config "async_replication" "1" diff --git a/docker/test/util/Dockerfile b/docker/test/util/Dockerfile index 396d5801be9..4f2dc9df849 100644 --- a/docker/test/util/Dockerfile +++ b/docker/test/util/Dockerfile @@ -41,20 +41,11 @@ RUN apt-get update \ bash \ bsdmainutils \ build-essential \ - clang-${LLVM_VERSION} \ - clang-tidy-${LLVM_VERSION} \ - cmake \ gdb \ git \ gperf \ - libclang-rt-${LLVM_VERSION}-dev \ - lld-${LLVM_VERSION} \ - llvm-${LLVM_VERSION} \ - llvm-${LLVM_VERSION}-dev \ - libclang-${LLVM_VERSION}-dev \ moreutils \ nasm \ - ninja-build \ pigz \ rename \ software-properties-common \ @@ -63,49 +54,4 @@ RUN apt-get update \ && apt-get clean \ && rm -rf /var/lib/apt/lists/* /var/cache/debconf /tmp/* -# This symlink is required by gcc to find the lld linker -RUN ln -s /usr/bin/lld-${LLVM_VERSION} /usr/bin/ld.lld -# for external_symbolizer_path -RUN ln -s /usr/bin/llvm-symbolizer-${LLVM_VERSION} /usr/bin/llvm-symbolizer -# FIXME: workaround for "The imported target "merge-fdata" references the file" error -# https://salsa.debian.org/pkg-llvm-team/llvm-toolchain/-/commit/992e52c0b156a5ba9c6a8a54f8c4857ddd3d371d -RUN sed -i '/_IMPORT_CHECK_FILES_FOR_\(mlir-\|llvm-bolt\|merge-fdata\|MLIR\)/ {s|^|#|}' /usr/lib/llvm-${LLVM_VERSION}/lib/cmake/llvm/LLVMExports-*.cmake - -ARG CCACHE_VERSION=4.6.1 -RUN mkdir /tmp/ccache \ - && cd /tmp/ccache \ - && curl -L \ - -O https://github.com/ccache/ccache/releases/download/v$CCACHE_VERSION/ccache-$CCACHE_VERSION.tar.xz \ - -O https://github.com/ccache/ccache/releases/download/v$CCACHE_VERSION/ccache-$CCACHE_VERSION.tar.xz.asc \ - && gpg --recv-keys --keyserver hkps://keyserver.ubuntu.com 5A939A71A46792CF57866A51996DDA075594ADB8 \ - && gpg --verify ccache-4.6.1.tar.xz.asc \ - && tar xf ccache-$CCACHE_VERSION.tar.xz \ - && cd /tmp/ccache/ccache-$CCACHE_VERSION \ - && cmake -DCMAKE_INSTALL_PREFIX=/usr \ - -DCMAKE_BUILD_TYPE=None \ - -DZSTD_FROM_INTERNET=ON \ - -DREDIS_STORAGE_BACKEND=OFF \ - -Wno-dev \ - -B build \ - -S . \ - && make VERBOSE=1 -C build \ - && make install -C build \ - && cd / \ - && rm -rf /tmp/ccache - -ARG TARGETARCH -ARG SCCACHE_VERSION=v0.5.4 -ENV SCCACHE_IGNORE_SERVER_IO_ERROR=1 -# sccache requires a value for the region. So by default we use The Default Region -ENV SCCACHE_REGION=us-east-1 -RUN arch=${TARGETARCH:-amd64} \ - && case $arch in \ - amd64) rarch=x86_64 ;; \ - arm64) rarch=aarch64 ;; \ - esac \ - && curl -Ls "https://github.com/mozilla/sccache/releases/download/$SCCACHE_VERSION/sccache-$SCCACHE_VERSION-$rarch-unknown-linux-musl.tar.gz" | \ - tar xz -C /tmp \ - && mv "/tmp/sccache-$SCCACHE_VERSION-$rarch-unknown-linux-musl/sccache" /usr/bin \ - && rm "/tmp/sccache-$SCCACHE_VERSION-$rarch-unknown-linux-musl" -r - COPY process_functional_tests_result.py / diff --git a/docs/changelogs/v23.3.20.27-lts.md b/docs/changelogs/v23.3.20.27-lts.md new file mode 100644 index 00000000000..9f49e47f0bc --- /dev/null +++ b/docs/changelogs/v23.3.20.27-lts.md @@ -0,0 +1,29 @@ +--- +sidebar_position: 1 +sidebar_label: 2024 +--- + +# 2024 Changelog + +### ClickHouse release v23.3.20.27-lts (cc974ba4f81) FIXME as compared to v23.3.19.32-lts (c4d4ca8ec02) + +#### Improvement +* Backported in [#58818](https://github.com/ClickHouse/ClickHouse/issues/58818): Add `SYSTEM JEMALLOC PURGE` for purging unused jemalloc pages, `SYSTEM JEMALLOC [ ENABLE | DISABLE | FLUSH ] PROFILE` for controlling jemalloc profile if the profiler is enabled. Add jemalloc-related 4LW command in Keeper: `jmst` for dumping jemalloc stats, `jmfp`, `jmep`, `jmdp` for controlling jemalloc profile if the profiler is enabled. [#58665](https://github.com/ClickHouse/ClickHouse/pull/58665) ([Antonio Andelic](https://github.com/antonio2368)). + +#### Build/Testing/Packaging Improvement +* Backported in [#59877](https://github.com/ClickHouse/ClickHouse/issues/59877): If you want to run initdb scripts every time when ClickHouse container is starting you shoud initialize environment varible CLICKHOUSE_ALWAYS_RUN_INITDB_SCRIPTS. [#59808](https://github.com/ClickHouse/ClickHouse/pull/59808) ([Alexander Nikolaev](https://github.com/AlexNik)). + +#### Bug Fix (user-visible misbehavior in an official stable release) + +* Fix working with read buffers in StreamingFormatExecutor [#57438](https://github.com/ClickHouse/ClickHouse/pull/57438) ([Kruglov Pavel](https://github.com/Avogar)). +* Fix double destroy call on exception throw in addBatchLookupTable8 [#58745](https://github.com/ClickHouse/ClickHouse/pull/58745) ([Raúl Marín](https://github.com/Algunenano)). +* Fix: LIMIT BY and LIMIT in distributed query [#59153](https://github.com/ClickHouse/ClickHouse/pull/59153) ([Igor Nikonov](https://github.com/devcrafter)). +* Fix translate() with FixedString input [#59356](https://github.com/ClickHouse/ClickHouse/pull/59356) ([Raúl Marín](https://github.com/Algunenano)). +* Fix leftPad / rightPad function with FixedString input [#59739](https://github.com/ClickHouse/ClickHouse/pull/59739) ([Raúl Marín](https://github.com/Algunenano)). +* Fix cosineDistance crash with Nullable [#60150](https://github.com/ClickHouse/ClickHouse/pull/60150) ([Raúl Marín](https://github.com/Algunenano)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Fix possible race in ManyAggregatedData dtor. [#58624](https://github.com/ClickHouse/ClickHouse/pull/58624) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Make ZooKeeper actually sequentialy consistent [#59735](https://github.com/ClickHouse/ClickHouse/pull/59735) ([Alexander Tokmakov](https://github.com/tavplubix)). + diff --git a/docs/changelogs/v23.8.10.43-lts.md b/docs/changelogs/v23.8.10.43-lts.md new file mode 100644 index 00000000000..0093467d129 --- /dev/null +++ b/docs/changelogs/v23.8.10.43-lts.md @@ -0,0 +1,39 @@ +--- +sidebar_position: 1 +sidebar_label: 2024 +--- + +# 2024 Changelog + +### ClickHouse release v23.8.10.43-lts (a278225bba9) FIXME as compared to v23.8.9.54-lts (192a1d231fa) + +#### Improvement +* Backported in [#58819](https://github.com/ClickHouse/ClickHouse/issues/58819): Add `SYSTEM JEMALLOC PURGE` for purging unused jemalloc pages, `SYSTEM JEMALLOC [ ENABLE | DISABLE | FLUSH ] PROFILE` for controlling jemalloc profile if the profiler is enabled. Add jemalloc-related 4LW command in Keeper: `jmst` for dumping jemalloc stats, `jmfp`, `jmep`, `jmdp` for controlling jemalloc profile if the profiler is enabled. [#58665](https://github.com/ClickHouse/ClickHouse/pull/58665) ([Antonio Andelic](https://github.com/antonio2368)). +* Backported in [#60286](https://github.com/ClickHouse/ClickHouse/issues/60286): Copy S3 file GCP fallback to buffer copy in case GCP returned `Internal Error` with `GATEWAY_TIMEOUT` HTTP error code. [#60164](https://github.com/ClickHouse/ClickHouse/pull/60164) ([Maksim Kita](https://github.com/kitaisreal)). + +#### Build/Testing/Packaging Improvement +* Backported in [#59879](https://github.com/ClickHouse/ClickHouse/issues/59879): If you want to run initdb scripts every time when ClickHouse container is starting you shoud initialize environment varible CLICKHOUSE_ALWAYS_RUN_INITDB_SCRIPTS. [#59808](https://github.com/ClickHouse/ClickHouse/pull/59808) ([Alexander Nikolaev](https://github.com/AlexNik)). + +#### Bug Fix (user-visible misbehavior in an official stable release) + +* Background merges correctly use temporary data storage in the cache [#57275](https://github.com/ClickHouse/ClickHouse/pull/57275) ([vdimir](https://github.com/vdimir)). +* MergeTree mutations reuse source part index granularity [#57352](https://github.com/ClickHouse/ClickHouse/pull/57352) ([Maksim Kita](https://github.com/kitaisreal)). +* Fix double destroy call on exception throw in addBatchLookupTable8 [#58745](https://github.com/ClickHouse/ClickHouse/pull/58745) ([Raúl Marín](https://github.com/Algunenano)). +* Fix JSONExtract function for LowCardinality(Nullable) columns [#58808](https://github.com/ClickHouse/ClickHouse/pull/58808) ([vdimir](https://github.com/vdimir)). +* Fix: LIMIT BY and LIMIT in distributed query [#59153](https://github.com/ClickHouse/ClickHouse/pull/59153) ([Igor Nikonov](https://github.com/devcrafter)). +* Fix translate() with FixedString input [#59356](https://github.com/ClickHouse/ClickHouse/pull/59356) ([Raúl Marín](https://github.com/Algunenano)). +* Fix error "Read beyond last offset" for AsynchronousBoundedReadBuffer [#59630](https://github.com/ClickHouse/ClickHouse/pull/59630) ([Vitaly Baranov](https://github.com/vitlibar)). +* Fix query start time on non initial queries [#59662](https://github.com/ClickHouse/ClickHouse/pull/59662) ([Raúl Marín](https://github.com/Algunenano)). +* Fix leftPad / rightPad function with FixedString input [#59739](https://github.com/ClickHouse/ClickHouse/pull/59739) ([Raúl Marín](https://github.com/Algunenano)). +* rabbitmq: fix having neither acked nor nacked messages [#59775](https://github.com/ClickHouse/ClickHouse/pull/59775) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix cosineDistance crash with Nullable [#60150](https://github.com/ClickHouse/ClickHouse/pull/60150) ([Raúl Marín](https://github.com/Algunenano)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Fix rare race in external sort/aggregation with temporary data in cache [#58013](https://github.com/ClickHouse/ClickHouse/pull/58013) ([Anton Popov](https://github.com/CurtizJ)). +* Fix possible race in ManyAggregatedData dtor. [#58624](https://github.com/ClickHouse/ClickHouse/pull/58624) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fix 02720_row_policy_column_with_dots [#59453](https://github.com/ClickHouse/ClickHouse/pull/59453) ([Duc Canh Le](https://github.com/canhld94)). +* Pin python dependencies in stateless tests [#59663](https://github.com/ClickHouse/ClickHouse/pull/59663) ([Raúl Marín](https://github.com/Algunenano)). +* Make ZooKeeper actually sequentialy consistent [#59735](https://github.com/ClickHouse/ClickHouse/pull/59735) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Remove broken test while we fix it [#60547](https://github.com/ClickHouse/ClickHouse/pull/60547) ([Raúl Marín](https://github.com/Algunenano)). + diff --git a/docs/en/interfaces/schema-inference.md b/docs/en/interfaces/schema-inference.md index 39ae69eaef4..05fae994cbe 100644 --- a/docs/en/interfaces/schema-inference.md +++ b/docs/en/interfaces/schema-inference.md @@ -549,6 +549,48 @@ Result: └───────┴─────────────────────────────────────────────────────────────────────────────────────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘ ``` +##### input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects + +Enabling this setting allows to use String type for ambiguous paths during named tuples inference from JSON objects (when `input_format_json_try_infer_named_tuples_from_objects` is enabled) instead of an exception. +It allows to read JSON objects as named Tuples even if there are ambiguous paths. + +Disabled by default. + +**Examples** + +With disabled setting: +```sql +SET input_format_json_try_infer_named_tuples_from_objects = 1; +SET input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects = 0; +DESC format(JSONEachRow, '{"obj" : {"a" : 42}}, {"obj" : {"a" : {"b" : "Hello"}}}'); +``` +Result: + +```text +Code: 636. DB::Exception: The table structure cannot be extracted from a JSONEachRow format file. Error: +Code: 117. DB::Exception: JSON objects have ambiguous data: in some objects path 'a' has type 'Int64' and in some - 'Tuple(b String)'. You can enable setting input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects to use String type for path 'a'. (INCORRECT_DATA) (version 24.3.1.1). +You can specify the structure manually. (CANNOT_EXTRACT_TABLE_STRUCTURE) +``` + +With enabled setting: +```sql +SET input_format_json_try_infer_named_tuples_from_objects = 1; +SET input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects = 1; +DESC format(JSONEachRow, '{"obj" : "a" : 42}, {"obj" : {"a" : {"b" : "Hello"}}}'); +SELECT * FROM format(JSONEachRow, '{"obj" : {"a" : 42}}, {"obj" : {"a" : {"b" : "Hello"}}}'); +``` + +Result: +```text +┌─name─┬─type──────────────────────────┬─default_type─┬─default_expression─┬─comment─┬─codec_expression─┬─ttl_expression─┠+│ obj │ Tuple(a Nullable(String)) │ │ │ │ │ │ +└──────┴───────────────────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘ +┌─obj─────────────────┠+│ ('42') │ +│ ('{"b" : "Hello"}') │ +└─────────────────────┘ +``` + ##### input_format_json_read_objects_as_strings Enabling this setting allows reading nested JSON objects as strings. @@ -1554,6 +1596,28 @@ DESC format(JSONEachRow, $$ └──────┴──────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘ ``` +#### input_format_try_infer_exponent_floats + +If enabled, ClickHouse will try to infer floats in exponential form for text formats (except JSON where numbers in exponential form are always inferred). + +Disabled by default. + +**Example** + +```sql +SET input_format_try_infer_exponent_floats = 1; +DESC format(CSV, +$$1.1E10 +2.3e-12 +42E00 +$$) +``` +```response +┌─name─┬─type──────────────┬─default_type─┬─default_expression─┬─comment─┬─codec_expression─┬─ttl_expression─┠+│ c1 │ Nullable(Float64) │ │ │ │ │ │ +└──────┴───────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘ +``` + ## Self describing formats {#self-describing-formats} Self-describing formats contain information about the structure of the data in the data itself, diff --git a/docs/en/operations/settings/settings-formats.md b/docs/en/operations/settings/settings-formats.md index f7d9586dd5b..477fbf94625 100644 --- a/docs/en/operations/settings/settings-formats.md +++ b/docs/en/operations/settings/settings-formats.md @@ -467,7 +467,7 @@ Enabled by default. Allow to use String type for JSON keys that contain only `Null`/`{}`/`[]` in data sample during schema inference. In JSON formats any value can be read as String, and we can avoid errors like `Cannot determine type for column 'column_name' by first 25000 rows of data, most likely this column contains only Nulls or empty Arrays/Maps` during schema inference -by using String type for keys with unknown types. +by using String type for keys with unknown types. Example: @@ -891,7 +891,7 @@ Default value: `,`. If it is set to true, allow strings in single quotes. -Enabled by default. +Disabled by default. ### format_csv_allow_double_quotes {#format_csv_allow_double_quotes} @@ -1605,7 +1605,7 @@ possible values: - `1` — Enabled. Pretty formats will use ANSI escape sequences except for `NoEscapes` formats. - `auto` - Enabled if `stdout` is a terminal except for `NoEscapes` formats. -Default value is `auto`. +Default value is `auto`. ### output_format_pretty_grid_charset {#output_format_pretty_grid_charset} diff --git a/docs/en/operations/settings/settings.md b/docs/en/operations/settings/settings.md index 622644a1543..3f4dec9dc93 100644 --- a/docs/en/operations/settings/settings.md +++ b/docs/en/operations/settings/settings.md @@ -3954,6 +3954,7 @@ Possible values: - `none` — Is similar to throw, but distributed DDL query returns no result set. - `null_status_on_timeout` — Returns `NULL` as execution status in some rows of result set instead of throwing `TIMEOUT_EXCEEDED` if query is not finished on the corresponding hosts. - `never_throw` — Do not throw `TIMEOUT_EXCEEDED` and do not rethrow exceptions if query has failed on some hosts. +- `none_only_active` - similar to `none`, but doesn't wait for inactive replicas of the `Replicated` database. Note: with this mode it's impossible to figure out that the query was not executed on some replica and will be executed in background. - `null_status_on_timeout_only_active` — similar to `null_status_on_timeout`, but doesn't wait for inactive replicas of the `Replicated` database - `throw_only_active` — similar to `throw`, but doesn't wait for inactive replicas of the `Replicated` database diff --git a/docs/en/operations/utilities/clickhouse-local.md b/docs/en/operations/utilities/clickhouse-local.md index 437a5f0fff0..93a3fecf3c6 100644 --- a/docs/en/operations/utilities/clickhouse-local.md +++ b/docs/en/operations/utilities/clickhouse-local.md @@ -201,12 +201,12 @@ Arguments: - `-S`, `--structure` — table structure for input data. - `--input-format` — input format, `TSV` by default. -- `-f`, `--file` — path to data, `stdin` by default. +- `-F`, `--file` — path to data, `stdin` by default. - `-q`, `--query` — queries to execute with `;` as delimiter. `--query` can be specified multiple times, e.g. `--query "SELECT 1" --query "SELECT 2"`. Cannot be used simultaneously with `--queries-file`. - `--queries-file` - file path with queries to execute. `--queries-file` can be specified multiple times, e.g. `--query queries1.sql --query queries2.sql`. Cannot be used simultaneously with `--query`. - `--multiquery, -n` – If specified, multiple queries separated by semicolons can be listed after the `--query` option. For convenience, it is also possible to omit `--query` and pass the queries directly after `--multiquery`. - `-N`, `--table` — table name where to put output data, `table` by default. -- `--format`, `--output-format` — output format, `TSV` by default. +- `-f`, `--format`, `--output-format` — output format, `TSV` by default. - `-d`, `--database` — default database, `_local` by default. - `--stacktrace` — whether to dump debug output in case of exception. - `--echo` — print query before execution. diff --git a/docs/en/sql-reference/aggregate-functions/reference/grouparraysorted.md b/docs/en/sql-reference/aggregate-functions/reference/grouparraysorted.md index cc601c097fe..9bee0c29e7a 100644 --- a/docs/en/sql-reference/aggregate-functions/reference/grouparraysorted.md +++ b/docs/en/sql-reference/aggregate-functions/reference/grouparraysorted.md @@ -14,8 +14,6 @@ - `N` – The number of elements to return. - If the parameter is omitted, default value is the size of input. - - `column` – The value (Integer, String, Float and other Generic types). **Example** @@ -36,13 +34,12 @@ Gets all the String implementations of all numbers in column: ``` sql -SELECT groupArraySorted(str) FROM (SELECT toString(number) as str FROM numbers(5)); +SELECT groupArraySorted(5)(str) FROM (SELECT toString(number) as str FROM numbers(5)); ``` ``` text - ┌─groupArraySorted(str)────────┠- │ ['0','1','2','3','4'] │ - └──────────────────────────────┘ - ``` - \ No newline at end of file +┌─groupArraySorted(5)(str)─┠+│ ['0','1','2','3','4'] │ +└──────────────────────────┘ + ``` \ No newline at end of file diff --git a/docs/en/sql-reference/functions/date-time-functions.md b/docs/en/sql-reference/functions/date-time-functions.md index 41503abfa2f..ba7695af3fa 100644 --- a/docs/en/sql-reference/functions/date-time-functions.md +++ b/docs/en/sql-reference/functions/date-time-functions.md @@ -394,8 +394,7 @@ Result: ## toYear -Converts a date or date with time to the year number (AD) as `UInt16` value. - +Returns the year component (AD) of a date or date with time. **Syntax** @@ -431,7 +430,7 @@ Result: ## toQuarter -Converts a date or date with time to the quarter number (1-4) as `UInt8` value. +Returns the quarter (1-4) of a date or date with time. **Syntax** @@ -465,10 +464,9 @@ Result: └──────────────────────────────────────────────┘ ``` - ## toMonth -Converts a date or date with time to the month number (1-12) as `UInt8` value. +Returns the month component (1-12) of a date or date with time. **Syntax** @@ -504,7 +502,7 @@ Result: ## toDayOfYear -Converts a date or date with time to the number of the day of the year (1-366) as `UInt16` value. +Returns the number of the day within the year (1-366) of a date or date with time. **Syntax** @@ -540,7 +538,7 @@ Result: ## toDayOfMonth -Converts a date or date with time to the number of the day in the month (1-31) as `UInt8` value. +Returns the number of the day within the month (1-31) of a date or date with time. **Syntax** @@ -576,7 +574,7 @@ Result: ## toDayOfWeek -Converts a date or date with time to the number of the day in the week as `UInt8` value. +Returns the number of the day within the week of a date or date with time. The two-argument form of `toDayOfWeek()` enables you to specify whether the week starts on Monday or Sunday, and whether the return value should be in the range from 0 to 6 or 1 to 7. If the mode argument is omitted, the default mode is 0. The time zone of the date can be specified as the third argument. @@ -627,7 +625,7 @@ Result: ## toHour -Converts a date with time to the number of the hour in 24-hour time (0-23) as `UInt8` value. +Returns the hour component (0-24) of a date with time. Assumes that if clocks are moved ahead, it is by one hour and occurs at 2 a.m., and if clocks are moved back, it is by one hour and occurs at 3 a.m. (which is not always exactly when it occurs - it depends on the timezone). @@ -641,7 +639,7 @@ Alias: `HOUR` **Arguments** -- `value` - a [Date](../data-types/date.md), [Date32](../data-types/date32.md), [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) +- `value` - a [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) **Returned value** @@ -665,7 +663,7 @@ Result: ## toMinute -Converts a date with time to the number of the minute of the hour (0-59) as `UInt8` value. +Returns the minute component (0-59) a date with time. **Syntax** @@ -677,7 +675,7 @@ Alias: `MINUTE` **Arguments** -- `value` - a [Date](../data-types/date.md), [Date32](../data-types/date32.md), [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) +- `value` - a [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) **Returned value** @@ -701,7 +699,7 @@ Result: ## toSecond -Converts a date with time to the second in the minute (0-59) as `UInt8` value. Leap seconds are not considered. +Returns the second component (0-59) of a date with time. Leap seconds are not considered. **Syntax** @@ -713,7 +711,7 @@ Alias: `SECOND` **Arguments** -- `value` - a [Date](../data-types/date.md), [Date32](../data-types/date32.md), [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) +- `value` - a [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) **Returned value** @@ -735,6 +733,40 @@ Result: └─────────────────────────────────────────────┘ ``` +## toMillisecond + +Returns the millisecond component (0-999) of a date with time. + +**Syntax** + +```sql +toMillisecond(value) +``` + +*Arguments** + +- `value` - [DateTime](../data-types/datetime.md) or [DateTime64](../data-types/datetime64.md) + +Alias: `MILLISECOND` + +```sql +SELECT toMillisecond(toDateTime64('2023-04-21 10:20:30.456', 3)) +``` + +Result: + +```response +┌──toMillisecond(toDateTime64('2023-04-21 10:20:30.456', 3))─┠+│ 456 │ +└────────────────────────────────────────────────────────────┘ +``` + +**Returned value** + +- The millisecond in the minute (0 - 59) of the given date/time + +Type: `UInt16` + ## toUnixTimestamp Converts a string, a date or a date with time to the [Unix Timestamp](https://en.wikipedia.org/wiki/Unix_time) in `UInt32` representation. diff --git a/docs/en/sql-reference/statements/alter/column.md b/docs/en/sql-reference/statements/alter/column.md index f6d9668e628..0989c151d18 100644 --- a/docs/en/sql-reference/statements/alter/column.md +++ b/docs/en/sql-reference/statements/alter/column.md @@ -272,10 +272,16 @@ ALTER TABLE table_name MODIFY COLUMN column_name RESET SETTING max_compress_bloc ## MATERIALIZE COLUMN -Materializes or updates a column with an expression for a default value (`DEFAULT` or `MATERIALIZED`). -It is used if it is necessary to add or update a column with a complicated expression, because evaluating such an expression directly on `SELECT` executing turns out to be expensive. +Materializes a column with a `DEFAULT` or `MATERIALIZED` value expression. +This statement can be used to rewrite existing column data after a `DEFAULT` or `MATERIALIZED` expression has been added or updated (which only updates the metadata but does not change existing data). Implemented as a [mutation](/docs/en/sql-reference/statements/alter/index.md#mutations). +For columns with a new or updated `MATERIALIZED` value expression, all existing rows are rewritten. + +For columns with a new or updated `DEFAULT` value expression, the behavior depends on the ClickHouse version: +- In ClickHouse < v24.2, all existing rows are rewritten. +- ClickHouse >= v24.2 distinguishes if a row value in a column with `DEFAULT` value expression was explicitly specified when it was inserted, or not, i.e. calculated from the `DEFAULT` value expression. If the value was explicitly specified, ClickHouse keeps it as is. If the value was was calculated, ClickHouse changes it to the new or updated `MATERIALIZED` value expression. + Syntax: ```sql diff --git a/docs/en/sql-reference/statements/grant.md b/docs/en/sql-reference/statements/grant.md index 4e5476210e3..a93db29e82c 100644 --- a/docs/en/sql-reference/statements/grant.md +++ b/docs/en/sql-reference/statements/grant.md @@ -202,6 +202,13 @@ Hierarchy of privileges: - `S3` - [dictGet](#grant-dictget) - [displaySecretsInShowAndSelect](#grant-display-secrets) +- [NAMED COLLECTION ADMIN](#grant-named-collection-admin) + - `CREATE NAMED COLLECTION` + - `DROP NAMED COLLECTION` + - `ALTER NAMED COLLECTION` + - `SHOW NAMED COLLECTIONS` + - `SHOW NAMED COLLECTIONS SECRETS` + - `NAMED COLLECTION` Examples of how this hierarchy is treated: @@ -498,6 +505,25 @@ and [`format_display_secrets_in_show_and_select` format setting](../../operations/settings/formats#format_display_secrets_in_show_and_select) are turned on. +### NAMED COLLECTION ADMIN + +Allows a certain operation on a specified named collection. Before version 23.7 it was called NAMED COLLECTION CONTROL, and after 23.7 NAMED COLLECTION ADMIN was added and NAMED COLLECTION CONTROL is preserved as an alias. + +- `NAMED COLLECTION ADMIN`. Level: `NAMED_COLLECTION`. Aliases: `NAMED COLLECTION CONTROL` + - `CREATE NAMED COLLECTION`. Level: `NAMED_COLLECTION` + - `DROP NAMED COLLECTION`. Level: `NAMED_COLLECTION` + - `ALTER NAMED COLLECTION`. Level: `NAMED_COLLECTION` + - `SHOW NAMED COLLECTIONS`. Level: `NAMED_COLLECTION`. Aliases: `SHOW NAMED COLLECTIONS` + - `SHOW NAMED COLLECTIONS SECRETS`. Level: `NAMED_COLLECTION`. Aliases: `SHOW NAMED COLLECTIONS SECRETS` + - `NAMED COLLECTION`. Level: `NAMED_COLLECTION`. Aliases: `NAMED COLLECTION USAGE, USE NAMED COLLECTION` + +Unlike all other grants (CREATE, DROP, ALTER, SHOW) grant NAMED COLLECTION was added only in 23.7, while all others were added earlier - in 22.12. + +**Examples** + +Assuming a named collection is called abc, we grant privilege CREATE NAMED COLLECTION to user john. +- `GRANT CREATE NAMED COLLECTION ON abc TO john` + ### ALL Grants all the privileges on regulated entity to a user account or a role. diff --git a/docs/en/sql-reference/window-functions/index.md b/docs/en/sql-reference/window-functions/index.md index 6340c369bff..9b2ded7b6ce 100644 --- a/docs/en/sql-reference/window-functions/index.md +++ b/docs/en/sql-reference/window-functions/index.md @@ -5,7 +5,12 @@ sidebar_label: Window Functions title: Window Functions --- -ClickHouse supports the standard grammar for defining windows and window functions. The following features are currently supported: +Windows functions let you perform calculations across a set of rows that are related to the current row. +Some of the calculations that you can do are similar to those that can be done with an aggregate function, but a window function doesn't cause rows to be grouped into a single output - the individual rows are still returned. + +## Standard Window Functions + +ClickHouse supports the standard grammar for defining windows and window functions. The table below indicates whether a feature is currently supported. | Feature | Support or workaround | |------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| @@ -25,6 +30,8 @@ ClickHouse supports the standard grammar for defining windows and window functio ## ClickHouse-specific Window Functions +There are also the following window function that's specific to ClickHouse: + ### nonNegativeDerivative(metric_column, timestamp_column[, INTERVAL X UNITS]) Finds non-negative derivative for given `metric_column` by `timestamp_column`. @@ -33,40 +40,6 @@ The computed value is the following for each row: - `0` for 1st row, - ${metric_i - metric_{i-1} \over timestamp_i - timestamp_{i-1}} * interval$ for $i_th$ row. -## References - -### GitHub Issues - -The roadmap for the initial support of window functions is [in this issue](https://github.com/ClickHouse/ClickHouse/issues/18097). - -All GitHub issues related to window functions have the [comp-window-functions](https://github.com/ClickHouse/ClickHouse/labels/comp-window-functions) tag. - -### Tests - -These tests contain the examples of the currently supported grammar: - -https://github.com/ClickHouse/ClickHouse/blob/master/tests/performance/window_functions.xml - -https://github.com/ClickHouse/ClickHouse/blob/master/tests/queries/0_stateless/01591_window_functions.sql - -### Postgres Docs - -https://www.postgresql.org/docs/current/sql-select.html#SQL-WINDOW - -https://www.postgresql.org/docs/devel/sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS - -https://www.postgresql.org/docs/devel/functions-window.html - -https://www.postgresql.org/docs/devel/tutorial-window.html - -### MySQL Docs - -https://dev.mysql.com/doc/refman/8.0/en/window-function-descriptions.html - -https://dev.mysql.com/doc/refman/8.0/en/window-functions-usage.html - -https://dev.mysql.com/doc/refman/8.0/en/window-functions-frames.html - ## Syntax ```text @@ -80,20 +53,7 @@ WINDOW window_name as ([[PARTITION BY grouping_column] [ORDER BY sorting_column] - `PARTITION BY` - defines how to break a resultset into groups. - `ORDER BY` - defines how to order rows inside the group during calculation aggregate_function. - `ROWS or RANGE` - defines bounds of a frame, aggregate_function is calculated within a frame. -- `WINDOW` - allows to reuse a window definition with multiple expressions. - -### Functions - -These functions can be used only as a window function. - -- `row_number()` - Number the current row within its partition starting from 1. -- `first_value(x)` - Return the first non-NULL value evaluated within its ordered frame. -- `last_value(x)` - Return the last non-NULL value evaluated within its ordered frame. -- `nth_value(x, offset)` - Return the first non-NULL value evaluated against the nth row (offset) in its ordered frame. -- `rank()` - Rank the current row within its partition with gaps. -- `dense_rank()` - Rank the current row within its partition without gaps. -- `lagInFrame(x)` - Return a value evaluated at the row that is at a specified physical offset row before the current row within the ordered frame. -- `leadInFrame(x)` - Return a value evaluated at the row that is offset rows after the current row within the ordered frame. +- `WINDOW` - allows multiple expressions to use the same window definition. ```text PARTITION @@ -112,8 +72,23 @@ These functions can be used only as a window function. └─────────────────┘ <--- UNBOUNDED FOLLOWING (END of the PARTITION) ``` +### Functions + +These functions can be used only as a window function. + +- `row_number()` - Number the current row within its partition starting from 1. +- `first_value(x)` - Return the first non-NULL value evaluated within its ordered frame. +- `last_value(x)` - Return the last non-NULL value evaluated within its ordered frame. +- `nth_value(x, offset)` - Return the first non-NULL value evaluated against the nth row (offset) in its ordered frame. +- `rank()` - Rank the current row within its partition with gaps. +- `dense_rank()` - Rank the current row within its partition without gaps. +- `lagInFrame(x)` - Return a value evaluated at the row that is at a specified physical offset row before the current row within the ordered frame. +- `leadInFrame(x)` - Return a value evaluated at the row that is offset rows after the current row within the ordered frame. + ## Examples +Let's have a look at some examples of how window functions can be used. + ```sql CREATE TABLE wf_partition ( @@ -589,6 +564,41 @@ ORDER BY └──────────────┴─────────────────────┴───────┴─────────────────────────┘ ``` +## References + +### GitHub Issues + +The roadmap for the initial support of window functions is [in this issue](https://github.com/ClickHouse/ClickHouse/issues/18097). + +All GitHub issues related to window functions have the [comp-window-functions](https://github.com/ClickHouse/ClickHouse/labels/comp-window-functions) tag. + +### Tests + +These tests contain the examples of the currently supported grammar: + +https://github.com/ClickHouse/ClickHouse/blob/master/tests/performance/window_functions.xml + +https://github.com/ClickHouse/ClickHouse/blob/master/tests/queries/0_stateless/01591_window_functions.sql + +### Postgres Docs + +https://www.postgresql.org/docs/current/sql-select.html#SQL-WINDOW + +https://www.postgresql.org/docs/devel/sql-expressions.html#SYNTAX-WINDOW-FUNCTIONS + +https://www.postgresql.org/docs/devel/functions-window.html + +https://www.postgresql.org/docs/devel/tutorial-window.html + +### MySQL Docs + +https://dev.mysql.com/doc/refman/8.0/en/window-function-descriptions.html + +https://dev.mysql.com/doc/refman/8.0/en/window-functions-usage.html + +https://dev.mysql.com/doc/refman/8.0/en/window-functions-frames.html + + ## Related Content - Blog: [Working with time series data in ClickHouse](https://clickhouse.com/blog/working-with-time-series-data-and-functions-ClickHouse) diff --git a/docs/ru/operations/clickhouse-keeper.md b/docs/ru/operations/clickhouse-keeper.md index 9f1301d817d..e1d21dd537c 100644 --- a/docs/ru/operations/clickhouse-keeper.md +++ b/docs/ru/operations/clickhouse-keeper.md @@ -38,6 +38,7 @@ ClickHouse Keeper может иÑпользоватьÑÑ ÐºÐ°Ðº равноце - `dead_session_check_period_ms` — чаÑтота, Ñ ÐºÐ¾Ñ‚Ð¾Ñ€Ð¾Ð¹ ClickHouse Keeper проверÑет мертвые ÑеÑÑии и удалÑет их, в миллиÑекундах (по умолчанию: 500). - `election_timeout_lower_bound_ms` — времÑ, поÑле которого поÑледователь может инициировать перевыбор лидера, еÑли не получил от него контрольный Ñигнал (по умолчанию: 1000). - `election_timeout_upper_bound_ms` — времÑ, поÑле которого поÑледователь должен инициировать перевыбор лидера, еÑли не получил от него контрольный Ñигнал (по умолчанию: 2000). +- `leadership_expiry_ms` — ЕÑли лидер не получает ответа от доÑтаточного количеÑтва поÑледователей в течение Ñтого промежутка времени, он добровольно отказываетÑÑ Ð¾Ñ‚ Ñвоего руководÑтва. При наÑтройке 0 автоматичеÑки уÑтанавливаетÑÑ 20 - кратное значение `heart_beat_interval_ms`, а при наÑтройке меньше 0 лидер не отказываетÑÑ Ð¾Ñ‚ лидерÑтва (по умолчанию 0). - `force_sync` — вызывать `fsync` при каждой запиÑи в журнал координации (по умолчанию: true). - `four_letter_word_white_list` — ÑпиÑок разрешенных 4-Ñ… буквенных команд (по умолчанию: "conf,cons,crst,envi,ruok,srst,srvr,stat,wchc,wchs,dirs,mntr,isro"). - `fresh_log_gap` — минимальное отÑтавание от лидера в количеÑтве запиÑей журнала поÑле которого поÑледователь Ñчитает ÑÐµÐ±Ñ Ð°ÐºÑ‚ÑƒÐ°Ð»ÑŒÐ½Ñ‹Ð¼ (по умолчанию: 200). @@ -209,6 +210,7 @@ dead_session_check_period_ms=500 heart_beat_interval_ms=500 election_timeout_lower_bound_ms=1000 election_timeout_upper_bound_ms=2000 +leadership_expiry_ms=0 reserved_log_items=1000000000000000 snapshot_distance=10000 auto_forwarding=true diff --git a/docs/zh/operations/clickhouse-keeper.md b/docs/zh/operations/clickhouse-keeper.md index 6d8a570aa12..e4412be2e30 100644 --- a/docs/zh/operations/clickhouse-keeper.md +++ b/docs/zh/operations/clickhouse-keeper.md @@ -45,6 +45,7 @@ ClickHouse Keeper 完全å¯ä»¥ä½œä¸ºZooKeeper的独立替代å“或者作为Click - `heart_beat_interval_ms` — ClickHouse Keeperçš„leaderå‘é€å¿ƒè·³é¢‘率(毫秒)(默认为500)。 - `election_timeout_lower_bound_ms` — 如果follower在此间隔内没有收到leader的心跳，那么它å¯ä»¥å¯åŠ¨leader选举(默认为1000). - `election_timeout_upper_bound_ms` — 如果follower在此间隔内没有收到leader的心跳，那么它必须å¯åŠ¨leader选举(默认为2000)。 +- `leadership_expiry_ms` — 如果leader在此间隔内没有收到足够的follower回å¤ï¼Œé‚£ä¹ˆä»–会主动放弃领导æƒã€‚当被设置为0时会自动设置为`heart_beat_interval_ms`çš„20å€ï¼Œå½“被设置å°äºŽ0æ—¶leaderä¸ä¼šä¸»åŠ¨æ”¾å¼ƒé¢†å¯¼æƒï¼ˆé»˜è®¤ä¸º0）。 - `rotate_log_storage_interval` — å•ä¸ªæ–‡ä»¶ä¸­å­˜å‚¨çš„日志记录数é‡(默认100000æ¡)。 - `reserved_log_items` — 在压缩之å‰éœ€è¦å­˜å‚¨å¤šå°‘å调日志记录(默认100000)。 - `snapshot_distance` — ClickHouse Keeper创建新快照的频率(以日志记录的数é‡ä¸ºå•ä½)(默认100000)。 @@ -214,6 +215,7 @@ dead_session_check_period_ms=500 heart_beat_interval_ms=500 election_timeout_lower_bound_ms=1000 election_timeout_upper_bound_ms=2000 +leadership_expiry_ms=0 reserved_log_items=1000000000000000 snapshot_distance=10000 auto_forwarding=true diff --git a/docs/zh/operations/settings/settings.md b/docs/zh/operations/settings/settings.md index 1874970ac95..c3b4194ed44 100644 --- a/docs/zh/operations/settings/settings.md +++ b/docs/zh/operations/settings/settings.md @@ -649,11 +649,22 @@ log_query_threads=1 ## max_query_size {#settings-max_query_size} -查询的最大部分，å¯ä»¥è¢«å¸¦åˆ°RAM用于使用SQL解æžå™¨è¿›è¡Œè§£æžã€‚ -æ’入查询还包å«ç”±å•ç‹¬çš„æµè§£æžå™¨ï¼ˆæ¶ˆè€—O(1)RAM）处ç†çš„æ’入数æ®ï¼Œè¿™äº›æ•°æ®ä¸åŒ…å«åœ¨æ­¤é™åˆ¶ä¸­ã€‚ +SQL 解æžå™¨è§£æžçš„查询字符串的最大字节数。 INSERT 查询的 VALUES å­å¥ä¸­çš„æ•°æ®ç”±å•ç‹¬çš„æµè§£æžå™¨ï¼ˆæ¶ˆè€— O(1) RAM）处ç†ï¼Œå¹¶ä¸”ä¸å—æ­¤é™åˆ¶çš„å½±å“。 默认值：256KiB。 + +## max_parser_depth {#max_parser_depth} + +é™åˆ¶é€’归下é™è§£æžå™¨ä¸­çš„最大递归深度。å…许控制堆栈大å°ã€‚ + +å¯èƒ½çš„值： + +- 正整数。 +- 0 — 递归深度ä¸å—é™åˆ¶ã€‚ + +默认值：1000。 + ## interactive_delay {#interactive-delay} 以微秒为å•ä½çš„间隔，用于检查请求执行是å¦å·²è¢«å–消并å‘é€è¿›åº¦ã€‚ @@ -1064,6 +1075,28 @@ ClickHouse生æˆå¼‚常 默认值：0。 +## optimize_functions_to_subcolumns {#optimize_functions_to_subcolumns} + +å¯ç”¨æˆ–ç¦ç”¨é€šè¿‡å°†æŸäº›å‡½æ•°è½¬æ¢ä¸ºè¯»å–å­åˆ—的优化。这å‡å°‘了è¦è¯»å–çš„æ•°æ®é‡ã€‚ + +这些函数å¯ä»¥è½¬åŒ–为： + +- [length](../../sql-reference/functions/array-functions.md/#array_functions-length) è¯»å– [size0](../../sql-reference/data-types/array.md/#array-size）å­åˆ—。 +- [empty](../../sql-reference/functions/array-functions.md/#empty函数) è¯»å– [size0](../../sql-reference/data-types/array.md/#array-size）å­åˆ—。 +- [notEmpty](../../sql-reference/functions/array-functions.md/#notempty函数) è¯»å– [size0](../../sql-reference/data-types/array.md/#array-size）å­åˆ—。 +- [isNull](../../sql-reference/operators/index.md#operator-is-null) è¯»å– [null](../../sql-reference/data-types/nullable. md/#finding-null) å­åˆ—。 +- [isNotNull](../../sql-reference/operators/index.md#is-not-null) è¯»å– [null](../../sql-reference/data-types/nullable. md/#finding-null) å­åˆ—。 +- [count](../../sql-reference/aggregate-functions/reference/count.md) è¯»å– [null](../../sql-reference/data-types/nullable.md/#finding-null) å­åˆ—。 +- [mapKeys](../../sql-reference/functions/tuple-map-functions.mdx/#mapkeys) è¯»å– [keys](../../sql-reference/data-types/map.md/#map-subcolumns) å­åˆ—。 +- [mapValues](../../sql-reference/functions/tuple-map-functions.mdx/#mapvalues) è¯»å– [values](../../sql-reference/data-types/map.md/#map-subcolumns) å­åˆ—。 + +å¯èƒ½çš„值： + +- 0 — ç¦ç”¨ä¼˜åŒ–。 +- 1 — 优化已å¯ç”¨ã€‚ + +默认值：`0`。 + ## distributed_replica_error_half_life {#settings-distributed_replica_error_half_life} - 类型：秒 diff --git a/docs/zh/sql-reference/data-types/array.md b/docs/zh/sql-reference/data-types/array.md index 46c40b889ad..da4cea65101 100644 --- a/docs/zh/sql-reference/data-types/array.md +++ b/docs/zh/sql-reference/data-types/array.md @@ -1,7 +1,7 @@ --- slug: /zh/sql-reference/data-types/array --- -# 阵列(T) {#data-type-array} +# 数组(T) {#data-type-array} ç”± `T` 类型元素组æˆçš„数组。 @@ -66,3 +66,27 @@ SELECT array(1, 'a') Received exception from server (version 1.1.54388): Code: 386. DB::Exception: Received from localhost:9000, 127.0.0.1. DB::Exception: There is no supertype for types UInt8, String because some of them are String/FixedString and some of them are not. ``` + +## æ•°ç»„å¤§å° {#array-size} + +å¯ä»¥ä½¿ç”¨ `size0` å­åˆ—找到数组的大å°ï¼Œè€Œæ— éœ€è¯»å–整个列。对于多维数组，您å¯ä»¥ä½¿ç”¨ `sizeN-1`，其中 `N` 是所需的维度。 + +**例å­** + +SQL查询： + +```sql +CREATE TABLE t_arr (`arr` Array(Array(Array(UInt32)))) ENGINE = MergeTree ORDER BY tuple(); + +INSERT INTO t_arr VALUES ([[[12, 13, 0, 1],[12]]]); + +SELECT arr.size0, arr.size1, arr.size2 FROM t_arr; +``` + +结果： + +``` text +┌─arr.size0─┬─arr.size1─┬─arr.size2─┠+│ 1 │ [2] │ [[4,1]] │ +└───────────┴───────────┴───────────┘ +``` diff --git a/docs/zh/sql-reference/data-types/nullable.md b/docs/zh/sql-reference/data-types/nullable.md index 94311f8298a..b1cc9dd7bae 100644 --- a/docs/zh/sql-reference/data-types/nullable.md +++ b/docs/zh/sql-reference/data-types/nullable.md @@ -20,6 +20,33 @@ slug: /zh/sql-reference/data-types/nullable 掩ç æ–‡ä»¶ä¸­çš„æ¡ç›®å…许ClickHouse区分æ¯ä¸ªè¡¨è¡Œçš„对应数æ®ç±»åž‹çš„«NULL»和默认值由于有é¢å¤–的文件，«Nullable»列比普通列消耗更多的存储空间 +## nullå­åˆ— {#finding-null} + +通过使用 `null` å­åˆ—å¯ä»¥åœ¨åˆ—中查找 `NULL` 值，而无需读å–整个列。如果对应的值为 `NULL`，则返回 `1`，å¦åˆ™è¿”回 `0`。 + +**示例** + +SQL查询: + +``` sql +CREATE TABLE nullable (`n` Nullable(UInt32)) ENGINE = MergeTree ORDER BY tuple(); + +INSERT INTO nullable VALUES (1) (NULL) (2) (NULL); + +SELECT n.null FROM nullable; +``` + +结果: + +``` text +┌─n.null─┠+│ 0 │ +│ 1 │ +│ 0 │ +│ 1 │ +└────────┘ +``` + ## 用法示例 {#yong-fa-shi-li} ``` sql diff --git a/programs/copier/Internals.cpp b/programs/copier/Internals.cpp index 0cfff7e3f6c..dcd199c6b38 100644 --- a/programs/copier/Internals.cpp +++ b/programs/copier/Internals.cpp @@ -259,7 +259,7 @@ ShardPriority getReplicasPriority(const Cluster::Addresses & replicas, const std res.is_remote = 1; for (const auto & replica : replicas) { - if (isLocalAddress(DNSResolver::instance().resolveHost(replica.host_name))) + if (isLocalAddress(DNSResolver::instance().resolveHostAllInOriginOrder(replica.host_name).front())) { res.is_remote = 0; break; diff --git a/programs/copier/Internals.h b/programs/copier/Internals.h index 48f4b0fab09..27fedd5d9e8 100644 --- a/programs/copier/Internals.h +++ b/programs/copier/Internals.h @@ -102,7 +102,7 @@ struct TaskStateWithOwner return TaskStateWithOwner(state, owner).toString(); } - String toString() + String toString() const { WriteBufferFromOwnString wb; wb << static_cast(state) << "\n" << escape << owner; diff --git a/programs/copier/ZooKeeperStaff.h b/programs/copier/ZooKeeperStaff.h index bbdec230d2d..c15db73f060 100644 --- a/programs/copier/ZooKeeperStaff.h +++ b/programs/copier/ZooKeeperStaff.h @@ -180,7 +180,7 @@ public: auto logger = getLogger("ClusterCopier"); if (rsp.error == Coordination::Error::ZOK) { - switch (rsp.type) + switch (rsp.type) /// NOLINT(bugprone-switch-missing-default-case) { case Coordination::CREATED: LOG_DEBUG(logger, "CleanStateClock change: CREATED, at {}", rsp.path); diff --git a/programs/library-bridge/LibraryBridgeHandlers.cpp b/programs/library-bridge/LibraryBridgeHandlers.cpp index ab146f458df..26d887cfc98 100644 --- a/programs/library-bridge/LibraryBridgeHandlers.cpp +++ b/programs/library-bridge/LibraryBridgeHandlers.cpp @@ -1,6 +1,5 @@ #include "LibraryBridgeHandlers.h" -#include "CatBoostLibraryHandler.h" #include "CatBoostLibraryHandlerFactory.h" #include "Common/ProfileEvents.h" #include "ExternalDictionaryLibraryHandler.h" @@ -11,10 +10,8 @@ #include #include #include -#include #include #include -#include #include #include #include diff --git a/programs/local/LocalServer.cpp b/programs/local/LocalServer.cpp index 867e7a69a6d..68f0e52ce08 100644 --- a/programs/local/LocalServer.cpp +++ b/programs/local/LocalServer.cpp @@ -841,7 +841,7 @@ void LocalServer::addOptions(OptionsDescription & options_description) /// If structure argument is omitted then initial query is not generated ("structure,S", po::value(), "structure of the initial table (list of column and type names)") - ("file,f", po::value(), "path to file with data of the initial table (stdin if not specified)") + ("file,F", po::value(), "path to file with data of the initial table (stdin if not specified)") ("input-format", po::value(), "input format of the initial table data") ("output-format", po::value(), "default output format") diff --git a/programs/odbc-bridge/ColumnInfoHandler.cpp b/programs/odbc-bridge/ColumnInfoHandler.cpp index 774883657b7..4cb15de3b2c 100644 --- a/programs/odbc-bridge/ColumnInfoHandler.cpp +++ b/programs/odbc-bridge/ColumnInfoHandler.cpp @@ -8,13 +8,11 @@ #include #include #include -#include #include #include #include #include #include -#include #include #include #include "getIdentifierQuote.h" diff --git a/programs/odbc-bridge/IdentifierQuoteHandler.cpp b/programs/odbc-bridge/IdentifierQuoteHandler.cpp index a23efb112de..cf5acdc4534 100644 --- a/programs/odbc-bridge/IdentifierQuoteHandler.cpp +++ b/programs/odbc-bridge/IdentifierQuoteHandler.cpp @@ -7,13 +7,10 @@ #include #include #include -#include -#include #include #include #include #include -#include #include "getIdentifierQuote.h" #include "validateODBCConnectionString.h" #include "ODBCPooledConnectionFactory.h" diff --git a/programs/odbc-bridge/ODBCBlockInputStream.cpp b/programs/odbc-bridge/ODBCBlockInputStream.cpp index c46144c3dc8..59a5deac960 100644 --- a/programs/odbc-bridge/ODBCBlockInputStream.cpp +++ b/programs/odbc-bridge/ODBCBlockInputStream.cpp @@ -3,12 +3,8 @@ #include #include #include -#include -#include -#include #include #include -#include #include @@ -54,21 +50,7 @@ Chunk ODBCSource::generate() const auto & sample = description.sample_block.getByPosition(idx); if (!result.is_null(idx)) - { - bool is_nullable = description.types[idx].second; - - if (is_nullable) - { - ColumnNullable & column_nullable = assert_cast(*columns[idx]); - const auto & data_type = assert_cast(*sample.type); - insertValue(column_nullable.getNestedColumn(), data_type.getNestedType(), description.types[idx].first, result, idx); - column_nullable.getNullMapData().emplace_back(0); - } - else - { - insertValue(*columns[idx], sample.type, description.types[idx].first, result, idx); - } - } + insertValue(*columns[idx], removeNullable(sample.type), description.types[idx].first, result, idx); else insertDefaultValue(*columns[idx], *sample.column); } @@ -87,59 +69,60 @@ void ODBCSource::insertValue( switch (type) { case ValueType::vtUInt8: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtUInt16: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtUInt32: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtUInt64: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtInt8: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtInt16: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtInt32: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtInt64: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtFloat32: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtFloat64: - assert_cast(column).insertValue(row.get(idx)); + column.insert(row.get(idx)); break; - case ValueType::vtFixedString:[[fallthrough]]; + case ValueType::vtFixedString: case ValueType::vtEnum8: case ValueType::vtEnum16: case ValueType::vtString: - assert_cast(column).insert(row.get(idx)); + column.insert(row.get(idx)); break; case ValueType::vtUUID: { auto value = row.get(idx); - assert_cast(column).insert(parse(value.data(), value.size())); + column.insert(parse(value.data(), value.size())); break; } case ValueType::vtDate: - assert_cast(column).insertValue(UInt16{LocalDate{row.get(idx)}.getDayNum()}); + column.insert(UInt16{LocalDate{row.get(idx)}.getDayNum()}); break; case ValueType::vtDateTime: { auto value = row.get(idx); ReadBufferFromString in(value); time_t time = 0; - readDateTimeText(time, in, assert_cast(data_type.get())->getTimeZone()); + const DataTypeDateTime & datetime_type = assert_cast(*data_type); + readDateTimeText(time, in, datetime_type.getTimeZone()); if (time < 0) time = 0; - assert_cast(column).insertValue(static_cast(time)); + column.insert(static_cast(time)); break; } case ValueType::vtDateTime64: @@ -147,14 +130,14 @@ void ODBCSource::insertValue( auto value = row.get(idx); ReadBufferFromString in(value); DateTime64 time = 0; - const auto * datetime_type = assert_cast(data_type.get()); - readDateTime64Text(time, datetime_type->getScale(), in, datetime_type->getTimeZone()); - assert_cast(column).insertValue(time); + const DataTypeDateTime64 & datetime_type = assert_cast(*data_type); + readDateTime64Text(time, datetime_type.getScale(), in, datetime_type.getTimeZone()); + column.insert(time); break; } - case ValueType::vtDecimal32: [[fallthrough]]; - case ValueType::vtDecimal64: [[fallthrough]]; - case ValueType::vtDecimal128: [[fallthrough]]; + case ValueType::vtDecimal32: + case ValueType::vtDecimal64: + case ValueType::vtDecimal128: case ValueType::vtDecimal256: { auto value = row.get(idx); diff --git a/programs/odbc-bridge/ODBCBlockOutputStream.cpp b/programs/odbc-bridge/ODBCBlockOutputStream.cpp index 87c09d1e757..37b70023169 100644 --- a/programs/odbc-bridge/ODBCBlockOutputStream.cpp +++ b/programs/odbc-bridge/ODBCBlockOutputStream.cpp @@ -1,6 +1,5 @@ #include "ODBCBlockOutputStream.h" -#include #include #include #include diff --git a/programs/odbc-bridge/ODBCPooledConnectionFactory.h b/programs/odbc-bridge/ODBCPooledConnectionFactory.h index b70e45f2b9d..c4e3d4c12c6 100644 --- a/programs/odbc-bridge/ODBCPooledConnectionFactory.h +++ b/programs/odbc-bridge/ODBCPooledConnectionFactory.h @@ -40,7 +40,6 @@ public: explicit ConnectionHolder(const String & connection_string_) : pool(nullptr) - , connection() , connection_string(connection_string_) { updateConnection(); @@ -143,7 +142,7 @@ public: { std::lock_guard lock(mutex); - if (!factory.count(connection_string)) + if (!factory.contains(connection_string)) factory.emplace(std::make_pair(connection_string, std::make_shared(pool_size))); auto & pool = factory[connection_string]; diff --git a/programs/server/.gitignore b/programs/server/.gitignore index ddc480e4b29..34a774bde9d 100644 --- a/programs/server/.gitignore +++ b/programs/server/.gitignore @@ -2,6 +2,7 @@ /metadata_dropped /data /store +/disks /access /flags /dictionaries_lib diff --git a/programs/server/Server.cpp b/programs/server/Server.cpp index c45291ba52c..7636f221ab5 100644 --- a/programs/server/Server.cpp +++ b/programs/server/Server.cpp @@ -46,6 +46,7 @@ #include #include #include +#include #include #include #include @@ -184,7 +185,7 @@ static bool jemallocOptionEnabled(const char *name) return value; } #else -static bool jemallocOptionEnabled(const char *) { return 0; } +static bool jemallocOptionEnabled(const char *) { return false; } #endif int mainEntryClickHouseServer(int argc, char ** argv) @@ -1547,6 +1548,23 @@ try FileCacheFactory::instance().updateSettingsFromConfig(*config); + HTTPConnectionPools::instance().setLimits( + HTTPConnectionPools::Limits{ + new_server_settings.disk_connections_soft_limit, + new_server_settings.disk_connections_warn_limit, + new_server_settings.disk_connections_store_limit, + }, + HTTPConnectionPools::Limits{ + new_server_settings.storage_connections_soft_limit, + new_server_settings.storage_connections_warn_limit, + new_server_settings.storage_connections_store_limit, + }, + HTTPConnectionPools::Limits{ + new_server_settings.http_connections_soft_limit, + new_server_settings.http_connections_warn_limit, + new_server_settings.http_connections_store_limit, + }); + ProfileEvents::increment(ProfileEvents::MainConfigLoads); /// Must be the last. diff --git a/programs/static-files-disk-uploader/static-files-disk-uploader.cpp b/programs/static-files-disk-uploader/static-files-disk-uploader.cpp index d54a2d2da6e..dfe68c819fc 100644 --- a/programs/static-files-disk-uploader/static-files-disk-uploader.cpp +++ b/programs/static-files-disk-uploader/static-files-disk-uploader.cpp @@ -65,7 +65,7 @@ void processFile(const fs::path & file_path, const fs::path & dst_path, bool tes /// test mode for integration tests. if (test_mode) - dst_buf = std::make_shared(Poco::URI(dst_file_path), Poco::Net::HTTPRequest::HTTP_PUT); + dst_buf = std::make_shared(HTTPConnectionGroupType::HTTP, Poco::URI(dst_file_path), Poco::Net::HTTPRequest::HTTP_PUT); else dst_buf = std::make_shared(dst_file_path); @@ -88,7 +88,7 @@ void processTableFiles(const fs::path & data_path, fs::path dst_path, bool test_ { dst_path /= "store"; auto files_root = dst_path / prefix; - root_meta = std::make_shared(Poco::URI(files_root / ".index"), Poco::Net::HTTPRequest::HTTP_PUT); + root_meta = std::make_shared(HTTPConnectionGroupType::HTTP, Poco::URI(files_root / ".index"), Poco::Net::HTTPRequest::HTTP_PUT); } else { @@ -112,7 +112,7 @@ void processTableFiles(const fs::path & data_path, fs::path dst_path, bool test_ if (test_mode) { auto files_root = dst_path / prefix; - directory_meta = std::make_shared(Poco::URI(dst_path / directory_prefix / ".index"), Poco::Net::HTTPRequest::HTTP_PUT); + directory_meta = std::make_shared(HTTPConnectionGroupType::HTTP, Poco::URI(dst_path / directory_prefix / ".index"), Poco::Net::HTTPRequest::HTTP_PUT); } else { diff --git a/.pylintrc b/pyproject.toml similarity index 68% rename from .pylintrc rename to pyproject.toml index b672cbfdfad..3d05abd9ec2 100644 --- a/.pylintrc +++ b/pyproject.toml @@ -1,6 +1,4 @@ -# vim: ft=config - -[BASIC] +[tool.pylint.BASIC] max-module-lines=2000 # due to SQL max-line-length=200 @@ -9,11 +7,13 @@ max-branches=50 max-nested-blocks=10 max-statements=200 -[FORMAT] -ignore-long-lines = (# )??$ +[tool.pylint.FORMAT] +#ignore-long-lines = (# )??$ -[MESSAGES CONTROL] -disable = missing-docstring, +[tool.pylint.'MESSAGES CONTROL'] + # pytest.mark.parametrize is not callable (not-callable) +disable = ''' + missing-docstring, too-few-public-methods, invalid-name, too-many-arguments, @@ -26,18 +26,15 @@ disable = missing-docstring, wildcard-import, unused-wildcard-import, singleton-comparison, - # pytest.mark.parametrize is not callable (not-callable) not-callable, - # https://github.com/PyCQA/pylint/issues/3882 - # [Python 3.9] Value 'Optional' is unsubscriptable (unsubscriptable-object) (also Union) - unsubscriptable-object, - # Drop them one day: redefined-outer-name, broad-except, bare-except, no-else-return, global-statement + ''' -[SIMILARITIES] +[tool.pylint.SIMILARITIES] # due to SQL min-similarity-lines=1000 + diff --git a/rust/Cargo.lock b/rust/Cargo.lock index dbbe184228c..a242a8243b5 100644 --- a/rust/Cargo.lock +++ b/rust/Cargo.lock @@ -6,6 +6,7 @@ version = 3 name = "_ch_rust_prql" version = "0.1.0" dependencies = [ + "anstream", "prqlc", "serde_json", ] @@ -698,9 +699,9 @@ checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a" [[package]] name = "miniz_oxide" -version = "0.7.1" +version = "0.7.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "e7810e0be55b428ada41041c41f32c9f1a42817901b4ccf45fa3d4b6561e74c7" +checksum = "9d811f3e15f28568be3407c8e7fdb6514c1cda3cb30683f15b6a1a1dc4ea14a7" dependencies = [ "adler", ] @@ -751,9 +752,9 @@ dependencies = [ [[package]] name = "object" -version = "0.32.1" +version = "0.32.2" source = "registry+https://github.com/rust-lang/crates.io-index" -checksum = "9cf5f9dd3933bd50a9e1f149ec995f39ae2c496d31fd772c1fd45ebc27e902b0" +checksum = "a6a622008b6e321afc04970976f62ee297fdbaa6f95318ca343e3eebb9648441" dependencies = [ "memchr", ] diff --git a/rust/prql/Cargo.toml b/rust/prql/Cargo.toml index e70a49658e3..514b5176d48 100644 --- a/rust/prql/Cargo.toml +++ b/rust/prql/Cargo.toml @@ -4,6 +4,7 @@ name = "_ch_rust_prql" version = "0.1.0" [dependencies] +anstream = {version = "0.6.12"} prqlc = {version = "0.11.3", default-features = false} serde_json = "1.0" diff --git a/rust/prql/src/lib.rs b/rust/prql/src/lib.rs index 9e4b0ae194a..2e5b2061fcb 100644 --- a/rust/prql/src/lib.rs +++ b/rust/prql/src/lib.rs @@ -39,6 +39,11 @@ pub unsafe extern "C" fn prql_to_sql_impl( }; if let Ok(sql_str) = prqlc::compile(&query_str, &opts) { + // NOTE: Over at PRQL we're considering to un-deprecate & re-enable the + // `color: false` option. If that happens, we can remove the `strip_str` + // here, which strips color codes from the output. + use anstream::adapter::strip_str; + let sql_str = strip_str(&sql_str).to_string(); set_output(sql_str, out, out_size); 0 } else { @@ -54,17 +59,50 @@ pub unsafe extern "C" fn prql_to_sql( out: *mut *mut u8, out_size: *mut u64, ) -> i64 { - let ret = panic::catch_unwind(|| { - return prql_to_sql_impl(query, size, out, out_size); - }); - return match ret { - // NOTE: using cxxbridge we can return proper Result<> type. - Err(_err) => 1, - Ok(res) => res, - } + // NOTE: using cxxbridge we can return proper Result<> type. + panic::catch_unwind(|| prql_to_sql_impl(query, size, out, out_size)).unwrap_or_else(|_| { + set_output("prqlc panicked".to_string(), out, out_size); + 1 + }) } #[no_mangle] pub unsafe extern "C" fn prql_free_pointer(ptr_to_free: *mut u8) { std::mem::drop(CString::from_raw(ptr_to_free as *mut c_char)); } + +#[cfg(test)] +mod tests { + use super::*; + use std::ffi::{CStr, CString}; + + /// A test helper to offer a rust interface to the C bindings + fn run_compile(query: &str) -> (String, i64) { + let query_cstr = CString::new(query).unwrap(); + let query_ptr = query_cstr.as_ptr() as *const u8; + let query_size = query_cstr.to_bytes_with_nul().len() as u64 - 1; // Excluding the null terminator + + let mut out: *mut u8 = std::ptr::null_mut(); + let mut out_size = 0_u64; + + unsafe { + let success = prql_to_sql(query_ptr, query_size, &mut out, &mut out_size); + let output = CStr::from_ptr(out as *const i8) + .to_str() + .unwrap() + .to_string(); + prql_free_pointer(out); + (output, success) + } + } + + #[test] + fn test_prql_to_sql() { + assert!(run_compile("from x").0.contains("SELECT")); + assert!(run_compile("asdf").1 == 1); + // In prqlc 0.11.3, this is a panic, so that allows us to test that the + // panic is caught. When we upgrade prqlc, it won't be a panic any + // longer. + assert!(run_compile("x -> y").1 == 1); + } +} diff --git a/src/Access/Common/AccessType.h b/src/Access/Common/AccessType.h index 87f96ca48be..6095f8ce6d3 100644 --- a/src/Access/Common/AccessType.h +++ b/src/Access/Common/AccessType.h @@ -154,7 +154,8 @@ enum class AccessType M(SET_DEFINER, "", USER_NAME, ALL) \ \ M(SYSTEM_SHUTDOWN, "SYSTEM KILL, SHUTDOWN", GLOBAL, SYSTEM) \ - M(SYSTEM_DROP_DNS_CACHE, "SYSTEM DROP DNS, DROP DNS CACHE, DROP DNS", GLOBAL, SYSTEM_DROP_CACHE) \ + M(SYSTEM_DROP_DNS_CACHE, "SYSTEM DROP DNS, DROP DNS CACHE, DROP DNS", GLOBAL, SYSTEM_DROP_CACHE) \ + M(SYSTEM_DROP_CONNECTIONS_CACHE, "SYSTEM DROP CONNECTIONS CACHE, DROP CONNECTIONS CACHE", GLOBAL, SYSTEM_DROP_CACHE) \ M(SYSTEM_DROP_MARK_CACHE, "SYSTEM DROP MARK, DROP MARK CACHE, DROP MARKS", GLOBAL, SYSTEM_DROP_CACHE) \ M(SYSTEM_DROP_UNCOMPRESSED_CACHE, "SYSTEM DROP UNCOMPRESSED, DROP UNCOMPRESSED CACHE, DROP UNCOMPRESSED", GLOBAL, SYSTEM_DROP_CACHE) \ M(SYSTEM_DROP_MMAP_CACHE, "SYSTEM DROP MMAP, DROP MMAP CACHE, DROP MMAP", GLOBAL, SYSTEM_DROP_CACHE) \ diff --git a/src/Access/Common/AllowedClientHosts.cpp b/src/Access/Common/AllowedClientHosts.cpp index c677465a7a1..bee0cdd7264 100644 --- a/src/Access/Common/AllowedClientHosts.cpp +++ b/src/Access/Common/AllowedClientHosts.cpp @@ -55,7 +55,7 @@ namespace { IPAddress addr_v6 = toIPv6(address); - auto host_addresses = DNSResolver::instance().resolveHostAll(host); + auto host_addresses = DNSResolver::instance().resolveHostAllInOriginOrder(host); for (const auto & addr : host_addresses) { diff --git a/src/Access/Common/QuotaDefs.cpp b/src/Access/Common/QuotaDefs.cpp index 0e9a4d5a365..04c16a562d2 100644 --- a/src/Access/Common/QuotaDefs.cpp +++ b/src/Access/Common/QuotaDefs.cpp @@ -49,71 +49,135 @@ String QuotaTypeInfo::valueToStringWithName(QuotaValue value) const const QuotaTypeInfo & QuotaTypeInfo::get(QuotaType type) { - static constexpr auto make_info = [](const char * raw_name_, UInt64 output_denominator_) + static constexpr auto make_info = [](const char * raw_name_, String current_usage_description_, String max_allowed_usage_description_, UInt64 output_denominator_) { String init_name = raw_name_; boost::to_lower(init_name); String init_keyword = raw_name_; boost::replace_all(init_keyword, "_", " "); bool init_output_as_float = (output_denominator_ != 1); - return QuotaTypeInfo{raw_name_, std::move(init_name), std::move(init_keyword), init_output_as_float, output_denominator_}; + return QuotaTypeInfo + { + .raw_name = raw_name_, + .name = std::move(init_name), + .keyword = std::move(init_keyword), + .current_usage_description = std::move(current_usage_description_), + .max_allowed_usage_description = std::move(max_allowed_usage_description_), + .output_as_float = init_output_as_float, + .output_denominator = output_denominator_ + }; }; switch (type) { case QuotaType::QUERIES: { - static const auto info = make_info("QUERIES", 1); + static const auto info = make_info( + "QUERIES", + "The current number of executed queries.", + "The maximum allowed number of queries of all types allowed to be executed.", + 1 + ); return info; } case QuotaType::QUERY_SELECTS: { - static const auto info = make_info("QUERY_SELECTS", 1); + static const auto info = make_info( + "QUERY_SELECTS", + "The current number of executed SELECT queries.", + "The maximum allowed number of SELECT queries allowed to be executed.", + 1 + ); return info; } case QuotaType::QUERY_INSERTS: { - static const auto info = make_info("QUERY_INSERTS", 1); + static const auto info = make_info( + "QUERY_INSERTS", + "The current number of executed INSERT queries.", + "The maximum allowed number of INSERT queries allowed to be executed.", + 1 + ); return info; } case QuotaType::ERRORS: { - static const auto info = make_info("ERRORS", 1); + static const auto info = make_info( + "ERRORS", + "The current number of queries resulted in an error.", + "The maximum number of queries resulted in an error allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::RESULT_ROWS: { - static const auto info = make_info("RESULT_ROWS", 1); + static const auto info = make_info( + "RESULT_ROWS", + "The current total number of rows in the result set of all queries within the current period of time.", + "The maximum total number of rows in the result set of all queries allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::RESULT_BYTES: { - static const auto info = make_info("RESULT_BYTES", 1); + static const auto info = make_info( + "RESULT_BYTES", + "The current total number of bytes in the result set of all queries within the current period of time.", + "The maximum total number of bytes in the result set of all queries allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::READ_ROWS: { - static const auto info = make_info("READ_ROWS", 1); + static const auto info = make_info( + "READ_ROWS", + "The current total number of rows read during execution of all queries within the current period of time.", + "The maximum number of rows to read during execution of all queries allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::READ_BYTES: { - static const auto info = make_info("READ_BYTES", 1); + static const auto info = make_info( + "READ_BYTES", + "The current total number of bytes read during execution of all queries within the current period of time.", + "The maximum number of bytes to read during execution of all queries allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::EXECUTION_TIME: { - static const auto info = make_info("EXECUTION_TIME", 1000000000 /* execution_time is stored in nanoseconds */); + static const auto info = make_info( + "EXECUTION_TIME", + "The current total amount of time (in nanoseconds) spent to execute queries within the current period of time", + "The maximum amount of time (in nanoseconds) allowed for all queries to execute within the specified period of time", + 1000000000 /* execution_time is stored in nanoseconds */ + ); return info; } case QuotaType::WRITTEN_BYTES: { - static const auto info = make_info("WRITTEN_BYTES", 1); + static const auto info = make_info( + "WRITTEN_BYTES", + "The current total number of bytes written during execution of all queries within the current period of time.", + "The maximum number of bytes to written during execution of all queries allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::FAILED_SEQUENTIAL_AUTHENTICATIONS: { - static const auto info = make_info("FAILED_SEQUENTIAL_AUTHENTICATIONS", 1); + static const auto info = make_info( + "FAILED_SEQUENTIAL_AUtheNTICATIONS", + "The current number of consecutive authentication failures within the current period of time.", + "The maximum number of consecutive authentication failures allowed within the specified period of time.", + 1 + ); return info; } case QuotaType::MAX: break; diff --git a/src/Access/Common/QuotaDefs.h b/src/Access/Common/QuotaDefs.h index 4f849a72b43..6618f01c8f9 100644 --- a/src/Access/Common/QuotaDefs.h +++ b/src/Access/Common/QuotaDefs.h @@ -33,6 +33,8 @@ struct QuotaTypeInfo const char * const raw_name = ""; const String name; /// Lowercased with underscores, e.g. "result_rows". const String keyword; /// Uppercased with spaces, e.g. "RESULT ROWS". + const String current_usage_description; + const String max_allowed_usage_description; const bool output_as_float = false; const UInt64 output_denominator = 1; String valueToString(QuotaValue value) const; diff --git a/src/Access/Common/RowPolicyDefs.cpp b/src/Access/Common/RowPolicyDefs.cpp index b1f882fe971..a9509b6dd76 100644 --- a/src/Access/Common/RowPolicyDefs.cpp +++ b/src/Access/Common/RowPolicyDefs.cpp @@ -33,7 +33,7 @@ String toString(RowPolicyFilterType type) const RowPolicyFilterTypeInfo & RowPolicyFilterTypeInfo::get(RowPolicyFilterType type_) { - static constexpr auto make_info = [](const char * raw_name_) + static constexpr auto make_info = [](const char * raw_name_, const String & comment_) { String init_name = raw_name_; boost::to_lower(init_name); @@ -41,14 +41,17 @@ const RowPolicyFilterTypeInfo & RowPolicyFilterTypeInfo::get(RowPolicyFilterType String init_command = init_name.substr(0, underscore_pos); boost::to_upper(init_command); bool init_is_check = (std::string_view{init_name}.substr(underscore_pos + 1) == "check"); - return RowPolicyFilterTypeInfo{raw_name_, std::move(init_name), std::move(init_command), init_is_check}; + return RowPolicyFilterTypeInfo{raw_name_, std::move(init_name), std::move(init_command), comment_, init_is_check}; }; switch (type_) { case RowPolicyFilterType::SELECT_FILTER: { - static const auto info = make_info("SELECT_FILTER"); + static const auto info = make_info( + "SELECT_FILTER", + "Expression which is used for filtering in SELECT queries." + ); return info; } #if 0 /// Row-level security for INSERT, UPDATE, DELETE is not implemented yet. diff --git a/src/Access/Common/RowPolicyDefs.h b/src/Access/Common/RowPolicyDefs.h index 7ffc99e1272..bf2f632e98b 100644 --- a/src/Access/Common/RowPolicyDefs.h +++ b/src/Access/Common/RowPolicyDefs.h @@ -52,6 +52,7 @@ struct RowPolicyFilterTypeInfo const char * const raw_name; const String name; /// Lowercased with underscores, e.g. "select_filter". const String command; /// Uppercased without last word, e.g. "SELECT". + const String description; const bool is_check; /// E.g. false for SELECT_FILTER. static const RowPolicyFilterTypeInfo & get(RowPolicyFilterType type); }; diff --git a/src/Access/HTTPAuthClient.h b/src/Access/HTTPAuthClient.h index caefe869005..a8b56cf05a7 100644 --- a/src/Access/HTTPAuthClient.h +++ b/src/Access/HTTPAuthClient.h @@ -36,7 +36,7 @@ public: Result authenticateRequest(Poco::Net::HTTPRequest & request) const { - auto session = makeHTTPSession(uri, timeouts); + auto session = makeHTTPSession(HTTPConnectionGroupType::HTTP, uri, timeouts); Poco::Net::HTTPResponse response; auto milliseconds_to_wait = retry_initial_backoff_ms; diff --git a/src/Analyzer/ConstantNode.cpp b/src/Analyzer/ConstantNode.cpp index 69bed3dbe90..e26500a9886 100644 --- a/src/Analyzer/ConstantNode.cpp +++ b/src/Analyzer/ConstantNode.cpp @@ -1,5 +1,7 @@ #include +#include + #include #include #include @@ -38,52 +40,9 @@ ConstantNode::ConstantNode(Field value_) : ConstantNode(value_, applyVisitor(FieldToDataType(), value_)) {} -void ConstantNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const -{ - buffer << std::string(indent, ' ') << "CONSTANT id: " << format_state.getNodeId(this); - - if (hasAlias()) - buffer << ", alias: " << getAlias(); - - buffer << ", constant_value: " << constant_value->getValue().dump(); - buffer << ", constant_value_type: " << constant_value->getType()->getName(); - - if (getSourceExpression()) - { - buffer << '\n' << std::string(indent + 2, ' ') << "EXPRESSION" << '\n'; - getSourceExpression()->dumpTreeImpl(buffer, format_state, indent + 4); - } -} - -bool ConstantNode::isEqualImpl(const IQueryTreeNode & rhs) const -{ - const auto & rhs_typed = assert_cast(rhs); - return *constant_value == *rhs_typed.constant_value && value_string == rhs_typed.value_string; -} - -void ConstantNode::updateTreeHashImpl(HashState & hash_state) const -{ - auto type_name = constant_value->getType()->getName(); - hash_state.update(type_name.size()); - hash_state.update(type_name); - - hash_state.update(value_string.size()); - hash_state.update(value_string); -} - -QueryTreeNodePtr ConstantNode::cloneImpl() const -{ - return std::make_shared(constant_value, source_expression); -} - -ASTPtr ConstantNode::toASTImpl(const ConvertToASTOptions & options) const +bool ConstantNode::requiresCastCall() const { const auto & constant_value_literal = constant_value->getValue(); - auto constant_value_ast = std::make_shared(constant_value_literal); - - if (!options.add_cast_for_constants) - return constant_value_ast; - bool need_to_add_cast_function = false; auto constant_value_literal_type = constant_value_literal.getType(); WhichDataType constant_value_type(constant_value->getType()); @@ -131,7 +90,72 @@ ASTPtr ConstantNode::toASTImpl(const ConvertToASTOptions & options) const // Add cast if constant was created as a result of constant folding. // Constant folding may lead to type transformation and literal on shard // may have a different type. - if (need_to_add_cast_function || source_expression != nullptr) + return need_to_add_cast_function || source_expression != nullptr; +} + +bool ConstantNode::receivedFromInitiatorServer() const +{ + if (!hasSourceExpression()) + return false; + + auto * cast_function = getSourceExpression()->as(); + if (!cast_function || cast_function->getFunctionName() != "_CAST") + return false; + return true; +} + +void ConstantNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const +{ + buffer << std::string(indent, ' ') << "CONSTANT id: " << format_state.getNodeId(this); + + if (hasAlias()) + buffer << ", alias: " << getAlias(); + + buffer << ", constant_value: "; + if (mask_id) + buffer << "[HIDDEN id: " << mask_id << "]"; + else + buffer << constant_value->getValue().dump(); + + buffer << ", constant_value_type: " << constant_value->getType()->getName(); + + if (!mask_id && getSourceExpression()) + { + buffer << '\n' << std::string(indent + 2, ' ') << "EXPRESSION" << '\n'; + getSourceExpression()->dumpTreeImpl(buffer, format_state, indent + 4); + } +} + +bool ConstantNode::isEqualImpl(const IQueryTreeNode & rhs) const +{ + const auto & rhs_typed = assert_cast(rhs); + return *constant_value == *rhs_typed.constant_value && value_string == rhs_typed.value_string; +} + +void ConstantNode::updateTreeHashImpl(HashState & hash_state) const +{ + auto type_name = constant_value->getType()->getName(); + hash_state.update(type_name.size()); + hash_state.update(type_name); + + hash_state.update(value_string.size()); + hash_state.update(value_string); +} + +QueryTreeNodePtr ConstantNode::cloneImpl() const +{ + return std::make_shared(constant_value, source_expression); +} + +ASTPtr ConstantNode::toASTImpl(const ConvertToASTOptions & options) const +{ + const auto & constant_value_literal = constant_value->getValue(); + auto constant_value_ast = std::make_shared(constant_value_literal); + + if (!options.add_cast_for_constants) + return constant_value_ast; + + if (requiresCastCall()) { auto constant_type_name_ast = std::make_shared(constant_value->getType()->getName()); return makeASTFunction("_CAST", std::move(constant_value_ast), std::move(constant_type_name_ast)); diff --git a/src/Analyzer/ConstantNode.h b/src/Analyzer/ConstantNode.h index 51c98a4a3b3..b065853e315 100644 --- a/src/Analyzer/ConstantNode.h +++ b/src/Analyzer/ConstantNode.h @@ -75,6 +75,17 @@ public: return constant_value->getType(); } + /// Check if conversion to AST requires wrapping with _CAST function. + bool requiresCastCall() const; + + /// Check if constant is a result of _CAST function constant folding. + bool receivedFromInitiatorServer() const; + + void setMaskId(size_t id) + { + mask_id = id; + } + void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: @@ -90,6 +101,7 @@ private: ConstantValuePtr constant_value; String value_string; QueryTreeNodePtr source_expression; + size_t mask_id = 0; static constexpr size_t children_size = 0; }; diff --git a/src/Analyzer/FunctionSecretArgumentsFinderTreeNode.h b/src/Analyzer/FunctionSecretArgumentsFinderTreeNode.h new file mode 100644 index 00000000000..439ddffe5e5 --- /dev/null +++ b/src/Analyzer/FunctionSecretArgumentsFinderTreeNode.h @@ -0,0 +1,372 @@ +#pragma once + +#include +#include +#include +#include +#include +#include +#include +#include + +#include + + +namespace DB +{ + + +/// Finds arguments of a specified function which should not be displayed for most users for security reasons. +/// That involves passwords and secret keys. +class FunctionSecretArgumentsFinderTreeNode +{ +public: + explicit FunctionSecretArgumentsFinderTreeNode(const FunctionNode & function_) : function(function_), arguments(function.getArguments()) + { + if (arguments.getNodes().empty()) + return; + + findFunctionSecretArguments(); + } + + struct Result + { + /// Result constructed by default means no arguments will be hidden. + size_t start = static_cast(-1); + size_t count = 0; /// Mostly it's either 0 or 1. There are only a few cases where `count` can be greater than 1 (e.g. see `encrypt`). + /// In all known cases secret arguments are consecutive + bool are_named = false; /// Arguments like `password = 'password'` are considered as named arguments. + /// E.g. "headers" in `url('..', headers('foo' = '[HIDDEN]'))` + std::vector nested_maps; + + bool hasSecrets() const + { + return count != 0 || !nested_maps.empty(); + } + }; + + FunctionSecretArgumentsFinder::Result getResult() const { return result; } + +private: + const FunctionNode & function; + const ListNode & arguments; + FunctionSecretArgumentsFinder::Result result; + + void markSecretArgument(size_t index, bool argument_is_named = false) + { + if (index >= arguments.getNodes().size()) + return; + if (!result.count) + { + result.start = index; + result.are_named = argument_is_named; + } + chassert(index >= result.start); /// We always check arguments consecutively + result.count = index + 1 - result.start; + if (!argument_is_named) + result.are_named = false; + } + + void findFunctionSecretArguments() + { + const auto & name = function.getFunctionName(); + + if ((name == "mysql") || (name == "postgresql") || (name == "mongodb")) + { + /// mysql('host:port', 'database', 'table', 'user', 'password', ...) + /// postgresql('host:port', 'database', 'table', 'user', 'password', ...) + /// mongodb('host:port', 'database', 'collection', 'user', 'password', ...) + findMySQLFunctionSecretArguments(); + } + else if ((name == "s3") || (name == "cosn") || (name == "oss") || + (name == "deltaLake") || (name == "hudi") || (name == "iceberg")) + { + /// s3('url', 'aws_access_key_id', 'aws_secret_access_key', ...) + findS3FunctionSecretArguments(/* is_cluster_function= */ false); + } + else if (name == "s3Cluster") + { + /// s3Cluster('cluster_name', 'url', 'aws_access_key_id', 'aws_secret_access_key', ...) + findS3FunctionSecretArguments(/* is_cluster_function= */ true); + } + else if ((name == "remote") || (name == "remoteSecure")) + { + /// remote('addresses_expr', 'db', 'table', 'user', 'password', ...) + findRemoteFunctionSecretArguments(); + } + else if ((name == "encrypt") || (name == "decrypt") || + (name == "aes_encrypt_mysql") || (name == "aes_decrypt_mysql") || + (name == "tryDecrypt")) + { + /// encrypt('mode', 'plaintext', 'key' [, iv, aad]) + findEncryptionFunctionSecretArguments(); + } + else if (name == "url") + { + findURLSecretArguments(); + } + } + + void findMySQLFunctionSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// mysql(named_collection, ..., password = 'password', ...) + findSecretNamedArgument("password", 1); + } + else + { + /// mysql('host:port', 'database', 'table', 'user', 'password', ...) + markSecretArgument(4); + } + } + + /// Returns the number of arguments excluding "headers" and "extra_credentials" (which should + /// always be at the end). Marks "headers" as secret, if found. + size_t excludeS3OrURLNestedMaps() + { + const auto & nodes = arguments.getNodes(); + size_t count = nodes.size(); + while (count > 0) + { + const FunctionNode * f = nodes.at(count - 1)->as(); + if (!f) + break; + if (f->getFunctionName() == "headers") + result.nested_maps.push_back(f->getFunctionName()); + else if (f->getFunctionName() != "extra_credentials") + break; + count -= 1; + } + return count; + } + + void findS3FunctionSecretArguments(bool is_cluster_function) + { + /// s3Cluster('cluster_name', 'url', ...) has 'url' as its second argument. + size_t url_arg_idx = is_cluster_function ? 1 : 0; + + if (!is_cluster_function && isNamedCollectionName(0)) + { + /// s3(named_collection, ..., secret_access_key = 'secret_access_key', ...) + findSecretNamedArgument("secret_access_key", 1); + return; + } + + /// We should check other arguments first because we don't need to do any replacement in case of + /// s3('url', NOSIGN, 'format' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) + /// s3('url', 'format', 'structure' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) + size_t count = excludeS3OrURLNestedMaps(); + if ((url_arg_idx + 3 <= count) && (count <= url_arg_idx + 4)) + { + String second_arg; + if (tryGetStringFromArgument(url_arg_idx + 1, &second_arg)) + { + if (boost::iequals(second_arg, "NOSIGN")) + return; /// The argument after 'url' is "NOSIGN". + + if (second_arg == "auto" || KnownFormatNames::instance().exists(second_arg)) + return; /// The argument after 'url' is a format: s3('url', 'format', ...) + } + } + + /// We're going to replace 'aws_secret_access_key' with '[HIDDEN]' for the following signatures: + /// s3('url', 'aws_access_key_id', 'aws_secret_access_key', ...) + /// s3Cluster('cluster_name', 'url', 'aws_access_key_id', 'aws_secret_access_key', 'format', 'compression') + if (url_arg_idx + 2 < count) + markSecretArgument(url_arg_idx + 2); + } + + void findURLSecretArguments() + { + if (!isNamedCollectionName(0)) + excludeS3OrURLNestedMaps(); + } + + bool tryGetStringFromArgument(size_t arg_idx, String * res, bool allow_identifier = true) const + { + if (arg_idx >= arguments.getNodes().size()) + return false; + + return tryGetStringFromArgument(arguments.getNodes()[arg_idx], res, allow_identifier); + } + + static bool tryGetStringFromArgument(const QueryTreeNodePtr argument, String * res, bool allow_identifier = true) + { + if (const auto * literal = argument->as()) + { + if (literal->getValue().getType() != Field::Types::String) + return false; + if (res) + *res = literal->getValue().safeGet(); + return true; + } + + if (allow_identifier) + { + if (const auto * id = argument->as()) + { + if (res) + *res = id->getIdentifier().getFullName(); + return true; + } + } + + return false; + } + + void findRemoteFunctionSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// remote(named_collection, ..., password = 'password', ...) + findSecretNamedArgument("password", 1); + return; + } + + /// We're going to replace 'password' with '[HIDDEN'] for the following signatures: + /// remote('addresses_expr', db.table, 'user' [, 'password'] [, sharding_key]) + /// remote('addresses_expr', 'db', 'table', 'user' [, 'password'] [, sharding_key]) + /// remote('addresses_expr', table_function(), 'user' [, 'password'] [, sharding_key]) + + /// But we should check the number of arguments first because we don't need to do any replacements in case of + /// remote('addresses_expr', db.table) + if (arguments.getNodes().size() < 3) + return; + + size_t arg_num = 1; + + /// Skip 1 or 2 arguments with table_function() or db.table or 'db', 'table'. + const auto * table_function = arguments.getNodes()[arg_num]->as(); + if (table_function && KnownTableFunctionNames::instance().exists(table_function->getFunctionName())) + { + ++arg_num; + } + else + { + std::optional database; + std::optional qualified_table_name; + if (!tryGetDatabaseNameOrQualifiedTableName(arg_num, database, qualified_table_name)) + { + /// We couldn't evaluate the argument so we don't know whether it is 'db.table' or just 'db'. + /// Hence we can't figure out whether we should skip one argument 'user' or two arguments 'table', 'user' + /// before the argument 'password'. So it's safer to wipe two arguments just in case. + /// The last argument can be also a `sharding_key`, so we need to check that argument is a literal string + /// before wiping it (because the `password` argument is always a literal string). + if (tryGetStringFromArgument(arg_num + 2, nullptr, /* allow_identifier= */ false)) + { + /// Wipe either `password` or `user`. + markSecretArgument(arg_num + 2); + } + if (tryGetStringFromArgument(arg_num + 3, nullptr, /* allow_identifier= */ false)) + { + /// Wipe either `password` or `sharding_key`. + markSecretArgument(arg_num + 3); + } + return; + } + + /// Skip the current argument (which is either a database name or a qualified table name). + ++arg_num; + if (database) + { + /// Skip the 'table' argument if the previous argument was a database name. + ++arg_num; + } + } + + /// Skip username. + ++arg_num; + + /// Do our replacement: + /// remote('addresses_expr', db.table, 'user', 'password', ...) -> remote('addresses_expr', db.table, 'user', '[HIDDEN]', ...) + /// The last argument can be also a `sharding_key`, so we need to check that argument is a literal string + /// before wiping it (because the `password` argument is always a literal string). + bool can_be_password = tryGetStringFromArgument(arg_num, nullptr, /* allow_identifier= */ false); + if (can_be_password) + markSecretArgument(arg_num); + } + + /// Tries to get either a database name or a qualified table name from an argument. + /// Empty string is also allowed (it means the default database). + /// The function is used by findRemoteFunctionSecretArguments() to determine how many arguments to skip before a password. + bool tryGetDatabaseNameOrQualifiedTableName( + size_t arg_idx, + std::optional & res_database, + std::optional & res_qualified_table_name) const + { + res_database.reset(); + res_qualified_table_name.reset(); + + String str; + if (!tryGetStringFromArgument(arg_idx, &str, /* allow_identifier= */ true)) + return false; + + if (str.empty()) + { + res_database = ""; + return true; + } + + auto qualified_table_name = QualifiedTableName::tryParseFromString(str); + if (!qualified_table_name) + return false; + + if (qualified_table_name->database.empty()) + res_database = std::move(qualified_table_name->table); + else + res_qualified_table_name = std::move(qualified_table_name); + return true; + } + + void findEncryptionFunctionSecretArguments() + { + if (arguments.getNodes().empty()) + return; + + /// We replace all arguments after 'mode' with '[HIDDEN]': + /// encrypt('mode', 'plaintext', 'key' [, iv, aad]) -> encrypt('mode', '[HIDDEN]') + result.start = 1; + result.count = arguments.getNodes().size() - 1; + } + + + /// Whether a specified argument can be the name of a named collection? + bool isNamedCollectionName(size_t arg_idx) const + { + if (arguments.getNodes().size() <= arg_idx) + return false; + + const auto * identifier = arguments.getNodes()[arg_idx]->as(); + return identifier != nullptr; + } + + /// Looks for a secret argument with a specified name. This function looks for arguments in format `key=value` where the key is specified. + void findSecretNamedArgument(const std::string_view & key, size_t start = 0) + { + for (size_t i = start; i < arguments.getNodes().size(); ++i) + { + const auto & argument = arguments.getNodes()[i]; + const auto * equals_func = argument->as(); + if (!equals_func || (equals_func->getFunctionName() != "equals")) + continue; + + const auto * expr_list = equals_func->getArguments().as(); + if (!expr_list) + continue; + + const auto & equal_args = expr_list->getNodes(); + if (equal_args.size() != 2) + continue; + + String found_key; + if (!tryGetStringFromArgument(equal_args[0], &found_key)) + continue; + + if (found_key == key) + markSecretArgument(i, /* argument_is_named= */ true); + } + } +}; + +} diff --git a/src/Analyzer/Passes/OptimizeDateOrDateTimeConverterWithPreimagePass.cpp b/src/Analyzer/Passes/OptimizeDateOrDateTimeConverterWithPreimagePass.cpp index cc334cde9c7..0c37749c706 100644 --- a/src/Analyzer/Passes/OptimizeDateOrDateTimeConverterWithPreimagePass.cpp +++ b/src/Analyzer/Passes/OptimizeDateOrDateTimeConverterWithPreimagePass.cpp @@ -94,7 +94,8 @@ public: if (!func_node || func_node->getArguments().getNodes().size() != 1) return; - const auto * column_id = func_node->getArguments().getNodes()[0]->as(); + const auto & argument_node = func_node->getArguments().getNodes()[0]; + const auto * column_id = argument_node->as(); if (!column_id) return; @@ -119,7 +120,7 @@ public: if (!preimage_range) return; - const auto new_node = generateOptimizedDateFilter(comparator, *column_id, *preimage_range); + const auto new_node = generateOptimizedDateFilter(comparator, argument_node, *preimage_range); if (!new_node) return; @@ -128,20 +129,22 @@ public: } private: - QueryTreeNodePtr - generateOptimizedDateFilter(const String & comparator, const ColumnNode & column_node, const std::pair & range) const + QueryTreeNodePtr generateOptimizedDateFilter( + const String & comparator, const QueryTreeNodePtr & column_node, const std::pair & range) const { const DateLUTImpl & date_lut = DateLUT::instance("UTC"); String start_date_or_date_time; String end_date_or_date_time; - if (isDateOrDate32(column_node.getColumnType().get())) + const auto & column_node_typed = column_node->as(); + const auto & column_type = column_node_typed.getColumnType().get(); + if (isDateOrDate32(column_type)) { start_date_or_date_time = date_lut.dateToString(range.first.get()); end_date_or_date_time = date_lut.dateToString(range.second.get()); } - else if (isDateTime(column_node.getColumnType().get()) || isDateTime64(column_node.getColumnType().get())) + else if (isDateTime(column_type) || isDateTime64(column_type)) { start_date_or_date_time = date_lut.timeToString(range.first.get()); end_date_or_date_time = date_lut.timeToString(range.second.get()); @@ -151,69 +154,29 @@ private: if (comparator == "equals") { - const auto lhs = std::make_shared("greaterOrEquals"); - lhs->getArguments().getNodes().push_back(std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - lhs->getArguments().getNodes().push_back(std::make_shared(start_date_or_date_time)); - resolveOrdinaryFunctionNode(*lhs, lhs->getFunctionName()); - - const auto rhs = std::make_shared("less"); - rhs->getArguments().getNodes().push_back(std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - rhs->getArguments().getNodes().push_back(std::make_shared(end_date_or_date_time)); - resolveOrdinaryFunctionNode(*rhs, rhs->getFunctionName()); - - const auto new_date_filter = std::make_shared("and"); - new_date_filter->getArguments().getNodes() = {lhs, rhs}; - resolveOrdinaryFunctionNode(*new_date_filter, new_date_filter->getFunctionName()); - - return new_date_filter; + return createFunctionNode( + "and", + createFunctionNode("greaterOrEquals", column_node, std::make_shared(start_date_or_date_time)), + createFunctionNode("less", column_node, std::make_shared(end_date_or_date_time))); } else if (comparator == "notEquals") { - const auto lhs = std::make_shared("less"); - lhs->getArguments().getNodes().push_back(std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - lhs->getArguments().getNodes().push_back(std::make_shared(start_date_or_date_time)); - resolveOrdinaryFunctionNode(*lhs, lhs->getFunctionName()); - - const auto rhs = std::make_shared("greaterOrEquals"); - rhs->getArguments().getNodes().push_back(std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - rhs->getArguments().getNodes().push_back(std::make_shared(end_date_or_date_time)); - resolveOrdinaryFunctionNode(*rhs, rhs->getFunctionName()); - - const auto new_date_filter = std::make_shared("or"); - new_date_filter->getArguments().getNodes() = {lhs, rhs}; - resolveOrdinaryFunctionNode(*new_date_filter, new_date_filter->getFunctionName()); - - return new_date_filter; + return createFunctionNode( + "or", + createFunctionNode("less", column_node, std::make_shared(start_date_or_date_time)), + createFunctionNode("greaterOrEquals", column_node, std::make_shared(end_date_or_date_time))); } else if (comparator == "greater") { - const auto new_date_filter = std::make_shared("greaterOrEquals"); - new_date_filter->getArguments().getNodes().push_back( - std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - new_date_filter->getArguments().getNodes().push_back(std::make_shared(end_date_or_date_time)); - resolveOrdinaryFunctionNode(*new_date_filter, new_date_filter->getFunctionName()); - - return new_date_filter; + return createFunctionNode("greaterOrEquals", column_node, std::make_shared(end_date_or_date_time)); } else if (comparator == "lessOrEquals") { - const auto new_date_filter = std::make_shared("less"); - new_date_filter->getArguments().getNodes().push_back( - std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - new_date_filter->getArguments().getNodes().push_back(std::make_shared(end_date_or_date_time)); - resolveOrdinaryFunctionNode(*new_date_filter, new_date_filter->getFunctionName()); - - return new_date_filter; + return createFunctionNode("less", column_node, std::make_shared(end_date_or_date_time)); } else if (comparator == "less" || comparator == "greaterOrEquals") { - const auto new_date_filter = std::make_shared(comparator); - new_date_filter->getArguments().getNodes().push_back( - std::make_shared(column_node.getColumn(), column_node.getColumnSource())); - new_date_filter->getArguments().getNodes().push_back(std::make_shared(start_date_or_date_time)); - resolveOrdinaryFunctionNode(*new_date_filter, new_date_filter->getFunctionName()); - - return new_date_filter; + return createFunctionNode(comparator, column_node, std::make_shared(start_date_or_date_time)); } else [[unlikely]] { @@ -224,10 +187,17 @@ private: } } - void resolveOrdinaryFunctionNode(FunctionNode & function_node, const String & function_name) const + template + QueryTreeNodePtr createFunctionNode(const String & function_name, Args &&... args) const { auto function = FunctionFactory::instance().get(function_name, getContext()); - function_node.resolveAsFunction(function->build(function_node.getArgumentColumns())); + const auto function_node = std::make_shared(function_name); + auto & new_arguments = function_node->getArguments().getNodes(); + new_arguments.reserve(sizeof...(args)); + (new_arguments.push_back(std::forward(args)), ...); + function_node->resolveAsFunction(function->build(function_node->getArgumentColumns())); + + return function_node; } }; diff --git a/src/Analyzer/Passes/QueryAnalysisPass.cpp b/src/Analyzer/Passes/QueryAnalysisPass.cpp index 907a732493d..c62641ca05c 100644 --- a/src/Analyzer/Passes/QueryAnalysisPass.cpp +++ b/src/Analyzer/Passes/QueryAnalysisPass.cpp @@ -3,6 +3,7 @@ #include #include #include +#include #include #include @@ -706,7 +707,10 @@ struct IdentifierResolveScope { subquery_depth = parent_scope->subquery_depth; context = parent_scope->context; + projection_mask_map = parent_scope->projection_mask_map; } + else + projection_mask_map = std::make_shared>(); if (auto * union_node = scope_node->as()) { @@ -718,6 +722,11 @@ struct IdentifierResolveScope group_by_use_nulls = context->getSettingsRef().group_by_use_nulls && (query_node->isGroupByWithGroupingSets() || query_node->isGroupByWithRollup() || query_node->isGroupByWithCube()); } + + if (context) + join_use_nulls = context->getSettingsRef().join_use_nulls; + else if (parent_scope) + join_use_nulls = parent_scope->join_use_nulls; } QueryTreeNodePtr scope_node; @@ -772,6 +781,8 @@ struct IdentifierResolveScope /// Apply nullability to aggregation keys bool group_by_use_nulls = false; + /// Join retutns NULLs instead of default values + bool join_use_nulls = false; /// JOINs count size_t joins_count = 0; @@ -784,6 +795,9 @@ struct IdentifierResolveScope */ QueryTreeNodePtr expression_join_tree_node; + /// Node hash to mask id map + std::shared_ptr> projection_mask_map; + [[maybe_unused]] const IdentifierResolveScope * getNearestQueryScope() const { const IdentifierResolveScope * scope_to_check = this; @@ -1068,6 +1082,8 @@ private: class QueryAnalyzer { public: + explicit QueryAnalyzer(bool only_analyze_) : only_analyze(only_analyze_) {} + void resolve(QueryTreeNodePtr & node, const QueryTreeNodePtr & table_expression, ContextPtr context) { IdentifierResolveScope scope(node, nullptr /*parent_scope*/); @@ -1430,6 +1446,7 @@ private: /// Global scalar subquery to scalar value map std::unordered_map scalar_subquery_to_scalar_value; + const bool only_analyze; }; /// Utility functions implementation @@ -1977,80 +1994,96 @@ void QueryAnalyzer::evaluateScalarSubqueryIfNeeded(QueryTreeNodePtr & node, Iden auto interpreter = std::make_unique(node->toAST(), subquery_context, subquery_context->getViewSource(), options); auto io = interpreter->execute(); - PullingAsyncPipelineExecutor executor(io.pipeline); io.pipeline.setProgressCallback(context->getProgressCallback()); io.pipeline.setProcessListElement(context->getProcessListElement()); - Block block; - - while (block.rows() == 0 && executor.pull(block)) + if (only_analyze) { - } - - if (block.rows() == 0) - { - auto types = interpreter->getSampleBlock().getDataTypes(); - if (types.size() != 1) - types = {std::make_shared(types)}; - - auto & type = types[0]; - if (!type->isNullable()) + /// If query is only analyzed, then constants are not correct. + scalar_block = interpreter->getSampleBlock(); + for (auto & column : scalar_block) { - if (!type->canBeInsideNullable()) - throw Exception(ErrorCodes::INCORRECT_RESULT_OF_SCALAR_SUBQUERY, - "Scalar subquery returned empty result of type {} which cannot be Nullable", - type->getName()); - - type = makeNullable(type); + if (column.column->empty()) + { + auto mut_col = column.column->cloneEmpty(); + mut_col->insertDefault(); + column.column = std::move(mut_col); + } } - - auto scalar_column = type->createColumn(); - scalar_column->insert(Null()); - scalar_block.insert({std::move(scalar_column), type, "null"}); } else { - if (block.rows() != 1) - throw Exception(ErrorCodes::INCORRECT_RESULT_OF_SCALAR_SUBQUERY, "Scalar subquery returned more than one row"); + Block block; - Block tmp_block; - while (tmp_block.rows() == 0 && executor.pull(tmp_block)) + while (block.rows() == 0 && executor.pull(block)) { } - if (tmp_block.rows() != 0) - throw Exception(ErrorCodes::INCORRECT_RESULT_OF_SCALAR_SUBQUERY, "Scalar subquery returned more than one row"); - - block = materializeBlock(block); - size_t columns = block.columns(); - - if (columns == 1) + if (block.rows() == 0) { - auto & column = block.getByPosition(0); - /// Here we wrap type to nullable if we can. - /// It is needed cause if subquery return no rows, it's result will be Null. - /// In case of many columns, do not check it cause tuple can't be nullable. - if (!column.type->isNullable() && column.type->canBeInsideNullable()) + auto types = interpreter->getSampleBlock().getDataTypes(); + if (types.size() != 1) + types = {std::make_shared(types)}; + + auto & type = types[0]; + if (!type->isNullable()) { - column.type = makeNullable(column.type); - column.column = makeNullable(column.column); + if (!type->canBeInsideNullable()) + throw Exception(ErrorCodes::INCORRECT_RESULT_OF_SCALAR_SUBQUERY, + "Scalar subquery returned empty result of type {} which cannot be Nullable", + type->getName()); + + type = makeNullable(type); } - scalar_block = block; + auto scalar_column = type->createColumn(); + scalar_column->insert(Null()); + scalar_block.insert({std::move(scalar_column), type, "null"}); } else { - /** Make unique column names for tuple. - * - * Example: SELECT (SELECT 2 AS x, x) - */ - makeUniqueColumnNamesInBlock(block); + if (block.rows() != 1) + throw Exception(ErrorCodes::INCORRECT_RESULT_OF_SCALAR_SUBQUERY, "Scalar subquery returned more than one row"); - scalar_block.insert({ - ColumnTuple::create(block.getColumns()), - std::make_shared(block.getDataTypes(), block.getNames()), - "tuple"}); + Block tmp_block; + while (tmp_block.rows() == 0 && executor.pull(tmp_block)) + { + } + + if (tmp_block.rows() != 0) + throw Exception(ErrorCodes::INCORRECT_RESULT_OF_SCALAR_SUBQUERY, "Scalar subquery returned more than one row"); + + block = materializeBlock(block); + size_t columns = block.columns(); + + if (columns == 1) + { + auto & column = block.getByPosition(0); + /// Here we wrap type to nullable if we can. + /// It is needed cause if subquery return no rows, it's result will be Null. + /// In case of many columns, do not check it cause tuple can't be nullable. + if (!column.type->isNullable() && column.type->canBeInsideNullable()) + { + column.type = makeNullable(column.type); + column.column = makeNullable(column.column); + } + + scalar_block = block; + } + else + { + /** Make unique column names for tuple. + * + * Example: SELECT (SELECT 2 AS x, x) + */ + makeUniqueColumnNamesInBlock(block); + + scalar_block.insert({ + ColumnTuple::create(block.getColumns()), + std::make_shared(block.getDataTypes(), block.getNames()), + "tuple"}); + } } } @@ -3286,7 +3319,6 @@ QueryTreeNodePtr QueryAnalyzer::tryResolveIdentifierFromJoin(const IdentifierLoo QueryTreeNodePtr resolved_identifier; JoinKind join_kind = from_join_node.getKind(); - bool join_use_nulls = scope.context->getSettingsRef().join_use_nulls; /// If columns from left or right table were missed Object(Nullable('json')) subcolumns, they will be replaced /// to ConstantNode(NULL), which can't be cast to ColumnNode, so we resolve it here. @@ -3451,7 +3483,7 @@ QueryTreeNodePtr QueryAnalyzer::tryResolveIdentifierFromJoin(const IdentifierLoo if (join_node_in_resolve_process || !resolved_identifier) return resolved_identifier; - if (join_use_nulls) + if (scope.join_use_nulls) { resolved_identifier = resolved_identifier->clone(); convertJoinedColumnTypeToNullIfNeeded(resolved_identifier, join_kind, resolved_side); @@ -4439,7 +4471,7 @@ ProjectionNames QueryAnalyzer::resolveMatcher(QueryTreeNodePtr & matcher_node, I else matched_expression_nodes_with_names = resolveUnqualifiedMatcher(matcher_node, scope); - if (scope.context->getSettingsRef().join_use_nulls) + if (scope.join_use_nulls) { /** If we are resolving matcher came from the result of JOIN and `join_use_nulls` is set, * we need to convert joined column type to Nullable. @@ -5124,22 +5156,31 @@ ProjectionNames QueryAnalyzer::resolveFunction(QueryTreeNodePtr & node, Identifi } /// Resolve function arguments - bool allow_table_expressions = is_special_function_in; auto arguments_projection_names = resolveExpressionNodeList(function_node_ptr->getArgumentsNode(), scope, true /*allow_lambda_expression*/, allow_table_expressions /*allow_table_expression*/); - if (function_node_ptr->toAST()->hasSecretParts()) + /// Mask arguments if needed + if (!scope.context->getSettingsRef().format_display_secrets_in_show_and_select) { - for (auto & argument : arguments_projection_names) + if (FunctionSecretArgumentsFinder::Result secret_arguments = FunctionSecretArgumentsFinderTreeNode(*function_node_ptr).getResult(); secret_arguments.count) { - SipHash hash; - hash.update(argument); - argument = getHexUIntLowercase(hash.get128()); + auto & argument_nodes = function_node_ptr->getArgumentsNode()->as().getNodes(); + + for (size_t n = secret_arguments.start; n < secret_arguments.start + secret_arguments.count; ++n) + { + if (auto * constant = argument_nodes[n]->as()) + { + auto mask = scope.projection_mask_map->insert({constant->getTreeHash(), scope.projection_mask_map->size() + 1}).first->second; + constant->setMaskId(mask); + arguments_projection_names[n] = "[HIDDEN id: " + std::to_string(mask) + "]"; + } + } } } + auto & function_node = *function_node_ptr; /// Replace right IN function argument if it is table or table function with subquery that read ordinary columns @@ -6651,7 +6692,6 @@ void QueryAnalyzer::initializeTableExpressionData(const QueryTreeNodePtr & table if (column_default && column_default->kind == ColumnDefaultKind::Alias) { auto alias_expression = buildQueryTree(column_default->expression, scope.context); - alias_expression = buildCastFunction(alias_expression, column_name_and_type.type, scope.context, false /*resolve*/); auto column_node = std::make_shared(column_name_and_type, std::move(alias_expression), table_expression_node); column_name_to_column_node.emplace(column_name_and_type.name, column_node); alias_columns_to_resolve.emplace_back(column_name_and_type.name, column_node); @@ -6684,7 +6724,9 @@ void QueryAnalyzer::initializeTableExpressionData(const QueryTreeNodePtr & table alias_column_resolve_scope, false /*allow_lambda_expression*/, false /*allow_table_expression*/); - + auto & resolved_expression = alias_column_to_resolve->getExpression(); + if (!resolved_expression->getResultType()->equals(*alias_column_to_resolve->getResultType())) + resolved_expression = buildCastFunction(resolved_expression, alias_column_to_resolve->getResultType(), scope.context, true); column_name_to_column_node = std::move(alias_column_resolve_scope.column_name_to_column_node); column_name_to_column_node[alias_column_to_resolve_name] = alias_column_to_resolve; } @@ -7558,8 +7600,22 @@ void QueryAnalyzer::resolveQuery(const QueryTreeNodePtr & query_node, Identifier } if (query_node_typed.getPrewhere()) + { + /** Expression in PREWHERE with JOIN should not be modified by join_use_nulls. + * Example: SELECT * FROM t1 JOIN t2 USING (id) PREWHERE a = 1 + * Column `a` should be resolved from table and should not change its type to Nullable. + */ + bool join_use_nulls = scope.join_use_nulls; + bool use_identifier_lookup_to_result_cache = scope.use_identifier_lookup_to_result_cache; + scope.join_use_nulls = false; + scope.use_identifier_lookup_to_result_cache = false; + resolveExpressionNode(query_node_typed.getPrewhere(), scope, false /*allow_lambda_expression*/, false /*allow_table_expression*/); + scope.join_use_nulls = join_use_nulls; + scope.use_identifier_lookup_to_result_cache = use_identifier_lookup_to_result_cache; + } + if (query_node_typed.getWhere()) resolveExpressionNode(query_node_typed.getWhere(), scope, false /*allow_lambda_expression*/, false /*allow_table_expression*/); @@ -7749,13 +7805,16 @@ void QueryAnalyzer::resolveUnion(const QueryTreeNodePtr & union_node, Identifier } -QueryAnalysisPass::QueryAnalysisPass(QueryTreeNodePtr table_expression_) +QueryAnalysisPass::QueryAnalysisPass(QueryTreeNodePtr table_expression_, bool only_analyze_) : table_expression(std::move(table_expression_)) + , only_analyze(only_analyze_) {} +QueryAnalysisPass::QueryAnalysisPass(bool only_analyze_) : only_analyze(only_analyze_) {} + void QueryAnalysisPass::run(QueryTreeNodePtr & query_tree_node, ContextPtr context) { - QueryAnalyzer analyzer; + QueryAnalyzer analyzer(only_analyze); analyzer.resolve(query_tree_node, table_expression, context); createUniqueTableAliases(query_tree_node, table_expression, context); } diff --git a/src/Analyzer/Passes/QueryAnalysisPass.h b/src/Analyzer/Passes/QueryAnalysisPass.h index 5d335d3e712..8c746833eee 100644 --- a/src/Analyzer/Passes/QueryAnalysisPass.h +++ b/src/Analyzer/Passes/QueryAnalysisPass.h @@ -71,13 +71,13 @@ public: /** Construct query analysis pass for query or union analysis. * Available columns are extracted from query node join tree. */ - QueryAnalysisPass() = default; + explicit QueryAnalysisPass(bool only_analyze_ = false); /** Construct query analysis pass for expression or list of expressions analysis. * Available expression columns are extracted from table expression. * Table expression node must have query, union, table, table function type. */ - explicit QueryAnalysisPass(QueryTreeNodePtr table_expression_); + QueryAnalysisPass(QueryTreeNodePtr table_expression_, bool only_analyze_ = false); String getName() override { @@ -93,6 +93,7 @@ public: private: QueryTreeNodePtr table_expression; + const bool only_analyze; }; } diff --git a/src/Analyzer/QueryTreePassManager.cpp b/src/Analyzer/QueryTreePassManager.cpp index 43bb534a44e..9c07884a464 100644 --- a/src/Analyzer/QueryTreePassManager.cpp +++ b/src/Analyzer/QueryTreePassManager.cpp @@ -246,9 +246,9 @@ void QueryTreePassManager::dump(WriteBuffer & buffer, size_t up_to_pass_index) } } -void addQueryTreePasses(QueryTreePassManager & manager) +void addQueryTreePasses(QueryTreePassManager & manager, bool only_analyze) { - manager.addPass(std::make_unique()); + manager.addPass(std::make_unique(only_analyze)); manager.addPass(std::make_unique()); manager.addPass(std::make_unique()); diff --git a/src/Analyzer/QueryTreePassManager.h b/src/Analyzer/QueryTreePassManager.h index 270563590ba..0a0d72a6698 100644 --- a/src/Analyzer/QueryTreePassManager.h +++ b/src/Analyzer/QueryTreePassManager.h @@ -47,6 +47,6 @@ private: std::vector passes; }; -void addQueryTreePasses(QueryTreePassManager & manager); +void addQueryTreePasses(QueryTreePassManager & manager, bool only_analyze = false); } diff --git a/src/Backups/BackupsWorker.cpp b/src/Backups/BackupsWorker.cpp index 7c6f0c74b1a..1b4279ca9a7 100644 --- a/src/Backups/BackupsWorker.cpp +++ b/src/Backups/BackupsWorker.cpp @@ -18,6 +18,7 @@ #include #include #include +#include #include #include #include @@ -486,7 +487,7 @@ OperationID BackupsWorker::startMakingBackup(const ASTPtr & query, const Context /// process_list_element_holder is used to make an element in ProcessList live while BACKUP is working asynchronously. auto process_list_element = context_in_use->getProcessListElement(); - scheduleFromThreadPool( + thread_pool.scheduleOrThrowOnError( [this, backup_query, backup_id, @@ -502,6 +503,8 @@ OperationID BackupsWorker::startMakingBackup(const ASTPtr & query, const Context BackupMutablePtr backup_async; try { + setThreadName("BackupWorker"); + CurrentThread::QueryScope query_scope(context_in_use); doBackup( backup_async, backup_query, @@ -517,8 +520,7 @@ OperationID BackupsWorker::startMakingBackup(const ASTPtr & query, const Context { on_exception(backup_async, backup_id, backup_name_for_logging, backup_settings, backup_coordination); } - }, - thread_pool, "BackupWorker"); + }); } else { @@ -864,7 +866,7 @@ OperationID BackupsWorker::startRestoring(const ASTPtr & query, ContextMutablePt /// process_list_element_holder is used to make an element in ProcessList live while RESTORE is working asynchronously. auto process_list_element = context_in_use->getProcessListElement(); - scheduleFromThreadPool( + thread_pool.scheduleOrThrowOnError( [this, restore_query, restore_id, @@ -878,6 +880,8 @@ OperationID BackupsWorker::startRestoring(const ASTPtr & query, ContextMutablePt { try { + setThreadName("RestorerWorker"); + CurrentThread::QueryScope query_scope(context_in_use); doRestore( restore_query, restore_id, @@ -891,9 +895,7 @@ OperationID BackupsWorker::startRestoring(const ASTPtr & query, ContextMutablePt { on_exception(restore_id, backup_name_for_logging, restore_settings, restore_coordination); } - }, - thread_pool, - "RestoreWorker"); + }); } else { diff --git a/src/Backups/registerBackupEngineS3.cpp b/src/Backups/registerBackupEngineS3.cpp index 86941040021..fed5c6b4d22 100644 --- a/src/Backups/registerBackupEngineS3.cpp +++ b/src/Backups/registerBackupEngineS3.cpp @@ -15,8 +15,6 @@ namespace DB { -namespace fs = std::filesystem; - namespace ErrorCodes { extern const int BAD_ARGUMENTS; @@ -65,13 +63,13 @@ void registerBackupEngineS3(BackupFactory & factory) secret_access_key = config.getString(config_prefix + ".secret_access_key", ""); if (config.has(config_prefix + ".filename")) - s3_uri = fs::path(s3_uri) / config.getString(config_prefix + ".filename"); + s3_uri = std::filesystem::path(s3_uri) / config.getString(config_prefix + ".filename"); if (args.size() > 1) throw Exception(ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH, "Backup S3 requires 1 or 2 arguments: named_collection, [filename]"); if (args.size() == 1) - s3_uri = fs::path(s3_uri) / args[0].safeGet(); + s3_uri = std::filesystem::path(s3_uri) / args[0].safeGet(); } else { diff --git a/src/BridgeHelper/CatBoostLibraryBridgeHelper.cpp b/src/BridgeHelper/CatBoostLibraryBridgeHelper.cpp index 4db1eb720ac..bb9d890a691 100644 --- a/src/BridgeHelper/CatBoostLibraryBridgeHelper.cpp +++ b/src/BridgeHelper/CatBoostLibraryBridgeHelper.cpp @@ -58,8 +58,12 @@ bool CatBoostLibraryBridgeHelper::bridgeHandShake() String result; try { - ReadWriteBufferFromHTTP buf(getPingURI(), Poco::Net::HTTPRequest::HTTP_GET, {}, http_timeouts, credentials); - readString(result, buf); + auto buf = BuilderRWBufferFromHTTP(getPingURI()) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withTimeouts(http_timeouts) + .create(credentials); + + readString(result, *buf); } catch (...) { @@ -79,29 +83,29 @@ ExternalModelInfos CatBoostLibraryBridgeHelper::listModels() { startBridgeSync(); - ReadWriteBufferFromHTTP buf( - createRequestURI(CATBOOST_LIST_METHOD), - Poco::Net::HTTPRequest::HTTP_POST, - [](std::ostream &) {}, - http_timeouts, credentials); + auto buf = BuilderRWBufferFromHTTP(createRequestURI(CATBOOST_LIST_METHOD)) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(http_timeouts) + .create(credentials); ExternalModelInfos result; UInt64 num_rows; - readIntBinary(num_rows, buf); + readIntBinary(num_rows, *buf); for (UInt64 i = 0; i < num_rows; ++i) { ExternalModelInfo info; - readStringBinary(info.model_path, buf); - readStringBinary(info.model_type, buf); + readStringBinary(info.model_path, *buf); + readStringBinary(info.model_type, *buf); UInt64 t; - readIntBinary(t, buf); + readIntBinary(t, *buf); info.loading_start_time = std::chrono::system_clock::from_time_t(t); - readIntBinary(t, buf); + readIntBinary(t, *buf); info.loading_duration = std::chrono::milliseconds(t); result.push_back(info); @@ -116,17 +120,19 @@ void CatBoostLibraryBridgeHelper::removeModel() assert(model_path); - ReadWriteBufferFromHTTP buf( - createRequestURI(CATBOOST_REMOVEMODEL_METHOD), - Poco::Net::HTTPRequest::HTTP_POST, - [this](std::ostream & os) - { - os << "model_path=" << escapeForFileName(*model_path); - }, - http_timeouts, credentials); + auto buf = BuilderRWBufferFromHTTP(createRequestURI(CATBOOST_REMOVEMODEL_METHOD)) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(http_timeouts) + .withOutCallback( + [this](std::ostream & os) + { + os << "model_path=" << escapeForFileName(*model_path); + }) + .create(credentials); String result; - readStringBinary(result, buf); + readStringBinary(result, *buf); assert(result == "1"); } @@ -134,14 +140,14 @@ void CatBoostLibraryBridgeHelper::removeAllModels() { startBridgeSync(); - ReadWriteBufferFromHTTP buf( - createRequestURI(CATBOOST_REMOVEALLMODELS_METHOD), - Poco::Net::HTTPRequest::HTTP_POST, - [](std::ostream &){}, - http_timeouts, credentials); + auto buf = BuilderRWBufferFromHTTP(createRequestURI(CATBOOST_REMOVEALLMODELS_METHOD)) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(http_timeouts) + .create(credentials); String result; - readStringBinary(result, buf); + readStringBinary(result, *buf); assert(result == "1"); } @@ -151,18 +157,20 @@ size_t CatBoostLibraryBridgeHelper::getTreeCount() assert(model_path && library_path); - ReadWriteBufferFromHTTP buf( - createRequestURI(CATBOOST_GETTREECOUNT_METHOD), - Poco::Net::HTTPRequest::HTTP_POST, - [this](std::ostream & os) - { - os << "library_path=" << escapeForFileName(*library_path) << "&"; - os << "model_path=" << escapeForFileName(*model_path); - }, - http_timeouts, credentials); + auto buf = BuilderRWBufferFromHTTP(createRequestURI(CATBOOST_GETTREECOUNT_METHOD)) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(http_timeouts) + .withOutCallback( + [this](std::ostream & os) + { + os << "library_path=" << escapeForFileName(*library_path) << "&"; + os << "model_path=" << escapeForFileName(*model_path); + }) + .create(credentials); size_t result; - readIntBinary(result, buf); + readIntBinary(result, *buf); return result; } @@ -177,17 +185,19 @@ ColumnPtr CatBoostLibraryBridgeHelper::evaluate(const ColumnsWithTypeAndName & c assert(model_path); - ReadWriteBufferFromHTTP buf( - createRequestURI(CATBOOST_LIB_EVALUATE_METHOD), - Poco::Net::HTTPRequest::HTTP_POST, - [this, serialized = string_write_buf.str()](std::ostream & os) - { - os << "model_path=" << escapeForFileName(*model_path) << "&"; - os << "data=" << escapeForFileName(serialized); - }, - http_timeouts, credentials); + auto buf = BuilderRWBufferFromHTTP(createRequestURI(CATBOOST_LIB_EVALUATE_METHOD)) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(http_timeouts) + .withOutCallback( + [this, serialized = string_write_buf.str()](std::ostream & os) + { + os << "model_path=" << escapeForFileName(*model_path) << "&"; + os << "data=" << escapeForFileName(serialized); + }) + .create(credentials); - NativeReader deserializer(buf, /*server_revision*/ 0); + NativeReader deserializer(*buf, /*server_revision*/ 0); Block block_read = deserializer.read(); return block_read.getColumns()[0]; diff --git a/src/BridgeHelper/ExternalDictionaryLibraryBridgeHelper.cpp b/src/BridgeHelper/ExternalDictionaryLibraryBridgeHelper.cpp index fcb8ebd1f22..a37d2bd396f 100644 --- a/src/BridgeHelper/ExternalDictionaryLibraryBridgeHelper.cpp +++ b/src/BridgeHelper/ExternalDictionaryLibraryBridgeHelper.cpp @@ -71,8 +71,12 @@ bool ExternalDictionaryLibraryBridgeHelper::bridgeHandShake() String result; try { - ReadWriteBufferFromHTTP buf(getPingURI(), Poco::Net::HTTPRequest::HTTP_GET, {}, http_timeouts, credentials); - readString(result, buf); + auto buf = BuilderRWBufferFromHTTP(getPingURI()) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withTimeouts(http_timeouts) + .create(credentials); + + readString(result, *buf); } catch (...) { @@ -247,30 +251,28 @@ QueryPipeline ExternalDictionaryLibraryBridgeHelper::loadKeys(const Block & requ bool ExternalDictionaryLibraryBridgeHelper::executeRequest(const Poco::URI & uri, ReadWriteBufferFromHTTP::OutStreamCallback out_stream_callback) const { - ReadWriteBufferFromHTTP buf( - uri, - Poco::Net::HTTPRequest::HTTP_POST, - std::move(out_stream_callback), - http_timeouts, credentials); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(http_timeouts) + .withOutCallback(std::move(out_stream_callback)) + .create(credentials); bool res; - readBoolText(res, buf); + readBoolText(res, *buf); return res; } QueryPipeline ExternalDictionaryLibraryBridgeHelper::loadBase(const Poco::URI & uri, ReadWriteBufferFromHTTP::OutStreamCallback out_stream_callback) { - auto read_buf_ptr = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_POST, - std::move(out_stream_callback), - http_timeouts, - credentials, - 0, - DBMS_DEFAULT_BUFFER_SIZE, - getContext()->getReadSettings(), - HTTPHeaderEntries{}); + auto read_buf_ptr = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withSettings(getContext()->getReadSettings()) + .withTimeouts(http_timeouts) + .withOutCallback(std::move(out_stream_callback)) + .create(credentials); auto source = FormatFactory::instance().getInput(ExternalDictionaryLibraryBridgeHelper::DEFAULT_FORMAT, *read_buf_ptr, sample_block, getContext(), DEFAULT_BLOCK_SIZE); source->addBuffer(std::move(read_buf_ptr)); diff --git a/src/BridgeHelper/XDBCBridgeHelper.h b/src/BridgeHelper/XDBCBridgeHelper.h index d208b8ddab0..838be42357a 100644 --- a/src/BridgeHelper/XDBCBridgeHelper.h +++ b/src/BridgeHelper/XDBCBridgeHelper.h @@ -97,8 +97,12 @@ protected: { try { - ReadWriteBufferFromHTTP buf(getPingURI(), Poco::Net::HTTPRequest::HTTP_GET, {}, getHTTPTimeouts(), credentials); - return checkString(PING_OK_ANSWER, buf); + auto buf = BuilderRWBufferFromHTTP(getPingURI()) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withTimeouts(getHTTPTimeouts()) + .create(credentials); + + return checkString(PING_OK_ANSWER, *buf); } catch (...) { @@ -198,10 +202,14 @@ protected: uri.addQueryParameter("connection_string", getConnectionString()); uri.addQueryParameter("use_connection_pooling", toString(use_connection_pooling)); - ReadWriteBufferFromHTTP buf(uri, Poco::Net::HTTPRequest::HTTP_POST, {}, getHTTPTimeouts(), credentials); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(getHTTPTimeouts()) + .create(credentials); - bool res; - readBoolText(res, buf); + bool res = false; + readBoolText(res, *buf); is_schema_allowed = res; } @@ -220,10 +228,14 @@ protected: uri.addQueryParameter("connection_string", getConnectionString()); uri.addQueryParameter("use_connection_pooling", toString(use_connection_pooling)); - ReadWriteBufferFromHTTP buf(uri, Poco::Net::HTTPRequest::HTTP_POST, {}, getHTTPTimeouts(), credentials); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(getHTTPTimeouts()) + .create(credentials); std::string character; - readStringBinary(character, buf); + readStringBinary(character, *buf); if (character.length() > 1) throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Failed to parse quoting style from '{}' for service {}", character, BridgeHelperMixin::serviceAlias()); diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index dff70e06ce4..73aa409e995 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -174,6 +174,8 @@ endif () add_library(clickhouse_common_io ${clickhouse_common_io_headers} ${clickhouse_common_io_sources}) +set_source_files_properties(Common/ThreadFuzzer.cpp PROPERTIES COMPILE_FLAGS "-fomit-frame-pointer -momit-leaf-frame-pointer") + add_library (clickhouse_malloc OBJECT Common/malloc.cpp) set_source_files_properties(Common/malloc.cpp PROPERTIES COMPILE_FLAGS "-fno-builtin") diff --git a/src/Client/Connection.cpp b/src/Client/Connection.cpp index 4848f4c7ee5..180942e6b83 100644 --- a/src/Client/Connection.cpp +++ b/src/Client/Connection.cpp @@ -153,6 +153,12 @@ void Connection::connect(const ConnectionTimeouts & timeouts) current_resolved_address = *it; break; } + catch (DB::NetException &) + { + if (++it == addresses.end()) + throw; + continue; + } catch (Poco::Net::NetException &) { if (++it == addresses.end()) @@ -199,6 +205,17 @@ void Connection::connect(const ConnectionTimeouts & timeouts) LOG_TRACE(log_wrapper.get(), "Connected to {} server version {}.{}.{}.", server_name, server_version_major, server_version_minor, server_version_patch); } + catch (DB::NetException & e) + { + disconnect(); + + /// Remove this possible stale entry from cache + DNSResolver::instance().removeHostFromCache(host); + + /// Add server address to exception. Exception will preserve stack trace. + e.addMessage("({})", getDescription()); + throw; + } catch (Poco::Net::NetException & e) { disconnect(); @@ -206,7 +223,7 @@ void Connection::connect(const ConnectionTimeouts & timeouts) /// Remove this possible stale entry from cache DNSResolver::instance().removeHostFromCache(host); - /// Add server address to exception. Also Exception will remember stack trace. It's a pity that more precise exception type is lost. + /// Add server address to exception. Also Exception will remember new stack trace. It's a pity that more precise exception type is lost. throw NetException(ErrorCodes::NETWORK_ERROR, "{} ({})", e.displayText(), getDescription()); } catch (Poco::TimeoutException & e) @@ -216,7 +233,7 @@ void Connection::connect(const ConnectionTimeouts & timeouts) /// Remove this possible stale entry from cache DNSResolver::instance().removeHostFromCache(host); - /// Add server address to exception. Also Exception will remember stack trace. It's a pity that more precise exception type is lost. + /// Add server address to exception. Also Exception will remember new stack trace. It's a pity that more precise exception type is lost. /// This exception can only be thrown from socket->connect(), so add information about connection timeout. const auto & connection_timeout = static_cast(secure) ? timeouts.secure_connection_timeout : timeouts.connection_timeout; throw NetException( diff --git a/src/Client/ConnectionParameters.cpp b/src/Client/ConnectionParameters.cpp index 5c22b6c6d3f..16911f97e84 100644 --- a/src/Client/ConnectionParameters.cpp +++ b/src/Client/ConnectionParameters.cpp @@ -115,7 +115,7 @@ ConnectionParameters::ConnectionParameters(const Poco::Util::AbstractConfigurati /// At the same time, I want clickhouse-local to always work, regardless. /// TODO: get rid of glibc, or replace getaddrinfo to c-ares. - compression = config.getBool("compression", host != "localhost" && !isLocalAddress(DNSResolver::instance().resolveHost(host))) + compression = config.getBool("compression", host != "localhost" && !isLocalAddress(DNSResolver::instance().resolveHostAllInOriginOrder(host).front())) ? Protocol::Compression::Enable : Protocol::Compression::Disable; timeouts = ConnectionTimeouts() diff --git a/src/Client/QueryFuzzer.cpp b/src/Client/QueryFuzzer.cpp index 0597a7c1eed..7f1dce4b29a 100644 --- a/src/Client/QueryFuzzer.cpp +++ b/src/Client/QueryFuzzer.cpp @@ -232,7 +232,7 @@ ASTPtr QueryFuzzer::getRandomColumnLike() return nullptr; } - ASTPtr new_ast = column_like[fuzz_rand() % column_like.size()]->clone(); + ASTPtr new_ast = column_like[fuzz_rand() % column_like.size()].second->clone(); new_ast->setAlias(""); return new_ast; @@ -272,7 +272,7 @@ void QueryFuzzer::replaceWithTableLike(ASTPtr & ast) return; } - ASTPtr new_ast = table_like[fuzz_rand() % table_like.size()]->clone(); + ASTPtr new_ast = table_like[fuzz_rand() % table_like.size()].second->clone(); std::string old_alias = ast->tryGetAlias(); new_ast->setAlias(old_alias); @@ -1214,57 +1214,46 @@ void QueryFuzzer::fuzz(ASTPtr & ast) } } +#define AST_FUZZER_PART_TYPE_CAP 1000 + /* * This functions collects various parts of query that we can then substitute * to a query being fuzzed. - * - * TODO: we just stop remembering new parts after our corpus reaches certain size. - * This is boring, should implement a random replacement of existing parst with - * small probability. Do this after we add this fuzzer to CI and fix all the - * problems it can routinely find even in this boring version. */ void QueryFuzzer::collectFuzzInfoMain(ASTPtr ast) { collectFuzzInfoRecurse(ast); - - aliases.clear(); - for (const auto & alias : aliases_set) - { - aliases.push_back(alias); - } - - column_like.clear(); - for (const auto & [name, value] : column_like_map) - { - column_like.push_back(value); - } - - table_like.clear(); - for (const auto & [name, value] : table_like_map) - { - table_like.push_back(value); - } } void QueryFuzzer::addTableLike(ASTPtr ast) { - if (table_like_map.size() > 1000) + if (table_like_map.size() > AST_FUZZER_PART_TYPE_CAP) { - table_like_map.clear(); + const auto iter = std::next(table_like.begin(), fuzz_rand() % table_like.size()); + const auto ast_del = *iter; + table_like.erase(iter); + table_like_map.erase(ast_del.first); } const auto name = ast->formatForErrorMessage(); if (name.size() < 200) { - table_like_map.insert({name, ast}); + const auto res = table_like_map.insert({name, ast}); + if (res.second) + { + table_like.push_back({name, ast}); + } } } void QueryFuzzer::addColumnLike(ASTPtr ast) { - if (column_like_map.size() > 1000) + if (column_like_map.size() > AST_FUZZER_PART_TYPE_CAP) { - column_like_map.clear(); + const auto iter = std::next(column_like.begin(), fuzz_rand() % column_like.size()); + const auto ast_del = *iter; + column_like.erase(iter); + column_like_map.erase(ast_del.first); } const auto name = ast->formatForErrorMessage(); @@ -1279,22 +1268,16 @@ void QueryFuzzer::addColumnLike(ASTPtr ast) } if (name.size() < 200) { - column_like_map.insert({name, ast}); + const auto res = column_like_map.insert({name, ast}); + if (res.second) + { + column_like.push_back({name, ast}); + } } } void QueryFuzzer::collectFuzzInfoRecurse(ASTPtr ast) { - if (auto * impl = dynamic_cast(ast.get())) - { - if (aliases_set.size() > 1000) - { - aliases_set.clear(); - } - - aliases_set.insert(impl->alias); - } - if (typeid_cast(ast.get())) { addColumnLike(ast); diff --git a/src/Client/QueryFuzzer.h b/src/Client/QueryFuzzer.h index f5465626d96..6165e589cae 100644 --- a/src/Client/QueryFuzzer.h +++ b/src/Client/QueryFuzzer.h @@ -50,14 +50,12 @@ struct QueryFuzzer // we are currently fuzzing. We add some part from each new query we are asked // to fuzz, and keep this state between queries, so the fuzzing output becomes // more interesting over time, as the queries mix. - std::unordered_set aliases_set; - std::vector aliases; - + // The hash tables are used for collection, and the vectors are used for random access. std::unordered_map column_like_map; - std::vector column_like; + std::vector> column_like; std::unordered_map table_like_map; - std::vector table_like; + std::vector> table_like; // Some debug fields for detecting problematic ASTs with loops. // These are reset for each fuzzMain call. diff --git a/src/Columns/ColumnUnique.cpp b/src/Columns/ColumnUnique.cpp new file mode 100644 index 00000000000..edfee69a752 --- /dev/null +++ b/src/Columns/ColumnUnique.cpp @@ -0,0 +1,25 @@ +#include + +namespace DB +{ + +/// Explicit template instantiations. +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; +template class ColumnUnique; + +} diff --git a/src/Columns/ColumnUnique.h b/src/Columns/ColumnUnique.h index 34f1ffc15cd..76bbbbacdbf 100644 --- a/src/Columns/ColumnUnique.h +++ b/src/Columns/ColumnUnique.h @@ -15,6 +15,8 @@ #include #include #include +#include "Columns/ColumnsDateTime.h" +#include "Columns/ColumnsNumber.h" #include #include @@ -736,4 +738,23 @@ UInt128 ColumnUnique::IncrementalHash::getHash(const ColumnType & co return cur_hash; } + +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; +extern template class ColumnUnique; + } diff --git a/src/Common/CurrentMetrics.cpp b/src/Common/CurrentMetrics.cpp index 82da4c4bbad..dfbf6199361 100644 --- a/src/Common/CurrentMetrics.cpp +++ b/src/Common/CurrentMetrics.cpp @@ -274,7 +274,19 @@ M(DistrCacheUsedConnections, "Number of currently used connections to Distributed Cache") \ M(DistrCacheReadRequests, "Number of executed Read requests to Distributed Cache") \ M(DistrCacheWriteRequests, "Number of executed Write requests to Distributed Cache") \ - M(DistrCacheServerConnections, "Number of open connections to ClickHouse server from Distributed Cache") + M(DistrCacheServerConnections, "Number of open connections to ClickHouse server from Distributed Cache") \ + \ + M(StorageConnectionsStored, "Total count of sessions stored in the session pool for storages") \ + M(StorageConnectionsTotal, "Total count of all sessions: stored in the pool and actively used right now for storages") \ + \ + M(DiskConnectionsStored, "Total count of sessions stored in the session pool for disks") \ + M(DiskConnectionsTotal, "Total count of all sessions: stored in the pool and actively used right now for disks") \ + \ + M(HTTPConnectionsStored, "Total count of sessions stored in the session pool for http hosts") \ + M(HTTPConnectionsTotal, "Total count of all sessions: stored in the pool and actively used right now for http hosts") \ + \ + M(AddressesActive, "Total count of addresses which are used for creation connections with connection pools") \ + #ifdef APPLY_FOR_EXTERNAL_METRICS #define APPLY_FOR_METRICS(M) APPLY_FOR_BUILTIN_METRICS(M) APPLY_FOR_EXTERNAL_METRICS(M) diff --git a/src/Common/DNSResolver.cpp b/src/Common/DNSResolver.cpp index 250a8b3fc49..e36e1483da8 100644 --- a/src/Common/DNSResolver.cpp +++ b/src/Common/DNSResolver.cpp @@ -1,6 +1,7 @@ #include "DNSResolver.h" #include #include +#include #include #include #include @@ -108,7 +109,7 @@ DNSResolver::IPAddresses hostByName(const std::string & host) if (addresses.empty()) { ProfileEvents::increment(ProfileEvents::DNSError); - throw Exception(ErrorCodes::DNS_ERROR, "Not found address of host: {}", host); + throw DB::NetException(ErrorCodes::DNS_ERROR, "Not found address of host: {}", host); } return addresses; @@ -202,10 +203,10 @@ DNSResolver::DNSResolver() : impl(std::make_unique()), log(ge Poco::Net::IPAddress DNSResolver::resolveHost(const std::string & host) { - return pickAddress(resolveHostAll(host)); + return pickAddress(resolveHostAll(host)); // random order -> random pick } -DNSResolver::IPAddresses DNSResolver::resolveHostAll(const std::string & host) +DNSResolver::IPAddresses DNSResolver::resolveHostAllInOriginOrder(const std::string & host) { if (impl->disable_cache) return resolveIPAddressImpl(host); @@ -214,6 +215,13 @@ DNSResolver::IPAddresses DNSResolver::resolveHostAll(const std::string & host) return resolveIPAddressWithCache(impl->cache_host, host); } +DNSResolver::IPAddresses DNSResolver::resolveHostAll(const std::string & host) +{ + auto addresses = resolveHostAllInOriginOrder(host); + std::shuffle(addresses.begin(), addresses.end(), thread_local_rng); + return addresses; +} + Poco::Net::SocketAddress DNSResolver::resolveAddress(const std::string & host_and_port) { if (impl->disable_cache) diff --git a/src/Common/DNSResolver.h b/src/Common/DNSResolver.h index 27d81c9442a..e3030e51a96 100644 --- a/src/Common/DNSResolver.h +++ b/src/Common/DNSResolver.h @@ -34,6 +34,9 @@ public: Poco::Net::IPAddress resolveHost(const std::string & host); /// Accepts host names like 'example.com' or '127.0.0.1' or '::1' and resolves all its IPs + /// resolveHostAllInOriginOrder returns addresses with the same order as system call returns it + IPAddresses resolveHostAllInOriginOrder(const std::string & host); + /// resolveHostAll returns addresses in random order IPAddresses resolveHostAll(const std::string & host); /// Accepts host names like 'example.com:port' or '127.0.0.1:port' or '[::1]:port' and resolves its IP and port diff --git a/src/Common/DateLUTImpl.h b/src/Common/DateLUTImpl.h index 0e72b489ace..7bf66c0504a 100644 --- a/src/Common/DateLUTImpl.h +++ b/src/Common/DateLUTImpl.h @@ -3,13 +3,13 @@ #include #include #include +#include #include #include #include #include - #define DATE_SECONDS_PER_DAY 86400 /// Number of seconds in a day, 60 * 60 * 24 #define DATE_LUT_MIN_YEAR 1900 /// 1900 since majority of financial organizations consider 1900 as an initial year. @@ -280,9 +280,9 @@ private: static_assert(std::is_integral_v && std::is_integral_v); assert(divisor > 0); - if (likely(offset_is_whole_number_of_hours_during_epoch)) + if (offset_is_whole_number_of_hours_during_epoch) [[likely]] { - if (likely(x >= 0)) + if (x >= 0) [[likely]] return static_cast(x / divisor * divisor); /// Integer division for negative numbers rounds them towards zero (up). @@ -576,10 +576,10 @@ public: unsigned toSecond(Time t) const { - if (likely(offset_is_whole_number_of_minutes_during_epoch)) + if (offset_is_whole_number_of_minutes_during_epoch) [[likely]] { Time res = t % 60; - if (likely(res >= 0)) + if (res >= 0) [[likely]] return static_cast(res); return static_cast(res) + 60; } @@ -593,6 +593,30 @@ public: return time % 60; } + template + unsigned toMillisecond(const DateOrTime & datetime, Int64 scale_multiplier) const + { + constexpr Int64 millisecond_multiplier = 1'000; + constexpr Int64 microsecond_multiplier = 1'000 * millisecond_multiplier; + constexpr Int64 divider = microsecond_multiplier / millisecond_multiplier; + + auto components = DB::DecimalUtils::splitWithScaleMultiplier(datetime, scale_multiplier); + + if (datetime.value < 0 && components.fractional) + { + components.fractional = scale_multiplier + (components.whole ? Int64(-1) : Int64(1)) * components.fractional; + --components.whole; + } + Int64 fractional = components.fractional; + if (scale_multiplier > microsecond_multiplier) + fractional = fractional / (scale_multiplier / microsecond_multiplier); + else if (scale_multiplier < microsecond_multiplier) + fractional = fractional * (microsecond_multiplier / scale_multiplier); + + UInt16 millisecond = static_cast(fractional / divider); + return millisecond; + } + unsigned toMinute(Time t) const { if (t >= 0 && offset_is_whole_number_of_hours_during_epoch) @@ -1122,9 +1146,9 @@ public: DateOrTime toStartOfMinuteInterval(DateOrTime t, UInt64 minutes) const { Int64 divisor = 60 * minutes; - if (likely(offset_is_whole_number_of_minutes_during_epoch)) + if (offset_is_whole_number_of_minutes_during_epoch) [[likely]] { - if (likely(t >= 0)) + if (t >= 0) [[likely]] return static_cast(t / divisor * divisor); return static_cast((t + 1 - divisor) / divisor * divisor); } @@ -1339,7 +1363,7 @@ public: UInt8 saturateDayOfMonth(Int16 year, UInt8 month, UInt8 day_of_month) const { - if (likely(day_of_month <= 28)) + if (day_of_month <= 28) [[likely]] return day_of_month; UInt8 days_in_month = daysInMonth(year, month); diff --git a/src/Common/FieldVisitorConvertToNumber.cpp b/src/Common/FieldVisitorConvertToNumber.cpp new file mode 100644 index 00000000000..75b3fbfe02a --- /dev/null +++ b/src/Common/FieldVisitorConvertToNumber.cpp @@ -0,0 +1,23 @@ +#include +#include "base/Decimal.h" + +namespace DB +{ + +/// Explicit template instantiations. +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; +template class FieldVisitorConvertToNumber; + +} diff --git a/src/Common/FieldVisitorConvertToNumber.h b/src/Common/FieldVisitorConvertToNumber.h index 47a1e669969..86e03cb5d77 100644 --- a/src/Common/FieldVisitorConvertToNumber.h +++ b/src/Common/FieldVisitorConvertToNumber.h @@ -117,4 +117,19 @@ public: T operator() (const bool & x) const { return T(x); } }; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; +extern template class FieldVisitorConvertToNumber; + } diff --git a/src/Common/HTTPConnectionPool.cpp b/src/Common/HTTPConnectionPool.cpp new file mode 100644 index 00000000000..a21438a11a2 --- /dev/null +++ b/src/Common/HTTPConnectionPool.cpp @@ -0,0 +1,865 @@ +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include + +#include +#include + +#include "config.h" + +#if USE_SSL +#include +#endif + + +namespace ProfileEvents +{ + extern const Event StorageConnectionsCreated; + extern const Event StorageConnectionsReused; + extern const Event StorageConnectionsReset; + extern const Event StorageConnectionsPreserved; + extern const Event StorageConnectionsExpired; + extern const Event StorageConnectionsErrors; + extern const Event StorageConnectionsElapsedMicroseconds; + + extern const Event DiskConnectionsCreated; + extern const Event DiskConnectionsReused; + extern const Event DiskConnectionsReset; + extern const Event DiskConnectionsPreserved; + extern const Event DiskConnectionsExpired; + extern const Event DiskConnectionsErrors; + extern const Event DiskConnectionsElapsedMicroseconds; + + extern const Event HTTPConnectionsCreated; + extern const Event HTTPConnectionsReused; + extern const Event HTTPConnectionsReset; + extern const Event HTTPConnectionsPreserved; + extern const Event HTTPConnectionsExpired; + extern const Event HTTPConnectionsErrors; + extern const Event HTTPConnectionsElapsedMicroseconds; +} + + +namespace CurrentMetrics +{ + extern const Metric StorageConnectionsStored; + extern const Metric StorageConnectionsTotal; + + extern const Metric DiskConnectionsStored; + extern const Metric DiskConnectionsTotal; + + extern const Metric HTTPConnectionsStored; + extern const Metric HTTPConnectionsTotal; +} + +namespace +{ + Poco::Net::HTTPClientSession::ProxyConfig proxyConfigurationToPocoProxyConfig(const DB::ProxyConfiguration & proxy_configuration) + { + Poco::Net::HTTPClientSession::ProxyConfig poco_proxy_config; + + poco_proxy_config.host = proxy_configuration.host; + poco_proxy_config.port = proxy_configuration.port; + poco_proxy_config.protocol = DB::ProxyConfiguration::protocolToString(proxy_configuration.protocol); + poco_proxy_config.tunnel = proxy_configuration.tunneling; + poco_proxy_config.originalRequestProtocol = DB::ProxyConfiguration::protocolToString(proxy_configuration.original_request_protocol); + + return poco_proxy_config; + } + + + size_t roundUp(size_t x, size_t rounding) + { + chassert(rounding > 0); + return (x + (rounding - 1)) / rounding * rounding; + } + + + Poco::Timespan divide(const Poco::Timespan span, int divisor) + { + return Poco::Timespan(Poco::Timestamp::TimeDiff(span.totalMicroseconds() / divisor)); + } +} + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int SUPPORT_IS_DISABLED; + extern const int UNSUPPORTED_URI_SCHEME; +} + + +IHTTPConnectionPoolForEndpoint::Metrics getMetricsForStorageConnectionPool() +{ + return IHTTPConnectionPoolForEndpoint::Metrics{ + .created = ProfileEvents::StorageConnectionsCreated, + .reused = ProfileEvents::StorageConnectionsReused, + .reset = ProfileEvents::StorageConnectionsReset, + .preserved = ProfileEvents::StorageConnectionsPreserved, + .expired = ProfileEvents::StorageConnectionsExpired, + .errors = ProfileEvents::StorageConnectionsErrors, + .elapsed_microseconds = ProfileEvents::StorageConnectionsElapsedMicroseconds, + .stored_count = CurrentMetrics::StorageConnectionsStored, + .active_count = CurrentMetrics::StorageConnectionsTotal, + }; +} + + +IHTTPConnectionPoolForEndpoint::Metrics getMetricsForDiskConnectionPool() +{ + return IHTTPConnectionPoolForEndpoint::Metrics{ + .created = ProfileEvents::DiskConnectionsCreated, + .reused = ProfileEvents::DiskConnectionsReused, + .reset = ProfileEvents::DiskConnectionsReset, + .preserved = ProfileEvents::DiskConnectionsPreserved, + .expired = ProfileEvents::DiskConnectionsExpired, + .errors = ProfileEvents::DiskConnectionsErrors, + .elapsed_microseconds = ProfileEvents::DiskConnectionsElapsedMicroseconds, + .stored_count = CurrentMetrics::DiskConnectionsStored, + .active_count = CurrentMetrics::DiskConnectionsTotal, + }; +} + + +IHTTPConnectionPoolForEndpoint::Metrics getMetricsForHTTPConnectionPool() +{ + return IHTTPConnectionPoolForEndpoint::Metrics{ + .created = ProfileEvents::HTTPConnectionsCreated, + .reused = ProfileEvents::HTTPConnectionsReused, + .reset = ProfileEvents::HTTPConnectionsReset, + .preserved = ProfileEvents::HTTPConnectionsPreserved, + .expired = ProfileEvents::HTTPConnectionsExpired, + .errors = ProfileEvents::HTTPConnectionsErrors, + .elapsed_microseconds = ProfileEvents::HTTPConnectionsElapsedMicroseconds, + .stored_count = CurrentMetrics::HTTPConnectionsStored, + .active_count = CurrentMetrics::HTTPConnectionsTotal, + }; +} + + +IHTTPConnectionPoolForEndpoint::Metrics getConnectionPoolMetrics(HTTPConnectionGroupType type) +{ + switch (type) + { + case HTTPConnectionGroupType::STORAGE: + return getMetricsForStorageConnectionPool(); + case HTTPConnectionGroupType::DISK: + return getMetricsForDiskConnectionPool(); + case HTTPConnectionGroupType::HTTP: + return getMetricsForHTTPConnectionPool(); + } +} + + +class ConnectionGroup +{ +public: + using Ptr = std::shared_ptr; + + explicit ConnectionGroup(HTTPConnectionGroupType type_) : type(type_), metrics(getConnectionPoolMetrics(type_)) { } + + void setLimits(HTTPConnectionPools::Limits limits_) + { + std::lock_guard lock(mutex); + limits = std::move(limits_); + mute_warning_until = 0; + } + + bool isSoftLimitReached() const + { + std::lock_guard lock(mutex); + return total_connections_in_group >= limits.soft_limit; + } + + bool isStoreLimitReached() const + { + std::lock_guard lock(mutex); + return total_connections_in_group >= limits.store_limit; + } + + void atConnectionCreate() + { + std::lock_guard lock(mutex); + + ++total_connections_in_group; + + if (total_connections_in_group >= limits.warning_limit && total_connections_in_group >= mute_warning_until) + { + LOG_WARNING(log, "Too many active sessions in group {}, count {}, warning limit {}", type, total_connections_in_group, limits.warning_limit); + mute_warning_until = roundUp(total_connections_in_group, limits.warning_step); + } + } + + void atConnectionDestroy() + { + std::lock_guard lock(mutex); + + --total_connections_in_group; + + const size_t reduced_warning_limit = limits.warning_limit > 10 ? limits.warning_limit - 10 : 1; + if (mute_warning_until > 0 && total_connections_in_group < reduced_warning_limit) + { + LOG_WARNING(log, "Sessions count is OK in the group {}, count {}", type, total_connections_in_group); + mute_warning_until = 0; + } + } + + HTTPConnectionGroupType getType() const { return type; } + + const IHTTPConnectionPoolForEndpoint::Metrics & getMetrics() const { return metrics; } + +private: + const HTTPConnectionGroupType type; + const IHTTPConnectionPoolForEndpoint::Metrics metrics; + + LoggerPtr log = getLogger("ConnectionGroup"); + + mutable std::mutex mutex; + HTTPConnectionPools::Limits limits TSA_GUARDED_BY(mutex) = HTTPConnectionPools::Limits(); + size_t total_connections_in_group TSA_GUARDED_BY(mutex) = 0; + size_t mute_warning_until TSA_GUARDED_BY(mutex) = 0; +}; + + +class IExtendedPool : public IHTTPConnectionPoolForEndpoint +{ +public: + using Ptr = std::shared_ptr; + + virtual HTTPConnectionGroupType getGroupType() const = 0; + virtual size_t wipeExpired() = 0; +}; + + +// EndpointConnectionPool manage connections to the endpoint +// Features: +// - it uses HostResolver for address selecting. See Common/HostResolver.h for more info. +// - it minimizes number of `Session::connect()`/`Session::reconnect()` calls +// - stores only connected and ready to use sessions +// - connection could be reused even when limits are reached +// - soft limit, warn limit, store limit +// - `Session::reconnect()` uses the pool as well +// - comprehensive sensors +// - session is reused according its inner state, automatically + + +template +class EndpointConnectionPool : public std::enable_shared_from_this>, public IExtendedPool +{ +private: + friend class HTTPConnectionPools; + + using WeakPtr = std::weak_ptr>; + + class PooledConnection : public Session + { + public: + using Ptr = std::shared_ptr; + + void reconnect() override + { + ProfileEvents::increment(metrics.reset); + Session::close(); + + if (auto lock = pool.lock()) + { + auto timeouts = getTimeouts(*this); + auto new_connection = lock->getConnection(timeouts); + Session::assign(*new_connection); + } + else + { + auto timer = CurrentThread::getProfileEvents().timer(metrics.elapsed_microseconds); + Session::reconnect(); + ProfileEvents::increment(metrics.created); + } + } + + String getTarget() const + { + if (!Session::getProxyConfig().host.empty()) + return fmt::format("{} over proxy {}", Session::getHost(), Session::getProxyConfig().host); + return Session::getHost(); + } + + void flushRequest() override + { + if (bool(request_stream)) + { + request_stream->flush(); + + if (auto * fixed_steam = dynamic_cast(request_stream)) + { + request_stream_completed = fixed_steam->isComplete(); + } + else if (auto * chunked_steam = dynamic_cast(request_stream)) + { + chunked_steam->rdbuf()->close(); + request_stream_completed = chunked_steam->isComplete(); + } + else if (auto * http_stream = dynamic_cast(request_stream)) + { + request_stream_completed = http_stream->isComplete(); + } + else + { + request_stream_completed = false; + } + } + request_stream = nullptr; + + Session::flushRequest(); + } + + std::ostream & sendRequest(Poco::Net::HTTPRequest & request) override + { + std::ostream & result = Session::sendRequest(request); + result.exceptions(std::ios::badbit); + + request_stream = &result; + request_stream_completed = false; + + response_stream = nullptr; + response_stream_completed = false; + + return result; + } + + std::istream & receiveResponse(Poco::Net::HTTPResponse & response) override + { + std::istream & result = Session::receiveResponse(response); + result.exceptions(std::ios::badbit); + + response_stream = &result; + response_stream_completed = false; + + return result; + } + + void reset() override + { + request_stream = nullptr; + request_stream_completed = false; + + response_stream = nullptr; + response_stream_completed = false; + + Session::reset(); + } + + ~PooledConnection() override + { + if (bool(response_stream)) + { + if (auto * fixed_steam = dynamic_cast(response_stream)) + { + response_stream_completed = fixed_steam->isComplete(); + } + else if (auto * chunked_steam = dynamic_cast(response_stream)) + { + response_stream_completed = chunked_steam->isComplete(); + } + else if (auto * http_stream = dynamic_cast(response_stream)) + { + response_stream_completed = http_stream->isComplete(); + } + else + { + response_stream_completed = false; + } + } + response_stream = nullptr; + + if (auto lock = pool.lock()) + lock->atConnectionDestroy(*this); + else + ProfileEvents::increment(metrics.reset); + + CurrentMetrics::sub(metrics.active_count); + } + + private: + friend class EndpointConnectionPool; + + template + explicit PooledConnection(EndpointConnectionPool::WeakPtr pool_, IHTTPConnectionPoolForEndpoint::Metrics metrics_, Args &&... args) + : Session(args...), pool(std::move(pool_)), metrics(std::move(metrics_)) + { + CurrentMetrics::add(metrics.active_count); + } + + template + static Ptr create(Args &&... args) + { + /// Pool is global, we shouldn't attribute this memory to query/user. + MemoryTrackerSwitcher switcher{&total_memory_tracker}; + + struct make_shared_enabler : public PooledConnection + { + explicit make_shared_enabler(Args &&... args) : PooledConnection(std::forward(args)...) { } + }; + return std::make_shared(std::forward(args)...); + } + + void doConnect() + { + Session::reconnect(); + } + + bool isCompleted() const + { + return request_stream_completed && response_stream_completed; + } + + WeakPtr pool; + IHTTPConnectionPoolForEndpoint::Metrics metrics; + + Poco::Logger * log = &Poco::Logger::get("PooledConnection"); + + std::ostream * request_stream = nullptr; + std::istream * response_stream = nullptr; + + bool request_stream_completed = true; + bool response_stream_completed = true; + }; + + using Connection = PooledConnection; + using ConnectionPtr = PooledConnection::Ptr; + + struct GreaterByLastRequest + { + static bool operator()(const ConnectionPtr & l, const ConnectionPtr & r) + { + return l->getLastRequest() + l->getKeepAliveTimeout() > r->getLastRequest() + r->getKeepAliveTimeout(); + } + }; + + using ConnectionsMinHeap = std::priority_queue, GreaterByLastRequest>; + +public: + template + static Ptr create(Args &&... args) + { + struct make_shared_enabler : public EndpointConnectionPool + { + explicit make_shared_enabler(Args &&... args) : EndpointConnectionPool(std::forward(args)...) { } + }; + return std::make_shared(std::forward(args)...); + } + + ~EndpointConnectionPool() override + { + CurrentMetrics::sub(group->getMetrics().stored_count, stored_connections.size()); + } + + String getTarget() const + { + if (!proxy_configuration.isEmpty()) + return fmt::format("{} over proxy {}", host, proxy_configuration.host); + return host; + } + + IHTTPConnectionPoolForEndpoint::ConnectionPtr getConnection(const ConnectionTimeouts & timeouts) override + { + Poco::Timestamp now; + std::vector expired_connections; + + SCOPE_EXIT({ + MemoryTrackerSwitcher switcher{&total_memory_tracker}; + expired_connections.clear(); + }); + + { + std::lock_guard lock(mutex); + + wipeExpiredImpl(expired_connections, now); + + if (!stored_connections.empty()) + { + auto it = stored_connections.top(); + stored_connections.pop(); + + setTimeouts(*it, timeouts); + + ProfileEvents::increment(getMetrics().reused, 1); + CurrentMetrics::sub(getMetrics().stored_count, 1); + + return it; + } + } + + return prepareNewConnection(timeouts); + } + + const IHTTPConnectionPoolForEndpoint::Metrics & getMetrics() const override + { + return group->getMetrics(); + } + + HTTPConnectionGroupType getGroupType() const override + { + return group->getType(); + } + + size_t wipeExpired() override + { + Poco::Timestamp now; + std::vector expired_connections; + + SCOPE_EXIT({ + MemoryTrackerSwitcher switcher{&total_memory_tracker}; + expired_connections.clear(); + }); + + std::lock_guard lock(mutex); + return wipeExpiredImpl(expired_connections, now); + } + + size_t wipeExpiredImpl(std::vector & expired_connections, Poco::Timestamp now) TSA_REQUIRES(mutex) + { + while (!stored_connections.empty()) + { + auto connection = stored_connections.top(); + + if (!isExpired(now, connection)) + return stored_connections.size(); + + stored_connections.pop(); + expired_connections.push_back(connection); + } + + CurrentMetrics::sub(getMetrics().stored_count, expired_connections.size()); + ProfileEvents::increment(getMetrics().expired, expired_connections.size()); + + return stored_connections.size(); + } + +private: + EndpointConnectionPool(ConnectionGroup::Ptr group_, String host_, UInt16 port_, bool https_, ProxyConfiguration proxy_configuration_) + : host(std::move(host_)) + , port(port_) + , https(https_) + , proxy_configuration(std::move(proxy_configuration_)) + , group(group_) + { + } + + WeakPtr getWeakFromThis() { return EndpointConnectionPool::weak_from_this(); } + + bool isExpired(Poco::Timestamp & now, ConnectionPtr connection) + { + if (group->isSoftLimitReached()) + return now > (connection->getLastRequest() + divide(connection->getKeepAliveTimeout(), 10)); + return now > connection->getLastRequest() + connection->getKeepAliveTimeout(); + } + + ConnectionPtr allocateNewConnection() + { + ConnectionPtr connection = PooledConnection::create(this->getWeakFromThis(), getMetrics(), host, port); + connection->setKeepAlive(true); + + if (!proxy_configuration.isEmpty()) + { + connection->setProxyConfig(proxyConfigurationToPocoProxyConfig(proxy_configuration)); + } + + group->atConnectionCreate(); + + return connection; + } + + ConnectionPtr prepareNewConnection(const ConnectionTimeouts & timeouts) + { + auto address = HostResolversPool::instance().getResolver(host)->resolve(); + + auto session = allocateNewConnection(); + + setTimeouts(*session, timeouts); + session->setResolvedHost(*address); + + try + { + auto timer = CurrentThread::getProfileEvents().timer(getMetrics().elapsed_microseconds); + session->doConnect(); + } + catch (...) + { + address.setFail(); + ProfileEvents::increment(getMetrics().errors); + session->reset(); + throw; + } + + ProfileEvents::increment(getMetrics().created); + return session; + } + + void atConnectionDestroy(PooledConnection & connection) + { + group->atConnectionDestroy(); + + if (!connection.connected() || connection.mustReconnect() || !connection.isCompleted() || connection.buffered() + || group->isStoreLimitReached()) + { + ProfileEvents::increment(getMetrics().reset, 1); + return; + } + + auto connection_to_store = allocateNewConnection(); + connection_to_store->assign(connection); + + CurrentMetrics::add(getMetrics().stored_count, 1); + ProfileEvents::increment(getMetrics().preserved, 1); + + { + MemoryTrackerSwitcher switcher{&total_memory_tracker}; + std::lock_guard lock(mutex); + stored_connections.push(connection_to_store); + } + } + + + const std::string host; + const UInt16 port; + const bool https; + const ProxyConfiguration proxy_configuration; + const ConnectionGroup::Ptr group; + + std::mutex mutex; + ConnectionsMinHeap stored_connections TSA_GUARDED_BY(mutex); +}; + +struct EndpointPoolKey +{ + HTTPConnectionGroupType connection_group; + String target_host; + UInt16 target_port; + bool is_target_https; + ProxyConfiguration proxy_config; + + bool operator==(const EndpointPoolKey & rhs) const + { + return std::tie( + connection_group, + target_host, + target_port, + is_target_https, + proxy_config.host, + proxy_config.port, + proxy_config.protocol, + proxy_config.tunneling, + proxy_config.original_request_protocol) + == std::tie( + rhs.connection_group, + rhs.target_host, + rhs.target_port, + rhs.is_target_https, + rhs.proxy_config.host, + rhs.proxy_config.port, + rhs.proxy_config.protocol, + rhs.proxy_config.tunneling, + rhs.proxy_config.original_request_protocol); + } +}; + +struct Hasher +{ + size_t operator()(const EndpointPoolKey & k) const + { + SipHash s; + s.update(k.connection_group); + s.update(k.target_host); + s.update(k.target_port); + s.update(k.is_target_https); + s.update(k.proxy_config.host); + s.update(k.proxy_config.port); + s.update(k.proxy_config.protocol); + s.update(k.proxy_config.tunneling); + s.update(k.proxy_config.original_request_protocol); + return s.get64(); + } +}; + +IExtendedPool::Ptr +createConnectionPool(ConnectionGroup::Ptr group, std::string host, UInt16 port, bool secure, ProxyConfiguration proxy_configuration) +{ + if (secure) + { +#if USE_SSL + return EndpointConnectionPool::create( + group, std::move(host), port, secure, std::move(proxy_configuration)); +#else + throw Exception( + ErrorCodes::SUPPORT_IS_DISABLED, "Inter-server secret support is disabled, because ClickHouse was built without SSL library"); +#endif + } + else + { + return EndpointConnectionPool::create( + group, std::move(host), port, secure, std::move(proxy_configuration)); + } +} + +class HTTPConnectionPools::Impl +{ +private: + const size_t DEFAULT_WIPE_TIMEOUT_SECONDS = 5 * 60; + const Poco::Timespan wipe_timeout = Poco::Timespan(DEFAULT_WIPE_TIMEOUT_SECONDS, 0); + + ConnectionGroup::Ptr disk_group = std::make_shared(HTTPConnectionGroupType::DISK); + ConnectionGroup::Ptr storage_group = std::make_shared(HTTPConnectionGroupType::STORAGE); + ConnectionGroup::Ptr http_group = std::make_shared(HTTPConnectionGroupType::HTTP); + + + /// If multiple mutexes are held simultaneously, + /// they should be locked in this order: + /// HTTPConnectionPools::mutex, then EndpointConnectionPool::mutex, then ConnectionGroup::mutex. + std::mutex mutex; + + std::unordered_map endpoints_pool TSA_GUARDED_BY(mutex); + Poco::Timestamp last_wipe_time TSA_GUARDED_BY(mutex); + +public: + IHTTPConnectionPoolForEndpoint::Ptr getPool(HTTPConnectionGroupType type, const Poco::URI & uri, const ProxyConfiguration & proxy_configuration) + { + Poco::Timestamp now; + + std::lock_guard lock(mutex); + + if (now - last_wipe_time > wipe_timeout) + { + wipeExpired(); + last_wipe_time = now; + } + + return getPoolImpl(type, uri, proxy_configuration); + } + + void setLimits(HTTPConnectionPools::Limits disk, HTTPConnectionPools::Limits storage, HTTPConnectionPools::Limits http) + { + disk_group->setLimits(std::move(disk)); + storage_group->setLimits(std::move(storage)); + http_group->setLimits(std::move(http)); + } + + void dropCache() + { + std::lock_guard lock(mutex); + endpoints_pool.clear(); + } + +protected: + ConnectionGroup::Ptr & getGroup(HTTPConnectionGroupType type) + { + switch (type) + { + case HTTPConnectionGroupType::DISK: + return disk_group; + case HTTPConnectionGroupType::STORAGE: + return storage_group; + case HTTPConnectionGroupType::HTTP: + return http_group; + } + } + + IExtendedPool::Ptr getPoolImpl(HTTPConnectionGroupType type, const Poco::URI & uri, const ProxyConfiguration & proxy_configuration) + TSA_REQUIRES(mutex) + { + auto [host, port, secure] = getHostPortSecure(uri, proxy_configuration); + auto key = EndpointPoolKey{type, host, port, secure, proxy_configuration}; + + auto it = endpoints_pool.find(key); + if (it != endpoints_pool.end()) + return it->second; + + it = endpoints_pool.emplace(key, createConnectionPool(getGroup(type), std::move(host), port, secure, proxy_configuration)).first; + + return it->second; + } + + void wipeExpired() TSA_REQUIRES(mutex) + { + std::vector keys_to_drop; + + for (auto & [key, pool] : endpoints_pool) + { + auto left_connections = pool->wipeExpired(); + if (left_connections == 0 && pool->getGroupType() != HTTPConnectionGroupType::DISK) + keys_to_drop.push_back(key); + } + + for (const auto & key : keys_to_drop) + endpoints_pool.erase(key); + } + + static bool useSecureConnection(const Poco::URI & uri, const ProxyConfiguration & proxy_configuration) + { + if (uri.getScheme() == "http") + return false; + + if (uri.getScheme() != "https") + throw Exception(ErrorCodes::UNSUPPORTED_URI_SCHEME, "Unsupported scheme in URI '{}'", uri.toString()); + + if (!proxy_configuration.isEmpty()) + { + if (ProxyConfiguration::Protocol::HTTP == proxy_configuration.protocol && !proxy_configuration.tunneling) + { + // If it is an HTTPS request, proxy server is HTTP and user opted for tunneling off, we must not create an HTTPS request. + // The desired flow is: HTTP request to the proxy server, then proxy server will initiate an HTTPS request to the target server. + // There is a weak link in the security, but that's what the user opted for. + return false; + } + } + + return true; + } + + static std::tuple getHostPortSecure(const Poco::URI & uri, const ProxyConfiguration & proxy_configuration) + { + return std::make_tuple(uri.getHost(), uri.getPort(), useSecureConnection(uri, proxy_configuration)); + } +}; + +HTTPConnectionPools::HTTPConnectionPools() + : impl(std::make_unique()) +{ +} + +HTTPConnectionPools & HTTPConnectionPools::instance() +{ + static HTTPConnectionPools instance; + return instance; +} + +void HTTPConnectionPools::setLimits(HTTPConnectionPools::Limits disk, HTTPConnectionPools::Limits storage, HTTPConnectionPools::Limits http) +{ + impl->setLimits(std::move(disk), std::move(storage), std::move(http)); +} + +void HTTPConnectionPools::dropCache() +{ + impl->dropCache(); +} + +IHTTPConnectionPoolForEndpoint::Ptr +HTTPConnectionPools::getPool(HTTPConnectionGroupType type, const Poco::URI & uri, const ProxyConfiguration & proxy_configuration) +{ + return impl->getPool(type, uri, proxy_configuration); +} +} diff --git a/src/Common/HTTPConnectionPool.h b/src/Common/HTTPConnectionPool.h new file mode 100644 index 00000000000..7fd0847f5a7 --- /dev/null +++ b/src/Common/HTTPConnectionPool.h @@ -0,0 +1,91 @@ +#pragma once + +#include + +#include +#include +#include +#include + +#include + +#include +#include + +#include +#include + +namespace DB +{ + +class IHTTPConnectionPoolForEndpoint +{ +public: + struct Metrics + { + const ProfileEvents::Event created = ProfileEvents::end(); + const ProfileEvents::Event reused = ProfileEvents::end(); + const ProfileEvents::Event reset = ProfileEvents::end(); + const ProfileEvents::Event preserved = ProfileEvents::end(); + const ProfileEvents::Event expired = ProfileEvents::end(); + const ProfileEvents::Event errors = ProfileEvents::end(); + const ProfileEvents::Event elapsed_microseconds = ProfileEvents::end(); + + const CurrentMetrics::Metric stored_count = CurrentMetrics::end(); + const CurrentMetrics::Metric active_count = CurrentMetrics::end(); + }; + + using Ptr = std::shared_ptr; + using Connection = Poco::Net::HTTPClientSession; + using ConnectionPtr = std::shared_ptr; + + /// can throw Poco::Net::Exception, DB::NetException, DB::Exception + virtual ConnectionPtr getConnection(const ConnectionTimeouts & timeouts) = 0; + virtual const Metrics & getMetrics() const = 0; + virtual ~IHTTPConnectionPoolForEndpoint() = default; + +protected: + IHTTPConnectionPoolForEndpoint() = default; + + IHTTPConnectionPoolForEndpoint(const IHTTPConnectionPoolForEndpoint &) = delete; + IHTTPConnectionPoolForEndpoint & operator=(const IHTTPConnectionPoolForEndpoint &) = delete; +}; + +enum class HTTPConnectionGroupType +{ + DISK, + STORAGE, + HTTP, +}; + +class HTTPConnectionPools +{ +public: + struct Limits + { + size_t soft_limit = 100; + size_t warning_limit = 1000; + size_t store_limit = 10000; + + static constexpr size_t warning_step = 100; + }; + +private: + HTTPConnectionPools(); + HTTPConnectionPools(const HTTPConnectionPools &) = delete; + HTTPConnectionPools & operator=(const HTTPConnectionPools &) = delete; + +public: + static HTTPConnectionPools & instance(); + + void setLimits(Limits disk, Limits storage, Limits http); + void dropCache(); + + IHTTPConnectionPoolForEndpoint::Ptr getPool(HTTPConnectionGroupType type, const Poco::URI & uri, const ProxyConfiguration & proxy_configuration); + +private: + class Impl; + std::unique_ptr impl; +}; + +} diff --git a/src/Common/HostResolvePool.cpp b/src/Common/HostResolvePool.cpp new file mode 100644 index 00000000000..6db28edc07e --- /dev/null +++ b/src/Common/HostResolvePool.cpp @@ -0,0 +1,298 @@ +#include + +#include +#include +#include +#include +#include +#include + +#include + +namespace ProfileEvents +{ + extern const Event AddressesDiscovered; + extern const Event AddressesExpired; + extern const Event AddressesMarkedAsFailed; +} + +namespace CurrentMetrics +{ + extern const Metric AddressesActive; +} + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int DNS_ERROR; +} + +HostResolverMetrics HostResolver::getMetrics() +{ + return HostResolverMetrics{ + .discovered = ProfileEvents::AddressesDiscovered, + .expired = ProfileEvents::AddressesExpired, + .failed = ProfileEvents::AddressesMarkedAsFailed, + .active_count = CurrentMetrics::AddressesActive, + }; +} + +HostResolver::WeakPtr HostResolver::getWeakFromThis() +{ + return weak_from_this(); +} + +HostResolver::HostResolver(String host_, Poco::Timespan history_) + : host(std::move(host_)) + , history(history_) + , resolve_function([](const String & host_to_resolve) { return DNSResolver::instance().resolveHostAll(host_to_resolve); }) +{ + update(); +} + +HostResolver::HostResolver( + ResolveFunction && resolve_function_, String host_, Poco::Timespan history_) + : host(std::move(host_)), history(history_), resolve_function(std::move(resolve_function_)) +{ + update(); +} + +HostResolver::~HostResolver() +{ + std::lock_guard lock(mutex); + CurrentMetrics::sub(metrics.active_count, records.size()); + records.clear(); +} + +void HostResolver::Entry::setFail() +{ + fail = true; + + if (auto lock = pool.lock()) + lock->setFail(address); +} + +HostResolver::Entry::~Entry() +{ + if (!fail) + { + if (auto lock = pool.lock()) + lock->setSuccess(address); + } +} + +void HostResolver::update() +{ + MemoryTrackerSwitcher switcher{&total_memory_tracker}; + + auto next_gen = resolve_function(host); + if (next_gen.empty()) + throw NetException(ErrorCodes::DNS_ERROR, "no endpoints resolved for host {}", host); + + std::sort(next_gen.begin(), next_gen.end()); + + Poco::Timestamp now; + + std::lock_guard lock(mutex); + updateImpl(now, next_gen); +} + +void HostResolver::reset() +{ + std::lock_guard lock(mutex); + + CurrentMetrics::sub(metrics.active_count, records.size()); + records.clear(); +} + +void HostResolver::updateWeights() +{ + updateWeightsImpl(); + + if (getTotalWeight() == 0 && !records.empty()) + { + for (auto & rec : records) + rec.failed = false; + + updateWeightsImpl(); + } + + chassert((getTotalWeight() > 0 && !records.empty()) || records.empty()); +} + +HostResolver::Entry HostResolver::resolve() +{ + if (isUpdateNeeded()) + update(); + + std::lock_guard lock(mutex); + return Entry(*this, selectBest()); +} + +void HostResolver::setSuccess(const Poco::Net::IPAddress & address) +{ + std::lock_guard lock(mutex); + + auto it = find(address); + if (it == records.end()) + return; + + auto old_weight = it->getWeight(); + ++it->usage; + auto new_weight = it->getWeight(); + + if (old_weight != new_weight) + updateWeights(); +} + +void HostResolver::setFail(const Poco::Net::IPAddress & address) +{ + Poco::Timestamp now; + + { + std::lock_guard lock(mutex); + + auto it = find(address); + if (it == records.end()) + return; + + it->failed = true; + it->fail_time = now; + } + + ProfileEvents::increment(metrics.failed); + update(); +} + +Poco::Net::IPAddress HostResolver::selectBest() +{ + chassert(!records.empty()); + auto random_weight_picker = std::uniform_int_distribution(0, getTotalWeight() - 1); + size_t weight = random_weight_picker(thread_local_rng); + auto it = std::partition_point(records.begin(), records.end(), [&](const Record & rec) { return rec.weight_prefix_sum <= weight; }); + chassert(it != records.end()); + return it->address; +} + +HostResolver::Records::iterator HostResolver::find(const Poco::Net::IPAddress & addr) TSA_REQUIRES(mutex) +{ + auto it = std::lower_bound( + records.begin(), records.end(), addr, [](const Record & rec, const Poco::Net::IPAddress & value) { return rec.address < value; }); + + if (it != records.end() && it->address != addr) + return records.end(); + + return it; +} + +bool HostResolver::isUpdateNeeded() +{ + Poco::Timestamp now; + + std::lock_guard lock(mutex); + return last_resolve_time + history < now || records.empty(); +} + +void HostResolver::updateImpl(Poco::Timestamp now, std::vector & next_gen) + TSA_REQUIRES(mutex) +{ + const auto last_effective_resolve = now - history; + + Records merged; + merged.reserve(records.size() + next_gen.size()); + + auto it_before = records.begin(); + auto it_next = next_gen.begin(); + + while (it_before != records.end() || it_next != next_gen.end()) + { + if (it_next == next_gen.end() || (it_before != records.end() && it_before->address < *it_next)) + { + if (it_before->resolve_time >= last_effective_resolve) + merged.push_back(*it_before); + else + { + CurrentMetrics::sub(metrics.active_count, 1); + ProfileEvents::increment(metrics.expired, 1); + } + ++it_before; + } + else if (it_before == records.end() || (it_next != next_gen.end() && *it_next < it_before->address)) + { + CurrentMetrics::add(metrics.active_count, 1); + ProfileEvents::increment(metrics.discovered, 1); + merged.push_back(Record(*it_next, now)); + ++it_next; + } + else + { + merged.push_back(*it_before); + merged.back().resolve_time = now; + + ++it_before; + ++it_next; + } + } + + for (auto & rec : merged) + if (rec.failed && rec.fail_time < last_effective_resolve) + rec.failed = false; + + chassert(std::is_sorted(merged.begin(), merged.end())); + + last_resolve_time = now; + records.swap(merged); + + if (records.empty()) + throw NetException(ErrorCodes::DNS_ERROR, "no endpoints resolved for host {}", host); + + updateWeights(); +} + +size_t HostResolver::getTotalWeight() const +{ + if (records.empty()) + return 0; + return records.back().weight_prefix_sum; +} + + +void HostResolver::updateWeightsImpl() +{ + size_t total_weight_next = 0; + + for (auto & rec: records) + { + total_weight_next += rec.getWeight(); + rec.weight_prefix_sum = total_weight_next; + } +} + +HostResolversPool & HostResolversPool::instance() +{ + static HostResolversPool instance; + return instance; +} + +void HostResolversPool::dropCache() +{ + std::lock_guard lock(mutex); + host_pools.clear(); +} + +HostResolver::Ptr HostResolversPool::getResolver(const String & host) +{ + std::lock_guard lock(mutex); + + auto it = host_pools.find(host); + if (it != host_pools.end()) + return it->second; + + it = host_pools.emplace(host, HostResolver::create(host)).first; + + return it->second; +} + +} diff --git a/src/Common/HostResolvePool.h b/src/Common/HostResolvePool.h new file mode 100644 index 00000000000..4f127f05253 --- /dev/null +++ b/src/Common/HostResolvePool.h @@ -0,0 +1,216 @@ +#pragma once + +#include +#include +#include +#include + +#include + +#include + +#include +#include + +// That class resolves host into multiply addresses +// Features: +// - balance address usage. +// `selectBest()` chooses the address by random with weights. +// The more ip is used the lesser weight it has. When new address is happened, it takes more weight. +// But still not all requests are assigned to the new address. +// - join resolve results +// In case when host is resolved into different set of addresses, this class join all that addresses and use them. +// An address expires after `history_` time. +// - failed address pessimization +// If an address marked with `setFail()` it is marked as faulty. Such address won't be selected until either +// a) it still occurs in resolve set after `history_` time or b) all other addresses are pessimized as well. +// - resolve schedule +// Addresses are resolved through `DB::DNSResolver::instance()`. +// Usually it does not happen more often than once in `history_` time. +// But also new resolve performed each `setFail()` call. + +namespace DB +{ + +struct HostResolverMetrics +{ + const ProfileEvents::Event discovered = ProfileEvents::end(); + const ProfileEvents::Event expired = ProfileEvents::end(); + const ProfileEvents::Event failed = ProfileEvents::end(); + + const CurrentMetrics::Metric active_count = CurrentMetrics::end(); +}; + +constexpr size_t DEFAULT_RESOLVE_TIME_HISTORY_SECONDS = 2*60; + + +class HostResolver : public std::enable_shared_from_this +{ +private: + using WeakPtr = std::weak_ptr; + +public: + using Ptr = std::shared_ptr; + + template + static Ptr create(Args&&... args) + { + struct make_shared_enabler : public HostResolver + { + explicit make_shared_enabler(Args&&... args) : HostResolver(std::forward(args)...) {} + }; + return std::make_shared(std::forward(args)...); + } + + virtual ~HostResolver(); + + class Entry + { + public: + explicit Entry(Entry && entry) = default; + explicit Entry(Entry & entry) = delete; + + // no access as r-value + const String * operator->() && = delete; + const String * operator->() const && = delete; + const String & operator*() && = delete; + const String & operator*() const && = delete; + + const String * operator->() & { return &resolved_host; } + const String * operator->() const & { return &resolved_host; } + const String & operator*() & { return resolved_host; } + const String & operator*() const & { return resolved_host; } + + void setFail(); + ~Entry(); + + private: + friend class HostResolver; + + Entry(HostResolver & pool_, Poco::Net::IPAddress address_) + : pool(pool_.getWeakFromThis()) + , address(std::move(address_)) + , resolved_host(address.toString()) + { } + + HostResolver::WeakPtr pool; + const Poco::Net::IPAddress address; + const String resolved_host; + + bool fail = false; + }; + + /// can throw NetException(ErrorCodes::DNS_ERROR, ...), Exception(ErrorCodes::BAD_ARGUMENTS, ...) + Entry resolve(); + void update(); + void reset(); + + static HostResolverMetrics getMetrics(); + +protected: + explicit HostResolver( + String host_, + Poco::Timespan history_ = Poco::Timespan(DEFAULT_RESOLVE_TIME_HISTORY_SECONDS, 0)); + + using ResolveFunction = std::function (const String & host)>; + HostResolver(ResolveFunction && resolve_function_, + String host_, + Poco::Timespan history_); + + friend class Entry; + WeakPtr getWeakFromThis(); + + void setSuccess(const Poco::Net::IPAddress & address); + void setFail(const Poco::Net::IPAddress & address); + + struct Record + { + Record(Poco::Net::IPAddress address_, Poco::Timestamp resolve_time_) + : address(std::move(address_)) + , resolve_time(resolve_time_) + {} + + explicit Record(Record && rec) = default; + Record& operator=(Record && s) = default; + + explicit Record(const Record & rec) = default; + Record& operator=(const Record & s) = default; + + Poco::Net::IPAddress address; + Poco::Timestamp resolve_time; + size_t usage = 0; + bool failed = false; + Poco::Timestamp fail_time = 0; + + size_t weight_prefix_sum; + + bool operator <(const Record & r) const + { + return address < r.address; + } + + size_t getWeight() const + { + if (failed) + return 0; + + /// There is no goal to make usage's distribution ideally even + /// The goal is to chose more often new address, but still use old addresses as well + /// when all addresses have usage counter greater than 10000, + /// no more corrections are needed, just random choice is ok + if (usage > 10000) + return 1; + if (usage > 1000) + return 5; + if (usage > 100) + return 8; + return 10; + } + }; + + using Records = std::vector; + + Poco::Net::IPAddress selectBest() TSA_REQUIRES(mutex); + Records::iterator find(const Poco::Net::IPAddress & address) TSA_REQUIRES(mutex); + bool isUpdateNeeded(); + + void updateImpl(Poco::Timestamp now, std::vector & next_gen) TSA_REQUIRES(mutex); + void updateWeights() TSA_REQUIRES(mutex); + void updateWeightsImpl() TSA_REQUIRES(mutex); + size_t getTotalWeight() const TSA_REQUIRES(mutex); + + const String host; + const Poco::Timespan history; + const HostResolverMetrics metrics = getMetrics(); + + // for tests purpose + const ResolveFunction resolve_function; + + std::mutex mutex; + + Poco::Timestamp last_resolve_time TSA_GUARDED_BY(mutex); + Records records TSA_GUARDED_BY(mutex); + + Poco::Logger * log = &Poco::Logger::get("ConnectionPool"); +}; + +class HostResolversPool +{ +private: + HostResolversPool() = default; + HostResolversPool(const HostResolversPool &) = delete; + HostResolversPool & operator=(const HostResolversPool &) = delete; + +public: + static HostResolversPool & instance(); + + void dropCache(); + + HostResolver::Ptr getResolver(const String & host); +private: + std::mutex mutex; + std::unordered_map host_pools TSA_GUARDED_BY(mutex); +}; + +} + diff --git a/src/Common/PageCache.cpp b/src/Common/PageCache.cpp index 511ec23d431..d4598d4683b 100644 --- a/src/Common/PageCache.cpp +++ b/src/Common/PageCache.cpp @@ -191,7 +191,7 @@ size_t PageCache::maxChunks() const { return chunks_per_mmap_target * max_mmaps; size_t PageCache::getPinnedSize() const { - std::unique_lock lock(global_mutex); + std::lock_guard lock(global_mutex); return (total_chunks - lru.size()) * bytes_per_page * pages_per_chunk; } @@ -202,8 +202,11 @@ PageCache::MemoryStats PageCache::getResidentSetSize() const if (use_madv_free) { std::unordered_set cache_mmap_addrs; - for (const auto & m : mmaps) - cache_mmap_addrs.insert(reinterpret_cast(m.ptr)); + { + std::lock_guard lock(global_mutex); + for (const auto & m : mmaps) + cache_mmap_addrs.insert(reinterpret_cast(m.ptr)); + } ReadBufferFromFile in("/proc/self/smaps"); @@ -283,6 +286,7 @@ PageCache::MemoryStats PageCache::getResidentSetSize() const } #endif + std::lock_guard lock(global_mutex); stats.page_cache_rss = bytes_per_page * pages_per_chunk * total_chunks; return stats; } @@ -294,12 +298,12 @@ PinnedPageChunk PageCache::getOrSet(PageCacheKey key, bool detached_if_missing, bool incremented_profile_events = false; { - std::unique_lock lock(global_mutex); + std::lock_guard lock(global_mutex); auto * it = chunk_by_key.find(key); if (it == chunk_by_key.end()) { - chunk = getFreeChunk(lock); + chunk = getFreeChunk(); chassert(!chunk->key.has_value()); if (!detached_if_missing) @@ -331,14 +335,14 @@ PinnedPageChunk PageCache::getOrSet(PageCacheKey key, bool detached_if_missing, /// otherwise we may detach a chunk pinned by someone else, which may be unexpected /// for that someone else. Or maybe the latter is fine, dropCache() already does it.) if (chunk->pages_populated.get(0) && reinterpret_cast*>(chunk->data)->load(std::memory_order_relaxed) == 0) - evictChunk(chunk, lock); + evictChunk(chunk); } if (inject_eviction && chunk->key.has_value() && rng() % 10 == 0) { /// Simulate eviction of the chunk or some of its pages. if (rng() % 2 == 0) - evictChunk(chunk, lock); + evictChunk(chunk); else for (size_t i = 0; i < 20; ++i) chunk->pages_populated.unset(rng() % (chunk->size / chunk->page_size)); @@ -353,7 +357,7 @@ PinnedPageChunk PageCache::getOrSet(PageCacheKey key, bool detached_if_missing, } { - std::unique_lock chunk_lock(chunk->chunk_mutex); + std::lock_guard chunk_lock(chunk->chunk_mutex); if (chunk->pages_state == PageChunkState::Limbo) { @@ -383,7 +387,7 @@ void PageCache::removeRef(PageChunk * chunk) noexcept return; { - std::unique_lock lock(global_mutex); + std::lock_guard lock(global_mutex); prev_pin_count = chunk->pin_count.fetch_sub(1); if (prev_pin_count > 1) @@ -398,7 +402,7 @@ void PageCache::removeRef(PageChunk * chunk) noexcept } { - std::unique_lock chunk_lock(chunk->chunk_mutex); + std::lock_guard chunk_lock(chunk->chunk_mutex); /// Need to be extra careful here because we unlocked global_mutex above, so other /// getOrSet()/removeRef() calls could have happened during this brief period. @@ -421,7 +425,7 @@ static void logUnexpectedSyscallError(std::string name) #endif } -void PageCache::sendChunkToLimbo(PageChunk * chunk [[maybe_unused]], std::unique_lock & /* chunk_mutex */) const noexcept +void PageCache::sendChunkToLimbo(PageChunk * chunk [[maybe_unused]], std::lock_guard & /* chunk_mutex */) const noexcept { #ifdef MADV_FREE // if we're not on a very old version of Linux chassert(chunk->size == bytes_per_page * pages_per_chunk); @@ -454,7 +458,7 @@ void PageCache::sendChunkToLimbo(PageChunk * chunk [[maybe_unused]], std::unique #endif } -std::pair PageCache::restoreChunkFromLimbo(PageChunk * chunk, std::unique_lock & /* chunk_mutex */) const noexcept +std::pair PageCache::restoreChunkFromLimbo(PageChunk * chunk, std::lock_guard & /* chunk_mutex */) const noexcept { static_assert(sizeof(std::atomic) == 1, "char is not atomic?"); // Make sure our strategic memory reads/writes are not reordered or optimized out. @@ -505,10 +509,10 @@ std::pair PageCache::restoreChunkFromLimbo(PageChunk * chunk, st return {pages_restored, pages_evicted}; } -PageChunk * PageCache::getFreeChunk(std::unique_lock & lock /* global_mutex */) +PageChunk * PageCache::getFreeChunk() { if (lru.empty() || (mmaps.size() < max_mmaps && lru.front().key.has_value())) - addMmap(lock); + addMmap(); if (lru.empty()) throw Exception(ErrorCodes::MEMORY_LIMIT_EXCEEDED, "All chunks in the entire page cache ({:.3} GiB) are pinned.", bytes_per_page * pages_per_chunk * total_chunks * 1. / (1l << 30)); @@ -519,12 +523,12 @@ PageChunk * PageCache::getFreeChunk(std::unique_lock & lock /* globa size_t prev_pin_count = chunk->pin_count.fetch_add(1); chassert(prev_pin_count == 0); - evictChunk(chunk, lock); + evictChunk(chunk); return chunk; } -void PageCache::evictChunk(PageChunk * chunk, std::unique_lock & /* global_mutex */) +void PageCache::evictChunk(PageChunk * chunk) { if (chunk->key.has_value()) { @@ -548,7 +552,7 @@ void PageCache::evictChunk(PageChunk * chunk, std::unique_lock & /* chunk->pages_populated.unsetAll(); } -void PageCache::addMmap(std::unique_lock & /* global_mutex */) +void PageCache::addMmap() { /// ASLR by hand. void * address_hint = reinterpret_cast(std::uniform_int_distribution(0x100000000000UL, 0x700000000000UL)(rng)); @@ -564,13 +568,13 @@ void PageCache::addMmap(std::unique_lock & /* global_mutex */) void PageCache::dropCache() { - std::unique_lock lock(global_mutex); + std::lock_guard lock(global_mutex); /// Detach and free unpinned chunks. bool logged_error = false; for (PageChunk & chunk : lru) { - evictChunk(&chunk, lock); + evictChunk(&chunk); if (use_madv_free) { diff --git a/src/Common/PageCache.h b/src/Common/PageCache.h index 7ff376baa6b..04411e6fc24 100644 --- a/src/Common/PageCache.h +++ b/src/Common/PageCache.h @@ -75,7 +75,7 @@ struct FileChunkAddress /// E.g. "s3:/" std::string path; /// Optional string with ETag, or file modification time, or anything else. - std::string file_version; + std::string file_version{}; size_t offset = 0; PageCacheKey hash() const; @@ -270,28 +270,28 @@ private: mutable std::mutex global_mutex; - pcg64 rng; + pcg64 rng TSA_GUARDED_BY(global_mutex); - std::vector mmaps; - size_t total_chunks = 0; + std::vector mmaps TSA_GUARDED_BY(global_mutex); + size_t total_chunks TSA_GUARDED_BY(global_mutex) = 0; /// All non-pinned chunks, including ones not assigned to any file. Least recently used is begin(). - boost::intrusive::list, boost::intrusive::constant_time_size> lru; + boost::intrusive::list, boost::intrusive::constant_time_size> lru TSA_GUARDED_BY(global_mutex); - HashMap chunk_by_key; + HashMap chunk_by_key TSA_GUARDED_BY(global_mutex); /// Get a usable chunk, doing eviction or allocation if needed. /// Caller is responsible for clearing pages_populated. - PageChunk * getFreeChunk(std::unique_lock & /* global_mutex */); - void addMmap(std::unique_lock & /* global_mutex */); - void evictChunk(PageChunk * chunk, std::unique_lock & /* global_mutex */); + PageChunk * getFreeChunk() TSA_REQUIRES(global_mutex); + void addMmap() TSA_REQUIRES(global_mutex); + void evictChunk(PageChunk * chunk) TSA_REQUIRES(global_mutex); void removeRef(PageChunk * chunk) noexcept; /// These may run in parallel with getFreeChunk(), so be very careful about which fields of the PageChunk we touch here. - void sendChunkToLimbo(PageChunk * chunk, std::unique_lock & /* chunk_mutex */) const noexcept; + void sendChunkToLimbo(PageChunk * chunk, std::lock_guard & /* chunk_mutex */) const noexcept; /// Returns {pages_restored, pages_evicted}. - std::pair restoreChunkFromLimbo(PageChunk * chunk, std::unique_lock & /* chunk_mutex */) const noexcept; + std::pair restoreChunkFromLimbo(PageChunk * chunk, std::lock_guard & /* chunk_mutex */) const noexcept; }; using PageCachePtr = std::shared_ptr; diff --git a/src/Common/ProfileEvents.cpp b/src/Common/ProfileEvents.cpp index 87fcf220ff0..c1ac3d08245 100644 --- a/src/Common/ProfileEvents.cpp +++ b/src/Common/ProfileEvents.cpp @@ -475,6 +475,7 @@ The server successfully detected this situation and will download merged part fr M(FileSegmentUseMicroseconds, "File segment use() time") \ M(FileSegmentRemoveMicroseconds, "File segment remove() time") \ M(FileSegmentHolderCompleteMicroseconds, "File segments holder complete() time") \ + M(FileSegmentFailToIncreasePriority, "Number of times the priority was not increased due to a high contention on the cache lock") \ M(FilesystemCacheHoldFileSegments, "Filesystem cache file segments count, which were hold") \ M(FilesystemCacheUnusedHoldFileSegments, "Filesystem cache file segments count, which were hold, but not used (because of seek or LIMIT n, etc)") \ \ @@ -695,6 +696,35 @@ The server successfully detected this situation and will download merged part fr \ M(ParallelReplicasAvailableCount, "Number of replicas available to execute a query with task-based parallel replicas") \ M(ParallelReplicasUnavailableCount, "Number of replicas which was chosen, but found to be unavailable during query execution with task-based parallel replicas") \ + \ + M(StorageConnectionsCreated, "Number of created connections for storages") \ + M(StorageConnectionsReused, "Number of reused connections for storages") \ + M(StorageConnectionsReset, "Number of reset connections for storages") \ + M(StorageConnectionsPreserved, "Number of preserved connections for storages") \ + M(StorageConnectionsExpired, "Number of expired connections for storages") \ + M(StorageConnectionsErrors, "Number of cases when creation of a connection for storage is failed") \ + M(StorageConnectionsElapsedMicroseconds, "Total time spend on creating connections for storages") \ + \ + M(DiskConnectionsCreated, "Number of created connections for disk") \ + M(DiskConnectionsReused, "Number of reused connections for disk") \ + M(DiskConnectionsReset, "Number of reset connections for disk") \ + M(DiskConnectionsPreserved, "Number of preserved connections for disk") \ + M(DiskConnectionsExpired, "Number of expired connections for disk") \ + M(DiskConnectionsErrors, "Number of cases when creation of a connection for disk is failed") \ + M(DiskConnectionsElapsedMicroseconds, "Total time spend on creating connections for disk") \ + \ + M(HTTPConnectionsCreated, "Number of created http connections") \ + M(HTTPConnectionsReused, "Number of reused http connections") \ + M(HTTPConnectionsReset, "Number of reset http connections") \ + M(HTTPConnectionsPreserved, "Number of preserved http connections") \ + M(HTTPConnectionsExpired, "Number of expired http connections") \ + M(HTTPConnectionsErrors, "Number of cases when creation of a http connection failed") \ + M(HTTPConnectionsElapsedMicroseconds, "Total time spend on creating http connections") \ + \ + M(AddressesDiscovered, "Total count of new addresses in dns resolve results for http connections") \ + M(AddressesExpired, "Total count of expired addresses which is no longer presented in dns resolve results for http connections") \ + M(AddressesMarkedAsFailed, "Total count of addresses which has been marked as faulty due to connection errors for http connections") \ + #ifdef APPLY_FOR_EXTERNAL_EVENTS #define APPLY_FOR_EVENTS(M) APPLY_FOR_BUILTIN_EVENTS(M) APPLY_FOR_EXTERNAL_EVENTS(M) diff --git a/src/Common/ProxyConfiguration.h b/src/Common/ProxyConfiguration.h index 53e569bf6e4..11a09cb5924 100644 --- a/src/Common/ProxyConfiguration.h +++ b/src/Common/ProxyConfiguration.h @@ -44,11 +44,13 @@ struct ProxyConfiguration } } - std::string host; - Protocol protocol; - uint16_t port; - bool tunneling; - Protocol original_request_protocol; + std::string host = std::string{}; + Protocol protocol = Protocol::HTTP; + uint16_t port = 0; + bool tunneling = false; + Protocol original_request_protocol = Protocol::HTTP; + + bool isEmpty() const { return host.size() == 0; } }; } diff --git a/src/Common/ProxyListConfigurationResolver.cpp b/src/Common/ProxyListConfigurationResolver.cpp index 01a6f52185f..c9b8923929a 100644 --- a/src/Common/ProxyListConfigurationResolver.cpp +++ b/src/Common/ProxyListConfigurationResolver.cpp @@ -26,8 +26,6 @@ ProxyConfiguration ProxyListConfigurationResolver::resolve() auto & proxy = proxies[index]; - LOG_DEBUG(getLogger("ProxyListConfigurationResolver"), "Use proxy: {}", proxies[index].toString()); - return ProxyConfiguration { proxy.getHost(), ProxyConfiguration::protocolFromString(proxy.getScheme()), diff --git a/src/Common/RemoteProxyConfigurationResolver.cpp b/src/Common/RemoteProxyConfigurationResolver.cpp index 117c8a34dbb..ef972a8e318 100644 --- a/src/Common/RemoteProxyConfigurationResolver.cpp +++ b/src/Common/RemoteProxyConfigurationResolver.cpp @@ -69,7 +69,7 @@ ProxyConfiguration RemoteProxyConfigurationResolver::resolve() { auto resolved_endpoint = endpoint; resolved_endpoint.setHost(resolved_hosts[i].toString()); - session = makeHTTPSession(resolved_endpoint, timeouts); + session = makeHTTPSession(HTTPConnectionGroupType::HTTP, resolved_endpoint, timeouts); try { diff --git a/src/Common/ThreadFuzzer.cpp b/src/Common/ThreadFuzzer.cpp index 0868613d880..1d944f4a458 100644 --- a/src/Common/ThreadFuzzer.cpp +++ b/src/Common/ThreadFuzzer.cpp @@ -51,7 +51,11 @@ ThreadFuzzer::ThreadFuzzer() { initConfiguration(); if (!isEffective()) + { + /// It has no effect - disable it + stop(); return; + } setup(); } @@ -172,6 +176,8 @@ void ThreadFuzzer::stop() void ThreadFuzzer::start() { + if (!instance().isEffective()) + return; started.store(true, std::memory_order_relaxed); } @@ -180,11 +186,11 @@ bool ThreadFuzzer::isStarted() return started.load(std::memory_order_relaxed); } -static void injection( +static void injectionImpl( double yield_probability, double migrate_probability, double sleep_probability, - double sleep_time_us [[maybe_unused]]) + double sleep_time_us) { DENY_ALLOCATIONS_IN_SCOPE; if (!ThreadFuzzer::isStarted()) @@ -222,6 +228,19 @@ static void injection( } } +static ALWAYS_INLINE void injection( + double yield_probability, + double migrate_probability, + double sleep_probability, + double sleep_time_us) +{ + DENY_ALLOCATIONS_IN_SCOPE; + if (!ThreadFuzzer::isStarted()) + return; + + injectionImpl(yield_probability, migrate_probability, sleep_probability, sleep_time_us); +} + void ThreadFuzzer::maybeInjectSleep() { auto & fuzzer = ThreadFuzzer::instance(); @@ -286,13 +305,13 @@ void ThreadFuzzer::setup() const #if THREAD_FUZZER_WRAP_PTHREAD #define INJECTION_BEFORE(NAME) \ - injection( \ + injectionImpl( \ NAME##_before_yield_probability.load(std::memory_order_relaxed), \ NAME##_before_migrate_probability.load(std::memory_order_relaxed), \ NAME##_before_sleep_probability.load(std::memory_order_relaxed), \ NAME##_before_sleep_time_us.load(std::memory_order_relaxed)); #define INJECTION_AFTER(NAME) \ - injection( \ + injectionImpl( \ NAME##_after_yield_probability.load(std::memory_order_relaxed), \ NAME##_after_migrate_probability.load(std::memory_order_relaxed), \ NAME##_after_sleep_probability.load(std::memory_order_relaxed), \ @@ -383,13 +402,16 @@ static void * getFunctionAddress(const char * name) static constinit RET(*real_##NAME)(__VA_ARGS__) = nullptr; \ extern "C" RET NAME(__VA_ARGS__) \ { \ - INJECTION_BEFORE(NAME); \ + bool thread_fuzzer_enabled = ThreadFuzzer::isStarted(); \ + if (thread_fuzzer_enabled) \ + INJECTION_BEFORE(NAME); \ if (unlikely(!real_##NAME)) { \ real_##NAME = \ reinterpret_cast(getFunctionAddress(#NAME)); \ } \ auto && ret{real_##NAME(arg)}; \ - INJECTION_AFTER(NAME); \ + if (thread_fuzzer_enabled) \ + INJECTION_AFTER(NAME); \ return ret; \ } FOR_EACH_WRAPPED_FUNCTION(MAKE_WRAPPER_USING_DLSYM) @@ -399,10 +421,17 @@ FOR_EACH_WRAPPED_FUNCTION(MAKE_WRAPPER_USING_DLSYM) extern "C" RET __##NAME(__VA_ARGS__); \ extern "C" RET NAME(__VA_ARGS__) \ { \ - INJECTION_BEFORE(NAME); \ - auto && ret{__##NAME(arg)}; \ - INJECTION_AFTER(NAME); \ - return ret; \ + if (!ThreadFuzzer::isStarted()) \ + { \ + return __##NAME(arg); \ + } \ + else \ + { \ + INJECTION_BEFORE(NAME); \ + auto && ret{__##NAME(arg)}; \ + INJECTION_AFTER(NAME); \ + return ret; \ + } \ } FOR_EACH_WRAPPED_FUNCTION(MAKE_WRAPPER_USING_INTERNAL_SYMBOLS) #undef MAKE_WRAPPER_USING_INTERNAL_SYMBOLS diff --git a/src/Common/ThreadFuzzer.h b/src/Common/ThreadFuzzer.h index 9dd55fe7995..1cff27a7588 100644 --- a/src/Common/ThreadFuzzer.h +++ b/src/Common/ThreadFuzzer.h @@ -1,6 +1,7 @@ #pragma once #include #include +#include namespace DB { @@ -56,7 +57,7 @@ public: static void stop(); static void start(); - static bool isStarted(); + static bool ALWAYS_INLINE isStarted(); static void maybeInjectSleep(); static void maybeInjectMemoryLimitException(); diff --git a/src/Common/ZooKeeper/ZooKeeperRetries.h b/src/Common/ZooKeeper/ZooKeeperRetries.h index d411549346a..d8d234ba913 100644 --- a/src/Common/ZooKeeper/ZooKeeperRetries.h +++ b/src/Common/ZooKeeper/ZooKeeperRetries.h @@ -5,8 +5,6 @@ #include #include -#include - namespace DB { @@ -220,8 +218,8 @@ private: return false; } - if (process_list_element && !process_list_element->checkTimeLimitSoft()) - return false; + if (process_list_element) + process_list_element->checkTimeLimit(); /// retries logLastError("will retry due to error"); diff --git a/src/Common/tests/gtest_connection_pool.cpp b/src/Common/tests/gtest_connection_pool.cpp new file mode 100644 index 00000000000..c271cc0e2ec --- /dev/null +++ b/src/Common/tests/gtest_connection_pool.cpp @@ -0,0 +1,585 @@ +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include + +namespace +{ + +size_t stream_copy_n(std::istream & in, std::ostream & out, std::size_t count = std::numeric_limits::max()) +{ + const size_t buffer_size = 4096; + char buffer[buffer_size]; + + size_t total_read = 0; + + while (count > buffer_size) + { + in.read(buffer, buffer_size); + size_t read = in.gcount(); + out.write(buffer, read); + count -= read; + total_read += read; + + if (read == 0) + return total_read; + } + + in.read(buffer, count); + size_t read = in.gcount(); + out.write(buffer, read); + total_read += read; + + return total_read; +} + +class MockRequestHandler : public Poco::Net::HTTPRequestHandler +{ +public: + explicit MockRequestHandler(std::shared_ptr> slowdown_) + : slowdown(std::move(slowdown_)) + { + } + + void handleRequest(Poco::Net::HTTPServerRequest & request, Poco::Net::HTTPServerResponse & response) override + { + response.setStatus(Poco::Net::HTTPResponse::HTTP_OK); + auto size = request.getContentLength(); + if (size > 0) + response.setContentLength(size); // ContentLength is required for keep alive + else + response.setChunkedTransferEncoding(true); // or chunk encoding + + sleepForSeconds(*slowdown); + + stream_copy_n(request.stream(), response.send(), size); + } + + std::shared_ptr> slowdown; +}; + +class HTTPRequestHandlerFactory : public Poco::Net::HTTPRequestHandlerFactory +{ +public: + explicit HTTPRequestHandlerFactory(std::shared_ptr> slowdown_) + : slowdown(std::move(slowdown_)) + { + } + + Poco::Net::HTTPRequestHandler * createRequestHandler(const Poco::Net::HTTPServerRequest &) override + { + return new MockRequestHandler(slowdown); + } + + std::shared_ptr> slowdown; +}; + +} + +using HTTPSession = Poco::Net::HTTPClientSession; +using HTTPSessionPtr = std::shared_ptr; + +class ConnectionPoolTest : public testing::Test { +protected: + ConnectionPoolTest() + { + startServer(); + } + + void SetUp() override { + timeouts = DB::ConnectionTimeouts(); + DB::HTTPConnectionPools::Limits def_limits{}; + DB::HTTPConnectionPools::instance().setLimits(def_limits, def_limits, def_limits); + + setSlowDown(0); + + DB::HTTPConnectionPools::instance().dropCache(); + DB::CurrentThread::getProfileEvents().reset(); + // Code here will be called immediately after the constructor (right + // before each test). + } + + void TearDown() override { + // Code here will be called immediately after each test (right + // before the destructor). + } + + DB::IHTTPConnectionPoolForEndpoint::Ptr getPool() + { + auto uri = Poco::URI(getServerUrl()); + return DB::HTTPConnectionPools::instance().getPool(DB::HTTPConnectionGroupType::HTTP, uri, DB::ProxyConfiguration{}); + } + + std::string getServerUrl() const + { + return "http://" + server_data.socket->address().toString(); + } + + void startServer() + { + server_data.reset(); + server_data.params = new Poco::Net::HTTPServerParams(); + server_data.socket = std::make_unique(server_data.port); + server_data.handler_factory = new HTTPRequestHandlerFactory(slowdown_receive); + server_data.server = std::make_unique( + server_data.handler_factory, *server_data.socket, server_data.params); + + server_data.server->start(); + } + + Poco::Net::HTTPServer & getServer() const + { + return *server_data.server; + } + + void setSlowDown(size_t seconds) + { + *slowdown_receive = seconds; + } + + DB::ConnectionTimeouts timeouts; + std::shared_ptr> slowdown_receive = std::make_shared>(0); + + struct ServerData + { + // just some port to avoid collisions with others tests + UInt16 port = 9871; + Poco::Net::HTTPServerParams::Ptr params; + std::unique_ptr socket; + HTTPRequestHandlerFactory::Ptr handler_factory; + std::unique_ptr server; + + ServerData() = default; + ServerData(ServerData &&) = default; + ServerData & operator =(ServerData &&) = delete; + + void reset() + { + if (server) + server->stop(); + + server = nullptr; + handler_factory = nullptr; + socket = nullptr; + params = nullptr; + } + + ~ServerData() { + reset(); + } + }; + + ServerData server_data; +}; + + +void wait_until(std::function pred) +{ + while (!pred()) + sleepForMilliseconds(250); +} + +void echoRequest(String data, HTTPSession & session) +{ + { + Poco::Net::HTTPRequest request(Poco::Net::HTTPRequest::HTTP_PUT, "/", "HTTP/1.1"); // HTTP/1.1 is required for keep alive + request.setContentLength(data.size()); + std::ostream & ostream = session.sendRequest(request); + ostream << data; + } + + { + std::stringstream result; + Poco::Net::HTTPResponse response; + std::istream & istream = session.receiveResponse(response); + ASSERT_EQ(response.getStatus(), Poco::Net::HTTPResponse::HTTP_OK); + + stream_copy_n(istream, result); + ASSERT_EQ(data, result.str()); + } +} + +TEST_F(ConnectionPoolTest, CanConnect) +{ + auto pool = getPool(); + auto connection = pool->getConnection(timeouts); + + ASSERT_TRUE(connection->connected()); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + + ASSERT_EQ(1, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().stored_count)); + + wait_until([&] () { return getServer().currentConnections() == 1; }); + ASSERT_EQ(1, getServer().currentConnections()); + ASSERT_EQ(1, getServer().totalConnections()); + + connection->reset(); + + wait_until([&] () { return getServer().currentConnections() == 0; }); + ASSERT_EQ(0, getServer().currentConnections()); + ASSERT_EQ(1, getServer().totalConnections()); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); +} + +TEST_F(ConnectionPoolTest, CanRequest) +{ + auto pool = getPool(); + auto connection = pool->getConnection(timeouts); + + echoRequest("Hello", *connection); + + ASSERT_EQ(1, getServer().totalConnections()); + ASSERT_EQ(1, getServer().currentConnections()); + + connection->reset(); + + wait_until([&] () { return getServer().currentConnections() == 0; }); + ASSERT_EQ(0, getServer().currentConnections()); + ASSERT_EQ(1, getServer().totalConnections()); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); +} + +TEST_F(ConnectionPoolTest, CanPreserve) +{ + auto pool = getPool(); + + { + auto connection = pool->getConnection(timeouts); + } + + ASSERT_EQ(1, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(1, CurrentMetrics::get(pool->getMetrics().stored_count)); + + wait_until([&] () { return getServer().currentConnections() == 1; }); + ASSERT_EQ(1, getServer().currentConnections()); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); +} + +TEST_F(ConnectionPoolTest, CanReuse) +{ + auto pool = getPool(); + + { + auto connection = pool->getConnection(timeouts); + // DB::setReuseTag(*connection); + } + + ASSERT_EQ(1, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(1, CurrentMetrics::get(pool->getMetrics().stored_count)); + + { + auto connection = pool->getConnection(timeouts); + + ASSERT_EQ(1, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().stored_count)); + + wait_until([&] () { return getServer().currentConnections() == 1; }); + ASSERT_EQ(1, getServer().currentConnections()); + + echoRequest("Hello", *connection); + + ASSERT_EQ(1, getServer().totalConnections()); + ASSERT_EQ(1, getServer().currentConnections()); + + connection->reset(); + } + + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().stored_count)); + + wait_until([&] () { return getServer().currentConnections() == 0; }); + ASSERT_EQ(0, getServer().currentConnections()); + ASSERT_EQ(1, getServer().totalConnections()); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); +} + +TEST_F(ConnectionPoolTest, CanReuse10) +{ + auto pool = getPool(); + + + for (int i = 0; i < 10; ++i) + { + auto connection = pool->getConnection(timeouts); + echoRequest("Hello", *connection); + } + + { + auto connection = pool->getConnection(timeouts); + connection->reset(); // reset just not to wait its expiration here + } + + wait_until([&] () { return getServer().currentConnections() == 0; }); + ASSERT_EQ(0, getServer().currentConnections()); + ASSERT_EQ(1, getServer().totalConnections()); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(10, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(10, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); +} + +TEST_F(ConnectionPoolTest, CanReuse5) +{ + timeouts.withHTTPKeepAliveTimeout(1); + + auto pool = getPool(); + + std::vector connections; + connections.reserve(5); + for (int i = 0; i < 5; ++i) + { + connections.push_back(pool->getConnection(timeouts)); + } + connections.clear(); + + ASSERT_EQ(5, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(5, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); + ASSERT_EQ(5, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(5, CurrentMetrics::get(pool->getMetrics().stored_count)); + + wait_until([&] () { return getServer().currentConnections() == 5; }); + ASSERT_EQ(5, getServer().currentConnections()); + ASSERT_EQ(5, getServer().totalConnections()); + + for (int i = 0; i < 5; ++i) + { + auto connection = pool->getConnection(timeouts); + echoRequest("Hello", *connection); + } + + ASSERT_EQ(5, getServer().totalConnections()); + + ASSERT_EQ(5, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(10, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(5, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); + ASSERT_EQ(5, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(5, CurrentMetrics::get(pool->getMetrics().stored_count)); +} + +TEST_F(ConnectionPoolTest, CanReconnectAndCreate) +{ + auto pool = getPool(); + + std::vector in_use; + + const size_t count = 2; + for (int i = 0; i < count; ++i) + { + auto connection = pool->getConnection(timeouts); + // DB::setReuseTag(*connection); + in_use.push_back(connection); + } + + ASSERT_EQ(count, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); + + ASSERT_EQ(count, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().stored_count)); + + auto connection = std::move(in_use.back()); + in_use.pop_back(); + + echoRequest("Hello", *connection); + + connection->abort(); // further usage requires reconnect, new connection + + echoRequest("Hello", *connection); + + connection->reset(); + + wait_until([&] () { return getServer().currentConnections() == 1; }); + ASSERT_EQ(1, getServer().currentConnections()); + ASSERT_EQ(count+1, getServer().totalConnections()); + + ASSERT_EQ(count+1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); +} + +TEST_F(ConnectionPoolTest, CanReconnectAndReuse) +{ + auto pool = getPool(); + + std::vector in_use; + + const size_t count = 2; + for (int i = 0; i < count; ++i) + { + auto connection = pool->getConnection(timeouts); + // DB::setReuseTag(*connection); + in_use.push_back(std::move(connection)); + } + + auto connection = std::move(in_use.back()); + in_use.pop_back(); + in_use.clear(); // other connection will be reused + + echoRequest("Hello", *connection); + + connection->abort(); // further usage requires reconnect, reuse connection from pool + + echoRequest("Hello", *connection); + + connection->reset(); + + wait_until([&] () { return getServer().currentConnections() == 0; }); + ASSERT_EQ(0, getServer().currentConnections()); + ASSERT_EQ(2, getServer().totalConnections()); + + ASSERT_EQ(count, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); +} + +TEST_F(ConnectionPoolTest, ReceiveTimeout) +{ + setSlowDown(2); + timeouts.withReceiveTimeout(1); + + auto pool = getPool(); + + { + auto connection = pool->getConnection(timeouts); + ASSERT_ANY_THROW( + echoRequest("Hello", *connection); + ); + } + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reset]); + + { + timeouts.withReceiveTimeout(3); + auto connection = pool->getConnection(timeouts); + ASSERT_NO_THROW( + echoRequest("Hello", *connection); + ); + } + + ASSERT_EQ(2, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reset]); + + { + /// timeouts have effect for reused session + timeouts.withReceiveTimeout(1); + auto connection = pool->getConnection(timeouts); + ASSERT_ANY_THROW( + echoRequest("Hello", *connection); + ); + } + + ASSERT_EQ(2, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reused]); + ASSERT_EQ(2, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reset]); +} + +TEST_F(ConnectionPoolTest, ReadWriteBufferFromHTTP) +{ + std::string_view message = "Hello ReadWriteBufferFromHTTP"; + auto uri = Poco::URI(getServerUrl()); + auto metrics = DB::HTTPConnectionPools::instance().getPool(DB::HTTPConnectionGroupType::HTTP, uri, DB::ProxyConfiguration{})->getMetrics(); + Poco::Net::HTTPBasicCredentials empty_creds; + auto buf_from_http = DB::BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(DB::HTTPConnectionGroupType::HTTP) + .withOutCallback( + [&] (std::ostream & in) + { + in << message; + }) + .withDelayInit(false) + .create(empty_creds); + + ASSERT_EQ(1, CurrentMetrics::get(metrics.active_count)); + ASSERT_EQ(0, CurrentMetrics::get(metrics.stored_count)); + + char buf[256]; + std::fill(buf, buf + sizeof(buf), 0); + + buf_from_http->readStrict(buf, message.size()); + ASSERT_EQ(std::string_view(buf), message); + ASSERT_TRUE(buf_from_http->eof()); + + buf_from_http.reset(); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[metrics.created]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[metrics.preserved]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[metrics.reused]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[metrics.reset]); + + ASSERT_EQ(1, CurrentMetrics::get(metrics.active_count)); + ASSERT_EQ(1, CurrentMetrics::get(metrics.stored_count)); +} + +TEST_F(ConnectionPoolTest, HardLimit) +{ + DB::HTTPConnectionPools::Limits zero_limits {0, 0, 0}; + DB::HTTPConnectionPools::instance().setLimits(zero_limits, zero_limits, zero_limits); + + auto pool = getPool(); + + { + auto connection = pool->getConnection(timeouts); + } + + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().stored_count)); + + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reset]); +} + +TEST_F(ConnectionPoolTest, NoReceiveCall) +{ + auto pool = getPool(); + + { + auto connection = pool->getConnection(timeouts); + + { + auto data = String("Hello"); + Poco::Net::HTTPRequest request(Poco::Net::HTTPRequest::HTTP_PUT, "/", "HTTP/1.1"); // HTTP/1.1 is required for keep alive + request.setContentLength(data.size()); + std::ostream & ostream = connection->sendRequest(request); + ostream << data; + } + + connection->flushRequest(); + } + + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().active_count)); + ASSERT_EQ(0, CurrentMetrics::get(pool->getMetrics().stored_count)); + + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().created]); + ASSERT_EQ(0, DB::CurrentThread::getProfileEvents()[pool->getMetrics().preserved]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[pool->getMetrics().reset]); +} diff --git a/src/Common/tests/gtest_resolve_pool.cpp b/src/Common/tests/gtest_resolve_pool.cpp new file mode 100644 index 00000000000..25e867fdebc --- /dev/null +++ b/src/Common/tests/gtest_resolve_pool.cpp @@ -0,0 +1,278 @@ +#include +#include +#include + +#include +#include + +class ResolvePoolMock : public DB::HostResolver +{ +public: + using ResolveFunction = DB::HostResolver::ResolveFunction; + + ResolvePoolMock(String host_, Poco::Timespan history_, ResolveFunction && func) + : DB::HostResolver(std::move(func), std::move(host_), history_) + { + } +}; + +class ResolvePoolTest : public testing::Test +{ +protected: + ResolvePoolTest() + { + DB::HostResolversPool::instance().dropCache(); + } + + void SetUp() override { + DB::CurrentThread::getProfileEvents().reset(); + + ASSERT_EQ(0, CurrentMetrics::get(metrics.active_count)); + + addresses = std::set{"127.0.0.1", "127.0.0.2", "127.0.0.3"}; + // Code here will be called immediately after the constructor (right + // before each test). + } + + void TearDown() override { + // Code here will be called immediately after each test (right + // before the destructor). + } + + DB::HostResolver::Ptr make_resolver(size_t history_ms = 200) + { + auto resolve_func = [&] (const String &) + { + std::vector result; + result.reserve(addresses.size()); + for (const auto & item : addresses) + { + result.push_back(Poco::Net::IPAddress(item)); + } + return result; + }; + + + return std::make_shared("some_host", Poco::Timespan(history_ms * 1000), std::move(resolve_func)); + } + + DB::HostResolverMetrics metrics = DB::HostResolver::getMetrics(); + std::set addresses; +}; + +TEST_F(ResolvePoolTest, CanResolve) +{ + auto resolver = make_resolver(); + auto address = resolver->resolve(); + + ASSERT_TRUE(addresses.contains(*address)); + + ASSERT_EQ(addresses.size(), DB::CurrentThread::getProfileEvents()[metrics.discovered]); + ASSERT_EQ(addresses.size(), CurrentMetrics::get(metrics.active_count)); +} + +TEST_F(ResolvePoolTest, CanResolveAll) +{ + auto resolver = make_resolver(); + + std::set results; + while (results.size() != addresses.size()) + { + auto next_addr = resolver->resolve(); + results.insert(*next_addr); + } + + ASSERT_EQ(addresses.size(), DB::CurrentThread::getProfileEvents()[metrics.discovered]); +} + +size_t getSum(std::map container) +{ + size_t sum = 0; + for (auto & [_, val] : container) + { + sum += val; + } + return sum; +} + +size_t getMin(std::map container) +{ + if (container.empty()) + return 0; + + size_t min_val = container.begin()->second; + for (auto & [_, val] : container) + { + min_val = std::min(min_val, val); + } + return min_val; +} + +double getMean(std::map container) +{ + return 1.0 * getSum(container) / container.size(); +} + +double getMaxDiff(std::map container, double ref_val) +{ + double diff = 0.0; + for (auto & [_, val] : container) + { + diff = std::max(std::fabs(val - ref_val), diff); + } + + return diff; +} + +TEST_F(ResolvePoolTest, CanResolveEvenly) +{ + auto resolver = make_resolver(); + + std::map results; + + for (size_t i = 0; i < 50000; ++i) + { + auto next_addr = resolver->resolve(); + if (results.contains(*next_addr)) + { + results[*next_addr] += 1; + } + else + { + results[*next_addr] = 1; + } + } + + auto mean = getMean(results); + auto diff = getMaxDiff(results, mean); + + ASSERT_GT(0.3 * mean, diff); +} + +TEST_F(ResolvePoolTest, CanMerge) +{ + auto resolver = make_resolver(100000); + auto address = resolver->resolve(); + + ASSERT_TRUE(addresses.contains(*address)); + + ASSERT_EQ(addresses.size(), DB::CurrentThread::getProfileEvents()[metrics.discovered]); + + auto old_addresses = addresses; + addresses = std::set{"127.0.0.4", "127.0.0.5"}; + + + resolver->update(); + ASSERT_EQ(addresses.size() + old_addresses.size(), DB::CurrentThread::getProfileEvents()[metrics.discovered]); + ASSERT_EQ(addresses.size() + old_addresses.size(), CurrentMetrics::get(metrics.active_count)); + + std::set results; + while (results.size() != addresses.size() + old_addresses.size()) + { + auto next_addr = resolver->resolve(); + results.insert(*next_addr); + } +} + +TEST_F(ResolvePoolTest, CanGainEven) +{ + auto resolver = make_resolver(); + auto address = resolver->resolve(); + + std::map results; + for (size_t i = 0; i < 40000; ++i) + { + auto next_addr = resolver->resolve(); + if (results.contains(*next_addr)) + { + results[*next_addr] += 1; + } + else + { + results[*next_addr] = 1; + } + } + + ASSERT_GT(getMin(results), 10000); + + addresses.insert("127.0.0.4"); + addresses.insert("127.0.0.5"); + + resolver->update(); + + /// return mostly new addresses + for (size_t i = 0; i < 3000; ++i) + { + auto next_addr = resolver->resolve(); + if (results.contains(*next_addr)) + { + results[*next_addr] += 1; + } + else + { + results[*next_addr] = 1; + } + } + + ASSERT_EQ(results.size(), 5); + + ASSERT_GT(getMin(results), 1000); +} + +TEST_F(ResolvePoolTest, CanFail) +{ + auto resolver = make_resolver(10000); + + auto failed_addr = resolver->resolve(); + failed_addr.setFail(); + + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[metrics.failed]); + ASSERT_EQ(addresses.size(), CurrentMetrics::get(metrics.active_count)); + ASSERT_EQ(addresses.size(), DB::CurrentThread::getProfileEvents()[metrics.discovered]); + + for (size_t i = 0; i < 1000; ++i) + { + auto next_addr = resolver->resolve(); + + ASSERT_TRUE(addresses.contains(*next_addr)); + ASSERT_NE(*next_addr, *failed_addr); + } +} + +TEST_F(ResolvePoolTest, CanFailAndHeal) +{ + auto resolver = make_resolver(); + + auto failed_addr = resolver->resolve(); + failed_addr.setFail(); + + while (true) + { + auto next_addr = resolver->resolve(); + if (*failed_addr == *next_addr) + break; + } +} + + +TEST_F(ResolvePoolTest, CanExpire) +{ + auto resolver = make_resolver(); + + auto expired_addr = resolver->resolve(); + ASSERT_TRUE(addresses.contains(*expired_addr)); + + addresses.erase(*expired_addr); + sleepForSeconds(1); + + for (size_t i = 0; i < 1000; ++i) + { + auto next_addr = resolver->resolve(); + + ASSERT_TRUE(addresses.contains(*next_addr)); + ASSERT_NE(*next_addr, *expired_addr); + } + + ASSERT_EQ(addresses.size() + 1, DB::CurrentThread::getProfileEvents()[metrics.discovered]); + ASSERT_EQ(1, DB::CurrentThread::getProfileEvents()[metrics.expired]); +} diff --git a/src/Common/tests/gtest_thread_fuzzer.cpp b/src/Common/tests/gtest_thread_fuzzer.cpp new file mode 100644 index 00000000000..ac2ccebebc6 --- /dev/null +++ b/src/Common/tests/gtest_thread_fuzzer.cpp @@ -0,0 +1,36 @@ +#include +#include +#include +#include +#include +#include + +TEST(ThreadFuzzer, mutex) +{ + /// Initialize ThreadFuzzer::started + DB::ThreadFuzzer::instance().isEffective(); + + std::mutex mutex; + std::atomic elapsed_ns = 0; + + auto func = [&]() + { + Stopwatch watch; + for (size_t i = 0; i < 1e6; ++i) + { + mutex.lock(); + mutex.unlock(); + } + elapsed_ns += watch.elapsedNanoseconds(); + }; + + std::vector> threads(10); + + for (auto & thread : threads) + thread.emplace(func); + + for (auto & thread : threads) + thread->join(); + + std::cout << "elapsed: " << elapsed_ns/1e9 << "\n"; +} diff --git a/src/Coordination/CoordinationSettings.cpp b/src/Coordination/CoordinationSettings.cpp index ea1acf02450..05f691ca76b 100644 --- a/src/Coordination/CoordinationSettings.cpp +++ b/src/Coordination/CoordinationSettings.cpp @@ -114,6 +114,8 @@ void KeeperConfigurationAndSettings::dump(WriteBufferFromOwnString & buf) const write_int(static_cast(coordination_settings->election_timeout_lower_bound_ms)); writeText("election_timeout_upper_bound_ms=", buf); write_int(static_cast(coordination_settings->election_timeout_upper_bound_ms)); + writeText("leadership_expiry_ms=", buf); + write_int(static_cast(coordination_settings->leadership_expiry_ms)); writeText("reserved_log_items=", buf); write_int(coordination_settings->reserved_log_items); diff --git a/src/Coordination/CoordinationSettings.h b/src/Coordination/CoordinationSettings.h index ed0490681b7..a32552616ee 100644 --- a/src/Coordination/CoordinationSettings.h +++ b/src/Coordination/CoordinationSettings.h @@ -26,6 +26,7 @@ struct Settings; M(Milliseconds, heart_beat_interval_ms, 500, "Heartbeat interval between quorum nodes", 0) \ M(Milliseconds, election_timeout_lower_bound_ms, 1000, "Lower bound of election timer (avoid too often leader elections)", 0) \ M(Milliseconds, election_timeout_upper_bound_ms, 2000, "Upper bound of election timer (avoid too often leader elections)", 0) \ + M(Milliseconds, leadership_expiry_ms, 0, "Duration after which a leader will expire if it fails to receive responses from peers. Set it lower or equal to election_timeout_lower_bound_ms to avoid multiple leaders.", 0) \ M(UInt64, reserved_log_items, 100000, "How many log items to store (don't remove during compaction)", 0) \ M(UInt64, snapshot_distance, 100000, "How many log items we have to collect to write new snapshot", 0) \ M(Bool, auto_forwarding, true, "Allow to forward write requests from followers to leader", 0) \ diff --git a/src/Coordination/KeeperServer.cpp b/src/Coordination/KeeperServer.cpp index cf4399461ab..57dc9596038 100644 --- a/src/Coordination/KeeperServer.cpp +++ b/src/Coordination/KeeperServer.cpp @@ -316,6 +316,18 @@ void KeeperServer::launchRaftServer(const Poco::Util::AbstractConfiguration & co } } + params.leadership_expiry_ = getValueOrMaxInt32AndLogWarning( + coordination_settings->leadership_expiry_ms.totalMilliseconds(), "leadership_expiry_ms", log); + + if (params.leadership_expiry_ > 0 && params.leadership_expiry_ <= params.election_timeout_lower_bound_) + { + LOG_INFO( + log, + "leadership_expiry_ is smaller than or equal to election_timeout_lower_bound_ms, which can avoid multiple leaders. " + "Notice that too small leadership_expiry_ may make Raft group sensitive to network status. " + ); + } + params.reserved_log_items_ = getValueOrMaxInt32AndLogWarning(coordination_settings->reserved_log_items, "reserved_log_items", log); params.snapshot_distance_ = getValueOrMaxInt32AndLogWarning(coordination_settings->snapshot_distance, "snapshot_distance", log); diff --git a/src/Coordination/KeeperStateManager.cpp b/src/Coordination/KeeperStateManager.cpp index 87c56909387..c30df0b6313 100644 --- a/src/Coordination/KeeperStateManager.cpp +++ b/src/Coordination/KeeperStateManager.cpp @@ -30,7 +30,7 @@ bool isLocalhost(const std::string & hostname) { try { - return isLocalAddress(DNSResolver::instance().resolveHost(hostname)); + return isLocalAddress(DNSResolver::instance().resolveHostAllInOriginOrder(hostname).front()); } catch (...) { diff --git a/src/Core/ServerSettings.h b/src/Core/ServerSettings.h index fc478ae4f41..c201bab6063 100644 --- a/src/Core/ServerSettings.h +++ b/src/Core/ServerSettings.h @@ -128,8 +128,17 @@ namespace DB M(Bool, format_alter_operations_with_parentheses, false, "If enabled, each operation in alter queries will be surrounded with parentheses in formatted queries to make them less ambiguous.", 0) \ M(String, default_replica_path, "/clickhouse/tables/{uuid}/{shard}", "The path to the table in ZooKeeper", 0) \ M(String, default_replica_name, "{replica}", "The replica name in ZooKeeper", 0) \ + M(UInt64, disk_connections_soft_limit, 1000, "Connections above this limit have significantly shorter time to live. The limit applies to the disks connections.", 0) \ + M(UInt64, disk_connections_warn_limit, 10000, "Warning massages are written to the logs if number of in-use connections are higher than this limit. The limit applies to the disks connections.", 0) \ + M(UInt64, disk_connections_store_limit, 12000, "Connections above this limit reset after use. Set to 0 to turn connection cache off. The limit applies to the disks connections.", 0) \ + M(UInt64, storage_connections_soft_limit, 100, "Connections above this limit have significantly shorter time to live. The limit applies to the storages connections.", 0) \ + M(UInt64, storage_connections_warn_limit, 1000, "Warning massages are written to the logs if number of in-use connections are higher than this limit. The limit applies to the storages connections.", 0) \ + M(UInt64, storage_connections_store_limit, 5000, "Connections above this limit reset after use. Set to 0 to turn connection cache off. The limit applies to the storages connections.", 0) \ + M(UInt64, http_connections_soft_limit, 100, "Connections above this limit have significantly shorter time to live. The limit applies to the http connections which do not belong to any disk or storage.", 0) \ + M(UInt64, http_connections_warn_limit, 1000, "Warning massages are written to the logs if number of in-use connections are higher than this limit. The limit applies to the http connections which do not belong to any disk or storage.", 0) \ + M(UInt64, http_connections_store_limit, 5000, "Connections above this limit reset after use. Set to 0 to turn connection cache off. The limit applies to the http connections which do not belong to any disk or storage.", 0) \ - /// If you add a setting which can be updated at runtime, please update 'changeable_settings' map in StorageSystemServerSettings.cpp +/// If you add a setting which can be updated at runtime, please update 'changeable_settings' map in StorageSystemServerSettings.cpp DECLARE_SETTINGS_TRAITS(ServerSettingsTraits, SERVER_SETTINGS) diff --git a/src/Core/Settings.h b/src/Core/Settings.h index 30bc397cbbb..d70a6cf51c5 100644 --- a/src/Core/Settings.h +++ b/src/Core/Settings.h @@ -110,7 +110,7 @@ class IColumn; M(Bool, s3_disable_checksum, false, "Do not calculate a checksum when sending a file to S3. This speeds up writes by avoiding excessive processing passes on a file. It is mostly safe as the data of MergeTree tables is checksummed by ClickHouse anyway, and when S3 is accessed with HTTPS, the TLS layer already provides integrity while transferring through the network. While additional checksums on S3 give defense in depth.", 0) \ M(UInt64, s3_retry_attempts, 100, "Setting for Aws::Client::RetryStrategy, Aws::Client does retries itself, 0 means no retries", 0) \ M(UInt64, s3_request_timeout_ms, 30000, "Idleness timeout for sending and receiving data to/from S3. Fail if a single TCP read or write call blocks for this long.", 0) \ - M(UInt64, s3_http_connection_pool_size, 1000, "How many reusable open connections to keep per S3 endpoint. This only applies to the S3 table engine and table function, not to S3 disks (for disks, use disk config instead). Global setting, can only be set in config, overriding it per session or per query has no effect.", 0) \ + M(UInt64, s3_connect_timeout_ms, 1000, "Connection timeout for host from s3 disks.", 0) \ M(Bool, enable_s3_requests_logging, false, "Enable very explicit logging of S3 requests. Makes sense for debug only.", 0) \ M(String, s3queue_default_zookeeper_path, "/clickhouse/s3queue/", "Default zookeeper path prefix for S3Queue engine", 0) \ M(Bool, s3queue_enable_logging_to_s3queue_log, false, "Enable writing to system.s3queue_log. The value can be overwritten per table with table settings", 0) \ @@ -186,6 +186,7 @@ class IColumn; \ M(String, cluster_for_parallel_replicas, "", "Cluster for a shard in which current server is located", 0) \ M(UInt64, allow_experimental_parallel_reading_from_replicas, 0, "Use all the replicas from a shard for SELECT query execution. Reading is parallelized and coordinated dynamically. 0 - disabled, 1 - enabled, silently disable them in case of failure, 2 - enabled, throw an exception in case of failure", 0) \ + M(Bool, parallel_replicas_allow_in_with_subquery, true, "If true, subquery for IN will be executed on every follower replica.", 0) \ M(Float, parallel_replicas_single_task_marks_count_multiplier, 2, "A multiplier which will be added during calculation for minimal number of marks to retrieve from coordinator. This will be applied only for remote replicas.", 0) \ M(Bool, parallel_replicas_for_non_replicated_merge_tree, false, "If true, ClickHouse will use parallel replicas algorithm also for non-replicated MergeTree tables", 0) \ M(UInt64, parallel_replicas_min_number_of_rows_per_replica, 0, "Limit the number of replicas used in a query to (estimated rows to read / min_number_of_rows_per_replica). The max is still limited by 'max_parallel_replicas'", 0) \ @@ -597,6 +598,7 @@ class IColumn; M(Bool, normalize_function_names, true, "Normalize function names to their canonical names", 0) \ M(Bool, enable_early_constant_folding, true, "Enable query optimization where we analyze function and subqueries results and rewrite query if there are constants there", 0) \ M(Bool, deduplicate_blocks_in_dependent_materialized_views, false, "Should deduplicate blocks for materialized views if the block is not a duplicate for the table. Use true to always deduplicate in dependent tables.", 0) \ + M(Bool, throw_if_deduplication_in_dependent_materialized_views_enabled_with_async_insert, true, "Throw exception on INSERT query when the setting `deduplicate_blocks_in_dependent_materialized_views` is enabled along with `async_insert`. It guarantees correctness, because these features can't work together.", 0) \ M(Bool, update_insert_deduplication_token_in_dependent_materialized_views, false, "Should update insert deduplication token with table identifier during insert in dependent materialized views.", 0) \ M(Bool, materialized_views_ignore_errors, false, "Allows to ignore errors for MATERIALIZED VIEW, and deliver original block to the table regardless of MVs", 0) \ M(Bool, ignore_materialized_views_with_dropped_target_table, false, "Ignore MVs with dropped target table during pushing to views", 0) \ @@ -697,7 +699,7 @@ class IColumn; M(Bool, database_replicated_allow_replicated_engine_arguments, true, "Allow to create only Replicated tables in database with engine Replicated with explicit arguments", 0) \ M(Bool, cloud_mode, false, "Only available in ClickHouse Cloud", 0) \ M(UInt64, cloud_mode_engine, 1, "Only available in ClickHouse Cloud", 0) \ - M(DistributedDDLOutputMode, distributed_ddl_output_mode, DistributedDDLOutputMode::THROW, "Format of distributed DDL query result, one of: 'none', 'throw', 'null_status_on_timeout', 'never_throw'", 0) \ + M(DistributedDDLOutputMode, distributed_ddl_output_mode, DistributedDDLOutputMode::THROW, "Format of distributed DDL query result, one of: 'none', 'throw', 'null_status_on_timeout', 'never_throw', 'none_only_active', 'throw_only_active', 'null_status_on_timeout_only_active'", 0) \ M(UInt64, distributed_ddl_entry_format_version, 5, "Compatibility version of distributed DDL (ON CLUSTER) queries", 0) \ \ M(UInt64, external_storage_max_read_rows, 0, "Limit maximum number of rows when table with external engine should flush history data. Now supported only for MySQL table engine, database engine, dictionary and MaterializedMySQL. If equal to 0, this setting is disabled", 0) \ @@ -1022,6 +1024,7 @@ class IColumn; M(Bool, input_format_json_read_objects_as_strings, true, "Allow to parse JSON objects as strings in JSON input formats", 0) \ M(Bool, input_format_json_read_arrays_as_strings, true, "Allow to parse JSON arrays as strings in JSON input formats", 0) \ M(Bool, input_format_json_try_infer_named_tuples_from_objects, true, "Try to infer named tuples from JSON objects in JSON input formats", 0) \ + M(Bool, input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects, false, "Use String type instead of an exception in case of ambiguous paths in JSON objects during named tuples inference", 0) \ M(Bool, input_format_json_infer_incomplete_types_as_strings, true, "Use type String for keys that contains only Nulls or empty objects/arrays during schema inference in JSON input formats", 0) \ M(Bool, input_format_json_named_tuples_as_objects, true, "Deserialize named tuple columns as JSON objects", 0) \ M(Bool, input_format_json_ignore_unknown_keys_in_named_tuple, true, "Ignore unknown keys in json object for named tuples", 0) \ @@ -1029,7 +1032,7 @@ class IColumn; M(Bool, input_format_try_infer_integers, true, "Try to infer integers instead of floats while schema inference in text formats", 0) \ M(Bool, input_format_try_infer_dates, true, "Try to infer dates from string fields while schema inference in text formats", 0) \ M(Bool, input_format_try_infer_datetimes, true, "Try to infer datetimes from string fields while schema inference in text formats", 0) \ - M(Bool, input_format_try_infer_exponent_floats, false, "Try to infer floats in exponential notation while schema inference in text formats", 0) \ + M(Bool, input_format_try_infer_exponent_floats, false, "Try to infer floats in exponential notation while schema inference in text formats (except JSON, where exponent numbers are always inferred)", 0) \ M(Bool, output_format_markdown_escape_special_characters, false, "Escape special characters in Markdown", 0) \ M(Bool, input_format_protobuf_flatten_google_wrappers, false, "Enable Google wrappers for regular non-nested columns, e.g. google.protobuf.StringValue 'str' for String column 'str'. For Nullable columns empty wrappers are recognized as defaults, and missing as nulls", 0) \ M(Bool, output_format_protobuf_nullables_with_google_wrappers, false, "When serializing Nullable columns with Google wrappers, serialize default values as empty wrappers. If turned off, default and null values are not serialized", 0) \ diff --git a/src/Core/SettingsChangesHistory.h b/src/Core/SettingsChangesHistory.h index 4a03320d427..e680c02671a 100644 --- a/src/Core/SettingsChangesHistory.h +++ b/src/Core/SettingsChangesHistory.h @@ -85,10 +85,14 @@ namespace SettingsChangesHistory /// It's used to implement `compatibility` setting (see https://github.com/ClickHouse/ClickHouse/issues/35972) static std::map settings_changes_history = { - {"24.3", {{"allow_experimental_shared_merge_tree", false, true, "The setting is obsolete"}, + {"24.3", {{"s3_connect_timeout_ms", 1000, 1000, "Introduce new dedicated setting for s3 connection timeout"}, + {"allow_experimental_shared_merge_tree", false, true, "The setting is obsolete"}, {"use_page_cache_for_disks_without_file_cache", false, false, "Added userspace page cache"}, {"read_from_page_cache_if_exists_otherwise_bypass_cache", false, false, "Added userspace page cache"}, {"page_cache_inject_eviction", false, false, "Added userspace page cache"}, + {"input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects", false, false, "Allow to use String type for ambiguous paths during named tuple inference from JSON objects"}, + {"throw_if_deduplication_in_dependent_materialized_views_enabled_with_async_insert", false, true, "Deduplication is dependent materialized view cannot work together with async inserts."}, + {"parallel_replicas_allow_in_with_subquery", false, true, "If true, subquery for IN will be executed on every follower replica"}, }}, {"24.2", {{"allow_suspicious_variant_types", true, false, "Don't allow creating Variant type with suspicious variants by default"}, {"validate_experimental_and_suspicious_types_inside_nested_types", false, true, "Validate usage of experimental and suspicious types inside nested types"}, diff --git a/src/Core/SettingsEnums.cpp b/src/Core/SettingsEnums.cpp index 04e1d0a18c8..de30d6d8eb5 100644 --- a/src/Core/SettingsEnums.cpp +++ b/src/Core/SettingsEnums.cpp @@ -118,6 +118,7 @@ IMPLEMENT_SETTING_ENUM(DistributedDDLOutputMode, ErrorCodes::BAD_ARGUMENTS, {"null_status_on_timeout", DistributedDDLOutputMode::NULL_STATUS_ON_TIMEOUT}, {"throw_only_active", DistributedDDLOutputMode::THROW_ONLY_ACTIVE}, {"null_status_on_timeout_only_active", DistributedDDLOutputMode::NULL_STATUS_ON_TIMEOUT_ONLY_ACTIVE}, + {"none_only_active", DistributedDDLOutputMode::NONE_ONLY_ACTIVE}, {"never_throw", DistributedDDLOutputMode::NEVER_THROW}}) IMPLEMENT_SETTING_ENUM(StreamingHandleErrorMode, ErrorCodes::BAD_ARGUMENTS, diff --git a/src/Core/SettingsEnums.h b/src/Core/SettingsEnums.h index 0d0138e6246..22fcf0389d8 100644 --- a/src/Core/SettingsEnums.h +++ b/src/Core/SettingsEnums.h @@ -177,6 +177,7 @@ enum class DistributedDDLOutputMode NEVER_THROW, THROW_ONLY_ACTIVE, NULL_STATUS_ON_TIMEOUT_ONLY_ACTIVE, + NONE_ONLY_ACTIVE, }; DECLARE_SETTING_ENUM(DistributedDDLOutputMode) diff --git a/src/DataTypes/DataTypeDecimalBase.h b/src/DataTypes/DataTypeDecimalBase.h index adbe9c95b14..9887dfabcdb 100644 --- a/src/DataTypes/DataTypeDecimalBase.h +++ b/src/DataTypes/DataTypeDecimalBase.h @@ -207,4 +207,10 @@ inline DataTypePtr createDecimal(UInt64 precision_value, UInt64 scale_value) return std::make_shared>(precision_value, scale_value); } +extern template class DataTypeDecimalBase; +extern template class DataTypeDecimalBase; +extern template class DataTypeDecimalBase; +extern template class DataTypeDecimalBase; +extern template class DataTypeDecimalBase; + } diff --git a/src/DataTypes/DataTypesDecimal.cpp b/src/DataTypes/DataTypesDecimal.cpp index 7ad9f0b6fd8..77a7a3e7237 100644 --- a/src/DataTypes/DataTypesDecimal.cpp +++ b/src/DataTypes/DataTypesDecimal.cpp @@ -112,6 +112,256 @@ static DataTypePtr createExact(const ASTPtr & arguments) return createDecimal(precision, scale); } +template +requires (IsDataTypeDecimal && IsDataTypeDecimal) +ReturnType convertDecimalsImpl(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename ToDataType::FieldType & result) +{ + using FromFieldType = typename FromDataType::FieldType; + using ToFieldType = typename ToDataType::FieldType; + using MaxFieldType = std::conditional_t<(sizeof(FromFieldType) > sizeof(ToFieldType)), FromFieldType, ToFieldType>; + using MaxNativeType = typename MaxFieldType::NativeType; + + static constexpr bool throw_exception = std::is_same_v; + + MaxNativeType converted_value; + if (scale_to > scale_from) + { + converted_value = DecimalUtils::scaleMultiplier(scale_to - scale_from); + if (common::mulOverflow(static_cast(value.value), converted_value, converted_value)) + { + if constexpr (throw_exception) + throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow while multiplying {} by scale {}", + std::string(ToDataType::family_name), toString(value.value), toString(converted_value)); + else + return ReturnType(false); + } + } + else if (scale_to == scale_from) + { + converted_value = value.value; + } + else + { + converted_value = value.value / DecimalUtils::scaleMultiplier(scale_from - scale_to); + } + + if constexpr (sizeof(FromFieldType) > sizeof(ToFieldType)) + { + if (converted_value < std::numeric_limits::min() || + converted_value > std::numeric_limits::max()) + { + if constexpr (throw_exception) + throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow: {} is not in range ({}, {})", + std::string(ToDataType::family_name), toString(converted_value), + toString(std::numeric_limits::min()), + toString(std::numeric_limits::max())); + else + return ReturnType(false); + } + } + + result = static_cast(converted_value); + + return ReturnType(true); +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template void convertDecimalsImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename TO_DATA_TYPE::FieldType & result); \ + template bool convertDecimalsImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef DISPATCH + + +template +requires (IsDataTypeDecimal && IsDataTypeDecimal) +typename ToDataType::FieldType convertDecimals(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to) +{ + using ToFieldType = typename ToDataType::FieldType; + ToFieldType result; + + convertDecimalsImpl(value, scale_from, scale_to, result); + + return result; +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template typename TO_DATA_TYPE::FieldType convertDecimals(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef DISPATCH + + +template +requires (IsDataTypeDecimal && IsDataTypeDecimal) +bool tryConvertDecimals(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename ToDataType::FieldType & result) +{ + return convertDecimalsImpl(value, scale_from, scale_to, result); +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template bool tryConvertDecimals(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef DISPATCH + + +template +requires (IsDataTypeDecimal && is_arithmetic_v) +ReturnType convertFromDecimalImpl(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType & result) +{ + using FromFieldType = typename FromDataType::FieldType; + using ToFieldType = typename ToDataType::FieldType; + + return DecimalUtils::convertToImpl(value, scale, result); +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template void convertFromDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); \ + template bool convertFromDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_ARITHMETIC_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + + +template +requires (IsDataTypeDecimal && is_arithmetic_v) +inline typename ToDataType::FieldType convertFromDecimal(const typename FromDataType::FieldType & value, UInt32 scale) +{ + typename ToDataType::FieldType result; + + convertFromDecimalImpl(value, scale, result); + + return result; +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template typename TO_DATA_TYPE::FieldType convertFromDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_ARITHMETIC_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + + +template +requires (IsDataTypeDecimal && is_arithmetic_v) +inline bool tryConvertFromDecimal(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result) +{ + return convertFromDecimalImpl(value, scale, result); +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template bool tryConvertFromDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType& result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_ARITHMETIC_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + + +template +requires (is_arithmetic_v && IsDataTypeDecimal) +ReturnType convertToDecimalImpl(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType & result) +{ + using FromFieldType = typename FromDataType::FieldType; + using ToFieldType = typename ToDataType::FieldType; + using ToNativeType = typename ToFieldType::NativeType; + + static constexpr bool throw_exception = std::is_same_v; + + if constexpr (std::is_floating_point_v) + { + if (!std::isfinite(value)) + { + if constexpr (throw_exception) + throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow. Cannot convert infinity or NaN to decimal", ToDataType::family_name); + else + return ReturnType(false); + } + + auto out = value * static_cast(DecimalUtils::scaleMultiplier(scale)); + + if (out <= static_cast(std::numeric_limits::min()) || + out >= static_cast(std::numeric_limits::max())) + { + if constexpr (throw_exception) + throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow. Float is out of Decimal range", ToDataType::family_name); + else + return ReturnType(false); + } + + result = static_cast(out); + return ReturnType(true); + } + else + { + if constexpr (is_big_int_v) + return ReturnType(convertDecimalsImpl, ToDataType, ReturnType>(static_cast(value), 0, scale, result)); + else if constexpr (std::is_same_v) + return ReturnType(convertDecimalsImpl, ToDataType, ReturnType>(static_cast(value), 0, scale, result)); + else + return ReturnType(convertDecimalsImpl, ToDataType, ReturnType>(static_cast(value), 0, scale, result)); + } +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template void convertToDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); \ + template bool convertToDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_ARITHMETIC_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + + +template +requires (is_arithmetic_v && IsDataTypeDecimal) +inline typename ToDataType::FieldType convertToDecimal(const typename FromDataType::FieldType & value, UInt32 scale) +{ + typename ToDataType::FieldType result; + convertToDecimalImpl(value, scale, result); + return result; +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template typename TO_DATA_TYPE::FieldType convertToDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale); +#define INVOKE(X) FOR_EACH_ARITHMETIC_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + + +template +requires (is_arithmetic_v && IsDataTypeDecimal) +inline bool tryConvertToDecimal(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result) +{ + return convertToDecimalImpl(value, scale, result); +} + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + template bool tryConvertToDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType& result); +#define INVOKE(X) FOR_EACH_ARITHMETIC_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + + +template +DataTypePtr createDecimalMaxPrecision(UInt64 scale) +{ + return std::make_shared>(DecimalUtils::max_precision, scale); +} + +template DataTypePtr createDecimalMaxPrecision(UInt64 scale); +template DataTypePtr createDecimalMaxPrecision(UInt64 scale); +template DataTypePtr createDecimalMaxPrecision(UInt64 scale); +template DataTypePtr createDecimalMaxPrecision(UInt64 scale); + +/// Explicit template instantiations. +template class DataTypeDecimal; +template class DataTypeDecimal; +template class DataTypeDecimal; +template class DataTypeDecimal; + void registerDataTypeDecimal(DataTypeFactory & factory) { factory.registerDataType("Decimal32", createExact, DataTypeFactory::CaseInsensitive); @@ -125,10 +375,4 @@ void registerDataTypeDecimal(DataTypeFactory & factory) factory.registerAlias("FIXED", "Decimal", DataTypeFactory::CaseInsensitive); } -/// Explicit template instantiations. -template class DataTypeDecimal; -template class DataTypeDecimal; -template class DataTypeDecimal; -template class DataTypeDecimal; - } diff --git a/src/DataTypes/DataTypesDecimal.h b/src/DataTypes/DataTypesDecimal.h index e2b433cbe2f..badefc4c75a 100644 --- a/src/DataTypes/DataTypesDecimal.h +++ b/src/DataTypes/DataTypesDecimal.h @@ -3,7 +3,11 @@ #include #include #include +#include +#include #include +#include +#include #include #include @@ -13,7 +17,6 @@ namespace DB namespace ErrorCodes { - extern const int DECIMAL_OVERFLOW; extern const int LOGICAL_ERROR; } @@ -99,171 +102,145 @@ inline UInt32 getDecimalScale(const DataTypeDecimal & data_type) return data_type.getScale(); } +#define FOR_EACH_DECIMAL_TYPE(M) \ + M(DataTypeDecimal) \ + M(DataTypeDateTime64) \ + M(DataTypeDecimal32) \ + M(DataTypeDecimal64) \ + M(DataTypeDecimal128) \ + M(DataTypeDecimal256) + +#define FOR_EACH_DECIMAL_TYPE_PASS(M, X) \ + M(DataTypeDecimal, X) \ + M(DataTypeDateTime64, X) \ + M(DataTypeDecimal32, X) \ + M(DataTypeDecimal64, X) \ + M(DataTypeDecimal128, X) \ + M(DataTypeDecimal256, X) + + template requires (IsDataTypeDecimal && IsDataTypeDecimal) -inline ReturnType convertDecimalsImpl(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename ToDataType::FieldType & result) -{ - using FromFieldType = typename FromDataType::FieldType; - using ToFieldType = typename ToDataType::FieldType; - using MaxFieldType = std::conditional_t<(sizeof(FromFieldType) > sizeof(ToFieldType)), FromFieldType, ToFieldType>; - using MaxNativeType = typename MaxFieldType::NativeType; +ReturnType convertDecimalsImpl(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename ToDataType::FieldType & result); - static constexpr bool throw_exception = std::is_same_v; +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template void convertDecimalsImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename TO_DATA_TYPE::FieldType & result); \ + extern template bool convertDecimalsImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH - MaxNativeType converted_value; - if (scale_to > scale_from) - { - converted_value = DecimalUtils::scaleMultiplier(scale_to - scale_from); - if (common::mulOverflow(static_cast(value.value), converted_value, converted_value)) - { - if constexpr (throw_exception) - throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow while multiplying {} by scale {}", - std::string(ToDataType::family_name), toString(value.value), toString(converted_value)); - else - return ReturnType(false); - } - } - else if (scale_to == scale_from) - { - converted_value = value.value; - } - else - { - converted_value = value.value / DecimalUtils::scaleMultiplier(scale_from - scale_to); - } - - if constexpr (sizeof(FromFieldType) > sizeof(ToFieldType)) - { - if (converted_value < std::numeric_limits::min() || - converted_value > std::numeric_limits::max()) - { - if constexpr (throw_exception) - throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow: {} is not in range ({}, {})", - std::string(ToDataType::family_name), toString(converted_value), - toString(std::numeric_limits::min()), - toString(std::numeric_limits::max())); - else - return ReturnType(false); - } - } - - result = static_cast(converted_value); - - return ReturnType(true); -} template requires (IsDataTypeDecimal && IsDataTypeDecimal) -inline typename ToDataType::FieldType convertDecimals(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to) -{ - using ToFieldType = typename ToDataType::FieldType; - ToFieldType result; +typename ToDataType::FieldType convertDecimals(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to); - convertDecimalsImpl(value, scale_from, scale_to, result); +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template typename TO_DATA_TYPE::FieldType convertDecimals(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH - return result; -} template requires (IsDataTypeDecimal && IsDataTypeDecimal) -inline bool tryConvertDecimals(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename ToDataType::FieldType & result) -{ - return convertDecimalsImpl(value, scale_from, scale_to, result); -} +bool tryConvertDecimals(const typename FromDataType::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename ToDataType::FieldType & result); + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template bool tryConvertDecimals(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale_from, UInt32 scale_to, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + template requires (IsDataTypeDecimal && is_arithmetic_v) -inline ReturnType convertFromDecimalImpl(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result) -{ - using FromFieldType = typename FromDataType::FieldType; - using ToFieldType = typename ToDataType::FieldType; +ReturnType convertFromDecimalImpl(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType & result); + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template void convertFromDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); \ + extern template bool convertFromDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_ARITHMETIC_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH - return DecimalUtils::convertToImpl(value, scale, result); -} template requires (IsDataTypeDecimal && is_arithmetic_v) -inline typename ToDataType::FieldType convertFromDecimal(const typename FromDataType::FieldType & value, UInt32 scale) -{ - typename ToDataType::FieldType result; +typename ToDataType::FieldType convertFromDecimal(const typename FromDataType::FieldType & value, UInt32 scale); - convertFromDecimalImpl(value, scale, result); +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template typename TO_DATA_TYPE::FieldType convertFromDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_ARITHMETIC_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH - return result; -} template requires (IsDataTypeDecimal && is_arithmetic_v) -inline bool tryConvertFromDecimal(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result) -{ - return convertFromDecimalImpl(value, scale, result); -} +bool tryConvertFromDecimal(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result); + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template bool tryConvertFromDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType& result); +#define INVOKE(X) FOR_EACH_DECIMAL_TYPE_PASS(DISPATCH, X) +FOR_EACH_ARITHMETIC_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + template requires (is_arithmetic_v && IsDataTypeDecimal) -inline ReturnType convertToDecimalImpl(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result) -{ - using FromFieldType = typename FromDataType::FieldType; - using ToFieldType = typename ToDataType::FieldType; - using ToNativeType = typename ToFieldType::NativeType; +ReturnType convertToDecimalImpl(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result); - static constexpr bool throw_exception = std::is_same_v; +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template void convertToDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); \ + extern template bool convertToDecimalImpl(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType & result); +#define INVOKE(X) FOR_EACH_ARITHMETIC_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH - if constexpr (std::is_floating_point_v) - { - if (!std::isfinite(value)) - { - if constexpr (throw_exception) - throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow. Cannot convert infinity or NaN to decimal", ToDataType::family_name); - else - return ReturnType(false); - } - - auto out = value * static_cast(DecimalUtils::scaleMultiplier(scale)); - - if (out <= static_cast(std::numeric_limits::min()) || - out >= static_cast(std::numeric_limits::max())) - { - if constexpr (throw_exception) - throw Exception(ErrorCodes::DECIMAL_OVERFLOW, "{} convert overflow. Float is out of Decimal range", ToDataType::family_name); - else - return ReturnType(false); - } - - result = static_cast(out); - return ReturnType(true); - } - else - { - if constexpr (is_big_int_v) - return ReturnType(convertDecimalsImpl, ToDataType, ReturnType>(static_cast(value), 0, scale, result)); - else if constexpr (std::is_same_v) - return ReturnType(convertDecimalsImpl, ToDataType, ReturnType>(static_cast(value), 0, scale, result)); - else - return ReturnType(convertDecimalsImpl, ToDataType, ReturnType>(static_cast(value), 0, scale, result)); - } -} template requires (is_arithmetic_v && IsDataTypeDecimal) -inline typename ToDataType::FieldType convertToDecimal(const typename FromDataType::FieldType & value, UInt32 scale) -{ - typename ToDataType::FieldType result; - convertToDecimalImpl(value, scale, result); - return result; -} +typename ToDataType::FieldType convertToDecimal(const typename FromDataType::FieldType & value, UInt32 scale); + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template typename TO_DATA_TYPE::FieldType convertToDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale); +#define INVOKE(X) FOR_EACH_ARITHMETIC_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + template requires (is_arithmetic_v && IsDataTypeDecimal) -inline bool tryConvertToDecimal(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result) -{ - return convertToDecimalImpl(value, scale, result); -} +bool tryConvertToDecimal(const typename FromDataType::FieldType & value, UInt32 scale, typename ToDataType::FieldType& result); + +#define DISPATCH(FROM_DATA_TYPE, TO_DATA_TYPE) \ + extern template bool tryConvertToDecimal(const typename FROM_DATA_TYPE::FieldType & value, UInt32 scale, typename TO_DATA_TYPE::FieldType& result); +#define INVOKE(X) FOR_EACH_ARITHMETIC_TYPE_PASS(DISPATCH, X) +FOR_EACH_DECIMAL_TYPE(INVOKE); +#undef INVOKE +#undef DISPATCH + template -inline DataTypePtr createDecimalMaxPrecision(UInt64 scale) -{ - return std::make_shared>(DecimalUtils::max_precision, scale); -} +DataTypePtr createDecimalMaxPrecision(UInt64 scale); + +extern template DataTypePtr createDecimalMaxPrecision(UInt64 scale); +extern template DataTypePtr createDecimalMaxPrecision(UInt64 scale); +extern template DataTypePtr createDecimalMaxPrecision(UInt64 scale); +extern template DataTypePtr createDecimalMaxPrecision(UInt64 scale); + +extern template class DataTypeDecimal; +extern template class DataTypeDecimal; +extern template class DataTypeDecimal; +extern template class DataTypeDecimal; } diff --git a/src/DataTypes/DataTypesNumber.cpp b/src/DataTypes/DataTypesNumber.cpp index 008fa287064..99446d24eed 100644 --- a/src/DataTypes/DataTypesNumber.cpp +++ b/src/DataTypes/DataTypesNumber.cpp @@ -102,4 +102,21 @@ void registerDataTypeNumbers(DataTypeFactory & factory) factory.registerAlias("DOUBLE PRECISION", "Float64", DataTypeFactory::CaseInsensitive); } +/// Explicit template instantiations. +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; + +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; +template class DataTypeNumber; + } diff --git a/src/DataTypes/DataTypesNumber.h b/src/DataTypes/DataTypesNumber.h index 0c1f88a7925..d550ceababc 100644 --- a/src/DataTypes/DataTypesNumber.h +++ b/src/DataTypes/DataTypesNumber.h @@ -55,6 +55,22 @@ private: bool unsigned_can_be_signed = false; }; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; + +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; +extern template class DataTypeNumber; + using DataTypeUInt8 = DataTypeNumber; using DataTypeUInt16 = DataTypeNumber; using DataTypeUInt32 = DataTypeNumber; diff --git a/src/DataTypes/IDataType.cpp b/src/DataTypes/IDataType.cpp index 392c56343e3..40915418aea 100644 --- a/src/DataTypes/IDataType.cpp +++ b/src/DataTypes/IDataType.cpp @@ -267,4 +267,91 @@ SerializationPtr IDataType::getSerialization(const NameAndTypePair & column) return column.type->getDefaultSerialization(); } +#define FOR_TYPES_OF_TYPE(M) \ + M(TypeIndex) \ + M(const IDataType &) \ + M(const DataTypePtr &) \ + M(WhichDataType) + +#define DISPATCH(TYPE) \ +bool isUInt8(TYPE data_type) { return WhichDataType(data_type).isUInt8(); } \ +bool isUInt16(TYPE data_type) { return WhichDataType(data_type).isUInt16(); } \ +bool isUInt32(TYPE data_type) { return WhichDataType(data_type).isUInt32(); } \ +bool isUInt64(TYPE data_type) { return WhichDataType(data_type).isUInt64(); } \ +bool isNativeUInt(TYPE data_type) { return WhichDataType(data_type).isNativeUInt(); } \ +bool isUInt(TYPE data_type) { return WhichDataType(data_type).isUInt(); } \ +\ +bool isInt8(TYPE data_type) { return WhichDataType(data_type).isInt8(); } \ +bool isInt16(TYPE data_type) { return WhichDataType(data_type).isInt16(); } \ +bool isInt32(TYPE data_type) { return WhichDataType(data_type).isInt32(); } \ +bool isInt64(TYPE data_type) { return WhichDataType(data_type).isInt64(); } \ +bool isNativeInt(TYPE data_type) { return WhichDataType(data_type).isNativeInt(); } \ +bool isInt(TYPE data_type) { return WhichDataType(data_type).isInt(); } \ +\ +bool isInteger(TYPE data_type) { return WhichDataType(data_type).isInteger(); } \ +bool isNativeInteger(TYPE data_type) { return WhichDataType(data_type).isNativeInteger(); } \ +\ +bool isDecimal(TYPE data_type) { return WhichDataType(data_type).isDecimal(); } \ +\ +bool isFloat(TYPE data_type) { return WhichDataType(data_type).isFloat(); } \ +\ +bool isNativeNumber(TYPE data_type) { return WhichDataType(data_type).isNativeNumber(); } \ +bool isNumber(TYPE data_type) { return WhichDataType(data_type).isNumber(); } \ +\ +bool isEnum8(TYPE data_type) { return WhichDataType(data_type).isEnum8(); } \ +bool isEnum16(TYPE data_type) { return WhichDataType(data_type).isEnum16(); } \ +bool isEnum(TYPE data_type) { return WhichDataType(data_type).isEnum(); } \ +\ +bool isDate(TYPE data_type) { return WhichDataType(data_type).isDate(); } \ +bool isDate32(TYPE data_type) { return WhichDataType(data_type).isDate32(); } \ +bool isDateOrDate32(TYPE data_type) { return WhichDataType(data_type).isDateOrDate32(); } \ +bool isDateTime(TYPE data_type) { return WhichDataType(data_type).isDateTime(); } \ +bool isDateTime64(TYPE data_type) { return WhichDataType(data_type).isDateTime64(); } \ +bool isDateTimeOrDateTime64(TYPE data_type) { return WhichDataType(data_type).isDateTimeOrDateTime64(); } \ +bool isDateOrDate32OrDateTimeOrDateTime64(TYPE data_type) { return WhichDataType(data_type).isDateOrDate32OrDateTimeOrDateTime64(); } \ +\ +bool isString(TYPE data_type) { return WhichDataType(data_type).isString(); } \ +bool isFixedString(TYPE data_type) { return WhichDataType(data_type).isFixedString(); } \ +bool isStringOrFixedString(TYPE data_type) { return WhichDataType(data_type).isStringOrFixedString(); } \ +\ +bool isUUID(TYPE data_type) { return WhichDataType(data_type).isUUID(); } \ +bool isIPv4(TYPE data_type) { return WhichDataType(data_type).isIPv4(); } \ +bool isIPv6(TYPE data_type) { return WhichDataType(data_type).isIPv6(); } \ +bool isArray(TYPE data_type) { return WhichDataType(data_type).isArray(); } \ +bool isTuple(TYPE data_type) { return WhichDataType(data_type).isTuple(); } \ +bool isMap(TYPE data_type) {return WhichDataType(data_type).isMap(); } \ +bool isInterval(TYPE data_type) {return WhichDataType(data_type).isInterval(); } \ +bool isObject(TYPE data_type) { return WhichDataType(data_type).isObject(); } \ +bool isVariant(TYPE data_type) { return WhichDataType(data_type).isVariant(); } \ +bool isNothing(TYPE data_type) { return WhichDataType(data_type).isNothing(); } \ +\ +bool isColumnedAsNumber(TYPE data_type) \ +{ \ + WhichDataType which(data_type); \ + return which.isInteger() || which.isFloat() || which.isDateOrDate32OrDateTimeOrDateTime64() || which.isUUID() || which.isIPv4() || which.isIPv6(); \ +} \ +\ +bool isColumnedAsDecimal(TYPE data_type) \ +{ \ + WhichDataType which(data_type); \ + return which.isDecimal() || which.isDateTime64(); \ +} \ +\ +bool isNotCreatable(TYPE data_type) \ +{ \ + WhichDataType which(data_type); \ + return which.isNothing() || which.isFunction() || which.isSet(); \ +} \ +\ +bool isNotDecimalButComparableToDecimal(TYPE data_type) \ +{ \ + WhichDataType which(data_type); \ + return which.isInt() || which.isUInt() || which.isFloat(); \ +} \ + +FOR_TYPES_OF_TYPE(DISPATCH) + +#undef DISPATCH +#undef FOR_TYPES_OF_TYPE + } diff --git a/src/DataTypes/IDataType.h b/src/DataTypes/IDataType.h index 220658afda5..55f584ef1e0 100644 --- a/src/DataTypes/IDataType.h +++ b/src/DataTypes/IDataType.h @@ -424,71 +424,76 @@ struct WhichDataType /// IDataType helpers (alternative for IDataType virtual methods with single point of truth) -template inline bool isUInt8(const T & data_type) { return WhichDataType(data_type).isUInt8(); } -template inline bool isUInt16(const T & data_type) { return WhichDataType(data_type).isUInt16(); } -template inline bool isUInt32(const T & data_type) { return WhichDataType(data_type).isUInt32(); } -template inline bool isUInt64(const T & data_type) { return WhichDataType(data_type).isUInt64(); } -template inline bool isNativeUInt(const T & data_type) { return WhichDataType(data_type).isNativeUInt(); } -template inline bool isUInt(const T & data_type) { return WhichDataType(data_type).isUInt(); } +#define FOR_TYPES_OF_TYPE(M) \ + M(TypeIndex) \ + M(const IDataType &) \ + M(const DataTypePtr &) \ + M(WhichDataType) -template inline bool isInt8(const T & data_type) { return WhichDataType(data_type).isInt8(); } -template inline bool isInt16(const T & data_type) { return WhichDataType(data_type).isInt16(); } -template inline bool isInt32(const T & data_type) { return WhichDataType(data_type).isInt32(); } -template inline bool isInt64(const T & data_type) { return WhichDataType(data_type).isInt64(); } -template inline bool isNativeInt(const T & data_type) { return WhichDataType(data_type).isNativeInt(); } -template inline bool isInt(const T & data_type) { return WhichDataType(data_type).isInt(); } +#define DISPATCH(TYPE) \ +bool isUInt8(TYPE data_type); \ +bool isUInt16(TYPE data_type); \ +bool isUInt32(TYPE data_type); \ +bool isUInt64(TYPE data_type); \ +bool isNativeUInt(TYPE data_type); \ +bool isUInt(TYPE data_type); \ +\ +bool isInt8(TYPE data_type); \ +bool isInt16(TYPE data_type); \ +bool isInt32(TYPE data_type); \ +bool isInt64(TYPE data_type); \ +bool isNativeInt(TYPE data_type); \ +bool isInt(TYPE data_type); \ +\ +bool isInteger(TYPE data_type); \ +bool isNativeInteger(TYPE data_type); \ +\ +bool isDecimal(TYPE data_type); \ +\ +bool isFloat(TYPE data_type); \ +\ +bool isNativeNumber(TYPE data_type); \ +bool isNumber(TYPE data_type); \ +\ +bool isEnum8(TYPE data_type); \ +bool isEnum16(TYPE data_type); \ +bool isEnum(TYPE data_type); \ +\ +bool isDate(TYPE data_type); \ +bool isDate32(TYPE data_type); \ +bool isDateOrDate32(TYPE data_type); \ +bool isDateTime(TYPE data_type); \ +bool isDateTime64(TYPE data_type); \ +bool isDateTimeOrDateTime64(TYPE data_type); \ +bool isDateOrDate32OrDateTimeOrDateTime64(TYPE data_type); \ +\ +bool isString(TYPE data_type); \ +bool isFixedString(TYPE data_type); \ +bool isStringOrFixedString(TYPE data_type); \ +\ +bool isUUID(TYPE data_type); \ +bool isIPv4(TYPE data_type); \ +bool isIPv6(TYPE data_type); \ +bool isArray(TYPE data_type); \ +bool isTuple(TYPE data_type); \ +bool isMap(TYPE data_type); \ +bool isInterval(TYPE data_type); \ +bool isObject(TYPE data_type); \ +bool isVariant(TYPE data_type); \ +bool isNothing(TYPE data_type); \ +\ +bool isColumnedAsNumber(TYPE data_type); \ +\ +bool isColumnedAsDecimal(TYPE data_type); \ +\ +bool isNotCreatable(TYPE data_type); \ +\ +bool isNotDecimalButComparableToDecimal(TYPE data_type); \ -template inline bool isInteger(const T & data_type) { return WhichDataType(data_type).isInteger(); } -template inline bool isNativeInteger(const T & data_type) { return WhichDataType(data_type).isNativeInteger(); } +FOR_TYPES_OF_TYPE(DISPATCH) -template inline bool isDecimal(const T & data_type) { return WhichDataType(data_type).isDecimal(); } - -template inline bool isFloat(const T & data_type) { return WhichDataType(data_type).isFloat(); } - -template inline bool isNativeNumber(const T & data_type) { return WhichDataType(data_type).isNativeNumber(); } -template inline bool isNumber(const T & data_type) { return WhichDataType(data_type).isNumber(); } - -template inline bool isEnum8(const T & data_type) { return WhichDataType(data_type).isEnum8(); } -template inline bool isEnum16(const T & data_type) { return WhichDataType(data_type).isEnum16(); } -template inline bool isEnum(const T & data_type) { return WhichDataType(data_type).isEnum(); } - -template inline bool isDate(const T & data_type) { return WhichDataType(data_type).isDate(); } -template inline bool isDate32(const T & data_type) { return WhichDataType(data_type).isDate32(); } -template inline bool isDateOrDate32(const T & data_type) { return WhichDataType(data_type).isDateOrDate32(); } -template inline bool isDateTime(const T & data_type) { return WhichDataType(data_type).isDateTime(); } -template inline bool isDateTime64(const T & data_type) { return WhichDataType(data_type).isDateTime64(); } -template inline bool isDateTimeOrDateTime64(const T & data_type) { return WhichDataType(data_type).isDateTimeOrDateTime64(); } -template inline bool isDateOrDate32OrDateTimeOrDateTime64(const T & data_type) { return WhichDataType(data_type).isDateOrDate32OrDateTimeOrDateTime64(); } - -template inline bool isString(const T & data_type) { return WhichDataType(data_type).isString(); } -template inline bool isFixedString(const T & data_type) { return WhichDataType(data_type).isFixedString(); } -template inline bool isStringOrFixedString(const T & data_type) { return WhichDataType(data_type).isStringOrFixedString(); } - -template inline bool isUUID(const T & data_type) { return WhichDataType(data_type).isUUID(); } -template inline bool isIPv4(const T & data_type) { return WhichDataType(data_type).isIPv4(); } -template inline bool isIPv6(const T & data_type) { return WhichDataType(data_type).isIPv6(); } -template inline bool isArray(const T & data_type) { return WhichDataType(data_type).isArray(); } -template inline bool isTuple(const T & data_type) { return WhichDataType(data_type).isTuple(); } -template inline bool isMap(const T & data_type) {return WhichDataType(data_type).isMap(); } -template inline bool isInterval(const T & data_type) {return WhichDataType(data_type).isInterval(); } -template inline bool isObject(const T & data_type) { return WhichDataType(data_type).isObject(); } -template inline bool isVariant(const T & data_type) { return WhichDataType(data_type).isVariant(); } - -template inline bool isNothing(const T & data_type) { return WhichDataType(data_type).isNothing(); } - -template -inline bool isColumnedAsNumber(const T & data_type) -{ - WhichDataType which(data_type); - return which.isInteger() || which.isFloat() || which.isDateOrDate32OrDateTimeOrDateTime64() || which.isUUID() || which.isIPv4() || which.isIPv6(); -} - -template -inline bool isColumnedAsDecimal(const T & data_type) -{ - WhichDataType which(data_type); - return which.isDecimal() || which.isDateTime64(); -} +#undef DISPATCH +#undef FOR_TYPES_OF_TYPE // Same as isColumnedAsDecimal but also checks value type of underlyig column. template @@ -498,19 +503,6 @@ inline bool isColumnedAsDecimalT(const DataType & data_type) return (which.isDecimal() || which.isDateTime64()) && which.idx == TypeToTypeIndex; } -template -inline bool isNotCreatable(const T & data_type) -{ - WhichDataType which(data_type); - return which.isNothing() || which.isFunction() || which.isSet(); -} - -inline bool isNotDecimalButComparableToDecimal(const DataTypePtr & data_type) -{ - WhichDataType which(data_type); - return which.isInt() || which.isUInt() || which.isFloat(); -} - inline bool isBool(const DataTypePtr & data_type) { return data_type->getName() == "Bool"; diff --git a/src/DataTypes/Serializations/SerializationDecimalBase.h b/src/DataTypes/Serializations/SerializationDecimalBase.h index 08f963cedbb..5676280d34b 100644 --- a/src/DataTypes/Serializations/SerializationDecimalBase.h +++ b/src/DataTypes/Serializations/SerializationDecimalBase.h @@ -29,4 +29,10 @@ public: void deserializeBinaryBulk(IColumn & column, ReadBuffer & istr, size_t limit, double avg_value_size_hint) const override; }; +extern template class SerializationDecimalBase; +extern template class SerializationDecimalBase; +extern template class SerializationDecimalBase; +extern template class SerializationDecimalBase; +extern template class SerializationDecimalBase; + } diff --git a/src/Databases/DatabaseFilesystem.cpp b/src/Databases/DatabaseFilesystem.cpp index bc6714a6471..5af1e1ae0d2 100644 --- a/src/Databases/DatabaseFilesystem.cpp +++ b/src/Databases/DatabaseFilesystem.cpp @@ -146,20 +146,7 @@ StoragePtr DatabaseFilesystem::getTableImpl(const String & name, ContextPtr cont if (!checkTableFilePath(table_path, context_, throw_on_error)) return {}; - String format; - if (throw_on_error) - { - format = FormatFactory::instance().getFormatFromFileName(table_path); - } - else - { - auto format_maybe = FormatFactory::instance().tryGetFormatFromFileName(table_path); - if (!format_maybe) - return {}; - format = *format_maybe; - } - - auto ast_function_ptr = makeASTFunction("file", std::make_shared(table_path), std::make_shared(format)); + auto ast_function_ptr = makeASTFunction("file", std::make_shared(table_path)); auto table_function = TableFunctionFactory::instance().get(ast_function_ptr, context_); if (!table_function) diff --git a/src/Dictionaries/HTTPDictionarySource.cpp b/src/Dictionaries/HTTPDictionarySource.cpp index bf42b7931ed..dae8ec06d30 100644 --- a/src/Dictionaries/HTTPDictionarySource.cpp +++ b/src/Dictionaries/HTTPDictionarySource.cpp @@ -88,20 +88,18 @@ void HTTPDictionarySource::getUpdateFieldAndDate(Poco::URI & uri) QueryPipeline HTTPDictionarySource::loadAll() { LOG_TRACE(log, "loadAll {}", toString()); - Poco::URI uri(configuration.url); - auto in_ptr = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_GET, - ReadWriteBufferFromHTTP::OutStreamCallback(), - timeouts, - credentials, - 0, - DBMS_DEFAULT_BUFFER_SIZE, - context->getReadSettings(), - configuration.header_entries, - nullptr, false); - return createWrappedBuffer(std::move(in_ptr)); + Poco::URI uri(configuration.url); + + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withSettings(context->getReadSettings()) + .withTimeouts(timeouts) + .withHeaders(configuration.header_entries) + .withDelayInit(false) + .create(credentials); + + return createWrappedBuffer(std::move(buf)); } QueryPipeline HTTPDictionarySource::loadUpdatedAll() @@ -109,19 +107,16 @@ QueryPipeline HTTPDictionarySource::loadUpdatedAll() Poco::URI uri(configuration.url); getUpdateFieldAndDate(uri); LOG_TRACE(log, "loadUpdatedAll {}", uri.toString()); - auto in_ptr = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_GET, - ReadWriteBufferFromHTTP::OutStreamCallback(), - timeouts, - credentials, - 0, - DBMS_DEFAULT_BUFFER_SIZE, - context->getReadSettings(), - configuration.header_entries, - nullptr, false); - return createWrappedBuffer(std::move(in_ptr)); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withSettings(context->getReadSettings()) + .withTimeouts(timeouts) + .withHeaders(configuration.header_entries) + .withDelayInit(false) + .create(credentials); + + return createWrappedBuffer(std::move(buf)); } QueryPipeline HTTPDictionarySource::loadIds(const std::vector & ids) @@ -139,19 +134,18 @@ QueryPipeline HTTPDictionarySource::loadIds(const std::vector & ids) }; Poco::URI uri(configuration.url); - auto in_ptr = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_POST, - out_stream_callback, - timeouts, - credentials, - 0, - DBMS_DEFAULT_BUFFER_SIZE, - context->getReadSettings(), - configuration.header_entries, - nullptr, false); - return createWrappedBuffer(std::move(in_ptr)); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withSettings(context->getReadSettings()) + .withTimeouts(timeouts) + .withHeaders(configuration.header_entries) + .withOutCallback(std::move(out_stream_callback)) + .withDelayInit(false) + .create(credentials); + + return createWrappedBuffer(std::move(buf)); } QueryPipeline HTTPDictionarySource::loadKeys(const Columns & key_columns, const std::vector & requested_rows) @@ -169,19 +163,18 @@ QueryPipeline HTTPDictionarySource::loadKeys(const Columns & key_columns, const }; Poco::URI uri(configuration.url); - auto in_ptr = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_POST, - out_stream_callback, - timeouts, - credentials, - 0, - DBMS_DEFAULT_BUFFER_SIZE, - context->getReadSettings(), - configuration.header_entries, - nullptr, false); - return createWrappedBuffer(std::move(in_ptr)); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withSettings(context->getReadSettings()) + .withTimeouts(timeouts) + .withHeaders(configuration.header_entries) + .withOutCallback(std::move(out_stream_callback)) + .withDelayInit(false) + .create(credentials); + + return createWrappedBuffer(std::move(buf)); } bool HTTPDictionarySource::isModified() const diff --git a/src/Dictionaries/HashedArrayDictionary.cpp b/src/Dictionaries/HashedArrayDictionary.cpp index 56d42a38c8e..d09f402143e 100644 --- a/src/Dictionaries/HashedArrayDictionary.cpp +++ b/src/Dictionaries/HashedArrayDictionary.cpp @@ -1078,7 +1078,7 @@ void HashedArrayDictionary::calculateBytesAllocate bytes_allocated += container.allocated_bytes(); } - bucket_count = container.capacity(); + bucket_count += container.capacity(); } }; @@ -1089,6 +1089,13 @@ void HashedArrayDictionary::calculateBytesAllocate bytes_allocated += container.size(); } + /// `bucket_count` should be a sum over all shards, + /// but it should not be a sum over all attributes, since it is used to + /// calculate load_factor like this: `element_count / bucket_count` + /// While element_count is a sum over all shards, not over all attributes. + if (attributes.size()) + bucket_count /= attributes.size(); + if (update_field_loaded_block) bytes_allocated += update_field_loaded_block->allocatedBytes(); @@ -1167,17 +1174,24 @@ void registerDictionaryArrayHashed(DictionaryFactory & factory) if (shards <= 0 || 128 < shards) throw Exception(ErrorCodes::BAD_ARGUMENTS,"{}: SHARDS parameter should be within [1, 128]", full_name); - HashedArrayDictionaryStorageConfiguration configuration{require_nonempty, dict_lifetime, static_cast(shards)}; + Int64 shard_load_queue_backlog = config.getInt(config_prefix + dictionary_layout_prefix + ".shard_load_queue_backlog", 10000); + if (shard_load_queue_backlog <= 0) + throw Exception(ErrorCodes::BAD_ARGUMENTS, "{}: SHARD_LOAD_QUEUE_BACKLOG parameter should be greater then zero", full_name); if (source_ptr->hasUpdateField() && shards > 1) throw Exception(ErrorCodes::BAD_ARGUMENTS, "{}: SHARDS parameter does not supports for updatable source (UPDATE_FIELD)", full_name); + HashedArrayDictionaryStorageConfiguration configuration{require_nonempty, dict_lifetime, static_cast(shards), static_cast(shard_load_queue_backlog)}; + ContextMutablePtr context = copyContextAndApplySettingsFromDictionaryConfig(global_context, config, config_prefix); const auto & settings = context->getSettingsRef(); const auto * clickhouse_source = dynamic_cast(source_ptr.get()); configuration.use_async_executor = clickhouse_source && clickhouse_source->isLocal() && settings.dictionary_use_async_executor; + if (settings.max_execution_time.totalSeconds() > 0) + configuration.load_timeout = std::chrono::seconds(settings.max_execution_time.totalSeconds()); + if (dictionary_key_type == DictionaryKeyType::Simple) { if (shards > 1) diff --git a/src/Dictionaries/HashedArrayDictionary.h b/src/Dictionaries/HashedArrayDictionary.h index 4b2570ad928..9877d92d457 100644 --- a/src/Dictionaries/HashedArrayDictionary.h +++ b/src/Dictionaries/HashedArrayDictionary.h @@ -29,6 +29,7 @@ struct HashedArrayDictionaryStorageConfiguration size_t shards = 1; size_t shard_load_queue_backlog = 10000; bool use_async_executor = false; + std::chrono::seconds load_timeout{0}; }; template diff --git a/src/Dictionaries/HashedDictionary.h b/src/Dictionaries/HashedDictionary.h index 3a5e4ff6306..b3b8cc56868 100644 --- a/src/Dictionaries/HashedDictionary.h +++ b/src/Dictionaries/HashedDictionary.h @@ -67,6 +67,7 @@ struct HashedDictionaryConfiguration const bool require_nonempty; const DictionaryLifetime lifetime; bool use_async_executor = false; + const std::chrono::seconds load_timeout{0}; }; template diff --git a/src/Dictionaries/HashedDictionaryParallelLoader.h b/src/Dictionaries/HashedDictionaryParallelLoader.h index a256f6de0e0..d88ee88f9a9 100644 --- a/src/Dictionaries/HashedDictionaryParallelLoader.h +++ b/src/Dictionaries/HashedDictionaryParallelLoader.h @@ -31,6 +31,7 @@ template clas namespace ErrorCodes { extern const int LOGICAL_ERROR; + extern const int TIMEOUT_EXCEEDED; } } @@ -50,9 +51,10 @@ public: , shards(dictionary.configuration.shards) , pool(CurrentMetrics::HashedDictionaryThreads, CurrentMetrics::HashedDictionaryThreadsActive, CurrentMetrics::HashedDictionaryThreadsScheduled, shards) , shards_queues(shards) + , loading_timeout(dictionary.configuration.load_timeout) { UInt64 backlog = dictionary.configuration.shard_load_queue_backlog; - LOG_TRACE(dictionary.log, "Will load the {} dictionary using {} threads (with {} backlog)", dictionary_name, shards, backlog); + LOG_TRACE(dictionary.log, "Will load the {} dictionary using {} threads (with {} backlog and timeout {} sec)", dictionary_name, shards, backlog, loading_timeout.count()); shards_slots.resize(shards); iota(shards_slots.data(), shards_slots.size(), UInt64(0)); @@ -62,7 +64,11 @@ public: shards_queues[shard].emplace(backlog); pool.scheduleOrThrowOnError([this, shard, thread_group = CurrentThread::getGroup()] { + WorkerStatistic statistic; SCOPE_EXIT_SAFE( + LOG_TRACE(dictionary.log, "Finished worker for dictionary {} shard {}, processed {} blocks, {} rows, total time {}ms", + dictionary_name, shard, statistic.total_blocks, statistic.total_rows, statistic.total_elapsed_ms); + if (thread_group) CurrentThread::detachFromGroupIfNotDetached(); ); @@ -74,7 +80,9 @@ public: CurrentThread::attachToGroupIfDetached(thread_group); setThreadName("HashedDictLoad"); - threadWorker(shard); + LOG_TRACE(dictionary.log, "Starting worker for dictionary {}, shard {}", dictionary_name, shard); + + threadWorker(shard, statistic); }); } } @@ -87,8 +95,28 @@ public: for (size_t shard = 0; shard < shards; ++shard) { - if (!shards_queues[shard]->push(std::move(shards_blocks[shard]))) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Could not push to shards queue #{}", shard); + const auto & current_block = shards_blocks[shard]; + while (!shards_queues[shard]->tryPush(current_block, /* milliseconds= */ 100)) + { + if (shards_queues[shard]->isFinished()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Could not push to finished shards queue #{}, dictionary {}", shard, dictionary_name); + + /// We need to check if some workers failed + if (pool.active() != shards) + { + LOG_DEBUG(dictionary.log, "Some workers for dictionary {} failed, stopping all workers", dictionary_name); + stop_all_workers = true; + pool.wait(); /// We expect exception to be thrown from the failed worker thread + throw Exception(ErrorCodes::LOGICAL_ERROR, "Worker threads for dictionary {} are not active", dictionary_name); + } + + if (loading_timeout.count() && std::chrono::milliseconds(total_loading_time.elapsedMilliseconds()) > loading_timeout) + { + stop_all_workers = true; + pool.wait(); + throw Exception(ErrorCodes::TIMEOUT_EXCEEDED, "Timeout {} sec for dictionary {} loading is expired", loading_timeout.count(), dictionary_name); + } + } } } @@ -124,27 +152,49 @@ private: String dictionary_name; const size_t shards; ThreadPool pool; + std::atomic_bool stop_all_workers{false}; std::vector>> shards_queues; + std::chrono::seconds loading_timeout; + Stopwatch total_loading_time; + std::vector shards_slots; DictionaryKeysArenaHolder arena_holder; - void threadWorker(size_t shard) + struct WorkerStatistic + { + UInt64 total_elapsed_ms = 0; + UInt64 total_blocks = 0; + UInt64 total_rows = 0; + }; + + void threadWorker(size_t shard, WorkerStatistic & statistic) { Block block; DictionaryKeysArenaHolder arena_holder_; auto & shard_queue = *shards_queues[shard]; - while (shard_queue.pop(block)) + while (true) { + if (!shard_queue.tryPop(block, /* milliseconds= */ 100)) + { + /// Check if we need to stop + if (stop_all_workers || shard_queue.isFinished()) + break; + /// Timeout expired, but the queue is not finished yet, try again + continue; + } + Stopwatch watch; dictionary.blockToAttributes(block, arena_holder_, shard); UInt64 elapsed_ms = watch.elapsedMilliseconds(); - if (elapsed_ms > 1'000) - LOG_TRACE(dictionary.log, "Block processing for shard #{} is slow {}ms (rows {}).", shard, elapsed_ms, block.rows()); - } - if (!shard_queue.isFinished()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Could not pull non finished shards queue #{}", shard); + statistic.total_elapsed_ms += elapsed_ms; + statistic.total_blocks += 1; + statistic.total_rows += block.rows(); + + if (elapsed_ms > 1'000) + LOG_TRACE(dictionary.log, "Block processing for shard #{} is slow {}ms (rows {})", shard, elapsed_ms, block.rows()); + } } /// Split block to shards smaller block, using 'selector'. diff --git a/src/Dictionaries/XDBCDictionarySource.cpp b/src/Dictionaries/XDBCDictionarySource.cpp index 70fe889a8ea..1ebfc4a29b0 100644 --- a/src/Dictionaries/XDBCDictionarySource.cpp +++ b/src/Dictionaries/XDBCDictionarySource.cpp @@ -203,7 +203,7 @@ std::string XDBCDictionarySource::doInvalidateQuery(const std::string & request) } -QueryPipeline XDBCDictionarySource::loadFromQuery(const Poco::URI & url, const Block & required_sample_block, const std::string & query) const +QueryPipeline XDBCDictionarySource::loadFromQuery(const Poco::URI & uri, const Block & required_sample_block, const std::string & query) const { bridge_helper->startBridgeSync(); @@ -214,10 +214,15 @@ QueryPipeline XDBCDictionarySource::loadFromQuery(const Poco::URI & url, const B os << "query=" << escapeForFileName(query); }; - auto read_buf = std::make_unique( - url, Poco::Net::HTTPRequest::HTTP_POST, write_body_callback, timeouts, credentials); - auto format = getContext()->getInputFormat(IXDBCBridgeHelper::DEFAULT_FORMAT, *read_buf, required_sample_block, max_block_size); - format->addBuffer(std::move(read_buf)); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(timeouts) + .withOutCallback(std::move(write_body_callback)) + .create(credentials); + + auto format = getContext()->getInputFormat(IXDBCBridgeHelper::DEFAULT_FORMAT, *buf, required_sample_block, max_block_size); + format->addBuffer(std::move(buf)); return QueryPipeline(std::move(format)); } diff --git a/src/Dictionaries/XDBCDictionarySource.h b/src/Dictionaries/XDBCDictionarySource.h index 6011563c522..64d22807254 100644 --- a/src/Dictionaries/XDBCDictionarySource.h +++ b/src/Dictionaries/XDBCDictionarySource.h @@ -74,7 +74,7 @@ private: // execute invalidate_query. expects single cell in result std::string doInvalidateQuery(const std::string & request) const; - QueryPipeline loadFromQuery(const Poco::URI & url, const Block & required_sample_block, const std::string & query) const; + QueryPipeline loadFromQuery(const Poco::URI & uri, const Block & required_sample_block, const std::string & query) const; LoggerPtr log; diff --git a/src/Dictionaries/registerHashedDictionary.cpp b/src/Dictionaries/registerHashedDictionary.cpp index 6b980e2d534..5fc4f5d5cb6 100644 --- a/src/Dictionaries/registerHashedDictionary.cpp +++ b/src/Dictionaries/registerHashedDictionary.cpp @@ -77,6 +77,7 @@ void registerDictionaryHashed(DictionaryFactory & factory) require_nonempty, dict_lifetime, use_async_executor, + std::chrono::seconds(settings.max_execution_time.totalSeconds()), }; if (source_ptr->hasUpdateField() && shards > 1) diff --git a/src/Disks/DiskType.h b/src/Disks/DiskType.h index 15940ea9155..aacf0beeb1d 100644 --- a/src/Disks/DiskType.h +++ b/src/Disks/DiskType.h @@ -33,8 +33,6 @@ enum class MetadataStorageType Memory, }; -String toString(DataSourceType data_source_type); - struct DataSourceDescription { DataSourceType type; diff --git a/src/Disks/IO/ReadBufferFromWebServer.cpp b/src/Disks/IO/ReadBufferFromWebServer.cpp index cc872392738..03300cc0714 100644 --- a/src/Disks/IO/ReadBufferFromWebServer.cpp +++ b/src/Disks/IO/ReadBufferFromWebServer.cpp @@ -1,8 +1,6 @@ #include "ReadBufferFromWebServer.h" #include -#include -#include #include #include #include @@ -45,12 +43,6 @@ std::unique_ptr ReadBufferFromWebServer::initialize() { if (read_until_position < offset) throw Exception(ErrorCodes::LOGICAL_ERROR, "Attempt to read beyond right offset ({} > {})", offset, read_until_position - 1); - - LOG_DEBUG(log, "Reading with range: {}-{}", offset, read_until_position); - } - else - { - LOG_DEBUG(log, "Reading from offset: {}", offset); } const auto & settings = context->getSettingsRef(); @@ -60,19 +52,14 @@ std::unique_ptr ReadBufferFromWebServer::initialize() connection_timeouts.withConnectionTimeout(std::max(settings.http_connection_timeout, Poco::Timespan(20, 0))); connection_timeouts.withReceiveTimeout(std::max(settings.http_receive_timeout, Poco::Timespan(20, 0))); - auto res = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_GET, - ReadWriteBufferFromHTTP::OutStreamCallback(), - connection_timeouts, - credentials, - 0, - buf_size, - read_settings, - HTTPHeaderEntries{}, - &context->getRemoteHostFilter(), - /* delay_initialization */true, - use_external_buffer); + auto res = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::DISK) + .withSettings(read_settings) + .withTimeouts(connection_timeouts) + .withBufSize(buf_size) + .withHostFilter(&context->getRemoteHostFilter()) + .withExternalBuf(use_external_buffer) + .create(credentials); if (read_until_position) res->setReadUntilPosition(read_until_position); @@ -101,44 +88,43 @@ bool ReadBufferFromWebServer::nextImpl() throw Exception(ErrorCodes::LOGICAL_ERROR, "Attempt to read beyond right offset ({} > {})", offset, read_until_position - 1); } - if (impl) - { - if (!use_external_buffer) - { - /** - * impl was initialized before, pass position() to it to make - * sure there is no pending data which was not read, because - * this branch means we read sequentially. - */ - impl->position() = position(); - assert(!impl->hasPendingData()); - } - } - else + if (!impl) { impl = initialize(); + + if (!use_external_buffer) + { + BufferBase::set(impl->buffer().begin(), impl->buffer().size(), impl->offset()); + } } if (use_external_buffer) { - /** - * use_external_buffer -- means we read into the buffer which - * was passed to us from somewhere else. We do not check whether - * previously returned buffer was read or not, because this branch - * means we are prefetching data, each nextImpl() call we can fill - * a different buffer. - */ impl->set(internal_buffer.begin(), internal_buffer.size()); - assert(working_buffer.begin() != nullptr); - assert(!internal_buffer.empty()); + } + else + { + impl->position() = position(); } + chassert(available() == 0); + + chassert(pos >= working_buffer.begin()); + chassert(pos <= working_buffer.end()); + + chassert(working_buffer.begin() != nullptr); + chassert(impl->buffer().begin() != nullptr); + + chassert(impl->available() == 0); + auto result = impl->next(); + + BufferBase::set(impl->buffer().begin(), impl->buffer().size(), impl->offset()); + + chassert(working_buffer.begin() == impl->buffer().begin()); + if (result) - { - BufferBase::set(impl->buffer().begin(), impl->buffer().size(), impl->offset()); offset += working_buffer.size(); - } return result; } diff --git a/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp b/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp index de7a71e8dc1..d25add625e8 100644 --- a/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp +++ b/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp @@ -23,7 +23,6 @@ namespace ErrorCodes extern const int CANNOT_OPEN_FILE; extern const int FILE_DOESNT_EXIST; extern const int BAD_FILE_TYPE; - extern const int FILE_ALREADY_EXISTS; extern const int CANNOT_PARSE_INPUT_ASSERTION_FAILED; extern const int LOGICAL_ERROR; } @@ -593,14 +592,8 @@ void DiskObjectStorageTransaction::moveDirectory(const std::string & from_path, void DiskObjectStorageTransaction::moveFile(const String & from_path, const String & to_path) { operations_to_execute.emplace_back( - std::make_unique(object_storage, metadata_storage, [from_path, to_path, this](MetadataTransactionPtr tx) + std::make_unique(object_storage, metadata_storage, [from_path, to_path](MetadataTransactionPtr tx) { - if (metadata_storage.exists(to_path)) - throw Exception(ErrorCodes::FILE_ALREADY_EXISTS, "File already exists: {}", to_path); - - if (!metadata_storage.exists(from_path)) - throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "File {} doesn't exist, cannot move", from_path); - tx->moveFile(from_path, to_path); })); } diff --git a/src/Disks/ObjectStorages/ObjectStorageFactory.cpp b/src/Disks/ObjectStorages/ObjectStorageFactory.cpp index f4df579de73..4f198be64fe 100644 --- a/src/Disks/ObjectStorages/ObjectStorageFactory.cpp +++ b/src/Disks/ObjectStorages/ObjectStorageFactory.cpp @@ -96,10 +96,10 @@ S3::URI getS3URI(const Poco::Util::AbstractConfiguration & config, const std::st } void checkS3Capabilities( - S3ObjectStorage & storage, const S3Capabilities s3_capabilities, const String & name, const String & key_with_trailing_slash) + S3ObjectStorage & storage, const S3Capabilities s3_capabilities, const String & name) { /// If `support_batch_delete` is turned on (default), check and possibly switch it off. - if (s3_capabilities.support_batch_delete && !checkBatchRemove(storage, key_with_trailing_slash)) + if (s3_capabilities.support_batch_delete && !checkBatchRemove(storage)) { LOG_WARNING( getLogger("S3ObjectStorage"), @@ -134,7 +134,7 @@ void registerS3ObjectStorage(ObjectStorageFactory & factory) /// NOTE: should we still perform this check for clickhouse-disks? if (!skip_access_check) - checkS3Capabilities(*object_storage, s3_capabilities, name, uri.key); + checkS3Capabilities(*object_storage, s3_capabilities, name); return object_storage; }); @@ -170,7 +170,7 @@ void registerS3PlainObjectStorage(ObjectStorageFactory & factory) /// NOTE: should we still perform this check for clickhouse-disks? if (!skip_access_check) - checkS3Capabilities(*object_storage, s3_capabilities, name, uri.key); + checkS3Capabilities(*object_storage, s3_capabilities, name); return object_storage; }); @@ -206,7 +206,7 @@ void registerHDFSObjectStorage(ObjectStorageFactory & factory) #if USE_AZURE_BLOB_STORAGE && !defined(CLICKHOUSE_KEEPER_STANDALONE_BUILD) void registerAzureObjectStorage(ObjectStorageFactory & factory) { - factory.registerObjectStorageType("azure_blob_storage", []( + auto creator = []( const std::string & name, const Poco::Util::AbstractConfiguration & config, const std::string & config_prefix, @@ -220,7 +220,9 @@ void registerAzureObjectStorage(ObjectStorageFactory & factory) getAzureBlobStorageSettings(config, config_prefix, context), endpoint.prefix.empty() ? endpoint.container_name : endpoint.container_name + "/" + endpoint.prefix); - }); + }; + factory.registerObjectStorageType("azure_blob_storage", creator); + factory.registerObjectStorageType("azure", creator); } #endif @@ -254,7 +256,7 @@ void registerWebObjectStorage(ObjectStorageFactory & factory) void registerLocalObjectStorage(ObjectStorageFactory & factory) { - factory.registerObjectStorageType("local_blob_storage", []( + auto creator = []( const std::string & name, const Poco::Util::AbstractConfiguration & config, const std::string & config_prefix, @@ -267,7 +269,10 @@ void registerLocalObjectStorage(ObjectStorageFactory & factory) /// keys are mapped to the fs, object_key_prefix is a directory also fs::create_directories(object_key_prefix); return std::make_shared(object_key_prefix); - }); + }; + + factory.registerObjectStorageType("local_blob_storage", creator); + factory.registerObjectStorageType("local", creator); } #endif diff --git a/src/Disks/ObjectStorages/S3/DiskS3Utils.cpp b/src/Disks/ObjectStorages/S3/DiskS3Utils.cpp index bb7b53b2d22..529df59ca17 100644 --- a/src/Disks/ObjectStorages/S3/DiskS3Utils.cpp +++ b/src/Disks/ObjectStorages/S3/DiskS3Utils.cpp @@ -79,13 +79,14 @@ static String getServerUUID() return toString(server_uuid); } -bool checkBatchRemove(S3ObjectStorage & storage, const String & key_with_trailing_slash) +bool checkBatchRemove(S3ObjectStorage & storage) { - /// NOTE: key_with_trailing_slash is the disk prefix, it is required - /// because access is done via S3ObjectStorage not via IDisk interface - /// (since we don't have disk yet). - const String path = fmt::format("{}clickhouse_remove_objects_capability_{}", key_with_trailing_slash, getServerUUID()); - StoredObject object(path); + /// NOTE: Here we are going to write and later drop some key. + /// We are using generateObjectKeyForPath() which returns random object key. + /// That generated key is placed in a right directory where we should have write access. + const String path = fmt::format("clickhouse_remove_objects_capability_{}", getServerUUID()); + const auto key = storage.generateObjectKeyForPath(path); + StoredObject object(key.serialize(), path); try { auto file = storage.writeObject(object, WriteMode::Rewrite); diff --git a/src/Disks/ObjectStorages/S3/DiskS3Utils.h b/src/Disks/ObjectStorages/S3/DiskS3Utils.h index 29e39d4bc1b..ddd204d449d 100644 --- a/src/Disks/ObjectStorages/S3/DiskS3Utils.h +++ b/src/Disks/ObjectStorages/S3/DiskS3Utils.h @@ -18,7 +18,7 @@ ObjectStorageKeysGeneratorPtr getKeyGenerator( const String & config_prefix); class S3ObjectStorage; -bool checkBatchRemove(S3ObjectStorage & storage, const std::string & key_with_trailing_slash); +bool checkBatchRemove(S3ObjectStorage & storage); } diff --git a/src/Disks/ObjectStorages/S3/diskSettings.cpp b/src/Disks/ObjectStorages/S3/diskSettings.cpp index 4fd4b17aabe..7cc29bf1da2 100644 --- a/src/Disks/ObjectStorages/S3/diskSettings.cpp +++ b/src/Disks/ObjectStorages/S3/diskSettings.cpp @@ -67,10 +67,6 @@ std::unique_ptr getClient( client_configuration.requestTimeoutMs = config.getUInt(config_prefix + ".request_timeout_ms", S3::DEFAULT_REQUEST_TIMEOUT_MS); client_configuration.maxConnections = config.getUInt(config_prefix + ".max_connections", S3::DEFAULT_MAX_CONNECTIONS); client_configuration.endpointOverride = uri.endpoint; - client_configuration.http_keep_alive_timeout_ms = config.getUInt( - config_prefix + ".http_keep_alive_timeout_ms", DEFAULT_HTTP_KEEP_ALIVE_TIMEOUT * 1000); - client_configuration.http_connection_pool_size = config.getUInt(config_prefix + ".http_connection_pool_size", 1000); - client_configuration.wait_on_pool_size_limit = false; client_configuration.s3_use_adaptive_timeouts = config.getBool( config_prefix + ".use_adaptive_timeouts", client_configuration.s3_use_adaptive_timeouts); diff --git a/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp b/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp index 48de0bf4168..4adb92cf5c8 100644 --- a/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp +++ b/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp @@ -44,34 +44,34 @@ WebObjectStorage::loadFiles(const String & path, const std::unique_lockgetSettingsRef(), - getContext()->getServerSettings().keep_alive_timeout), - credentials, - /* max_redirects= */ 0, - /* buffer_size_= */ DBMS_DEFAULT_BUFFER_SIZE, - getContext()->getReadSettings()); + auto timeouts = ConnectionTimeouts::getHTTPTimeouts( + getContext()->getSettingsRef(), + getContext()->getServerSettings().keep_alive_timeout); + + auto metadata_buf = BuilderRWBufferFromHTTP(Poco::URI(fs::path(full_url) / ".index")) + .withConnectionGroup(HTTPConnectionGroupType::DISK) + .withSettings(getContext()->getReadSettings()) + .withTimeouts(timeouts) + .withHostFilter(&getContext()->getRemoteHostFilter()) + .withSkipNotFound(true) + .create(credentials); String file_name; - while (!metadata_buf.eof()) + while (!metadata_buf->eof()) { - readText(file_name, metadata_buf); - assertChar('\t', metadata_buf); + readText(file_name, *metadata_buf); + assertChar('\t', *metadata_buf); bool is_directory; - readBoolText(is_directory, metadata_buf); + readBoolText(is_directory, *metadata_buf); size_t size = 0; if (!is_directory) { - assertChar('\t', metadata_buf); - readIntText(size, metadata_buf); + assertChar('\t', *metadata_buf); + readIntText(size, *metadata_buf); } - assertChar('\n', metadata_buf); + assertChar('\n', *metadata_buf); FileDataPtr file_data = is_directory ? FileData::createDirectoryInfo(false) @@ -99,10 +99,6 @@ WebObjectStorage::loadFiles(const String & path, const std::unique_lockleaf_type = type; } - return root_node.getType(); + return root_node.getType(use_string_type_for_ambiguous_paths); } private: @@ -180,19 +180,8 @@ namespace /// Store path to this node for better exception message in case of ambiguous paths. String path; - DataTypePtr getType() const + DataTypePtr getType(bool use_string_type_for_ambiguous_paths) const { - /// Check if we have ambiguous paths. - /// For example: - /// 'a.b.c' : Int32 and 'a.b' : String - /// Also check if leaf type is Nothing, because the next situation is possible: - /// {"a" : {"b" : null}} -> 'a.b' : Nullable(Nothing) - /// {"a" : {"b" : {"c" : 42}}} -> 'a.b.c' : Int32 - /// And after merge we will have ambiguous paths 'a.b.c' : Int32 and 'a.b' : Nullable(Nothing), - /// but it's a valid case and we should ignore path 'a.b'. - if (leaf_type && !isNothing(removeNullable(leaf_type)) && !nodes.empty()) - throw Exception(ErrorCodes::INCORRECT_DATA, "JSON objects have ambiguous paths: '{}' with type {} and '{}'", path, leaf_type->getName(), nodes.begin()->second.path); - if (nodes.empty()) return leaf_type; @@ -203,10 +192,33 @@ namespace for (const auto & [name, node] : nodes) { node_names.push_back(name); - node_types.push_back(node.getType()); + node_types.push_back(node.getType(use_string_type_for_ambiguous_paths)); } - return std::make_shared(std::move(node_types), std::move(node_names)); + auto tuple_type = std::make_shared(std::move(node_types), std::move(node_names)); + + /// Check if we have ambiguous paths. + /// For example: + /// 'a.b.c' : Int32 and 'a.b' : String + /// Also check if leaf type is Nothing, because the next situation is possible: + /// {"a" : {"b" : null}} -> 'a.b' : Nullable(Nothing) + /// {"a" : {"b" : {"c" : 42}}} -> 'a.b.c' : Int32 + /// And after merge we will have ambiguous paths 'a.b.c' : Int32 and 'a.b' : Nullable(Nothing), + /// but it's a valid case and we should ignore path 'a.b'. + if (leaf_type && !isNothing(removeNullable(leaf_type)) && !nodes.empty()) + { + if (use_string_type_for_ambiguous_paths) + return std::make_shared(); + + throw Exception( + ErrorCodes::INCORRECT_DATA, + "JSON objects have ambiguous data: in some objects path '{}' has type '{}' and in some - '{}'. You can enable setting " + "input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects to use String type " + "for path '{}'", + path, leaf_type->getName(), tuple_type->getName(), path); + } + + return tuple_type; } }; @@ -866,13 +878,15 @@ namespace return std::make_shared(nested_types); } + template bool tryReadFloat(Float64 & value, ReadBuffer & buf, const FormatSettings & settings) { - if (settings.try_infer_exponent_floats) + if (is_json || settings.try_infer_exponent_floats) return tryReadFloatText(value, buf); return tryReadFloatTextNoExponent(value, buf); } + template DataTypePtr tryInferNumber(ReadBuffer & buf, const FormatSettings & settings) { if (buf.eof()) @@ -911,7 +925,7 @@ namespace buf.position() = number_start; } - if (tryReadFloat(tmp_float, buf, settings)) + if (tryReadFloat(tmp_float, buf, settings)) { if (read_int && buf.position() == int_end) return std::make_shared(); @@ -945,7 +959,7 @@ namespace peekable_buf.rollbackToCheckpoint(true); } - if (tryReadFloat(tmp_float, peekable_buf, settings)) + if (tryReadFloat(tmp_float, peekable_buf, settings)) { /// Float parsing reads no fewer bytes than integer parsing, /// so position of the buffer is either the same, or further. @@ -957,7 +971,7 @@ namespace return std::make_shared(); } } - else if (tryReadFloat(tmp_float, buf, settings)) + else if (tryReadFloat(tmp_float, buf, settings)) { return std::make_shared(); } @@ -966,6 +980,36 @@ namespace return nullptr; } + template + DataTypePtr tryInferNumberFromStringImpl(std::string_view field, const FormatSettings & settings) + { + ReadBufferFromString buf(field); + + if (settings.try_infer_integers) + { + Int64 tmp_int; + if (tryReadIntText(tmp_int, buf) && buf.eof()) + return std::make_shared(); + + /// We can safely get back to the start of buffer, because we read from a string and we didn't reach eof. + buf.position() = buf.buffer().begin(); + + /// In case of Int64 overflow, try to infer UInt64 + UInt64 tmp_uint; + if (tryReadIntText(tmp_uint, buf) && buf.eof()) + return std::make_shared(); + } + + /// We can safely get back to the start of buffer, because we read from a string and we didn't reach eof. + buf.position() = buf.buffer().begin(); + + Float64 tmp; + if (tryReadFloat(tmp, buf, settings) && buf.eof()) + return std::make_shared(); + + return nullptr; + } + template DataTypePtr tryInferString(ReadBuffer & buf, const FormatSettings & settings, JSONInferenceInfo * json_info) { @@ -995,7 +1039,7 @@ namespace { if (settings.json.try_infer_numbers_from_strings) { - if (auto number_type = tryInferNumberFromString(field, settings)) + if (auto number_type = tryInferNumberFromStringImpl(field, settings)) { json_info->numbers_parsed_from_json_strings.insert(number_type.get()); return number_type; @@ -1238,7 +1282,7 @@ namespace } /// Number - return tryInferNumber(buf, settings); + return tryInferNumber(buf, settings); } } @@ -1294,7 +1338,7 @@ void transformFinalInferredJSONTypeIfNeededImpl(DataTypePtr & data_type, const F return; } - data_type = json_paths->finalize(); + data_type = json_paths->finalize(settings.json.use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects); transformFinalInferredJSONTypeIfNeededImpl(data_type, settings, json_info, remain_nothing_types); return; } @@ -1377,31 +1421,7 @@ void transformFinalInferredJSONTypeIfNeeded(DataTypePtr & data_type, const Forma DataTypePtr tryInferNumberFromString(std::string_view field, const FormatSettings & settings) { - ReadBufferFromString buf(field); - - if (settings.try_infer_integers) - { - Int64 tmp_int; - if (tryReadIntText(tmp_int, buf) && buf.eof()) - return std::make_shared(); - - /// We can safely get back to the start of buffer, because we read from a string and we didn't reach eof. - buf.position() = buf.buffer().begin(); - - /// In case of Int64 overflow, try to infer UInt64 - UInt64 tmp_uint; - if (tryReadIntText(tmp_uint, buf) && buf.eof()) - return std::make_shared(); - } - - /// We can safely get back to the start of buffer, because we read from a string and we didn't reach eof. - buf.position() = buf.buffer().begin(); - - Float64 tmp; - if (tryReadFloat(tmp, buf, settings) && buf.eof()) - return std::make_shared(); - - return nullptr; + return tryInferNumberFromStringImpl(field, settings); } DataTypePtr tryInferDateOrDateTimeFromString(std::string_view field, const FormatSettings & settings) diff --git a/src/Functions/DateTimeTransforms.cpp b/src/Functions/DateTimeTransforms.cpp index 7ec13be9d6d..006d1e94ccd 100644 --- a/src/Functions/DateTimeTransforms.cpp +++ b/src/Functions/DateTimeTransforms.cpp @@ -10,16 +10,17 @@ namespace ErrorCodes void throwDateIsNotSupported(const char * name) { - throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type Date of argument for function {}", name); -} - -void throwDateTimeIsNotSupported(const char * name) -{ - throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type DateTime of argument for function {}", name); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal argument of type Date for function {}", name); } void throwDate32IsNotSupported(const char * name) { - throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type Date32 of argument for function {}", name); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal argument of type Date32 for function {}", name); } + +void throwDateTimeIsNotSupported(const char * name) +{ + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal argument of type DateTime for function {}", name); +} + } diff --git a/src/Functions/DateTimeTransforms.h b/src/Functions/DateTimeTransforms.h index cad40d0acb8..70b2a7a83b4 100644 --- a/src/Functions/DateTimeTransforms.h +++ b/src/Functions/DateTimeTransforms.h @@ -6,6 +6,7 @@ #include #include #include +#include "base/Decimal.h" #include #include #include @@ -54,8 +55,8 @@ constexpr time_t MAX_DATE_TIMESTAMP = 5662310399; // 2149-06-06 23:59:59 U constexpr time_t MAX_DATETIME_DAY_NUM = 49710; // 2106-02-07 [[noreturn]] void throwDateIsNotSupported(const char * name); -[[noreturn]] void throwDateTimeIsNotSupported(const char * name); [[noreturn]] void throwDate32IsNotSupported(const char * name); +[[noreturn]] void throwDateTimeIsNotSupported(const char * name); /// This factor transformation will say that the function is monotone everywhere. struct ZeroTransform @@ -481,7 +482,7 @@ struct ToStartOfInterval } static UInt32 execute(Int32, Int64, const DateLUTImpl &, Int64) { - throwDateIsNotSupported(TO_START_OF_INTERVAL_NAME); + throwDate32IsNotSupported(TO_START_OF_INTERVAL_NAME); } static UInt32 execute(UInt32, Int64, const DateLUTImpl &, Int64) { @@ -516,7 +517,7 @@ struct ToStartOfInterval } static UInt32 execute(Int32, Int64, const DateLUTImpl &, Int64) { - throwDateIsNotSupported(TO_START_OF_INTERVAL_NAME); + throwDate32IsNotSupported(TO_START_OF_INTERVAL_NAME); } static UInt32 execute(UInt32, Int64, const DateLUTImpl &, Int64) { @@ -538,7 +539,7 @@ struct ToStartOfInterval { Int64 scale_diff = scale_multiplier / static_cast(1000000); if (t >= 0) [[likely]] /// When we divide the `t` value we should round the result - return (t / microseconds + scale_diff / 2) / scale_diff * microseconds; + return (t + scale_diff / 2) / (microseconds * scale_diff) * microseconds; else return ((t + 1) / microseconds / scale_diff - 1) * microseconds; } @@ -559,7 +560,7 @@ struct ToStartOfInterval } static UInt32 execute(Int32, Int64, const DateLUTImpl &, Int64) { - throwDateIsNotSupported(TO_START_OF_INTERVAL_NAME); + throwDate32IsNotSupported(TO_START_OF_INTERVAL_NAME); } static UInt32 execute(UInt32, Int64, const DateLUTImpl &, Int64) { @@ -581,7 +582,7 @@ struct ToStartOfInterval { Int64 scale_diff = scale_multiplier / static_cast(1000); if (t >= 0) [[likely]] /// When we divide the `t` value we should round the result - return (t / milliseconds + scale_diff / 2) / scale_diff * milliseconds; + return (t + scale_diff / 2) / (milliseconds * scale_diff) * milliseconds; else return ((t + 1) / milliseconds / scale_diff - 1) * milliseconds; } @@ -602,7 +603,7 @@ struct ToStartOfInterval } static UInt32 execute(Int32, Int64, const DateLUTImpl &, Int64) { - throwDateIsNotSupported(TO_START_OF_INTERVAL_NAME); + throwDate32IsNotSupported(TO_START_OF_INTERVAL_NAME); } static UInt32 execute(UInt32 t, Int64 seconds, const DateLUTImpl & time_zone, Int64) { @@ -623,7 +624,7 @@ struct ToStartOfInterval } static UInt32 execute(Int32, Int64, const DateLUTImpl &, Int64) { - throwDateIsNotSupported(TO_START_OF_INTERVAL_NAME); + throwDate32IsNotSupported(TO_START_OF_INTERVAL_NAME); } static UInt32 execute(UInt32 t, Int64 minutes, const DateLUTImpl & time_zone, Int64) { @@ -644,7 +645,7 @@ struct ToStartOfInterval } static UInt32 execute(Int32, Int64, const DateLUTImpl &, Int64) { - throwDateIsNotSupported(TO_START_OF_INTERVAL_NAME); + throwDate32IsNotSupported(TO_START_OF_INTERVAL_NAME); } static UInt32 execute(UInt32 t, Int64 hours, const DateLUTImpl & time_zone, Int64) { @@ -777,7 +778,7 @@ struct ToTimeImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -802,7 +803,7 @@ struct ToStartOfMinuteImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -849,7 +850,7 @@ struct ToStartOfSecondImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -897,7 +898,7 @@ struct ToStartOfMillisecondImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -941,7 +942,7 @@ struct ToStartOfMicrosecondImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -979,7 +980,7 @@ struct ToStartOfNanosecondImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -1004,7 +1005,7 @@ struct ToStartOfFiveMinutesImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -1036,7 +1037,7 @@ struct ToStartOfTenMinutesImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -1068,7 +1069,7 @@ struct ToStartOfFifteenMinutesImpl } static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) { @@ -1103,7 +1104,7 @@ struct TimeSlotImpl static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) @@ -1142,7 +1143,7 @@ struct ToStartOfHourImpl static UInt32 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt32 execute(UInt16, const DateLUTImpl &) @@ -1429,7 +1430,7 @@ struct ToHourImpl } static UInt8 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt8 execute(UInt16, const DateLUTImpl &) { @@ -1456,7 +1457,7 @@ struct TimezoneOffsetImpl static time_t execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static time_t execute(UInt16, const DateLUTImpl &) @@ -1482,7 +1483,7 @@ struct ToMinuteImpl } static UInt8 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt8 execute(UInt16, const DateLUTImpl &) { @@ -1507,7 +1508,7 @@ struct ToSecondImpl } static UInt8 execute(Int32, const DateLUTImpl &) { - throwDateIsNotSupported(name); + throwDate32IsNotSupported(name); } static UInt8 execute(UInt16, const DateLUTImpl &) { @@ -1518,6 +1519,32 @@ struct ToSecondImpl using FactorTransform = ToStartOfMinuteImpl; }; +struct ToMillisecondImpl +{ + static constexpr auto name = "toMillisecond"; + + static UInt16 execute(const DateTime64 & datetime64, Int64 scale_multiplier, const DateLUTImpl & time_zone) + { + return time_zone.toMillisecond(datetime64, scale_multiplier); + } + + static UInt16 execute(UInt32, const DateLUTImpl &) + { + return 0; + } + static UInt16 execute(Int32, const DateLUTImpl &) + { + throwDate32IsNotSupported(name); + } + static UInt16 execute(UInt16, const DateLUTImpl &) + { + throwDateIsNotSupported(name); + } + static constexpr bool hasPreimage() { return false; } + + using FactorTransform = ZeroTransform; +}; + struct ToISOYearImpl { static constexpr auto name = "toISOYear"; diff --git a/src/Functions/FunctionBase64Conversion.h b/src/Functions/FunctionBase64Conversion.h index de922747ccd..979c589c64b 100644 --- a/src/Functions/FunctionBase64Conversion.h +++ b/src/Functions/FunctionBase64Conversion.h @@ -100,7 +100,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_arguments{ - {"value", &isStringOrFixedString, nullptr, "String or FixedString"} + {"value", static_cast(&isStringOrFixedString), nullptr, "String or FixedString"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_arguments); diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index d253095ca01..79e5ee442c2 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -170,7 +170,8 @@ public: /// DateTime, but if both operands are Dates, their type must be the same (e.g. Date - DateTime is invalid). using ResultDataType = Switch< /// Result must be Integer - Case::div_int || IsOperation::div_int_or_zero, DataTypeFromFieldType>, + Case::int_div || IsOperation::int_div_or_zero, + std::conditional_t && IsDataTypeDecimalOrNumber, DataTypeFromFieldType, InvalidType>>, /// Decimal cases Case || IsDataTypeDecimal, DecimalResultDataType>, Case< @@ -672,8 +673,8 @@ private: IsOperation::minus; static constexpr bool is_multiply = IsOperation::multiply; static constexpr bool is_float_division = IsOperation::div_floating; - static constexpr bool is_int_division = IsOperation::div_int || - IsOperation::div_int_or_zero; + static constexpr bool is_int_division = IsOperation::int_div || + IsOperation::int_div_or_zero; static constexpr bool is_division = is_float_division || is_int_division; static constexpr bool is_compare = IsOperation::least || IsOperation::greatest; @@ -781,8 +782,8 @@ class FunctionBinaryArithmetic : public IFunction static constexpr bool is_division = IsOperation::division; static constexpr bool is_bit_hamming_distance = IsOperation::bit_hamming_distance; static constexpr bool is_modulo = IsOperation::modulo; - static constexpr bool is_div_int = IsOperation::div_int; - static constexpr bool is_div_int_or_zero = IsOperation::div_int_or_zero; + static constexpr bool is_int_div = IsOperation::int_div; + static constexpr bool is_int_div_or_zero = IsOperation::int_div_or_zero; ContextPtr context; bool check_decimal_overflow = true; @@ -1007,11 +1008,11 @@ class FunctionBinaryArithmetic : public IFunction { function_name = "tupleModuloByNumber"; } - else if constexpr (is_div_int) + else if constexpr (is_int_div) { function_name = "tupleIntDivByNumber"; } - else if constexpr (is_div_int_or_zero) + else if constexpr (is_int_div_or_zero) { function_name = "tupleIntDivOrZeroByNumber"; } @@ -1466,7 +1467,7 @@ public: bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & arguments) const override { - return ((IsOperation::div_int || IsOperation::modulo || IsOperation::positive_modulo) && !arguments[1].is_const) + return ((IsOperation::int_div || IsOperation::modulo || IsOperation::positive_modulo) && !arguments[1].is_const) || (IsOperation::div_floating && (isDecimalOrNullableDecimal(arguments[0].type) || isDecimalOrNullableDecimal(arguments[1].type))); } @@ -1690,7 +1691,7 @@ public: if constexpr (!std::is_same_v) { - if constexpr (is_div_int || is_div_int_or_zero) + if constexpr (is_int_div || is_int_div_or_zero) type_res = std::make_shared(); else if constexpr (IsDataTypeDecimal && IsDataTypeDecimal) { @@ -2086,7 +2087,7 @@ ColumnPtr executeStringInteger(const ColumnsWithTypeAndName & arguments, const A right_nullmap); } /// Here we check if we have `intDiv` or `intDivOrZero` and at least one of the arguments is decimal, because in this case originally we had result as decimal, so we need to convert result into integer after calculations - else if constexpr (!decimal_with_float && (is_div_int || is_div_int_or_zero) && (IsDataTypeDecimal || IsDataTypeDecimal)) + else if constexpr (!decimal_with_float && (is_int_div || is_int_div_or_zero) && (IsDataTypeDecimal || IsDataTypeDecimal)) { if constexpr (!std::is_same_v) @@ -2624,7 +2625,7 @@ public: /// Check the case when operation is divide, intDiv or modulo and denominator is Nullable(Something). /// For divide operation we should check only Nullable(Decimal), because only this case can throw division by zero error. bool division_by_nullable = !arguments[0].type->onlyNull() && !arguments[1].type->onlyNull() && arguments[1].type->isNullable() - && (IsOperation::div_int || IsOperation::modulo || IsOperation::positive_modulo + && (IsOperation::int_div || IsOperation::modulo || IsOperation::positive_modulo || (IsOperation::div_floating && (isDecimalOrNullableDecimal(arguments[0].type) || isDecimalOrNullableDecimal(arguments[1].type)))); diff --git a/src/Functions/FunctionHelpers.h b/src/Functions/FunctionHelpers.h index 5619ebdae49..9f44d3e95c2 100644 --- a/src/Functions/FunctionHelpers.h +++ b/src/Functions/FunctionHelpers.h @@ -108,8 +108,10 @@ struct FunctionArgumentDescriptor { const char * argument_name; - std::function type_validator_func; - std::function column_validator_func; + using TypeValidator = bool (*)(const IDataType &); + TypeValidator type_validator_func; + using ColumnValidator = bool (*)(const IColumn &); + ColumnValidator column_validator_func; const char * expected_type_description; diff --git a/src/Functions/FunctionStringReplace.h b/src/Functions/FunctionStringReplace.h index 4d723a5632c..aee04a5969a 100644 --- a/src/Functions/FunctionStringReplace.h +++ b/src/Functions/FunctionStringReplace.h @@ -35,9 +35,9 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"haystack", &isStringOrFixedString, nullptr, "String or FixedString"}, - {"pattern", &isString, nullptr, "String"}, - {"replacement", &isString, nullptr, "String"} + {"haystack", static_cast(&isStringOrFixedString), nullptr, "String or FixedString"}, + {"pattern", static_cast(&isString), nullptr, "String"}, + {"replacement", static_cast(&isString), nullptr, "String"} }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/FunctionTokens.h b/src/Functions/FunctionTokens.h index 5c4e582c637..c80152bc71d 100644 --- a/src/Functions/FunctionTokens.h +++ b/src/Functions/FunctionTokens.h @@ -74,6 +74,8 @@ public: size_t getNumberOfArguments() const override { return Generator::getNumberOfArguments(); } + ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return Generator::getArgumentsThatAreAlwaysConstant(); } + DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { Generator::checkArguments(*this, arguments); @@ -184,12 +186,12 @@ static inline void checkArgumentsWithSeparatorAndOptionalMaxSubstrings( const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"separator", &isString, isColumnConst, "const String"}, - {"s", &isString, nullptr, "String"} + {"separator", static_cast(&isString), isColumnConst, "const String"}, + {"s", static_cast(&isString), nullptr, "String"} }; FunctionArgumentDescriptors optional_args{ - {"max_substrings", &isNativeInteger, isColumnConst, "const Number"}, + {"max_substrings", static_cast(&isNativeInteger), isColumnConst, "const Number"}, }; validateFunctionArgumentTypes(func, arguments, mandatory_args, optional_args); @@ -198,11 +200,11 @@ static inline void checkArgumentsWithSeparatorAndOptionalMaxSubstrings( static inline void checkArgumentsWithOptionalMaxSubstrings(const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"s", &isString, nullptr, "String"}, + {"s", static_cast(&isString), nullptr, "String"}, }; FunctionArgumentDescriptors optional_args{ - {"max_substrings", &isNativeInteger, isColumnConst, "const Number"}, + {"max_substrings", static_cast(&isNativeInteger), isColumnConst, "const Number"}, }; validateFunctionArgumentTypes(func, arguments, mandatory_args, optional_args); diff --git a/src/Functions/FunctionUnixTimestamp64.h b/src/Functions/FunctionUnixTimestamp64.h index d74237afd77..53421a565cb 100644 --- a/src/Functions/FunctionUnixTimestamp64.h +++ b/src/Functions/FunctionUnixTimestamp64.h @@ -45,7 +45,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"value", &isDateTime64, nullptr, "DateTime64"} + {"value", static_cast(&isDateTime64), nullptr, "DateTime64"} }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/FunctionsAES.h b/src/Functions/FunctionsAES.h index 4792c997f51..a03f0b602b9 100644 --- a/src/Functions/FunctionsAES.h +++ b/src/Functions/FunctionsAES.h @@ -154,21 +154,21 @@ private: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { auto optional_args = FunctionArgumentDescriptors{ - {"IV", &isStringOrFixedString, nullptr, "Initialization vector binary string"}, + {"IV", static_cast(&isStringOrFixedString), nullptr, "Initialization vector binary string"}, }; if constexpr (compatibility_mode == OpenSSLDetails::CompatibilityMode::OpenSSL) { optional_args.emplace_back(FunctionArgumentDescriptor{ - "AAD", &isStringOrFixedString, nullptr, "Additional authenticated data binary string for GCM mode" + "AAD", static_cast(&isStringOrFixedString), nullptr, "Additional authenticated data binary string for GCM mode" }); } validateFunctionArgumentTypes(*this, arguments, FunctionArgumentDescriptors{ - {"mode", &isStringOrFixedString, isColumnConst, "encryption mode string"}, - {"input", &isStringOrFixedString, {}, "plaintext"}, - {"key", &isStringOrFixedString, {}, "encryption key binary string"}, + {"mode", static_cast(&isStringOrFixedString), isColumnConst, "encryption mode string"}, + {"input", static_cast(&isStringOrFixedString), {}, "plaintext"}, + {"key", static_cast(&isStringOrFixedString), {}, "encryption key binary string"}, }, optional_args ); @@ -425,21 +425,21 @@ private: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { auto optional_args = FunctionArgumentDescriptors{ - {"IV", &isStringOrFixedString, nullptr, "Initialization vector binary string"}, + {"IV", static_cast(&isStringOrFixedString), nullptr, "Initialization vector binary string"}, }; if constexpr (compatibility_mode == OpenSSLDetails::CompatibilityMode::OpenSSL) { optional_args.emplace_back(FunctionArgumentDescriptor{ - "AAD", &isStringOrFixedString, nullptr, "Additional authenticated data binary string for GCM mode" + "AAD", static_cast(&isStringOrFixedString), nullptr, "Additional authenticated data binary string for GCM mode" }); } validateFunctionArgumentTypes(*this, arguments, FunctionArgumentDescriptors{ - {"mode", &isStringOrFixedString, isColumnConst, "decryption mode string"}, - {"input", &isStringOrFixedString, {}, "ciphertext"}, - {"key", &isStringOrFixedString, {}, "decryption key binary string"}, + {"mode", static_cast(&isStringOrFixedString), isColumnConst, "decryption mode string"}, + {"input", static_cast(&isStringOrFixedString), {}, "ciphertext"}, + {"key", static_cast(&isStringOrFixedString), {}, "decryption key binary string"}, }, optional_args ); diff --git a/src/Functions/FunctionsConversion.h b/src/Functions/FunctionsConversion.h index 19647e2f086..1522e76893e 100644 --- a/src/Functions/FunctionsConversion.h +++ b/src/Functions/FunctionsConversion.h @@ -2129,12 +2129,12 @@ public: if constexpr (to_decimal) { - mandatory_args.push_back({"scale", &isNativeInteger, &isColumnConst, "const Integer"}); + mandatory_args.push_back({"scale", static_cast(&isNativeInteger), &isColumnConst, "const Integer"}); } if (!to_decimal && isDateTime64(arguments)) { - mandatory_args.push_back({"scale", &isNativeInteger, &isColumnConst, "const Integer"}); + mandatory_args.push_back({"scale", static_cast(&isNativeInteger), &isColumnConst, "const Integer"}); } // toString(DateTime or DateTime64, [timezone: String]) @@ -2150,7 +2150,7 @@ public: // toDateTime64(value, scale : Integer[, timezone: String]) || std::is_same_v) { - optional_args.push_back({"timezone", &isString, nullptr, "String"}); + optional_args.push_back({"timezone", static_cast(&isString), nullptr, "String"}); } validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); @@ -2498,11 +2498,11 @@ public: if (isDateTime64(arguments)) { validateFunctionArgumentTypes(*this, arguments, - FunctionArgumentDescriptors{{"string", &isStringOrFixedString, nullptr, "String or FixedString"}}, + FunctionArgumentDescriptors{{"string", static_cast(&isStringOrFixedString), nullptr, "String or FixedString"}}, // optional FunctionArgumentDescriptors{ - {"precision", &isUInt8, isColumnConst, "const UInt8"}, - {"timezone", &isStringOrFixedString, isColumnConst, "const String or FixedString"}, + {"precision", static_cast(&isUInt8), isColumnConst, "const UInt8"}, + {"timezone", static_cast(&isStringOrFixedString), isColumnConst, "const String or FixedString"}, }); UInt64 scale = to_datetime64 ? DataTypeDateTime64::default_scale : 0; diff --git a/src/Functions/IsOperation.h b/src/Functions/IsOperation.h index b2c7a27d375..a74df8f4dd9 100644 --- a/src/Functions/IsOperation.h +++ b/src/Functions/IsOperation.h @@ -51,8 +51,8 @@ struct IsOperation static constexpr bool minus = IsSameOperation::value; static constexpr bool multiply = IsSameOperation::value; static constexpr bool div_floating = IsSameOperation::value; - static constexpr bool div_int = IsSameOperation::value; - static constexpr bool div_int_or_zero = IsSameOperation::value; + static constexpr bool int_div = IsSameOperation::value; + static constexpr bool int_div_or_zero = IsSameOperation::value; static constexpr bool modulo = IsSameOperation::value; static constexpr bool positive_modulo = IsSameOperation::value; static constexpr bool least = IsSameOperation::value; @@ -60,7 +60,7 @@ struct IsOperation static constexpr bool bit_hamming_distance = IsSameOperation::value; - static constexpr bool division = div_floating || div_int || div_int_or_zero || modulo; + static constexpr bool division = div_floating || int_div || int_div_or_zero || modulo; // NOTE: allow_decimal should not fully contain `division` because of divInt static constexpr bool allow_decimal = plus || minus || multiply || division || least || greatest; }; diff --git a/src/Functions/JSONArrayLength.cpp b/src/Functions/JSONArrayLength.cpp index a82c50360f9..84e87061398 100644 --- a/src/Functions/JSONArrayLength.cpp +++ b/src/Functions/JSONArrayLength.cpp @@ -45,7 +45,7 @@ namespace DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { auto args = FunctionArgumentDescriptors{ - {"json", &isString, nullptr, "String"}, + {"json", static_cast(&isString), nullptr, "String"}, }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/URL/URLHierarchy.cpp b/src/Functions/URL/URLHierarchy.cpp index 25c6c9ef40b..a0c78c5c1a2 100644 --- a/src/Functions/URL/URLHierarchy.cpp +++ b/src/Functions/URL/URLHierarchy.cpp @@ -24,10 +24,12 @@ public: static bool isVariadic() { return false; } static size_t getNumberOfArguments() { return 1; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"URL", &isString, nullptr, "String"}, + {"URL", static_cast(&isString), nullptr, "String"}, }; validateFunctionArgumentTypes(func, arguments, mandatory_args); diff --git a/src/Functions/URL/URLPathHierarchy.cpp b/src/Functions/URL/URLPathHierarchy.cpp index 9a60d4cf989..8f546ef6a56 100644 --- a/src/Functions/URL/URLPathHierarchy.cpp +++ b/src/Functions/URL/URLPathHierarchy.cpp @@ -22,10 +22,12 @@ public: static bool isVariadic() { return false; } static size_t getNumberOfArguments() { return 1; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"URL", &isString, nullptr, "String"}, + {"URL", static_cast(&isString), nullptr, "String"}, }; validateFunctionArgumentTypes(func, arguments, mandatory_args); diff --git a/src/Functions/URL/extractURLParameterNames.cpp b/src/Functions/URL/extractURLParameterNames.cpp index 08da148b43e..16ace36d39b 100644 --- a/src/Functions/URL/extractURLParameterNames.cpp +++ b/src/Functions/URL/extractURLParameterNames.cpp @@ -22,10 +22,12 @@ public: static bool isVariadic() { return false; } static size_t getNumberOfArguments() { return 1; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"URL", &isString, nullptr, "String"}, + {"URL", static_cast(&isString), nullptr, "String"}, }; validateFunctionArgumentTypes(func, arguments, mandatory_args); diff --git a/src/Functions/URL/extractURLParameters.cpp b/src/Functions/URL/extractURLParameters.cpp index 939622dd9d1..43079834872 100644 --- a/src/Functions/URL/extractURLParameters.cpp +++ b/src/Functions/URL/extractURLParameters.cpp @@ -23,10 +23,12 @@ public: static bool isVariadic() { return false; } static size_t getNumberOfArguments() { return 1; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"URL", &isString, nullptr, "String"}, + {"URL", static_cast(&isString), nullptr, "String"}, }; validateFunctionArgumentTypes(func, arguments, mandatory_args); diff --git a/src/Functions/alphaTokens.cpp b/src/Functions/alphaTokens.cpp index 35cacdbdbb8..35f434e7498 100644 --- a/src/Functions/alphaTokens.cpp +++ b/src/Functions/alphaTokens.cpp @@ -32,6 +32,8 @@ public: static size_t getNumberOfArguments() { return 0; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {1}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { checkArgumentsWithOptionalMaxSubstrings(func, arguments); diff --git a/src/Functions/array/arrayJaccardIndex.cpp b/src/Functions/array/arrayJaccardIndex.cpp index c2a4fee4845..9cb74a7aa62 100644 --- a/src/Functions/array/arrayJaccardIndex.cpp +++ b/src/Functions/array/arrayJaccardIndex.cpp @@ -84,8 +84,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"array_1", &isArray, nullptr, "Array"}, - {"array_2", &isArray, nullptr, "Array"}, + {"array_1", static_cast(&isArray), nullptr, "Array"}, + {"array_2", static_cast(&isArray), nullptr, "Array"}, }; validateFunctionArgumentTypes(*this, arguments, args); return std::make_shared>(); diff --git a/src/Functions/array/arrayRandomSample.cpp b/src/Functions/array/arrayRandomSample.cpp index 40344efb077..b08a73b93f3 100644 --- a/src/Functions/array/arrayRandomSample.cpp +++ b/src/Functions/array/arrayRandomSample.cpp @@ -36,8 +36,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"array", &isArray, nullptr, "Array"}, - {"samples", &isUInt, isColumnConst, "const UInt*"}, + {"array", static_cast(&isArray), nullptr, "Array"}, + {"samples", static_cast(&isUInt), isColumnConst, "const UInt*"}, }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/array/arrayShingles.cpp b/src/Functions/array/arrayShingles.cpp index ade1cb862f7..8932482c69c 100644 --- a/src/Functions/array/arrayShingles.cpp +++ b/src/Functions/array/arrayShingles.cpp @@ -28,8 +28,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"array", &isArray, nullptr, "Array"}, - {"length", &isInteger, nullptr, "Integer"} + {"array", static_cast(&isArray), nullptr, "Array"}, + {"length", static_cast(&isInteger), nullptr, "Integer"} }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/arrayStringConcat.cpp b/src/Functions/arrayStringConcat.cpp index 0194cc4871a..c186c0ca7e6 100644 --- a/src/Functions/arrayStringConcat.cpp +++ b/src/Functions/arrayStringConcat.cpp @@ -151,12 +151,12 @@ public: { FunctionArgumentDescriptors mandatory_args { - {"arr", &isArray, nullptr, "Array"}, + {"arr", static_cast(&isArray), nullptr, "Array"}, }; FunctionArgumentDescriptors optional_args { - {"separator", &isString, isColumnConst, "const String"}, + {"separator", static_cast(&isString), isColumnConst, "const String"}, }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); diff --git a/src/Functions/castOrDefault.cpp b/src/Functions/castOrDefault.cpp index 26eaf4f5613..970e6fd6f75 100644 --- a/src/Functions/castOrDefault.cpp +++ b/src/Functions/castOrDefault.cpp @@ -210,10 +210,10 @@ private: FunctionArgumentDescriptors optional_args; if constexpr (IsDataTypeDecimal) - mandatory_args.push_back({"scale", &isNativeInteger, &isColumnConst, "const Integer"}); + mandatory_args.push_back({"scale", static_cast(&isNativeInteger), &isColumnConst, "const Integer"}); if (std::is_same_v || std::is_same_v) - optional_args.push_back({"timezone", &isString, isColumnConst, "const String"}); + optional_args.push_back({"timezone", static_cast(&isString), isColumnConst, "const String"}); optional_args.push_back({"default_value", nullptr, nullptr, nullptr}); diff --git a/src/Functions/countMatches.h b/src/Functions/countMatches.h index e9880e6e93f..fbbb9d017ee 100644 --- a/src/Functions/countMatches.h +++ b/src/Functions/countMatches.h @@ -35,8 +35,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"haystack", &isStringOrFixedString, nullptr, "String or FixedString"}, - {"pattern", &isString, isColumnConst, "constant String"} + {"haystack", static_cast(&isStringOrFixedString), nullptr, "String or FixedString"}, + {"pattern", static_cast(&isString), isColumnConst, "constant String"} }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/extractAll.cpp b/src/Functions/extractAll.cpp index ad49f32f769..f0c18bf79b9 100644 --- a/src/Functions/extractAll.cpp +++ b/src/Functions/extractAll.cpp @@ -50,11 +50,13 @@ public: static bool isVariadic() { return false; } static size_t getNumberOfArguments() { return 2; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {1}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { FunctionArgumentDescriptors mandatory_args{ - {"haystack", &isString, nullptr, "String"}, - {"pattern", &isString, isColumnConst, "const String"} + {"haystack", static_cast(&isString), nullptr, "String"}, + {"pattern", static_cast(&isString), isColumnConst, "const String"} }; validateFunctionArgumentTypes(func, arguments, mandatory_args); diff --git a/src/Functions/extractAllGroups.h b/src/Functions/extractAllGroups.h index c64c9d6ccef..ac12cad1698 100644 --- a/src/Functions/extractAllGroups.h +++ b/src/Functions/extractAllGroups.h @@ -71,8 +71,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"haystack", &isStringOrFixedString, nullptr, "const String or const FixedString"}, - {"needle", &isStringOrFixedString, isColumnConst, "const String or const FixedString"}, + {"haystack", static_cast(&isStringOrFixedString), nullptr, "const String or const FixedString"}, + {"needle", static_cast(&isStringOrFixedString), isColumnConst, "const String or const FixedString"}, }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/extractGroups.cpp b/src/Functions/extractGroups.cpp index e22938f8565..f62352af0bd 100644 --- a/src/Functions/extractGroups.cpp +++ b/src/Functions/extractGroups.cpp @@ -45,8 +45,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"haystack", &isStringOrFixedString, nullptr, "const String or const FixedString"}, - {"needle", &isStringOrFixedString, isColumnConst, "const String or const FixedString"}, + {"haystack", static_cast(&isStringOrFixedString), nullptr, "const String or const FixedString"}, + {"needle", static_cast(&isStringOrFixedString), isColumnConst, "const String or const FixedString"}, }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/formatQuery.cpp b/src/Functions/formatQuery.cpp index 2f6bc6f9903..92403d2e88e 100644 --- a/src/Functions/formatQuery.cpp +++ b/src/Functions/formatQuery.cpp @@ -54,7 +54,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"query", &isString, nullptr, "String"} + {"query", static_cast(&isString), nullptr, "String"} }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/fromDaysSinceYearZero.cpp b/src/Functions/fromDaysSinceYearZero.cpp index a21d0cc25bf..b98c587d172 100644 --- a/src/Functions/fromDaysSinceYearZero.cpp +++ b/src/Functions/fromDaysSinceYearZero.cpp @@ -52,7 +52,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { - FunctionArgumentDescriptors args{{"days", &isNativeInteger, nullptr, "Integer"}}; + FunctionArgumentDescriptors args{{"days", static_cast(&isNativeInteger), nullptr, "Integer"}}; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/identity.cpp b/src/Functions/identity.cpp index 7174f1fd318..43cca76c801 100644 --- a/src/Functions/identity.cpp +++ b/src/Functions/identity.cpp @@ -9,4 +9,9 @@ REGISTER_FUNCTION(Identity) factory.registerFunction(); } +REGISTER_FUNCTION(ScalarSubqueryResult) +{ + factory.registerFunction(); +} + } diff --git a/src/Functions/identity.h b/src/Functions/identity.h index efee95841f5..c753625caa7 100644 --- a/src/Functions/identity.h +++ b/src/Functions/identity.h @@ -6,11 +6,12 @@ namespace DB { -class FunctionIdentity : public IFunction +template +class FunctionIdentityBase : public IFunction { public: - static constexpr auto name = "identity"; - static FunctionPtr create(ContextPtr) { return std::make_shared(); } + static constexpr auto name = Name::name; + static FunctionPtr create(ContextPtr) { return std::make_shared>(); } String getName() const override { return name; } size_t getNumberOfArguments() const override { return 1; } @@ -28,4 +29,17 @@ public: } }; +struct IdentityName +{ + static constexpr auto name = "identity"; +}; + +struct ScalarSubqueryResultName +{ + static constexpr auto name = "__scalarSubqueryResult"; +}; + +using FunctionIdentity = FunctionIdentityBase; +using FunctionScalarSubqueryResult = FunctionIdentityBase; + } diff --git a/src/Functions/makeDate.cpp b/src/Functions/makeDate.cpp index 987cf4eb1a9..c7f3c195578 100644 --- a/src/Functions/makeDate.cpp +++ b/src/Functions/makeDate.cpp @@ -82,17 +82,17 @@ public: if (is_year_month_variant) { FunctionArgumentDescriptors args{ - {mandatory_argument_names_year_month_day[0], &isNumber, nullptr, "Number"}, - {mandatory_argument_names_year_month_day[1], &isNumber, nullptr, "Number"}, - {mandatory_argument_names_year_month_day[2], &isNumber, nullptr, "Number"} + {mandatory_argument_names_year_month_day[0], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names_year_month_day[1], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names_year_month_day[2], static_cast(&isNumber), nullptr, "Number"} }; validateFunctionArgumentTypes(*this, arguments, args); } else { FunctionArgumentDescriptors args{ - {mandatory_argument_names_year_dayofyear[0], &isNumber, nullptr, "Number"}, - {mandatory_argument_names_year_dayofyear[1], &isNumber, nullptr, "Number"} + {mandatory_argument_names_year_dayofyear[0], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names_year_dayofyear[1], static_cast(&isNumber), nullptr, "Number"} }; validateFunctionArgumentTypes(*this, arguments, args); } @@ -189,7 +189,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {mandatory_argument_names[0], &isNumber, nullptr, "Number"} + {mandatory_argument_names[0], static_cast(&isNumber), nullptr, "Number"} }; validateFunctionArgumentTypes(*this, arguments, args); @@ -344,16 +344,16 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {mandatory_argument_names[0], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[1], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[2], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[3], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[4], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[5], &isNumber, nullptr, "Number"} + {mandatory_argument_names[0], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[1], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[2], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[3], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[4], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[5], static_cast(&isNumber), nullptr, "Number"} }; FunctionArgumentDescriptors optional_args{ - {optional_argument_names[0], &isString, isColumnConst, "const String"} + {optional_argument_names[0], static_cast(&isString), isColumnConst, "const String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); @@ -425,18 +425,18 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {mandatory_argument_names[0], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[1], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[2], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[3], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[4], &isNumber, nullptr, "Number"}, - {mandatory_argument_names[5], &isNumber, nullptr, "Number"} + {mandatory_argument_names[0], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[1], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[2], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[3], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[4], static_cast(&isNumber), nullptr, "Number"}, + {mandatory_argument_names[5], static_cast(&isNumber), nullptr, "Number"} }; FunctionArgumentDescriptors optional_args{ - {optional_argument_names[0], &isNumber, nullptr, "const Number"}, - {optional_argument_names[1], &isNumber, isColumnConst, "const Number"}, - {optional_argument_names[2], &isString, isColumnConst, "const String"} + {optional_argument_names[0], static_cast(&isNumber), nullptr, "const Number"}, + {optional_argument_names[1], static_cast(&isNumber), isColumnConst, "const Number"}, + {optional_argument_names[2], static_cast(&isString), isColumnConst, "const String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); @@ -564,11 +564,11 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {mandatory_argument_names[0], &isNumber, nullptr, "Number"} + {mandatory_argument_names[0], static_cast(&isNumber), nullptr, "Number"} }; FunctionArgumentDescriptors optional_args{ - {optional_argument_names[0], &isString, isColumnConst, "const String"} + {optional_argument_names[0], static_cast(&isString), isColumnConst, "const String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); @@ -643,12 +643,12 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {mandatory_argument_names[0], &isNumber, nullptr, "Number"} + {mandatory_argument_names[0], static_cast(&isNumber), nullptr, "Number"} }; FunctionArgumentDescriptors optional_args{ - {optional_argument_names[0], &isNumber, isColumnConst, "const Number"}, - {optional_argument_names[0], &isString, isColumnConst, "const String"} + {optional_argument_names[0], static_cast(&isNumber), isColumnConst, "const Number"}, + {optional_argument_names[0], static_cast(&isString), isColumnConst, "const String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); diff --git a/src/Functions/multiIf.cpp b/src/Functions/multiIf.cpp index 862ce9909e4..81304f3afbd 100644 --- a/src/Functions/multiIf.cpp +++ b/src/Functions/multiIf.cpp @@ -3,12 +3,20 @@ #include #include #include +#include #include #include #include #include #include #include +#include +#include +#include +#include +#include +#include +#include #include #include @@ -20,7 +28,7 @@ namespace ErrorCodes extern const int ILLEGAL_TYPE_OF_ARGUMENT; extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH; extern const int NOT_IMPLEMENTED; - extern const int LOGICAL_ERROR; + extern const int BAD_ARGUMENTS; } namespace @@ -149,6 +157,10 @@ public: ColumnPtr executeImpl(const ColumnsWithTypeAndName & args, const DataTypePtr & result_type, size_t input_rows_count) const override { + /// Fast path when data is empty + if (input_rows_count == 0) + return result_type->createColumn(); + ColumnsWithTypeAndName arguments = args; executeShortCircuitArguments(arguments); /** We will gather values from columns in branches to result column, @@ -249,64 +261,73 @@ public: } const WhichDataType which(removeNullable(result_type)); - bool execute_multiif_columnar - = allow_execute_multiif_columnar && !contains_short && (which.isInt() || which.isUInt() || which.isFloat()); + bool execute_multiif_columnar = allow_execute_multiif_columnar && !contains_short + && instructions.size() <= std::numeric_limits::max() + && (which.isInt() || which.isUInt() || which.isFloat() || which.isDecimal() || which.isDateOrDate32OrDateTimeOrDateTime64() + || which.isEnum() || which.isIPv4() || which.isIPv6()); size_t rows = input_rows_count; if (!execute_multiif_columnar) { MutableColumnPtr res = return_type->createColumn(); + res->reserve(rows); executeInstructions(instructions, rows, res); return std::move(res); } -#define EXECUTE_INSTRUCTIONS_COLUMNAR(TYPE, INDEX) \ +#define EXECUTE_INSTRUCTIONS_COLUMNAR(TYPE, FIELD, INDEX) \ if (which.is##TYPE()) \ { \ - MutableColumnPtr res = ColumnVector::create(rows); \ - MutableColumnPtr null_map = result_type->isNullable() ? ColumnUInt8::create(rows) : nullptr; \ - executeInstructionsColumnar(instructions, rows, res, null_map, result_type->isNullable()); \ - if (!result_type->isNullable()) \ - return std::move(res); \ + MutableColumnPtr res = result_type->createColumn(); \ + if (result_type->isNullable()) \ + { \ + auto & res_nullable = assert_cast(*res); \ + auto & res_data = assert_cast &>(res_nullable.getNestedColumn()).getData(); \ + auto & res_null_map = res_nullable.getNullMapData(); \ + executeInstructionsColumnar(instructions, rows, res_data, &res_null_map); \ + } \ else \ - return ColumnNullable::create(std::move(res), std::move(null_map)); \ + { \ + auto & res_data = assert_cast &>(*res).getData(); \ + executeInstructionsColumnar(instructions, rows, res_data, nullptr); \ + } \ + return std::move(res); \ } #define ENUMERATE_NUMERIC_TYPES(M, INDEX) \ - M(UInt8, INDEX) \ - M(UInt16, INDEX) \ - M(UInt32, INDEX) \ - M(UInt64, INDEX) \ - M(Int8, INDEX) \ - M(Int16, INDEX) \ - M(Int32, INDEX) \ - M(Int64, INDEX) \ - M(UInt128, INDEX) \ - M(UInt256, INDEX) \ - M(Int128, INDEX) \ - M(Int256, INDEX) \ - M(Float32, INDEX) \ - M(Float64, INDEX) \ + M(UInt8, UInt8, INDEX) \ + M(UInt16, UInt16, INDEX) \ + M(UInt32, UInt32, INDEX) \ + M(UInt64, UInt64, INDEX) \ + M(Int8, Int8, INDEX) \ + M(Int16, Int16, INDEX) \ + M(Int32, Int32, INDEX) \ + M(Int64, Int64, INDEX) \ + M(Float32, Float32, INDEX) \ + M(Float64, Float64, INDEX) \ + M(UInt128, UInt128, INDEX) \ + M(UInt256, UInt256, INDEX) \ + M(Int128, Int128, INDEX) \ + M(Int256, Int256, INDEX) \ + M(Decimal32, Decimal32, INDEX) \ + M(Decimal64, Decimal64, INDEX) \ + M(Decimal128, Decimal128, INDEX) \ + M(Decimal256, Decimal256, INDEX) \ + M(Date, UInt16, INDEX) \ + M(Date32, Int32, INDEX) \ + M(DateTime, UInt32, INDEX) \ + M(DateTime64, DateTime64, INDEX) \ + M(Enum8, Int8, INDEX) \ + M(Enum16, Int16, INDEX) \ + M(IPv4, IPv4, INDEX) \ + M(IPv6, IPv6, INDEX) \ throw Exception( \ ErrorCodes::NOT_IMPLEMENTED, "Columnar execution of function {} not implemented for type {}", getName(), result_type->getName()); - size_t num_instructions = instructions.size(); - if (num_instructions <= std::numeric_limits::max()) - { - ENUMERATE_NUMERIC_TYPES(EXECUTE_INSTRUCTIONS_COLUMNAR, Int16) - } - else if (num_instructions <= std::numeric_limits::max()) - { - ENUMERATE_NUMERIC_TYPES(EXECUTE_INSTRUCTIONS_COLUMNAR, Int32) - } - else if (num_instructions <= std::numeric_limits::max()) - { - ENUMERATE_NUMERIC_TYPES(EXECUTE_INSTRUCTIONS_COLUMNAR, Int64) - } - else - throw Exception( - ErrorCodes::LOGICAL_ERROR, "Instruction size({}) of function {} is out of range", getName(), result_type->getName()); + ENUMERATE_NUMERIC_TYPES(EXECUTE_INSTRUCTIONS_COLUMNAR, UInt8) } +#undef ENUMERATE_NUMERIC_TYPES +#undef EXECUTE_INSTRUCTIONS_COLUMNAR private: @@ -348,11 +369,11 @@ private: /// We should read source from which instruction on each row? template - static void calculateInserts(std::vector & instructions, size_t rows, PaddedPODArray & inserts) + static NO_INLINE void calculateInserts(const std::vector & instructions, size_t rows, PaddedPODArray & inserts) { - for (S i = static_cast(instructions.size() - 1); i >= 0; --i) + for (S i = instructions.size() - 1; i != static_cast(-1); --i) { - auto & instruction = instructions[i]; + const auto & instruction = instructions[i]; if (instruction.condition_always_true) { for (size_t row_i = 0; row_i < rows; ++row_i) @@ -388,60 +409,62 @@ private: } } - template - static void executeInstructionsColumnar(std::vector & instructions, size_t rows, const MutableColumnPtr & res, const MutableColumnPtr & null_map, bool nullable) + template + static NO_INLINE void executeInstructionsColumnar( + const std::vector & instructions, + size_t rows, + PaddedPODArray & res_data, + PaddedPODArray * res_null_map = nullptr) { PaddedPODArray inserts(rows, static_cast(instructions.size())); calculateInserts(instructions, rows, inserts); - PaddedPODArray & res_data = assert_cast &>(*res).getData(); - if (!nullable) + res_data.resize_exact(rows); + if constexpr (nullable_result) { - for (size_t row_i = 0; row_i < rows; ++row_i) + if (!res_null_map) + throw Exception(ErrorCodes::BAD_ARGUMENTS, "Invalid result null_map while result type is nullable"); + + res_null_map->resize_exact(rows); + } + + std::vector data_cols(instructions.size(), nullptr); + std::vector null_map_cols(instructions.size(), nullptr); + for (size_t i = 0; i < instructions.size(); ++i) + { + const auto & instruction = instructions[i]; + const IColumn * non_const_col = instructions[i].source_is_constant + ? &assert_cast(*instruction.source).getDataColumn() + : instruction.source.get(); + const ColumnNullable * nullable_col = checkAndGetColumn(non_const_col); + data_cols[i] = nullable_col ? assert_cast &>(nullable_col->getNestedColumn()).getData().data() + : assert_cast &>(*non_const_col).getData().data(); + null_map_cols[i] = nullable_col ? assert_cast(nullable_col->getNullMapColumn()).getData().data() : nullptr; + } + + std::unique_ptr> shared_null_map; + if constexpr (nullable_result) + { + for (auto & col : null_map_cols) { - auto & instruction = instructions[inserts[row_i]]; - auto ref = instruction.source->getDataAt(row_i); - res_data[row_i] = *reinterpret_cast(ref.data); + if (!col) + { + if (!shared_null_map) + shared_null_map = std::make_unique>(rows, 0); + + col = shared_null_map->data(); + } } } - else + + for (size_t row_i = 0; row_i < rows; ++row_i) { - PaddedPODArray & null_map_data = assert_cast(*null_map).getData(); - std::vector data_cols(instructions.size()); - std::vector null_map_cols(instructions.size()); - ColumnPtr shared_null_map_col = nullptr; - for (size_t i = 0; i < instructions.size(); ++i) - { - if (instructions[i].source->isNullable()) - { - const ColumnNullable * nullable_col; - if (!instructions[i].source_is_constant) - nullable_col = assert_cast(instructions[i].source.get()); - else - { - const ColumnPtr data_column = assert_cast(*instructions[i].source).getDataColumnPtr(); - nullable_col = assert_cast(data_column.get()); - } - null_map_cols[i] = assert_cast(*nullable_col->getNullMapColumnPtr()).getData().data(); - data_cols[i] = assert_cast &>(*nullable_col->getNestedColumnPtr()).getData().data(); - } - else - { - if (!shared_null_map_col) - { - shared_null_map_col = ColumnUInt8::create(rows, 0); - } - null_map_cols[i] = assert_cast(*shared_null_map_col).getData().data(); - data_cols[i] = assert_cast &>(*instructions[i].source).getData().data(); - } - } - for (size_t row_i = 0; row_i < rows; ++row_i) - { - auto & instruction = instructions[inserts[row_i]]; - size_t index = instruction.source_is_constant ? 0 : row_i; - res_data[row_i] = *(data_cols[inserts[row_i]] + index); - null_map_data[row_i] = *(null_map_cols[inserts[row_i]] + index); - } + S insert = inserts[row_i]; + const auto & instruction = instructions[insert]; + size_t index = instruction.source_is_constant ? 0 : row_i; + res_data[row_i] = *(data_cols[insert] + index); + if constexpr (nullable_result) + (*res_null_map)[row_i] = *(null_map_cols[insert] + index); } } diff --git a/src/Functions/parseDateTime.cpp b/src/Functions/parseDateTime.cpp index 860603dc503..18882177c90 100644 --- a/src/Functions/parseDateTime.cpp +++ b/src/Functions/parseDateTime.cpp @@ -489,12 +489,12 @@ namespace DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {"time", &isString, nullptr, "String"}, - {"format", &isString, nullptr, "String"} + {"time", static_cast(&isString), nullptr, "String"}, + {"format", static_cast(&isString), nullptr, "String"} }; FunctionArgumentDescriptors optional_args{ - {"timezone", &isString, &isColumnConst, "const String"} + {"timezone", static_cast(&isString), &isColumnConst, "const String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); diff --git a/src/Functions/regexpExtract.cpp b/src/Functions/regexpExtract.cpp index f6bbd2f96f2..cfb42580cb0 100644 --- a/src/Functions/regexpExtract.cpp +++ b/src/Functions/regexpExtract.cpp @@ -47,12 +47,12 @@ public: arguments.size()); FunctionArgumentDescriptors args{ - {"haystack", &isString, nullptr, "String"}, - {"pattern", &isString, isColumnConst, "const String"}, + {"haystack", static_cast(&isString), nullptr, "String"}, + {"pattern", static_cast(&isString), isColumnConst, "const String"}, }; if (arguments.size() == 3) - args.emplace_back(FunctionArgumentDescriptor{"index", &isInteger, nullptr, "Integer"}); + args.emplace_back(FunctionArgumentDescriptor{"index", static_cast(&isInteger), nullptr, "Integer"}); validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/repeat.cpp b/src/Functions/repeat.cpp index c1b553ac6b3..11a2ca37a3b 100644 --- a/src/Functions/repeat.cpp +++ b/src/Functions/repeat.cpp @@ -186,8 +186,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"s", &isString, nullptr, "String"}, - {"n", &isInteger, nullptr, "Integer"}, + {"s", static_cast(&isString), nullptr, "String"}, + {"n", static_cast(&isInteger), nullptr, "Integer"}, }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/seriesDecomposeSTL.cpp b/src/Functions/seriesDecomposeSTL.cpp index fbabc801913..618808b64ed 100644 --- a/src/Functions/seriesDecomposeSTL.cpp +++ b/src/Functions/seriesDecomposeSTL.cpp @@ -42,8 +42,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"time_series", &isArray, nullptr, "Array"}, - {"period", &isNativeUInt, nullptr, "Unsigned Integer"}, + {"time_series", static_cast(&isArray), nullptr, "Array"}, + {"period", static_cast(&isNativeUInt), nullptr, "Unsigned Integer"}, }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/seriesOutliersDetectTukey.cpp b/src/Functions/seriesOutliersDetectTukey.cpp index 5bc8edf3a54..da04d3b78d3 100644 --- a/src/Functions/seriesOutliersDetectTukey.cpp +++ b/src/Functions/seriesOutliersDetectTukey.cpp @@ -45,11 +45,11 @@ public: getName(), arguments.size()); - FunctionArgumentDescriptors mandatory_args{{"time_series", &isArray, nullptr, "Array"}}; + FunctionArgumentDescriptors mandatory_args{{"time_series", static_cast(&isArray), nullptr, "Array"}}; FunctionArgumentDescriptors optional_args{ - {"min_percentile", &isFloat, isColumnConst, "Number"}, - {"max_percentile", &isFloat, isColumnConst, "Number"}, - {"k", &isNativeNumber, isColumnConst, "Number"}}; + {"min_percentile", static_cast(&isFloat), isColumnConst, "Number"}, + {"max_percentile", static_cast(&isFloat), isColumnConst, "Number"}, + {"k", static_cast(&isNativeNumber), isColumnConst, "Number"}}; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); diff --git a/src/Functions/seriesPeriodDetectFFT.cpp b/src/Functions/seriesPeriodDetectFFT.cpp index c01f6b7f07b..fbaa2b14e64 100644 --- a/src/Functions/seriesPeriodDetectFFT.cpp +++ b/src/Functions/seriesPeriodDetectFFT.cpp @@ -52,7 +52,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { - FunctionArgumentDescriptors args{{"time_series", &isArray, nullptr, "Array"}}; + FunctionArgumentDescriptors args{{"time_series", static_cast(&isArray), nullptr, "Array"}}; validateFunctionArgumentTypes(*this, arguments, args); return std::make_shared(); diff --git a/src/Functions/snowflake.cpp b/src/Functions/snowflake.cpp index 6aafa2cb5cf..f2dd1f1c51d 100644 --- a/src/Functions/snowflake.cpp +++ b/src/Functions/snowflake.cpp @@ -47,7 +47,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"value", &isDateTime, nullptr, "DateTime"} + {"value", static_cast(&isDateTime), nullptr, "DateTime"} }; validateFunctionArgumentTypes(*this, arguments, args); @@ -91,10 +91,10 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {"value", &isInt64, nullptr, "Int64"} + {"value", static_cast(&isInt64), nullptr, "Int64"} }; FunctionArgumentDescriptors optional_args{ - {"time_zone", &isString, nullptr, "String"} + {"time_zone", static_cast(&isString), nullptr, "String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); @@ -151,7 +151,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"value", &isDateTime64, nullptr, "DateTime64"} + {"value", static_cast(&isDateTime64), nullptr, "DateTime64"} }; validateFunctionArgumentTypes(*this, arguments, args); @@ -203,10 +203,10 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {"value", &isInt64, nullptr, "Int64"} + {"value", static_cast(&isInt64), nullptr, "Int64"} }; FunctionArgumentDescriptors optional_args{ - {"time_zone", &isString, nullptr, "String"} + {"time_zone", static_cast(&isString), nullptr, "String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); diff --git a/src/Functions/space.cpp b/src/Functions/space.cpp index 009bc20e065..03dc0d06719 100644 --- a/src/Functions/space.cpp +++ b/src/Functions/space.cpp @@ -45,7 +45,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"n", &isInteger, nullptr, "Integer"} + {"n", static_cast(&isInteger), nullptr, "Integer"} }; validateFunctionArgumentTypes(*this, arguments, args); diff --git a/src/Functions/splitByChar.cpp b/src/Functions/splitByChar.cpp index d537039dc23..d3d5dc9fe4a 100644 --- a/src/Functions/splitByChar.cpp +++ b/src/Functions/splitByChar.cpp @@ -40,6 +40,8 @@ public: static bool isVariadic() { return true; } static size_t getNumberOfArguments() { return 0; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {0, 2}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { checkArgumentsWithSeparatorAndOptionalMaxSubstrings(func, arguments); diff --git a/src/Functions/splitByNonAlpha.cpp b/src/Functions/splitByNonAlpha.cpp index 467e7b0b5c3..4486a33aa88 100644 --- a/src/Functions/splitByNonAlpha.cpp +++ b/src/Functions/splitByNonAlpha.cpp @@ -42,6 +42,8 @@ public: static bool isVariadic() { return true; } static size_t getNumberOfArguments() { return 0; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {1}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { checkArgumentsWithOptionalMaxSubstrings(func, arguments); diff --git a/src/Functions/splitByRegexp.cpp b/src/Functions/splitByRegexp.cpp index 77328205c01..430089f14ee 100644 --- a/src/Functions/splitByRegexp.cpp +++ b/src/Functions/splitByRegexp.cpp @@ -44,6 +44,8 @@ public: static bool isVariadic() { return true; } static size_t getNumberOfArguments() { return 0; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {0, 2}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { checkArgumentsWithSeparatorAndOptionalMaxSubstrings(func, arguments); diff --git a/src/Functions/splitByString.cpp b/src/Functions/splitByString.cpp index 7d6803b2f27..5c97f9841e7 100644 --- a/src/Functions/splitByString.cpp +++ b/src/Functions/splitByString.cpp @@ -39,6 +39,8 @@ public: static bool isVariadic() { return true; } static size_t getNumberOfArguments() { return 0; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {0, 2}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { checkArgumentsWithSeparatorAndOptionalMaxSubstrings(func, arguments); diff --git a/src/Functions/splitByWhitespace.cpp b/src/Functions/splitByWhitespace.cpp index 168e429c6f5..cf21a218b15 100644 --- a/src/Functions/splitByWhitespace.cpp +++ b/src/Functions/splitByWhitespace.cpp @@ -30,6 +30,8 @@ public: static bool isVariadic() { return true; } static size_t getNumberOfArguments() { return 0; } + static ColumnNumbers getArgumentsThatAreAlwaysConstant() { return {1}; } + static void checkArguments(const IFunction & func, const ColumnsWithTypeAndName & arguments) { checkArgumentsWithOptionalMaxSubstrings(func, arguments); diff --git a/src/Functions/sqid.cpp b/src/Functions/sqid.cpp index cd3875e2607..6679646fef4 100644 --- a/src/Functions/sqid.cpp +++ b/src/Functions/sqid.cpp @@ -98,7 +98,7 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors args{ - {"sqid", &isString, nullptr, "String"} + {"sqid", static_cast(&isString), nullptr, "String"} }; validateFunctionArgumentTypes(*this, arguments, args); @@ -122,7 +122,7 @@ public: for (size_t i = 0; i < input_rows_count; ++i) { std::string_view sqid = col_non_const->getDataAt(i).toView(); - std::vector integers = sqids.decode(sqid); + std::vector integers = sqids.decode(String(sqid)); res_nested_data.insert(integers.begin(), integers.end()); res_offsets_data.push_back(integers.size()); } diff --git a/src/Functions/timestamp.cpp b/src/Functions/timestamp.cpp index 48012c1376f..fbca08b0968 100644 --- a/src/Functions/timestamp.cpp +++ b/src/Functions/timestamp.cpp @@ -41,10 +41,10 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args{ - {"timestamp", &isStringOrFixedString, nullptr, "String or FixedString"} + {"timestamp", static_cast(&isStringOrFixedString), nullptr, "String or FixedString"} }; FunctionArgumentDescriptors optional_args{ - {"time", &isString, nullptr, "String"} + {"time", static_cast(&isString), nullptr, "String"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, optional_args); diff --git a/src/Functions/toDecimalString.cpp b/src/Functions/toDecimalString.cpp index cc2de8df0d4..fc621b272de 100644 --- a/src/Functions/toDecimalString.cpp +++ b/src/Functions/toDecimalString.cpp @@ -39,8 +39,8 @@ public: DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override { FunctionArgumentDescriptors mandatory_args = { - {"Value", &isNumber, nullptr, "Number"}, - {"precision", &isNativeInteger, &isColumnConst, "const Integer"} + {"Value", static_cast(&isNumber), nullptr, "Number"}, + {"precision", static_cast(&isNativeInteger), &isColumnConst, "const Integer"} }; validateFunctionArgumentTypes(*this, arguments, mandatory_args, {}); diff --git a/src/Functions/toMillisecond.cpp b/src/Functions/toMillisecond.cpp new file mode 100644 index 00000000000..aaef517c996 --- /dev/null +++ b/src/Functions/toMillisecond.cpp @@ -0,0 +1,33 @@ +#include +#include +#include +#include + +namespace DB +{ + +using FunctionToMillisecond = FunctionDateOrDateTimeToSomething; + +REGISTER_FUNCTION(ToMillisecond) +{ + factory.registerFunction( + + + FunctionDocumentation{ + .description=R"( +Returns the millisecond component (0-999) of a date with time. + )", + .syntax="toMillisecond(value)", + .arguments={{"value", "DateTime or DateTime64"}}, + .returned_value="The millisecond in the minute (0 - 59) of the given date/time", + .examples{ + {"toMillisecond", "SELECT toMillisecond(toDateTime64('2023-04-21 10:20:30.456', 3)", "456"}}, + .categories{"Dates and Times"} + } + ); + + /// MySQL compatibility alias. + factory.registerAlias("MILLISECOND", "toMillisecond", FunctionFactory::CaseInsensitive); +} + +} diff --git a/src/IO/CachedInMemoryReadBufferFromFile.cpp b/src/IO/CachedInMemoryReadBufferFromFile.cpp index 384d2229f14..ceaf0ca4752 100644 --- a/src/IO/CachedInMemoryReadBufferFromFile.cpp +++ b/src/IO/CachedInMemoryReadBufferFromFile.cpp @@ -136,7 +136,6 @@ bool CachedInMemoryReadBufferFromFile::nextImpl() char * piece_start = chunk->getChunk()->data + pos; size_t piece_size = chunk_size - pos; in->set(piece_start, piece_size); - LOG_INFO(&Poco::Logger::get("asdqwe"), "this {:x}, in {:x}, path {}, size {}, offset {:x}, pos {:x}", reinterpret_cast(this), reinterpret_cast(in.get()), cache_key.path, file_size.value(), cache_key.offset, pos); if (pos == 0) in->seek(cache_key.offset, SEEK_SET); else @@ -155,7 +154,6 @@ bool CachedInMemoryReadBufferFromFile::nextImpl() memmove(piece_start, in->position(), n); in->position() += n; pos += n; - LOG_INFO(&Poco::Logger::get("asdqwe"), "this {:x}, got {:x} bytes", reinterpret_cast(this), n); } chunk->markPrefixPopulated(chunk_size); diff --git a/src/IO/ConnectionTimeouts.cpp b/src/IO/ConnectionTimeouts.cpp index f2db3169400..c4b636103fe 100644 --- a/src/IO/ConnectionTimeouts.cpp +++ b/src/IO/ConnectionTimeouts.cpp @@ -141,4 +141,19 @@ ConnectionTimeouts ConnectionTimeouts::getAdaptiveTimeouts(const String & method .withReceiveTimeout(saturate(recv, receive_timeout)); } +void setTimeouts(Poco::Net::HTTPClientSession & session, const ConnectionTimeouts & timeouts) +{ + session.setTimeout(timeouts.connection_timeout, timeouts.send_timeout, timeouts.receive_timeout); + session.setKeepAliveTimeout(timeouts.http_keep_alive_timeout); +} + +ConnectionTimeouts getTimeouts(const Poco::Net::HTTPClientSession & session) +{ + return ConnectionTimeouts() + .withConnectionTimeout(session.getConnectionTimeout()) + .withSendTimeout(session.getSendTimeout()) + .withReceiveTimeout(session.getReceiveTimeout()) + .withHTTPKeepAliveTimeout(session.getKeepAliveTimeout()); +} + } diff --git a/src/IO/ConnectionTimeouts.h b/src/IO/ConnectionTimeouts.h index 7fe97b5ec36..2b2ab0e7ab8 100644 --- a/src/IO/ConnectionTimeouts.h +++ b/src/IO/ConnectionTimeouts.h @@ -4,6 +4,7 @@ #include #include +#include #include namespace DB @@ -111,4 +112,7 @@ inline ConnectionTimeouts & ConnectionTimeouts::withConnectionTimeout(Poco::Time return *this; } +void setTimeouts(Poco::Net::HTTPClientSession & session, const ConnectionTimeouts & timeouts); +ConnectionTimeouts getTimeouts(const Poco::Net::HTTPClientSession & session); + } diff --git a/src/IO/HTTPCommon.cpp b/src/IO/HTTPCommon.cpp index c4468a1b896..09f7724d613 100644 --- a/src/IO/HTTPCommon.cpp +++ b/src/IO/HTTPCommon.cpp @@ -2,13 +2,7 @@ #include #include -#include -#include #include -#include -#include -#include -#include #include "config.h" @@ -25,338 +19,18 @@ #include -#include -#include +#include #include #include -namespace ProfileEvents -{ - extern const Event CreatedHTTPConnections; -} - namespace DB { + namespace ErrorCodes { extern const int RECEIVED_ERROR_FROM_REMOTE_IO_SERVER; extern const int RECEIVED_ERROR_TOO_MANY_REQUESTS; - extern const int FEATURE_IS_NOT_ENABLED_AT_BUILD_TIME; - extern const int UNSUPPORTED_URI_SCHEME; - extern const int LOGICAL_ERROR; -} - - -namespace -{ - Poco::Net::HTTPClientSession::ProxyConfig proxyConfigurationToPocoProxyConfig(const ProxyConfiguration & proxy_configuration) - { - Poco::Net::HTTPClientSession::ProxyConfig poco_proxy_config; - - poco_proxy_config.host = proxy_configuration.host; - poco_proxy_config.port = proxy_configuration.port; - poco_proxy_config.protocol = ProxyConfiguration::protocolToString(proxy_configuration.protocol); - poco_proxy_config.tunnel = proxy_configuration.tunneling; - poco_proxy_config.originalRequestProtocol = ProxyConfiguration::protocolToString(proxy_configuration.original_request_protocol); - - return poco_proxy_config; - } - - template - requires std::derived_from - class HTTPSessionAdapter : public Session - { - static_assert(std::has_virtual_destructor_v, "The base class must have a virtual destructor"); - - public: - HTTPSessionAdapter(const std::string & host, UInt16 port) : Session(host, port), log{getLogger("HTTPSessionAdapter")} { } - ~HTTPSessionAdapter() override = default; - - protected: - void reconnect() override - { - // First of all will try to establish connection with last used addr. - if (!Session::getResolvedHost().empty()) - { - try - { - Session::reconnect(); - return; - } - catch (...) - { - Session::close(); - LOG_TRACE( - log, - "Last ip ({}) is unreachable for {}:{}. Will try another resolved address.", - Session::getResolvedHost(), - Session::getHost(), - Session::getPort()); - } - } - - const auto endpoinds = DNSResolver::instance().resolveHostAll(Session::getHost()); - - for (auto it = endpoinds.begin();;) - { - try - { - Session::setResolvedHost(it->toString()); - Session::reconnect(); - - LOG_TRACE( - log, - "Created HTTP(S) session with {}:{} ({}:{})", - Session::getHost(), - Session::getPort(), - it->toString(), - Session::getPort()); - - break; - } - catch (...) - { - Session::close(); - if (++it == endpoinds.end()) - { - Session::setResolvedHost(""); - throw; - } - LOG_TRACE( - log, - "Failed to create connection with {}:{}, Will try another resolved address. {}", - Session::getResolvedHost(), - Session::getPort(), - getCurrentExceptionMessage(false)); - } - } - } - LoggerPtr log; - }; - - bool isHTTPS(const Poco::URI & uri) - { - if (uri.getScheme() == "https") - return true; - else if (uri.getScheme() == "http") - return false; - else - throw Exception(ErrorCodes::UNSUPPORTED_URI_SCHEME, "Unsupported scheme in URI '{}'", uri.toString()); - } - - HTTPSessionPtr makeHTTPSessionImpl( - const std::string & host, - UInt16 port, - bool https, - bool keep_alive, - DB::ProxyConfiguration proxy_configuration = {}) - { - HTTPSessionPtr session; - - if (!proxy_configuration.host.empty()) - { - bool is_proxy_http_and_is_tunneling_off = DB::ProxyConfiguration::Protocol::HTTP == proxy_configuration.protocol - && !proxy_configuration.tunneling; - - // If it is an HTTPS request, proxy server is HTTP and user opted for tunneling off, we must not create an HTTPS request. - // The desired flow is: HTTP request to the proxy server, then proxy server will initiate an HTTPS request to the target server. - // There is a weak link in the security, but that's what the user opted for. - if (https && is_proxy_http_and_is_tunneling_off) - { - https = false; - } - } - - if (https) - { -#if USE_SSL - session = std::make_shared>(host, port); -#else - throw Exception(ErrorCodes::FEATURE_IS_NOT_ENABLED_AT_BUILD_TIME, "ClickHouse was built without HTTPS support"); -#endif - } - else - { - session = std::make_shared>(host, port); - } - - ProfileEvents::increment(ProfileEvents::CreatedHTTPConnections); - - /// doesn't work properly without patch - session->setKeepAlive(keep_alive); - - if (!proxy_configuration.host.empty()) - { - session->setProxyConfig(proxyConfigurationToPocoProxyConfig(proxy_configuration)); - } - - return session; - } - - class SingleEndpointHTTPSessionPool : public PoolBase - { - private: - const std::string host; - const UInt16 port; - const bool https; - ProxyConfiguration proxy_config; - - using Base = PoolBase; - - ObjectPtr allocObject() override - { - /// Pool is global, we shouldn't attribute this memory to query/user. - MemoryTrackerSwitcher switcher{&total_memory_tracker}; - - auto session = makeHTTPSessionImpl(host, port, https, true, proxy_config); - return session; - } - - public: - SingleEndpointHTTPSessionPool( - const std::string & host_, - UInt16 port_, - bool https_, - ProxyConfiguration proxy_config_, - size_t max_pool_size_, - bool wait_on_pool_size_limit) - : Base( - static_cast(max_pool_size_), - getLogger("HTTPSessionPool"), - wait_on_pool_size_limit ? BehaviourOnLimit::Wait : BehaviourOnLimit::AllocateNewBypassingPool) - , host(host_) - , port(port_) - , https(https_) - , proxy_config(proxy_config_) - { - } - }; - - class HTTPSessionPool : private boost::noncopyable - { - public: - struct Key - { - String target_host; - UInt16 target_port; - bool is_target_https; - ProxyConfiguration proxy_config; - bool wait_on_pool_size_limit; - - bool operator ==(const Key & rhs) const - { - return std::tie( - target_host, - target_port, - is_target_https, - proxy_config.host, - proxy_config.port, - proxy_config.protocol, - proxy_config.tunneling, - proxy_config.original_request_protocol, - wait_on_pool_size_limit) - == std::tie( - rhs.target_host, - rhs.target_port, - rhs.is_target_https, - rhs.proxy_config.host, - rhs.proxy_config.port, - rhs.proxy_config.protocol, - rhs.proxy_config.tunneling, - rhs.proxy_config.original_request_protocol, - rhs.wait_on_pool_size_limit); - } - }; - - private: - using PoolPtr = std::shared_ptr; - using Entry = SingleEndpointHTTPSessionPool::Entry; - - struct Hasher - { - size_t operator()(const Key & k) const - { - SipHash s; - s.update(k.target_host); - s.update(k.target_port); - s.update(k.is_target_https); - s.update(k.proxy_config.host); - s.update(k.proxy_config.port); - s.update(k.proxy_config.protocol); - s.update(k.proxy_config.tunneling); - s.update(k.proxy_config.original_request_protocol); - s.update(k.wait_on_pool_size_limit); - return s.get64(); - } - }; - - std::mutex mutex; - std::unordered_map endpoints_pool; - - protected: - HTTPSessionPool() = default; - - public: - static auto & instance() - { - static HTTPSessionPool instance; - return instance; - } - - Entry getSession( - const Poco::URI & uri, - const ProxyConfiguration & proxy_config, - const ConnectionTimeouts & timeouts, - size_t max_connections_per_endpoint, - bool wait_on_pool_size_limit) - { - std::unique_lock lock(mutex); - const std::string & host = uri.getHost(); - UInt16 port = uri.getPort(); - bool https = isHTTPS(uri); - - HTTPSessionPool::Key key{host, port, https, proxy_config, wait_on_pool_size_limit}; - auto pool_ptr = endpoints_pool.find(key); - if (pool_ptr == endpoints_pool.end()) - std::tie(pool_ptr, std::ignore) = endpoints_pool.emplace( - key, - std::make_shared( - host, - port, - https, - proxy_config, - max_connections_per_endpoint, - wait_on_pool_size_limit)); - - /// Some routines held session objects until the end of its lifetime. Also this routines may create another sessions in this time frame. - /// If some other session holds `lock` because it waits on another lock inside `pool_ptr->second->get` it isn't possible to create any - /// new session and thus finish routine, return session to the pool and unlock the thread waiting inside `pool_ptr->second->get`. - /// To avoid such a deadlock we unlock `lock` before entering `pool_ptr->second->get`. - lock.unlock(); - - auto retry_timeout = timeouts.connection_timeout.totalMilliseconds(); - auto session = pool_ptr->second->get(retry_timeout); - - const auto & session_data = session->sessionData(); - if (session_data.empty() || !Poco::AnyCast(&session_data)) - { - /// Reset session if it is not reusable. See comment for HTTPSessionReuseTag. - session->reset(); - } - session->attachSessionData({}); - - setTimeouts(*session, timeouts); - - return session; - } - }; -} - -void setTimeouts(Poco::Net::HTTPClientSession & session, const ConnectionTimeouts & timeouts) -{ - session.setTimeout(timeouts.connection_timeout, timeouts.send_timeout, timeouts.receive_timeout); - session.setKeepAliveTimeout(timeouts.http_keep_alive_timeout); } void setResponseDefaultHeaders(HTTPServerResponse & response, size_t keep_alive_timeout) @@ -370,28 +44,13 @@ void setResponseDefaultHeaders(HTTPServerResponse & response, size_t keep_alive_ } HTTPSessionPtr makeHTTPSession( + HTTPConnectionGroupType group, const Poco::URI & uri, const ConnectionTimeouts & timeouts, - ProxyConfiguration proxy_configuration -) + ProxyConfiguration proxy_configuration) { - const std::string & host = uri.getHost(); - UInt16 port = uri.getPort(); - bool https = isHTTPS(uri); - - auto session = makeHTTPSessionImpl(host, port, https, false, proxy_configuration); - setTimeouts(*session, timeouts); - return session; -} - -PooledHTTPSessionPtr makePooledHTTPSession( - const Poco::URI & uri, - const ConnectionTimeouts & timeouts, - size_t per_endpoint_pool_size, - bool wait_on_pool_size_limit, - ProxyConfiguration proxy_config) -{ - return HTTPSessionPool::instance().getSession(uri, proxy_config, timeouts, per_endpoint_pool_size, wait_on_pool_size_limit); + auto connection_pool = HTTPConnectionPools::instance().getPool(group, uri, proxy_configuration); + return connection_pool->getConnection(timeouts); } bool isRedirect(const Poco::Net::HTTPResponse::HTTPStatus status) { return status == Poco::Net::HTTPResponse::HTTP_MOVED_PERMANENTLY || status == Poco::Net::HTTPResponse::HTTP_FOUND || status == Poco::Net::HTTPResponse::HTTP_SEE_OTHER || status == Poco::Net::HTTPResponse::HTTP_TEMPORARY_REDIRECT; } @@ -400,11 +59,11 @@ std::istream * receiveResponse( Poco::Net::HTTPClientSession & session, const Poco::Net::HTTPRequest & request, Poco::Net::HTTPResponse & response, const bool allow_redirects) { auto & istr = session.receiveResponse(response); - assertResponseIsOk(request, response, istr, allow_redirects); + assertResponseIsOk(request.getURI(), response, istr, allow_redirects); return &istr; } -void assertResponseIsOk(const Poco::Net::HTTPRequest & request, Poco::Net::HTTPResponse & response, std::istream & istr, const bool allow_redirects) +void assertResponseIsOk(const String & uri, Poco::Net::HTTPResponse & response, std::istream & istr, const bool allow_redirects) { auto status = response.getStatus(); @@ -422,7 +81,7 @@ void assertResponseIsOk(const Poco::Net::HTTPRequest & request, Poco::Net::HTTPR body.exceptions(std::ios::failbit); body << istr.rdbuf(); - throw HTTPException(code, request.getURI(), status, response.getReason(), body.str()); + throw HTTPException(code, uri, status, response.getReason(), body.str()); } } @@ -440,24 +99,4 @@ Exception HTTPException::makeExceptionMessage( uri, static_cast(http_status), reason, body); } -void markSessionForReuse(Poco::Net::HTTPSession & session) -{ - const auto & session_data = session.sessionData(); - if (!session_data.empty() && !Poco::AnyCast(&session_data)) - throw Exception( - ErrorCodes::LOGICAL_ERROR, "Data of an unexpected type ({}) is attached to the session", session_data.type().name()); - - session.attachSessionData(HTTPSessionReuseTag{}); -} - -void markSessionForReuse(HTTPSessionPtr session) -{ - markSessionForReuse(*session); -} - -void markSessionForReuse(PooledHTTPSessionPtr session) -{ - markSessionForReuse(static_cast(*session)); -} - } diff --git a/src/IO/HTTPCommon.h b/src/IO/HTTPCommon.h index c9968fc6915..e27269e2559 100644 --- a/src/IO/HTTPCommon.h +++ b/src/IO/HTTPCommon.h @@ -7,9 +7,9 @@ #include #include #include -#include -#include #include +#include +#include #include @@ -36,7 +36,7 @@ public: HTTPException * clone() const override { return new HTTPException(*this); } void rethrow() const override { throw *this; } - int getHTTPStatus() const { return http_status; } + Poco::Net::HTTPResponse::HTTPStatus getHTTPStatus() const { return http_status; } private: Poco::Net::HTTPResponse::HTTPStatus http_status{}; @@ -52,55 +52,18 @@ private: const char * className() const noexcept override { return "DB::HTTPException"; } }; -using PooledHTTPSessionPtr = PoolBase::Entry; // SingleEndpointHTTPSessionPool::Entry using HTTPSessionPtr = std::shared_ptr; -/// If a session have this tag attached, it will be reused without calling `reset()` on it. -/// All pooled sessions don't have this tag attached after being taken from a pool. -/// If the request and the response were fully written/read, the client code should add this tag -/// explicitly by calling `markSessionForReuse()`. -/// -/// Note that HTTP response may contain extra bytes after the last byte of the payload. Specifically, -/// when chunked encoding is used, there's an empty chunk at the end. Those extra bytes must also be -/// read before the session can be reused. So we usually put an `istr->ignore(INT64_MAX)` call -/// before `markSessionForReuse()`. -struct HTTPSessionReuseTag -{ -}; - -void markSessionForReuse(Poco::Net::HTTPSession & session); -void markSessionForReuse(HTTPSessionPtr session); -void markSessionForReuse(PooledHTTPSessionPtr session); - - void setResponseDefaultHeaders(HTTPServerResponse & response, size_t keep_alive_timeout); /// Create session object to perform requests and set required parameters. HTTPSessionPtr makeHTTPSession( + HTTPConnectionGroupType group, const Poco::URI & uri, const ConnectionTimeouts & timeouts, ProxyConfiguration proxy_config = {} ); -/// As previous method creates session, but takes it from pool, without and with proxy uri. -/// -/// The max_connections_per_endpoint parameter makes it look like the pool size can be different for -/// different requests (whatever that means), but actually we just assign the endpoint's connection -/// pool size when we see the endpoint for the first time, then we never change it. -/// We should probably change how this configuration works, and how this pooling works in general: -/// * Make the per_endpoint_pool_size be a global server setting instead of per-disk or per-query. -/// * Have boolean per-disk/per-query settings for enabling/disabling pooling. -/// * Add a limit on the number of endpoints and the total number of sessions across all endpoints. -/// * Enable pooling by default everywhere. In particular StorageURL and StorageS3. -/// (Enabling it for StorageURL is scary without the previous item - the user may query lots of -/// different endpoints. So currently pooling is mainly used for S3.) -PooledHTTPSessionPtr makePooledHTTPSession( - const Poco::URI & uri, - const ConnectionTimeouts & timeouts, - size_t per_endpoint_pool_size, - bool wait_on_pool_size_limit = true, - ProxyConfiguration proxy_config = {}); - bool isRedirect(Poco::Net::HTTPResponse::HTTPStatus status); /** Used to receive response (response headers and possibly body) @@ -112,7 +75,6 @@ std::istream * receiveResponse( Poco::Net::HTTPClientSession & session, const Poco::Net::HTTPRequest & request, Poco::Net::HTTPResponse & response, bool allow_redirects); void assertResponseIsOk( - const Poco::Net::HTTPRequest & request, Poco::Net::HTTPResponse & response, std::istream & istr, bool allow_redirects = false); + const String & uri, Poco::Net::HTTPResponse & response, std::istream & istr, bool allow_redirects = false); -void setTimeouts(Poco::Net::HTTPClientSession & session, const ConnectionTimeouts & timeouts); } diff --git a/src/IO/LimitReadBuffer.cpp b/src/IO/LimitReadBuffer.cpp index e14112f8d19..84c7ac86227 100644 --- a/src/IO/LimitReadBuffer.cpp +++ b/src/IO/LimitReadBuffer.cpp @@ -1,5 +1,4 @@ #include - #include @@ -15,7 +14,7 @@ namespace ErrorCodes bool LimitReadBuffer::nextImpl() { - assert(position() >= in->position()); + chassert(position() >= in->position()); /// Let underlying buffer calculate read bytes in `next()` call. in->position() = position(); @@ -39,20 +38,18 @@ bool LimitReadBuffer::nextImpl() if (exact_limit && bytes != *exact_limit) throw Exception(ErrorCodes::CANNOT_READ_ALL_DATA, "Unexpected EOF, got {} of {} bytes", bytes, *exact_limit); /// Clearing the buffer with existing data. - set(in->position(), 0); + BufferBase::set(in->position(), 0, 0); + return false; } - working_buffer = in->buffer(); - - if (limit - bytes < working_buffer.size()) - working_buffer.resize(limit - bytes); + BufferBase::set(in->position(), std::min(in->available(), limit - bytes), 0); return true; } -LimitReadBuffer::LimitReadBuffer(ReadBuffer * in_, bool owns, UInt64 limit_, bool throw_exception_, +LimitReadBuffer::LimitReadBuffer(ReadBuffer * in_, bool owns, size_t limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_) : ReadBuffer(in_ ? in_->position() : nullptr, 0) , in(in_) @@ -62,24 +59,20 @@ LimitReadBuffer::LimitReadBuffer(ReadBuffer * in_, bool owns, UInt64 limit_, boo , exact_limit(exact_limit_) , exception_message(std::move(exception_message_)) { - assert(in); + chassert(in); - size_t remaining_bytes_in_buffer = in->buffer().end() - in->position(); - if (remaining_bytes_in_buffer > limit) - remaining_bytes_in_buffer = limit; - - working_buffer = Buffer(in->position(), in->position() + remaining_bytes_in_buffer); + BufferBase::set(in->position(), std::min(in->available(), limit), 0); } -LimitReadBuffer::LimitReadBuffer(ReadBuffer & in_, UInt64 limit_, bool throw_exception_, +LimitReadBuffer::LimitReadBuffer(ReadBuffer & in_, size_t limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_) : LimitReadBuffer(&in_, false, limit_, throw_exception_, exact_limit_, exception_message_) { } -LimitReadBuffer::LimitReadBuffer(std::unique_ptr in_, UInt64 limit_, bool throw_exception_, +LimitReadBuffer::LimitReadBuffer(std::unique_ptr in_, size_t limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_) : LimitReadBuffer(in_.release(), true, limit_, throw_exception_, exact_limit_, exception_message_) { diff --git a/src/IO/LimitReadBuffer.h b/src/IO/LimitReadBuffer.h index 15885c1d850..b869f2935fb 100644 --- a/src/IO/LimitReadBuffer.h +++ b/src/IO/LimitReadBuffer.h @@ -13,22 +13,24 @@ namespace DB class LimitReadBuffer : public ReadBuffer { public: - LimitReadBuffer(ReadBuffer & in_, UInt64 limit_, bool throw_exception_, + LimitReadBuffer(ReadBuffer & in_, size_t limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_ = {}); - LimitReadBuffer(std::unique_ptr in_, UInt64 limit_, bool throw_exception_, std::optional exact_limit_, + LimitReadBuffer(std::unique_ptr in_, size_t limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_ = {}); ~LimitReadBuffer() override; private: ReadBuffer * in; - bool owns_in; + const bool owns_in; - UInt64 limit; - bool throw_exception; - std::optional exact_limit; - std::string exception_message; + const size_t limit; + const bool throw_exception; + const std::optional exact_limit; + const std::string exception_message; - LimitReadBuffer(ReadBuffer * in_, bool owns, UInt64 limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_); + LoggerPtr log; + + LimitReadBuffer(ReadBuffer * in_, bool owns, size_t limit_, bool throw_exception_, std::optional exact_limit_, std::string exception_message_); bool nextImpl() override; }; diff --git a/src/IO/MMapReadBufferFromFileDescriptor.cpp b/src/IO/MMapReadBufferFromFileDescriptor.cpp index 9b1c132cc01..f27828f71b2 100644 --- a/src/IO/MMapReadBufferFromFileDescriptor.cpp +++ b/src/IO/MMapReadBufferFromFileDescriptor.cpp @@ -92,7 +92,7 @@ size_t MMapReadBufferFromFileDescriptor::getFileSize() return getSizeFromFileDescriptor(getFD(), getFileName()); } -size_t MMapReadBufferFromFileDescriptor::readBigAt(char * to, size_t n, size_t offset, const std::function &) +size_t MMapReadBufferFromFileDescriptor::readBigAt(char * to, size_t n, size_t offset, const std::function &) const { if (offset >= mapped.getLength()) return 0; diff --git a/src/IO/MMapReadBufferFromFileDescriptor.h b/src/IO/MMapReadBufferFromFileDescriptor.h index 2a039e04971..f774538374a 100644 --- a/src/IO/MMapReadBufferFromFileDescriptor.h +++ b/src/IO/MMapReadBufferFromFileDescriptor.h @@ -40,7 +40,7 @@ public: size_t getFileSize() override; - size_t readBigAt(char * to, size_t n, size_t offset, const std::function &) override; + size_t readBigAt(char * to, size_t n, size_t offset, const std::function &) const override; bool supportsReadAt() override { return true; } }; diff --git a/src/IO/ReadBuffer.h b/src/IO/ReadBuffer.h index 00325734354..056e25a5fbe 100644 --- a/src/IO/ReadBuffer.h +++ b/src/IO/ReadBuffer.h @@ -63,21 +63,23 @@ public: */ bool next() { - assert(!hasPendingData()); - assert(position() <= working_buffer.end()); + chassert(!hasPendingData()); + chassert(position() <= working_buffer.end()); bytes += offset(); bool res = nextImpl(); if (!res) + { working_buffer = Buffer(pos, pos); + } else { - pos = working_buffer.begin() + nextimpl_working_buffer_offset; - assert(position() != working_buffer.end()); + pos = working_buffer.begin() + std::min(nextimpl_working_buffer_offset, working_buffer.size()); + chassert(position() < working_buffer.end()); } nextimpl_working_buffer_offset = 0; - assert(position() <= working_buffer.end()); + chassert(position() <= working_buffer.end()); return res; } diff --git a/src/IO/ReadBufferFromFileDescriptor.cpp b/src/IO/ReadBufferFromFileDescriptor.cpp index 3211f8eeb35..57442a15853 100644 --- a/src/IO/ReadBufferFromFileDescriptor.cpp +++ b/src/IO/ReadBufferFromFileDescriptor.cpp @@ -49,7 +49,7 @@ std::string ReadBufferFromFileDescriptor::getFileName() const } -size_t ReadBufferFromFileDescriptor::readImpl(char * to, size_t min_bytes, size_t max_bytes, size_t offset) +size_t ReadBufferFromFileDescriptor::readImpl(char * to, size_t min_bytes, size_t max_bytes, size_t offset) const { chassert(min_bytes <= max_bytes); @@ -265,7 +265,7 @@ bool ReadBufferFromFileDescriptor::checkIfActuallySeekable() return res == 0 && S_ISREG(stat.st_mode); } -size_t ReadBufferFromFileDescriptor::readBigAt(char * to, size_t n, size_t offset, const std::function &) +size_t ReadBufferFromFileDescriptor::readBigAt(char * to, size_t n, size_t offset, const std::function &) const { chassert(use_pread); return readImpl(to, n, n, offset); diff --git a/src/IO/ReadBufferFromFileDescriptor.h b/src/IO/ReadBufferFromFileDescriptor.h index 4762998c67b..db256ef91c7 100644 --- a/src/IO/ReadBufferFromFileDescriptor.h +++ b/src/IO/ReadBufferFromFileDescriptor.h @@ -34,7 +34,7 @@ protected: /// Doesn't seek (`offset` must match fd's position if !use_pread). /// Stops after min_bytes or eof. Returns 0 if eof. /// Thread safe. - size_t readImpl(char * to, size_t min_bytes, size_t max_bytes, size_t offset); + size_t readImpl(char * to, size_t min_bytes, size_t max_bytes, size_t offset) const; public: explicit ReadBufferFromFileDescriptor( @@ -73,7 +73,7 @@ public: bool checkIfActuallySeekable() override; - size_t readBigAt(char * to, size_t n, size_t offset, const std::function &) override; + size_t readBigAt(char * to, size_t n, size_t offset, const std::function &) const override; bool supportsReadAt() override { return use_pread; } }; diff --git a/src/IO/ReadBufferFromIStream.cpp b/src/IO/ReadBufferFromIStream.cpp index 52546f1703d..bc90ec7ed15 100644 --- a/src/IO/ReadBufferFromIStream.cpp +++ b/src/IO/ReadBufferFromIStream.cpp @@ -5,52 +5,44 @@ namespace DB { -namespace ErrorCodes -{ - extern const int CANNOT_READ_FROM_ISTREAM; -} - bool ReadBufferFromIStream::nextImpl() { if (eof) return false; + chassert(internal_buffer.begin() != nullptr); + chassert(!internal_buffer.empty()); + size_t bytes_read = 0; char * read_to = internal_buffer.begin(); /// It is necessary to read in a loop, since socket usually returns only data available at the moment. while (bytes_read < internal_buffer.size()) { - try + const auto bytes_read_last_time = stream_buf.readFromDevice(read_to, internal_buffer.size() - bytes_read); + if (bytes_read_last_time <= 0) { - const auto bytes_read_last_time = stream_buf.readFromDevice(read_to, internal_buffer.size() - bytes_read); - if (bytes_read_last_time <= 0) - { - eof = true; - break; - } + eof = true; + break; + } - bytes_read += bytes_read_last_time; - read_to += bytes_read_last_time; - } - catch (...) - { - throw Exception( - ErrorCodes::CANNOT_READ_FROM_ISTREAM, - "Cannot read from istream at offset {}: {}", - count(), - getCurrentExceptionMessage(/*with_stacktrace=*/true)); - } + bytes_read += bytes_read_last_time; + read_to += bytes_read_last_time; } if (bytes_read) + { + working_buffer = internal_buffer; working_buffer.resize(bytes_read); + } return bytes_read; } ReadBufferFromIStream::ReadBufferFromIStream(std::istream & istr_, size_t size) - : BufferWithOwnMemory(size), istr(istr_), stream_buf(dynamic_cast(*istr.rdbuf())) + : BufferWithOwnMemory(size) + , istr(istr_) + , stream_buf(dynamic_cast(*istr.rdbuf())) { } diff --git a/src/IO/ReadBufferFromS3.cpp b/src/IO/ReadBufferFromS3.cpp index 4529771e7b2..491ff253066 100644 --- a/src/IO/ReadBufferFromS3.cpp +++ b/src/IO/ReadBufferFromS3.cpp @@ -34,61 +34,6 @@ namespace ProfileEvents extern const Event RemoteReadThrottlerSleepMicroseconds; } -namespace -{ -DB::PooledHTTPSessionPtr getSession(Aws::S3::Model::GetObjectResult & read_result) -{ - if (auto * session_aware_stream = dynamic_cast *>(&read_result.GetBody())) - return static_cast(session_aware_stream->getSession()); - - if (dynamic_cast *>(&read_result.GetBody())) - return {}; - - /// accept result from S# mock in gtest_writebuffer_s3.cpp - if (dynamic_cast(&read_result.GetBody())) - return {}; - - throw DB::Exception(DB::ErrorCodes::LOGICAL_ERROR, "Session of unexpected type encountered"); -} - -void resetSession(Aws::S3::Model::GetObjectResult & read_result) -{ - if (auto session = getSession(read_result); !session.isNull()) - { - auto & http_session = static_cast(*session); - http_session.reset(); - } -} - -void resetSessionIfNeeded(bool read_all_range_successfully, std::optional & read_result) -{ - if (!read_result) - return; - - if (!read_all_range_successfully) - { - /// When we abandon a session with an ongoing GetObject request and there is another one trying to delete the same object this delete - /// operation will hang until GetObject's session idle timeouts. So we have to call `reset()` on GetObject's session session immediately. - resetSession(*read_result); - ProfileEvents::increment(ProfileEvents::ReadBufferFromS3ResetSessions); - } - else if (auto session = getSession(*read_result); !session.isNull()) - { - if (!session->getProxyHost().empty()) - { - /// Reset proxified sessions because proxy can change for every request. See ProxyConfigurationResolver. - resetSession(*read_result); - ProfileEvents::increment(ProfileEvents::ReadBufferFromS3ResetSessions); - } - else - { - DB::markSessionForReuse(session); - ProfileEvents::increment(ProfileEvents::ReadBufferFromS3PreservedSessions); - } - } -} -} - namespace DB { namespace ErrorCodes @@ -228,7 +173,7 @@ bool ReadBufferFromS3::nextImpl() } -size_t ReadBufferFromS3::readBigAt(char * to, size_t n, size_t range_begin, const std::function & progress_callback) +size_t ReadBufferFromS3::readBigAt(char * to, size_t n, size_t range_begin, const std::function & progress_callback) const { size_t initial_n = n; size_t sleep_time_with_backoff_milliseconds = 100; @@ -240,29 +185,6 @@ size_t ReadBufferFromS3::readBigAt(char * to, size_t n, size_t range_begin, cons ProfileEventTimeIncrement watch(ProfileEvents::ReadBufferFromS3Microseconds); std::optional result; - /// Connection is reusable if we've read the full response. - bool session_is_reusable = false; - SCOPE_EXIT( - { - if (!result.has_value()) - return; - if (session_is_reusable) - { - auto session = getSession(*result); - if (!session.isNull()) - { - DB::markSessionForReuse(session); - ProfileEvents::increment(ProfileEvents::ReadBufferFromS3PreservedSessions); - } - else - session_is_reusable = false; - } - if (!session_is_reusable) - { - resetSession(*result); - ProfileEvents::increment(ProfileEvents::ReadBufferFromS3ResetSessions); - } - }); try { @@ -276,9 +198,8 @@ size_t ReadBufferFromS3::readBigAt(char * to, size_t n, size_t range_begin, cons if (read_settings.remote_throttler) read_settings.remote_throttler->add(bytes_copied, ProfileEvents::RemoteReadThrottlerBytes, ProfileEvents::RemoteReadThrottlerSleepMicroseconds); - /// Read remaining bytes after the end of the payload, see HTTPSessionReuseTag. + /// Read remaining bytes after the end of the payload istr.ignore(INT64_MAX); - session_is_reusable = true; } catch (Poco::Exception & e) { @@ -451,21 +372,8 @@ bool ReadBufferFromS3::atEndOfRequestedRangeGuess() return false; } -ReadBufferFromS3::~ReadBufferFromS3() -{ - try - { - resetSessionIfNeeded(readAllRangeSuccessfully(), read_result); - } - catch (...) - { - tryLogCurrentException(log); - } -} - std::unique_ptr ReadBufferFromS3::initialize(size_t attempt) { - resetSessionIfNeeded(readAllRangeSuccessfully(), read_result); read_all_range_successfully = false; /** @@ -534,10 +442,6 @@ Aws::S3::Model::GetObjectResult ReadBufferFromS3::sendRequest(size_t attempt, si } } -bool ReadBufferFromS3::readAllRangeSuccessfully() const -{ - return read_until_position ? offset == read_until_position : read_all_range_successfully; -} } #endif diff --git a/src/IO/ReadBufferFromS3.h b/src/IO/ReadBufferFromS3.h index f28c23a71d7..003c88df7d2 100644 --- a/src/IO/ReadBufferFromS3.h +++ b/src/IO/ReadBufferFromS3.h @@ -55,7 +55,7 @@ public: bool restricted_seek_ = false, std::optional file_size = std::nullopt); - ~ReadBufferFromS3() override; + ~ReadBufferFromS3() override = default; bool nextImpl() override; @@ -74,7 +74,7 @@ public: String getFileName() const override { return bucket + "/" + key; } - size_t readBigAt(char * to, size_t n, size_t range_begin, const std::function & progress_callback) override; + size_t readBigAt(char * to, size_t n, size_t range_begin, const std::function & progress_callback) const override; bool supportsReadAt() override { return true; } @@ -90,8 +90,6 @@ private: Aws::S3::Model::GetObjectResult sendRequest(size_t attempt, size_t range_begin, std::optional range_end_incl) const; - bool readAllRangeSuccessfully() const; - ReadSettings read_settings; bool use_external_buffer; diff --git a/src/IO/ReadSettings.h b/src/IO/ReadSettings.h index f4dc7880be4..38904df4403 100644 --- a/src/IO/ReadSettings.h +++ b/src/IO/ReadSettings.h @@ -103,7 +103,6 @@ struct ReadSettings bool avoid_readthrough_cache_outside_query_context = true; size_t filesystem_cache_segments_batch_size = 20; - //asdqwe assign these two bool use_page_cache_for_disks_without_file_cache = false; bool read_from_page_cache_if_exists_otherwise_bypass_cache = false; bool page_cache_inject_eviction = false; @@ -123,7 +122,7 @@ struct ReadSettings // Resource to be used during reading ResourceLink resource_link; - size_t http_max_tries = 1; + size_t http_max_tries = 10; size_t http_retry_initial_backoff_ms = 100; size_t http_retry_max_backoff_ms = 1600; bool http_skip_not_found_url_for_globs = true; diff --git a/src/IO/ReadWriteBufferFromHTTP.cpp b/src/IO/ReadWriteBufferFromHTTP.cpp index 15c2a0a021b..bcbec97537a 100644 --- a/src/IO/ReadWriteBufferFromHTTP.cpp +++ b/src/IO/ReadWriteBufferFromHTTP.cpp @@ -1,13 +1,65 @@ #include "ReadWriteBufferFromHTTP.h" #include +#include +#include + namespace ProfileEvents { extern const Event ReadBufferSeekCancelConnection; - extern const Event ReadWriteBufferFromHTTPPreservedSessions; } + +namespace +{ + +bool isRetriableError(const Poco::Net::HTTPResponse::HTTPStatus http_status) noexcept +{ + static constexpr std::array non_retriable_errors{ + Poco::Net::HTTPResponse::HTTPStatus::HTTP_BAD_REQUEST, + Poco::Net::HTTPResponse::HTTPStatus::HTTP_UNAUTHORIZED, + Poco::Net::HTTPResponse::HTTPStatus::HTTP_NOT_FOUND, + Poco::Net::HTTPResponse::HTTPStatus::HTTP_FORBIDDEN, + Poco::Net::HTTPResponse::HTTPStatus::HTTP_NOT_IMPLEMENTED, + Poco::Net::HTTPResponse::HTTPStatus::HTTP_METHOD_NOT_ALLOWED}; + + return std::all_of( + non_retriable_errors.begin(), non_retriable_errors.end(), [&](const auto status) { return http_status != status; }); +} + +Poco::URI getUriAfterRedirect(const Poco::URI & prev_uri, Poco::Net::HTTPResponse & response) +{ + chassert(DB::isRedirect(response.getStatus())); + + auto location = response.get("Location"); + auto location_uri = Poco::URI(location); + if (!location_uri.isRelative()) + return location_uri; + /// Location header contains relative path. So we need to concatenate it + /// with path from the original URI and normalize it. + auto path = std::filesystem::weakly_canonical(std::filesystem::path(prev_uri.getPath()) / location); + location_uri = prev_uri; + location_uri.setPath(path); + return location_uri; +} + +class ReadBufferFromSessionResponse : public DB::ReadBufferFromIStream +{ +private: + DB::HTTPSessionPtr session; + +public: + ReadBufferFromSessionResponse(DB::HTTPSessionPtr && session_, std::istream & rstr, size_t size) + : ReadBufferFromIStream(rstr, size) + , session(std::move(session_)) + { + } +}; + +} + + namespace DB { @@ -21,94 +73,29 @@ namespace ErrorCodes extern const int UNKNOWN_FILE_SIZE; } -template -UpdatableSession::UpdatableSession(const Poco::URI & uri, UInt64 max_redirects_, std::shared_ptr session_factory_) - : max_redirects{max_redirects_} - , initial_uri(uri) - , session_factory(std::move(session_factory_)) +std::unique_ptr ReadWriteBufferFromHTTP::CallResult::transformToReadBuffer(size_t buf_size) && { - session = session_factory->buildNewSession(uri); + chassert(session); + return std::make_unique(std::move(session), *response_stream, buf_size); } -template -typename UpdatableSession::SessionPtr UpdatableSession::getSession() { return session; } - -template -void UpdatableSession::updateSession(const Poco::URI & uri) -{ - ++redirects; - if (redirects <= max_redirects) - session = session_factory->buildNewSession(uri); - else - throw Exception(ErrorCodes::TOO_MANY_REDIRECTS, - "Too many redirects while trying to access {}." - " You can {} redirects by changing the setting 'max_http_get_redirects'." - " Example: `SET max_http_get_redirects = 10`." - " Redirects are restricted to prevent possible attack when a malicious server redirects to an internal resource, bypassing the authentication or firewall.", - initial_uri.toString(), max_redirects ? "increase the allowed maximum number of" : "allow"); -} - -template -typename UpdatableSession::SessionPtr UpdatableSession::createDetachedSession(const Poco::URI & uri) -{ - return session_factory->buildNewSession(uri); -} - -template -std::shared_ptr> UpdatableSession::clone(const Poco::URI & uri) -{ - return std::make_shared>(uri, max_redirects, session_factory); -} - - -namespace detail -{ - -static bool isRetriableError(const Poco::Net::HTTPResponse::HTTPStatus http_status) noexcept -{ - static constexpr std::array non_retriable_errors{ - Poco::Net::HTTPResponse::HTTPStatus::HTTP_BAD_REQUEST, - Poco::Net::HTTPResponse::HTTPStatus::HTTP_UNAUTHORIZED, - Poco::Net::HTTPResponse::HTTPStatus::HTTP_NOT_FOUND, - Poco::Net::HTTPResponse::HTTPStatus::HTTP_FORBIDDEN, - Poco::Net::HTTPResponse::HTTPStatus::HTTP_NOT_IMPLEMENTED, - Poco::Net::HTTPResponse::HTTPStatus::HTTP_METHOD_NOT_ALLOWED}; - - return std::all_of( - non_retriable_errors.begin(), non_retriable_errors.end(), [&](const auto status) { return http_status != status; }); -} - -static Poco::URI getUriAfterRedirect(const Poco::URI & prev_uri, Poco::Net::HTTPResponse & response) -{ - auto location = response.get("Location"); - auto location_uri = Poco::URI(location); - if (!location_uri.isRelative()) - return location_uri; - /// Location header contains relative path. So we need to concatenate it - /// with path from the original URI and normalize it. - auto path = std::filesystem::weakly_canonical(std::filesystem::path(prev_uri.getPath()) / location); - location_uri = prev_uri; - location_uri.setPath(path); - return location_uri; -} - -template -bool ReadWriteBufferFromHTTPBase::withPartialContent(const HTTPRange & range) const +bool ReadWriteBufferFromHTTP::withPartialContent() const { /** * Add range header if we have some passed range * or if we want to retry GET request on purpose. */ - return range.begin || range.end || retry_with_range_header; + return read_range.begin || read_range.end || getOffset() > 0; } -template -size_t ReadWriteBufferFromHTTPBase::getOffset() const { return read_range.begin.value_or(0) + offset_from_begin_pos; } - -template -void ReadWriteBufferFromHTTPBase::prepareRequest(Poco::Net::HTTPRequest & request, Poco::URI uri_, std::optional range) const +size_t ReadWriteBufferFromHTTP::getOffset() const { - request.setHost(uri_.getHost()); // use original, not resolved host name in header + return read_range.begin.value_or(0) + offset_from_begin_pos; +} + +void ReadWriteBufferFromHTTP::prepareRequest(Poco::Net::HTTPRequest & request, std::optional range) const +{ + request.setHost(initial_uri.getHost()); // use original, not resolved host name in header if (out_stream_callback) request.setChunkedTransferEncoding(true); @@ -125,7 +112,6 @@ void ReadWriteBufferFromHTTPBase::prepareRequest(Poco::Net: range_header_value = fmt::format("bytes={}-{}", *range->begin, *range->end); else range_header_value = fmt::format("bytes={}-", *range->begin); - LOG_TEST(log, "Adding header: Range: {}", range_header_value); request.set("Range", range_header_value); } @@ -133,45 +119,7 @@ void ReadWriteBufferFromHTTPBase::prepareRequest(Poco::Net: credentials.authenticate(request); } -template -std::istream * ReadWriteBufferFromHTTPBase::callImpl( - UpdatableSessionPtr & current_session, Poco::URI uri_, Poco::Net::HTTPResponse & response, const std::string & method_, bool for_object_info) -{ - // With empty path poco will send "POST HTTP/1.1" its bug. - if (uri_.getPath().empty()) - uri_.setPath("/"); - - std::optional range; - if (!for_object_info) - { - if (withPartialContent(read_range)) - range = HTTPRange{getOffset(), read_range.end}; - } - - Poco::Net::HTTPRequest request(method_, uri_.getPathAndQuery(), Poco::Net::HTTPRequest::HTTP_1_1); - prepareRequest(request, uri_, range); - - LOG_TRACE(log, "Sending request to {}", uri_.toString()); - - auto sess = current_session->getSession(); - auto & stream_out = sess->sendRequest(request); - - if (out_stream_callback) - out_stream_callback(stream_out); - - auto result_istr = receiveResponse(*sess, request, response, true); - response.getCookies(cookies); - - /// we can fetch object info while the request is being processed - /// and we don't want to override any context used by it - if (!for_object_info) - content_encoding = response.get("Content-Encoding", ""); - - return result_istr; -} - -template -size_t ReadWriteBufferFromHTTPBase::getFileSize() +size_t ReadWriteBufferFromHTTP::getFileSize() { if (!file_info) file_info = getFileInfo(); @@ -179,243 +127,288 @@ size_t ReadWriteBufferFromHTTPBase::getFileSize() if (file_info->file_size) return *file_info->file_size; - throw Exception(ErrorCodes::UNKNOWN_FILE_SIZE, "Cannot find out file size for: {}", uri.toString()); + throw Exception(ErrorCodes::UNKNOWN_FILE_SIZE, "Cannot find out file size for: {}", initial_uri.toString()); } -template -bool ReadWriteBufferFromHTTPBase::supportsReadAt() +bool ReadWriteBufferFromHTTP::supportsReadAt() { if (!file_info) file_info = getFileInfo(); return method == Poco::Net::HTTPRequest::HTTP_GET && file_info->seekable; } -template -bool ReadWriteBufferFromHTTPBase::checkIfActuallySeekable() +bool ReadWriteBufferFromHTTP::checkIfActuallySeekable() { if (!file_info) file_info = getFileInfo(); return file_info->seekable; } -template -String ReadWriteBufferFromHTTPBase::getFileName() const { return uri.toString(); } - -template -void ReadWriteBufferFromHTTPBase::getHeadResponse(Poco::Net::HTTPResponse & response) +String ReadWriteBufferFromHTTP::getFileName() const { - for (size_t i = 0; i < settings.http_max_tries; ++i) - { - try - { - callWithRedirects(response, Poco::Net::HTTPRequest::HTTP_HEAD, true, true); - break; - } - catch (const Poco::Exception & e) - { - if (i == settings.http_max_tries - 1 || e.code() == ErrorCodes::TOO_MANY_REDIRECTS || !isRetriableError(response.getStatus())) - throw; - - LOG_ERROR(log, "Failed to make HTTP_HEAD request to {}. Error: {}", uri.toString(), e.displayText()); - } - } + return initial_uri.toString(); } -template -void ReadWriteBufferFromHTTPBase::setupExternalBuffer() +void ReadWriteBufferFromHTTP::getHeadResponse(Poco::Net::HTTPResponse & response) { - /** - * use_external_buffer -- means we read into the buffer which - * was passed to us from somewhere else. We do not check whether - * previously returned buffer was read or not (no hasPendingData() check is needed), - * because this branch means we are prefetching data, - * each nextImpl() call we can fill a different buffer. - */ - impl->set(internal_buffer.begin(), internal_buffer.size()); - assert(working_buffer.begin() != nullptr); - assert(!internal_buffer.empty()); + doWithRetries( + [&] () + { + callWithRedirects(response, Poco::Net::HTTPRequest::HTTP_HEAD, {}); + }, + /*on_retry=*/ nullptr, + /*mute_logging=*/ true); } -template -ReadWriteBufferFromHTTPBase::ReadWriteBufferFromHTTPBase( - UpdatableSessionPtr session_, - Poco::URI uri_, - const Poco::Net::HTTPBasicCredentials & credentials_, +ReadWriteBufferFromHTTP::ReadWriteBufferFromHTTP( + const HTTPConnectionGroupType & connection_group_, + const Poco::URI & uri_, const std::string & method_, - OutStreamCallback out_stream_callback_, - size_t buffer_size_, - const ReadSettings & settings_, - HTTPHeaderEntries http_header_entries_, + ProxyConfiguration proxy_config_, + ReadSettings read_settings_, + ConnectionTimeouts timeouts_, + const Poco::Net::HTTPBasicCredentials & credentials_, const RemoteHostFilter * remote_host_filter_, - bool delay_initialization, + size_t buffer_size_, + size_t max_redirects_, + OutStreamCallback out_stream_callback_, bool use_external_buffer_, bool http_skip_not_found_url_, - std::optional file_info_, - ProxyConfiguration proxy_config_) + HTTPHeaderEntries http_header_entries_, + bool delay_initialization, + std::optional file_info_) : SeekableReadBuffer(nullptr, 0) - , uri {uri_} - , method {!method_.empty() ? method_ : out_stream_callback_ ? Poco::Net::HTTPRequest::HTTP_POST : Poco::Net::HTTPRequest::HTTP_GET} - , session {session_} - , out_stream_callback {out_stream_callback_} - , credentials {credentials_} - , http_header_entries {std::move(http_header_entries_)} - , remote_host_filter {remote_host_filter_} - , buffer_size {buffer_size_} - , use_external_buffer {use_external_buffer_} - , file_info(file_info_) + , connection_group(connection_group_) + , initial_uri(uri_) + , method(!method_.empty() ? method_ : out_stream_callback_ ? Poco::Net::HTTPRequest::HTTP_POST : Poco::Net::HTTPRequest::HTTP_GET) + , proxy_config(std::move(proxy_config_)) + , read_settings(std::move(read_settings_)) + , timeouts(std::move(timeouts_)) + , credentials(credentials_) + , remote_host_filter(remote_host_filter_) + , buffer_size(buffer_size_) + , max_redirects(max_redirects_) + , use_external_buffer(use_external_buffer_) , http_skip_not_found_url(http_skip_not_found_url_) - , settings {settings_} + , out_stream_callback(std::move(out_stream_callback_)) + , redirects(0) + , http_header_entries {std::move(http_header_entries_)} + , file_info(file_info_) , log(getLogger("ReadWriteBufferFromHTTP")) - , proxy_config(proxy_config_) { - if (settings.http_max_tries <= 0 || settings.http_retry_initial_backoff_ms <= 0 - || settings.http_retry_initial_backoff_ms >= settings.http_retry_max_backoff_ms) + current_uri = initial_uri; + + if (current_uri.getPath().empty()) + current_uri.setPath("/"); + + if (read_settings.http_max_tries <= 0 || read_settings.http_retry_initial_backoff_ms <= 0 + || read_settings.http_retry_initial_backoff_ms >= read_settings.http_retry_max_backoff_ms) throw Exception( ErrorCodes::BAD_ARGUMENTS, "Invalid setting for http backoff, " "must be http_max_tries >= 1 (current is {}) and " "0 < http_retry_initial_backoff_ms < settings.http_retry_max_backoff_ms (now 0 < {} < {})", - settings.http_max_tries, - settings.http_retry_initial_backoff_ms, - settings.http_retry_max_backoff_ms); + read_settings.http_max_tries, + read_settings.http_retry_initial_backoff_ms, + read_settings.http_retry_max_backoff_ms); // Configure User-Agent if it not already set. const std::string user_agent = "User-Agent"; - auto iter = std::find_if( - http_header_entries.begin(), - http_header_entries.end(), - [&user_agent](const HTTPHeaderEntry & entry) { return entry.name == user_agent; }); + auto iter = std::find_if(http_header_entries.begin(), http_header_entries.end(), + [&user_agent] (const HTTPHeaderEntry & entry) { return entry.name == user_agent; }); if (iter == http_header_entries.end()) { - http_header_entries.emplace_back("User-Agent", fmt::format("ClickHouse/{}", VERSION_STRING)); + http_header_entries.emplace_back(user_agent, fmt::format("ClickHouse/{}", VERSION_STRING)); } + if (!delay_initialization && use_external_buffer) + throw Exception( + ErrorCodes::BAD_ARGUMENTS, + "Invalid setting for ReadWriteBufferFromHTTP" + "delay_initialization is false and use_external_buffer it true."); + if (!delay_initialization) { - initialize(); - if (exception) - std::rethrow_exception(exception); + next(); } } -template -void ReadWriteBufferFromHTTPBase::callWithRedirects(Poco::Net::HTTPResponse & response, const String & method_, bool throw_on_all_errors, bool for_object_info) +ReadWriteBufferFromHTTP::CallResult ReadWriteBufferFromHTTP::callImpl( + Poco::Net::HTTPResponse & response, const Poco::URI & uri_, const std::string & method_, const std::optional & range, bool allow_redirects) const { - UpdatableSessionPtr current_session = nullptr; + if (remote_host_filter) + remote_host_filter->checkURL(uri_); - /// we can fetch object info while the request is being processed - /// and we don't want to override any context used by it - if (for_object_info) - current_session = session->clone(uri); - else - current_session = session; + Poco::Net::HTTPRequest request(method_, uri_.getPathAndQuery(), Poco::Net::HTTPRequest::HTTP_1_1); + prepareRequest(request, range); - call(current_session, response, method_, throw_on_all_errors, for_object_info); - saved_uri_redirect = uri; + auto session = makeHTTPSession(connection_group, uri_, timeouts, proxy_config); + + auto & stream_out = session->sendRequest(request); + if (out_stream_callback) + out_stream_callback(stream_out); + + auto & resp_stream = session->receiveResponse(response); + + assertResponseIsOk(current_uri.toString(), response, resp_stream, allow_redirects); + + return ReadWriteBufferFromHTTP::CallResult(std::move(session), resp_stream); +} + +ReadWriteBufferFromHTTP::CallResult ReadWriteBufferFromHTTP::callWithRedirects( + Poco::Net::HTTPResponse & response, const String & method_, const std::optional & range) +{ + auto result = callImpl(response, current_uri, method_, range, true); while (isRedirect(response.getStatus())) { - Poco::URI uri_redirect = getUriAfterRedirect(*saved_uri_redirect, response); - saved_uri_redirect = uri_redirect; - if (remote_host_filter) - remote_host_filter->checkURL(uri_redirect); + Poco::URI uri_redirect = getUriAfterRedirect(current_uri, response); + ++redirects; + if (redirects > max_redirects) + throw Exception( + ErrorCodes::TOO_MANY_REDIRECTS, + "Too many redirects while trying to access {}." + " You can {} redirects by changing the setting 'max_http_get_redirects'." + " Example: `SET max_http_get_redirects = 10`." + " Redirects are restricted to prevent possible attack when a malicious server redirects to an internal resource, bypassing the authentication or firewall.", + initial_uri.toString(), max_redirects ? "increase the allowed maximum number of" : "allow"); - current_session->updateSession(uri_redirect); + current_uri = uri_redirect; - /// we can fetch object info while the request is being processed - /// and we don't want to override any context used by it - auto result_istr = callImpl(current_session, uri_redirect, response, method, for_object_info); - if (!for_object_info) - istr = result_istr; + result = callImpl(response, uri_redirect, method_, range, true); } + + return result; } -template -void ReadWriteBufferFromHTTPBase::call(UpdatableSessionPtr & current_session, Poco::Net::HTTPResponse & response, const String & method_, bool throw_on_all_errors, bool for_object_info) + +void ReadWriteBufferFromHTTP::doWithRetries(std::function && callable, + std::function on_retry, + bool mute_logging) const { - try + [[maybe_unused]] auto milliseconds_to_wait = read_settings.http_retry_initial_backoff_ms; + + bool is_retriable = true; + std::exception_ptr exception = nullptr; + + for (size_t attempt = 1; attempt <= read_settings.http_max_tries; ++attempt) { - /// we can fetch object info while the request is being processed - /// and we don't want to override any context used by it - auto result_istr = callImpl(current_session, saved_uri_redirect ? *saved_uri_redirect : uri, response, method_, for_object_info); - if (!for_object_info) - istr = result_istr; - } - catch (...) - { - /// we can fetch object info while the request is being processed - /// and we don't want to override any context used by it - if (for_object_info) - throw; + [[maybe_unused]] bool last_attempt = attempt + 1 > read_settings.http_max_tries; - if (throw_on_all_errors) - throw; + String error_message; - auto http_status = response.getStatus(); - - if (http_status == Poco::Net::HTTPResponse::HTTPStatus::HTTP_NOT_FOUND && http_skip_not_found_url) + try { - initialization_error = InitializeError::SKIP_NOT_FOUND_URL; + callable(); + return; } - else if (!isRetriableError(http_status)) + catch (Poco::Net::NetException & e) { - initialization_error = InitializeError::NON_RETRYABLE_ERROR; + error_message = e.displayText(); exception = std::current_exception(); } + catch (DB::NetException & e) + { + error_message = e.displayText(); + exception = std::current_exception(); + } + catch (DB::HTTPException & e) + { + if (!isRetriableError(e.getHTTPStatus())) + is_retriable = false; + + error_message = e.displayText(); + exception = std::current_exception(); + } + catch (DB::Exception & e) + { + is_retriable = false; + + error_message = e.displayText(); + exception = std::current_exception(); + } + catch (Poco::Exception & e) + { + if (e.code() == POCO_EMFILE) + is_retriable = false; + + error_message = e.displayText(); + exception = std::current_exception(); + } + + chassert(exception); + + if (last_attempt || !is_retriable) + { + if (!mute_logging) + LOG_ERROR(log, + "Failed to make request to '{}'. Error: '{}'. " + "Failed at try {}/{}.", + initial_uri.toString(), error_message, + attempt, read_settings.http_max_tries); + + std::rethrow_exception(exception); + } else { - throw; + if (on_retry) + on_retry(); + + if (!mute_logging) + LOG_INFO(log, + "Failed to make request to `{}`. Error: {}. " + "Failed at try {}/{}. " + "Will retry with current backoff wait is {}/{} ms.", + initial_uri.toString(), error_message, + attempt + 1, read_settings.http_max_tries, + milliseconds_to_wait, read_settings.http_retry_max_backoff_ms); + + sleepForMilliseconds(milliseconds_to_wait); + milliseconds_to_wait = std::min(milliseconds_to_wait * 2, read_settings.http_retry_max_backoff_ms); } } } -template -void ReadWriteBufferFromHTTPBase::initialize() + +std::unique_ptr ReadWriteBufferFromHTTP::initialize() { Poco::Net::HTTPResponse response; - call(session, response, method); - if (initialization_error != InitializeError::NONE) - return; + std::optional range; + if (withPartialContent()) + range = HTTPRange{getOffset(), read_range.end}; - while (isRedirect(response.getStatus())) - { - Poco::URI uri_redirect = getUriAfterRedirect(saved_uri_redirect.value_or(uri), response); - if (remote_host_filter) - remote_host_filter->checkURL(uri_redirect); + auto result = callWithRedirects(response, method, range); - session->updateSession(uri_redirect); - - istr = callImpl(session, uri_redirect, response, method); - saved_uri_redirect = uri_redirect; - } - - if (response.hasContentLength()) - LOG_DEBUG(log, "Received response with content length: {}", response.getContentLength()); - - if (withPartialContent(read_range) && response.getStatus() != Poco::Net::HTTPResponse::HTTPStatus::HTTP_PARTIAL_CONTENT) + if (range.has_value() && response.getStatus() != Poco::Net::HTTPResponse::HTTPStatus::HTTP_PARTIAL_CONTENT) { /// Having `200 OK` instead of `206 Partial Content` is acceptable in case we retried with range.begin == 0. if (getOffset() != 0) { - if (!exception) + /// Retry 200OK + if (response.getStatus() == Poco::Net::HTTPResponse::HTTPStatus::HTTP_OK) { - exception = std::make_exception_ptr(Exception( + String reason = fmt::format( + "Cannot read with range: [{}, {}] (response status: {}, reason: {}), will retry", + *read_range.begin, read_range.end ? toString(*read_range.end) : "-", + toString(response.getStatus()), response.getReason()); + + /// it is retriable error + throw HTTPException( + ErrorCodes::HTTP_RANGE_NOT_SATISFIABLE, + current_uri.toString(), + Poco::Net::HTTPResponse::HTTP_REQUESTED_RANGE_NOT_SATISFIABLE, + reason, + ""); + } + else + throw Exception( ErrorCodes::HTTP_RANGE_NOT_SATISFIABLE, "Cannot read with range: [{}, {}] (response status: {}, reason: {})", *read_range.begin, read_range.end ? toString(*read_range.end) : "-", - toString(response.getStatus()), response.getReason())); - } - - /// Retry 200OK - if (response.getStatus() == Poco::Net::HTTPResponse::HTTPStatus::HTTP_OK) - initialization_error = InitializeError::RETRYABLE_ERROR; - else - initialization_error = InitializeError::NON_RETRYABLE_ERROR; - - return; + toString(response.getStatus()), response.getReason()); } else if (read_range.end) { @@ -425,257 +418,140 @@ void ReadWriteBufferFromHTTPBase::initialize() } } + response.getCookies(cookies); + content_encoding = response.get("Content-Encoding", ""); + // Remember file size. It'll be used to report eof in next nextImpl() call. if (!read_range.end && response.hasContentLength()) - file_info = parseFileInfo(response, withPartialContent(read_range) ? getOffset() : 0); + file_info = parseFileInfo(response, range.has_value() ? getOffset() : 0); - impl = std::make_unique(*istr, buffer_size); - - if (use_external_buffer) - setupExternalBuffer(); + return std::move(result).transformToReadBuffer(use_external_buffer ? 0 : buffer_size); } -template -bool ReadWriteBufferFromHTTPBase::nextImpl() +bool ReadWriteBufferFromHTTP::nextImpl() { - if (initialization_error == InitializeError::SKIP_NOT_FOUND_URL) - return false; - assert(initialization_error == InitializeError::NONE); - if (next_callback) next_callback(count()); - if ((read_range.end && getOffset() > read_range.end.value()) || - (file_info && file_info->file_size && getOffset() >= file_info->file_size.value())) - { - /// Response was fully read. - markSessionForReuse(session->getSession()); - ProfileEvents::increment(ProfileEvents::ReadWriteBufferFromHTTPPreservedSessions); - return false; - } + bool next_result = false; - if (impl) - { - if (use_external_buffer) - { - setupExternalBuffer(); - } - else - { - /** - * impl was initialized before, pass position() to it to make - * sure there is no pending data which was not read. - */ - if (!working_buffer.empty()) - impl->position() = position(); - } - } - - bool result = false; - size_t milliseconds_to_wait = settings.http_retry_initial_backoff_ms; - bool last_attempt = false; - - auto on_retriable_error = [&]() - { - retry_with_range_header = true; - impl.reset(); - auto http_session = session->getSession(); - http_session->reset(); - if (!last_attempt) - { - sleepForMilliseconds(milliseconds_to_wait); - milliseconds_to_wait = std::min(milliseconds_to_wait * 2, settings.http_retry_max_backoff_ms); - } - }; - - for (size_t i = 0;; ++i) - { - if (last_attempt) - break; - last_attempt = i + 1 >= settings.http_max_tries; - - exception = nullptr; - initialization_error = InitializeError::NONE; - - try + doWithRetries( + /*callable=*/ [&] () { if (!impl) { - initialize(); - - if (initialization_error == InitializeError::NON_RETRYABLE_ERROR) + try { - assert(exception); - break; + impl = initialize(); } - else if (initialization_error == InitializeError::SKIP_NOT_FOUND_URL) + catch (HTTPException & e) { - return false; - } - else if (initialization_error == InitializeError::RETRYABLE_ERROR) - { - LOG_TRACE( - log, - "HTTP request to `{}` failed at try {}/{} with bytes read: {}/{}. " - "(Current backoff wait is {}/{} ms)", - uri.toString(), i + 1, settings.http_max_tries, getOffset(), - read_range.end ? toString(*read_range.end) : "unknown", - milliseconds_to_wait, settings.http_retry_max_backoff_ms); + if (http_skip_not_found_url && e.getHTTPStatus() == Poco::Net::HTTPResponse::HTTPStatus::HTTP_NOT_FOUND) + { + next_result = false; + return; + } - assert(exception); - on_retriable_error(); - continue; + throw; } - assert(!exception); - if (use_external_buffer) { - setupExternalBuffer(); + impl->set(internal_buffer.begin(), internal_buffer.size()); + } + else + { + BufferBase::set(impl->buffer().begin(), impl->buffer().size(), impl->offset()); } } - result = impl->next(); - exception = nullptr; - break; - } - catch (const Poco::Exception & e) + if (use_external_buffer) + { + impl->set(internal_buffer.begin(), internal_buffer.size()); + } + else + { + impl->position() = position(); + } + + next_result = impl->next(); + + BufferBase::set(impl->buffer().begin(), impl->buffer().size(), impl->offset()); + + offset_from_begin_pos += working_buffer.size(); + }, + /*on_retry=*/ [&] () { - /// Too many open files or redirects - non-retryable. - if (e.code() == POCO_EMFILE || e.code() == ErrorCodes::TOO_MANY_REDIRECTS) - throw; + impl.reset(); + }); - /** Retry request unconditionally if nothing has been read yet. - * Otherwise if it is GET method retry with range header. - */ - bool can_retry_request = !offset_from_begin_pos || method == Poco::Net::HTTPRequest::HTTP_GET; - if (!can_retry_request) - throw; - - LOG_INFO( - log, - "HTTP request to `{}` failed at try {}/{} with bytes read: {}/{}. " - "Error: {}. (Current backoff wait is {}/{} ms)", - uri.toString(), - i + 1, - settings.http_max_tries, - getOffset(), - read_range.end ? toString(*read_range.end) : "unknown", - e.displayText(), - milliseconds_to_wait, - settings.http_retry_max_backoff_ms); - - on_retriable_error(); - exception = std::current_exception(); - } - } - - if (exception) - std::rethrow_exception(exception); - - if (!result) - { - /// Eof is reached, i.e response was fully read. - markSessionForReuse(session->getSession()); - ProfileEvents::increment(ProfileEvents::ReadWriteBufferFromHTTPPreservedSessions); - return false; - } - - internal_buffer = impl->buffer(); - working_buffer = internal_buffer; - offset_from_begin_pos += working_buffer.size(); - return true; + return next_result; } -template -size_t ReadWriteBufferFromHTTPBase::readBigAt(char * to, size_t n, size_t offset, const std::function & progress_callback) +size_t ReadWriteBufferFromHTTP::readBigAt(char * to, size_t n, size_t offset, const std::function & progress_callback) const { /// Caller must have checked supportsReadAt(). - /// This ensures we've sent at least one HTTP request and populated saved_uri_redirect. + /// This ensures we've sent at least one HTTP request and populated current_uri. chassert(file_info && file_info->seekable); - Poco::URI uri_ = saved_uri_redirect.value_or(uri); - if (uri_.getPath().empty()) - uri_.setPath("/"); - size_t initial_n = n; - size_t milliseconds_to_wait = settings.http_retry_initial_backoff_ms; + size_t total_bytes_copied = 0; + size_t bytes_copied = 0; + bool is_canceled = false; - for (size_t attempt = 0; n > 0; ++attempt) - { - bool last_attempt = attempt + 1 >= settings.http_max_tries; - - Poco::Net::HTTPRequest request(method, uri_.getPathAndQuery(), Poco::Net::HTTPRequest::HTTP_1_1); - prepareRequest(request, uri_, HTTPRange { .begin = offset, .end = offset + n - 1}); - - LOG_TRACE(log, "Sending request to {} for range [{}, {})", uri_.toString(), offset, offset + n); - - auto sess = session->createDetachedSession(uri_); - - Poco::Net::HTTPResponse response; - std::istream * result_istr; - size_t bytes_copied = 0; - - try + doWithRetries( + /*callable=*/ [&] () { - sess->sendRequest(request); - result_istr = receiveResponse(*sess, request, response, /*allow_redirects*/ false); + auto range = HTTPRange{offset, offset + n - 1}; + + Poco::Net::HTTPResponse response; + auto result = callImpl(response, current_uri, method, range, false); if (response.getStatus() != Poco::Net::HTTPResponse::HTTPStatus::HTTP_PARTIAL_CONTENT && (offset != 0 || offset + n < *file_info->file_size)) - throw Exception( - ErrorCodes::HTTP_RANGE_NOT_SATISFIABLE, - "Expected 206 Partial Content, got {} when reading {} range [{}, {})", - toString(response.getStatus()), uri_.toString(), offset, offset + n); - - copyFromIStreamWithProgressCallback(*result_istr, to, n, progress_callback, &bytes_copied); - if (bytes_copied == n) { - result_istr->ignore(UINT64_MAX); - /// Response was fully read. - markSessionForReuse(*sess); - ProfileEvents::increment(ProfileEvents::ReadWriteBufferFromHTTPPreservedSessions); + String reason = fmt::format( + "When reading with readBigAt {}." + "Cannot read with range: [{}, {}] (response status: {}, reason: {}), will retry", + initial_uri.toString(), + *range.begin, *range.end, + toString(response.getStatus()), response.getReason()); + + throw HTTPException( + ErrorCodes::HTTP_RANGE_NOT_SATISFIABLE, + current_uri.toString(), + Poco::Net::HTTPResponse::HTTP_REQUESTED_RANGE_NOT_SATISFIABLE, + reason, + ""); } - } - catch (const Poco::Exception & e) + + copyFromIStreamWithProgressCallback(*result.response_stream, to, n, progress_callback, &bytes_copied, &is_canceled); + + offset += bytes_copied; + total_bytes_copied += bytes_copied; + to += bytes_copied; + n -= bytes_copied; + bytes_copied = 0; + }, + /*on_retry=*/ [&] () { - LOG_ERROR( - log, - "HTTP request (positioned) to `{}` with range [{}, {}) failed at try {}/{}: {}", - uri_.toString(), offset, offset + n, attempt + 1, settings.http_max_tries, - e.what()); + offset += bytes_copied; + total_bytes_copied += bytes_copied; + to += bytes_copied; + n -= bytes_copied; + bytes_copied = 0; + }); - /// Decide whether to retry. - - if (last_attempt) - throw; - - /// Too many open files - non-retryable. - if (e.code() == POCO_EMFILE) - throw; - - if (const auto * h = dynamic_cast(&e); - h && !isRetriableError(static_cast(h->getHTTPStatus()))) - throw; - - sleepForMilliseconds(milliseconds_to_wait); - milliseconds_to_wait = std::min(milliseconds_to_wait * 2, settings.http_retry_max_backoff_ms); - } - - /// Make sure retries don't re-read the bytes that we've already reported to progress_callback. - offset += bytes_copied; - to += bytes_copied; - n -= bytes_copied; - } - - return initial_n; + chassert(total_bytes_copied == initial_n || is_canceled); + return total_bytes_copied; } -template -off_t ReadWriteBufferFromHTTPBase::getPosition() { return getOffset() - available(); } +off_t ReadWriteBufferFromHTTP::getPosition() +{ + return getOffset() - available(); +} -template -off_t ReadWriteBufferFromHTTPBase::seek(off_t offset_, int whence) +off_t ReadWriteBufferFromHTTP::seek(off_t offset_, int whence) { if (whence != SEEK_SET) throw Exception(ErrorCodes::CANNOT_SEEK_THROUGH_FILE, "Only SEEK_SET mode is allowed."); @@ -688,8 +564,8 @@ off_t ReadWriteBufferFromHTTPBase::seek(off_t offset_, int if (!working_buffer.empty() && size_t(offset_) >= current_offset - working_buffer.size() && offset_ < current_offset) { pos = working_buffer.end() - (current_offset - offset_); - assert(pos >= working_buffer.begin()); - assert(pos < working_buffer.end()); + chassert(pos >= working_buffer.begin()); + chassert(pos < working_buffer.end()); return getPosition(); } @@ -700,7 +576,7 @@ off_t ReadWriteBufferFromHTTPBase::seek(off_t offset_, int if (offset_ > position) { size_t diff = offset_ - position; - if (diff < settings.remote_read_min_bytes_for_seek) + if (diff < read_settings.remote_read_min_bytes_for_seek) { ignore(diff); return offset_; @@ -709,6 +585,7 @@ off_t ReadWriteBufferFromHTTPBase::seek(off_t offset_, int if (!atEndOfRequestedRangeGuess()) ProfileEvents::increment(ProfileEvents::ReadBufferSeekCancelConnection); + impl.reset(); } @@ -719,8 +596,8 @@ off_t ReadWriteBufferFromHTTPBase::seek(off_t offset_, int return offset_; } -template -void ReadWriteBufferFromHTTPBase::setReadUntilPosition(size_t until) + +void ReadWriteBufferFromHTTP::setReadUntilPosition(size_t until) { until = std::max(until, 1ul); if (read_range.end && *read_range.end + 1 == until) @@ -736,8 +613,7 @@ void ReadWriteBufferFromHTTPBase::setReadUntilPosition(size } } -template -void ReadWriteBufferFromHTTPBase::setReadUntilEnd() +void ReadWriteBufferFromHTTP::setReadUntilEnd() { if (!read_range.end) return; @@ -752,11 +628,9 @@ void ReadWriteBufferFromHTTPBase::setReadUntilEnd() } } -template -bool ReadWriteBufferFromHTTPBase::supportsRightBoundedReads() const { return true; } +bool ReadWriteBufferFromHTTP::supportsRightBoundedReads() const { return true; } -template -bool ReadWriteBufferFromHTTPBase::atEndOfRequestedRangeGuess() +bool ReadWriteBufferFromHTTP::atEndOfRequestedRangeGuess() { if (!impl) return true; @@ -767,8 +641,7 @@ bool ReadWriteBufferFromHTTPBase::atEndOfRequestedRangeGues return false; } -template -std::string ReadWriteBufferFromHTTPBase::getResponseCookie(const std::string & name, const std::string & def) const +std::string ReadWriteBufferFromHTTP::getResponseCookie(const std::string & name, const std::string & def) const { for (const auto & cookie : cookies) if (cookie.getName() == name) @@ -776,19 +649,19 @@ std::string ReadWriteBufferFromHTTPBase::getResponseCookie( return def; } -template -void ReadWriteBufferFromHTTPBase::setNextCallback(NextCallback next_callback_) +void ReadWriteBufferFromHTTP::setNextCallback(NextCallback next_callback_) { next_callback = next_callback_; /// Some data maybe already read next_callback(count()); } -template -const std::string & ReadWriteBufferFromHTTPBase::getCompressionMethod() const { return content_encoding; } +const std::string & ReadWriteBufferFromHTTP::getCompressionMethod() const +{ + return content_encoding; +} -template -std::optional ReadWriteBufferFromHTTPBase::tryGetLastModificationTime() +std::optional ReadWriteBufferFromHTTP::tryGetLastModificationTime() { if (!file_info) { @@ -805,12 +678,11 @@ std::optional ReadWriteBufferFromHTTPBase::tryGetLa return file_info->last_modified; } -template -HTTPFileInfo ReadWriteBufferFromHTTPBase::getFileInfo() +ReadWriteBufferFromHTTP::HTTPFileInfo ReadWriteBufferFromHTTP::getFileInfo() { /// May be disabled in case the user knows in advance that the server doesn't support HEAD requests. /// Allows to avoid making unnecessary requests in such cases. - if (!settings.http_make_head_request) + if (!read_settings.http_make_head_request) return HTTPFileInfo{}; Poco::Net::HTTPResponse response; @@ -832,11 +704,11 @@ HTTPFileInfo ReadWriteBufferFromHTTPBase::getFileInfo() throw; } + return parseFileInfo(response, 0); } -template -HTTPFileInfo ReadWriteBufferFromHTTPBase::parseFileInfo(const Poco::Net::HTTPResponse & response, size_t requested_range_begin) +ReadWriteBufferFromHTTP::HTTPFileInfo ReadWriteBufferFromHTTP::parseFileInfo(const Poco::Net::HTTPResponse & response, size_t requested_range_begin) { HTTPFileInfo res; @@ -869,78 +741,3 @@ HTTPFileInfo ReadWriteBufferFromHTTPBase::parseFileInfo(con } -SessionFactory::SessionFactory(const ConnectionTimeouts & timeouts_, ProxyConfiguration proxy_config_) - : timeouts(timeouts_), proxy_config(proxy_config_) {} - -SessionFactory::SessionType SessionFactory::buildNewSession(const Poco::URI & uri) -{ - return makeHTTPSession(uri, timeouts, proxy_config); -} - -ReadWriteBufferFromHTTP::ReadWriteBufferFromHTTP( - Poco::URI uri_, - const std::string & method_, - OutStreamCallback out_stream_callback_, - const ConnectionTimeouts & timeouts, - const Poco::Net::HTTPBasicCredentials & credentials_, - const UInt64 max_redirects, - size_t buffer_size_, - const ReadSettings & settings_, - const HTTPHeaderEntries & http_header_entries_, - const RemoteHostFilter * remote_host_filter_, - bool delay_initialization_, - bool use_external_buffer_, - bool skip_not_found_url_, - std::optional file_info_, - ProxyConfiguration proxy_config_) - : Parent( - std::make_shared(uri_, max_redirects, std::make_shared(timeouts, proxy_config_)), - uri_, - credentials_, - method_, - out_stream_callback_, - buffer_size_, - settings_, - http_header_entries_, - remote_host_filter_, - delay_initialization_, - use_external_buffer_, - skip_not_found_url_, - file_info_, - proxy_config_) {} - - -PooledSessionFactory::PooledSessionFactory( - const ConnectionTimeouts & timeouts_, size_t per_endpoint_pool_size_) - : timeouts(timeouts_) - , per_endpoint_pool_size(per_endpoint_pool_size_) {} - -PooledSessionFactory::SessionType PooledSessionFactory::buildNewSession(const Poco::URI & uri) -{ - return makePooledHTTPSession(uri, timeouts, per_endpoint_pool_size); -} - - -PooledReadWriteBufferFromHTTP::PooledReadWriteBufferFromHTTP( - Poco::URI uri_, - const std::string & method_, - OutStreamCallback out_stream_callback_, - const Poco::Net::HTTPBasicCredentials & credentials_, - size_t buffer_size_, - const UInt64 max_redirects, - PooledSessionFactoryPtr session_factory) - : Parent( - std::make_shared(uri_, max_redirects, session_factory), - uri_, - credentials_, - method_, - out_stream_callback_, - buffer_size_) {} - - -template class UpdatableSession; -template class UpdatableSession; -template class detail::ReadWriteBufferFromHTTPBase>>; -template class detail::ReadWriteBufferFromHTTPBase>>; - -} diff --git a/src/IO/ReadWriteBufferFromHTTP.h b/src/IO/ReadWriteBufferFromHTTP.h index 63ca3e0417c..d0d792430c0 100644 --- a/src/IO/ReadWriteBufferFromHTTP.h +++ b/src/IO/ReadWriteBufferFromHTTP.h @@ -20,7 +20,6 @@ #include #include #include -#include #include #include "config.h" #include @@ -30,44 +29,19 @@ namespace DB { -template -class UpdatableSession +class ReadWriteBufferFromHTTP : public SeekableReadBuffer, public WithFileName, public WithFileSize { public: - using SessionPtr = typename TSessionFactory::SessionType; - - explicit UpdatableSession(const Poco::URI & uri, UInt64 max_redirects_, std::shared_ptr session_factory_); - - SessionPtr getSession(); - - void updateSession(const Poco::URI & uri); - - /// Thread safe. - SessionPtr createDetachedSession(const Poco::URI & uri); - - std::shared_ptr> clone(const Poco::URI & uri); + /// Information from HTTP response header. + struct HTTPFileInfo + { + // nullopt if the server doesn't report it. + std::optional file_size; + std::optional last_modified; + bool seekable = false; + }; private: - SessionPtr session; - UInt64 redirects{0}; - UInt64 max_redirects; - Poco::URI initial_uri; - std::shared_ptr session_factory; -}; - - -/// Information from HTTP response header. -struct HTTPFileInfo -{ - // nullopt if the server doesn't report it. - std::optional file_size; - std::optional last_modified; - bool seekable = false; -}; - - -namespace detail -{ /// Byte range, including right bound [begin, end]. struct HTTPRange { @@ -75,218 +49,210 @@ namespace detail std::optional end; }; - template - class ReadWriteBufferFromHTTPBase : public SeekableReadBuffer, public WithFileName, public WithFileSize + struct CallResult { - protected: - Poco::URI uri; - std::string method; - std::string content_encoding; + HTTPSessionPtr session; + std::istream * response_stream = nullptr; - UpdatableSessionPtr session; - std::istream * istr; /// owned by session - std::unique_ptr impl; - std::function out_stream_callback; - const Poco::Net::HTTPBasicCredentials & credentials; - std::vector cookies; - HTTPHeaderEntries http_header_entries; - const RemoteHostFilter * remote_host_filter = nullptr; - std::function next_callback; + CallResult(HTTPSessionPtr && session_, std::istream & response_stream_) + : session(session_) + , response_stream(&response_stream_) + {} + CallResult(CallResult &&) = default; + CallResult & operator= (CallResult &&) = default; - size_t buffer_size; - bool use_external_buffer; - - size_t offset_from_begin_pos = 0; - HTTPRange read_range; - std::optional file_info; - - /// Delayed exception in case retries with partial content are not satisfiable. - std::exception_ptr exception; - bool retry_with_range_header = false; - /// In case of redirects, save result uri to use it if we retry the request. - std::optional saved_uri_redirect; - - bool http_skip_not_found_url; - - ReadSettings settings; - LoggerPtr log; - - ProxyConfiguration proxy_config; - - bool withPartialContent(const HTTPRange & range) const; - - size_t getOffset() const; - - void prepareRequest(Poco::Net::HTTPRequest & request, Poco::URI uri_, std::optional range) const; - - std::istream * callImpl(UpdatableSessionPtr & current_session, Poco::URI uri_, Poco::Net::HTTPResponse & response, const std::string & method_, bool for_object_info = false); - - size_t getFileSize() override; - - bool supportsReadAt() override; - - bool checkIfActuallySeekable() override; - - String getFileName() const override; - - enum class InitializeError - { - RETRYABLE_ERROR, - /// If error is not retriable, `exception` variable must be set. - NON_RETRYABLE_ERROR, - /// Allows to skip not found urls for globs - SKIP_NOT_FOUND_URL, - NONE, - }; - - InitializeError initialization_error = InitializeError::NONE; - - private: - void getHeadResponse(Poco::Net::HTTPResponse & response); - - void setupExternalBuffer(); - - public: - using NextCallback = std::function; - using OutStreamCallback = std::function; - - explicit ReadWriteBufferFromHTTPBase( - UpdatableSessionPtr session_, - Poco::URI uri_, - const Poco::Net::HTTPBasicCredentials & credentials_, - const std::string & method_ = {}, - OutStreamCallback out_stream_callback_ = {}, - size_t buffer_size_ = DBMS_DEFAULT_BUFFER_SIZE, - const ReadSettings & settings_ = {}, - HTTPHeaderEntries http_header_entries_ = {}, - const RemoteHostFilter * remote_host_filter_ = nullptr, - bool delay_initialization = false, - bool use_external_buffer_ = false, - bool http_skip_not_found_url_ = false, - std::optional file_info_ = std::nullopt, - ProxyConfiguration proxy_config_ = {}); - - void callWithRedirects(Poco::Net::HTTPResponse & response, const String & method_, bool throw_on_all_errors = false, bool for_object_info = false); - - void call(UpdatableSessionPtr & current_session, Poco::Net::HTTPResponse & response, const String & method_, bool throw_on_all_errors = false, bool for_object_info = false); - - /** - * Throws if error is retryable, otherwise sets initialization_error = NON_RETRYABLE_ERROR and - * saves exception into `exception` variable. In case url is not found and skip_not_found_url == true, - * sets initialization_error = SKIP_NOT_FOUND_URL, otherwise throws. - */ - void initialize(); - - bool nextImpl() override; - - size_t readBigAt(char * to, size_t n, size_t offset, const std::function & progress_callback) override; - - off_t getPosition() override; - - off_t seek(off_t offset_, int whence) override; - - void setReadUntilPosition(size_t until) override; - - void setReadUntilEnd() override; - - bool supportsRightBoundedReads() const override; - - // If true, if we destroy impl now, no work was wasted. Just for metrics. - bool atEndOfRequestedRangeGuess(); - - std::string getResponseCookie(const std::string & name, const std::string & def) const; - - /// Set function to call on each nextImpl, useful when you need to track - /// progress. - /// NOTE: parameter on each call is not incremental -- it's all bytes count - /// passed through the buffer - void setNextCallback(NextCallback next_callback_); - - const std::string & getCompressionMethod() const; - - std::optional tryGetLastModificationTime(); - - HTTPFileInfo getFileInfo(); - - HTTPFileInfo parseFileInfo(const Poco::Net::HTTPResponse & response, size_t requested_range_begin); + std::unique_ptr transformToReadBuffer(size_t buf_size) &&; }; -} -class SessionFactory -{ -public: - explicit SessionFactory(const ConnectionTimeouts & timeouts_, ProxyConfiguration proxy_config_ = {}); + const HTTPConnectionGroupType connection_group; + const Poco::URI initial_uri; + const std::string method; + const ProxyConfiguration proxy_config; + const ReadSettings read_settings; + const ConnectionTimeouts timeouts; - using SessionType = HTTPSessionPtr; + const Poco::Net::HTTPBasicCredentials & credentials; + const RemoteHostFilter * remote_host_filter; - SessionType buildNewSession(const Poco::URI & uri); -private: - ConnectionTimeouts timeouts; - ProxyConfiguration proxy_config; -}; + const size_t buffer_size; + const size_t max_redirects; -class ReadWriteBufferFromHTTP : public detail::ReadWriteBufferFromHTTPBase>> -{ - using SessionType = UpdatableSession; - using Parent = detail::ReadWriteBufferFromHTTPBase>; + const bool use_external_buffer; + const bool http_skip_not_found_url; + + std::function out_stream_callback; + + Poco::URI current_uri; + size_t redirects = 0; + + std::string content_encoding; + std::unique_ptr impl; + + std::vector cookies; + HTTPHeaderEntries http_header_entries; + std::function next_callback; + + size_t offset_from_begin_pos = 0; + HTTPRange read_range; + std::optional file_info; + + LoggerPtr log; + + bool withPartialContent() const; + + void prepareRequest(Poco::Net::HTTPRequest & request, std::optional range) const; + + void doWithRetries(std::function && callable, std::function on_retry = nullptr, bool mute_logging = false) const; + + CallResult callImpl( + Poco::Net::HTTPResponse & response, + const Poco::URI & uri_, + const std::string & method_, + const std::optional & range, + bool allow_redirects) const; + + CallResult callWithRedirects( + Poco::Net::HTTPResponse & response, + const String & method_, + const std::optional & range); + + std::unique_ptr initialize(); + + size_t getFileSize() override; + + bool supportsReadAt() override; + + bool checkIfActuallySeekable() override; + + String getFileName() const override; + + void getHeadResponse(Poco::Net::HTTPResponse & response); + + void setupExternalBuffer(); + + size_t getOffset() const; + + // If true, if we destroy impl now, no work was wasted. Just for metrics. + bool atEndOfRequestedRangeGuess(); public: + using NextCallback = std::function; + using OutStreamCallback = std::function; + ReadWriteBufferFromHTTP( - Poco::URI uri_, + const HTTPConnectionGroupType & connection_group_, + const Poco::URI & uri_, const std::string & method_, - OutStreamCallback out_stream_callback_, - const ConnectionTimeouts & timeouts, - const Poco::Net::HTTPBasicCredentials & credentials_, - const UInt64 max_redirects = 0, - size_t buffer_size_ = DBMS_DEFAULT_BUFFER_SIZE, - const ReadSettings & settings_ = {}, - const HTTPHeaderEntries & http_header_entries_ = {}, - const RemoteHostFilter * remote_host_filter_ = nullptr, - bool delay_initialization_ = true, - bool use_external_buffer_ = false, - bool skip_not_found_url_ = false, - std::optional file_info_ = std::nullopt, - ProxyConfiguration proxy_config_ = {}); -}; - -class PooledSessionFactory -{ -public: - explicit PooledSessionFactory( - const ConnectionTimeouts & timeouts_, size_t per_endpoint_pool_size_); - - using SessionType = PooledHTTPSessionPtr; - - /// Thread safe. - SessionType buildNewSession(const Poco::URI & uri); - -private: - ConnectionTimeouts timeouts; - size_t per_endpoint_pool_size; -}; - -using PooledSessionFactoryPtr = std::shared_ptr; - -class PooledReadWriteBufferFromHTTP : public detail::ReadWriteBufferFromHTTPBase>> -{ - using SessionType = UpdatableSession; - using Parent = detail::ReadWriteBufferFromHTTPBase>; - -public: - explicit PooledReadWriteBufferFromHTTP( - Poco::URI uri_, - const std::string & method_, - OutStreamCallback out_stream_callback_, + ProxyConfiguration proxy_config_, + ReadSettings read_settings_, + ConnectionTimeouts timeouts_, const Poco::Net::HTTPBasicCredentials & credentials_, + const RemoteHostFilter * remote_host_filter_, size_t buffer_size_, - const UInt64 max_redirects, - PooledSessionFactoryPtr session_factory); + size_t max_redirects_, + OutStreamCallback out_stream_callback_, + bool use_external_buffer_, + bool http_skip_not_found_url_, + HTTPHeaderEntries http_header_entries_, + bool delay_initialization, + std::optional file_info_); + + bool nextImpl() override; + + size_t readBigAt(char * to, size_t n, size_t offset, const std::function & progress_callback) const override; + + off_t seek(off_t offset_, int whence) override; + + void setReadUntilPosition(size_t until) override; + + void setReadUntilEnd() override; + + bool supportsRightBoundedReads() const override; + + off_t getPosition() override; + + std::string getResponseCookie(const std::string & name, const std::string & def) const; + + /// Set function to call on each nextImpl, useful when you need to track + /// progress. + /// NOTE: parameter on each call is not incremental -- it's all bytes count + /// passed through the buffer + void setNextCallback(NextCallback next_callback_); + + const std::string & getCompressionMethod() const; + + std::optional tryGetLastModificationTime(); + + HTTPFileInfo getFileInfo(); + static HTTPFileInfo parseFileInfo(const Poco::Net::HTTPResponse & response, size_t requested_range_begin); }; +using ReadWriteBufferFromHTTPPtr = std::unique_ptr; -extern template class UpdatableSession; -extern template class UpdatableSession; -extern template class detail::ReadWriteBufferFromHTTPBase>>; -extern template class detail::ReadWriteBufferFromHTTPBase>>; +class BuilderRWBufferFromHTTP +{ + Poco::URI uri; + std::string method = Poco::Net::HTTPRequest::HTTP_GET; + HTTPConnectionGroupType connection_group = HTTPConnectionGroupType::HTTP; + ProxyConfiguration proxy_config{}; + ReadSettings read_settings{}; + ConnectionTimeouts timeouts{}; + const RemoteHostFilter * remote_host_filter = nullptr; + size_t buffer_size = DBMS_DEFAULT_BUFFER_SIZE; + size_t max_redirects = 0; + ReadWriteBufferFromHTTP::OutStreamCallback out_stream_callback = nullptr; + bool use_external_buffer = false; + bool http_skip_not_found_url = false; + HTTPHeaderEntries http_header_entries{}; + bool delay_initialization = true; + +public: + BuilderRWBufferFromHTTP(Poco::URI uri_) + : uri(uri_) + {} + +#define setterMember(name, member) \ + BuilderRWBufferFromHTTP & name(decltype(BuilderRWBufferFromHTTP::member) arg_##member) \ + { \ + member = std::move(arg_##member); \ + return *this; \ + } + + setterMember(withConnectionGroup, connection_group) + setterMember(withMethod, method) + setterMember(withProxy, proxy_config) + setterMember(withSettings, read_settings) + setterMember(withTimeouts, timeouts) + setterMember(withHostFilter, remote_host_filter) + setterMember(withBufSize, buffer_size) + setterMember(withRedirects, max_redirects) + setterMember(withOutCallback, out_stream_callback) + setterMember(withHeaders, http_header_entries) + setterMember(withExternalBuf, use_external_buffer) + setterMember(withDelayInit, delay_initialization) + setterMember(withSkipNotFound, http_skip_not_found_url) +#undef setterMember + + ReadWriteBufferFromHTTPPtr create(const Poco::Net::HTTPBasicCredentials & credentials_) + { + return std::make_unique( + connection_group, + uri, + method, + proxy_config, + read_settings, + timeouts, + credentials_, + remote_host_filter, + buffer_size, + max_redirects, + out_stream_callback, + use_external_buffer, + http_skip_not_found_url, + http_header_entries, + delay_initialization, + /*file_info_=*/ std::nullopt); + } +}; } diff --git a/src/IO/S3/PocoHTTPClient.cpp b/src/IO/S3/PocoHTTPClient.cpp index dbb93e63143..a29a4b0b8ee 100644 --- a/src/IO/S3/PocoHTTPClient.cpp +++ b/src/IO/S3/PocoHTTPClient.cpp @@ -1,5 +1,4 @@ #include -#include "Common/DNSResolver.h" #include "config.h" #if USE_AWS_S3 @@ -147,9 +146,7 @@ ConnectionTimeouts getTimeoutsFromConfiguration(const PocoHTTPClientConfiguratio .withSendTimeout(Poco::Timespan(client_configuration.requestTimeoutMs * 1000)) .withReceiveTimeout(Poco::Timespan(client_configuration.requestTimeoutMs * 1000)) .withTCPKeepAliveTimeout(Poco::Timespan( - client_configuration.enableTcpKeepAlive ? client_configuration.tcpKeepAliveIntervalMs * 1000 : 0)) - .withHTTPKeepAliveTimeout(Poco::Timespan( - client_configuration.http_keep_alive_timeout_ms * 1000)); /// flag indicating whether keep-alive is enabled is set to each session upon creation + client_configuration.enableTcpKeepAlive ? client_configuration.tcpKeepAliveIntervalMs * 1000 : 0)); } PocoHTTPClient::PocoHTTPClient(const PocoHTTPClientConfiguration & client_configuration) @@ -164,8 +161,6 @@ PocoHTTPClient::PocoHTTPClient(const PocoHTTPClientConfiguration & client_config , get_request_throttler(client_configuration.get_request_throttler) , put_request_throttler(client_configuration.put_request_throttler) , extra_headers(client_configuration.extra_headers) - , http_connection_pool_size(client_configuration.http_connection_pool_size) - , wait_on_pool_size_limit(client_configuration.wait_on_pool_size_limit) { } @@ -308,12 +303,8 @@ void PocoHTTPClient::makeRequestInternal( Aws::Utils::RateLimits::RateLimiterInterface * readLimiter, Aws::Utils::RateLimits::RateLimiterInterface * writeLimiter) const { - /// Most sessions in pool are already connected and it is not possible to set proxy host/port to a connected session. const auto request_configuration = per_request_configuration(); - if (http_connection_pool_size) - makeRequestInternalImpl(request, request_configuration, response, readLimiter, writeLimiter); - else - makeRequestInternalImpl(request, request_configuration, response, readLimiter, writeLimiter); + makeRequestInternalImpl(request, request_configuration, response, readLimiter, writeLimiter); } String getMethod(const Aws::Http::HttpRequest & request) @@ -335,7 +326,6 @@ String getMethod(const Aws::Http::HttpRequest & request) } } -template void PocoHTTPClient::makeRequestInternalImpl( Aws::Http::HttpRequest & request, const DB::ProxyConfiguration & proxy_configuration, @@ -343,8 +333,6 @@ void PocoHTTPClient::makeRequestInternalImpl( Aws::Utils::RateLimits::RateLimiterInterface *, Aws::Utils::RateLimits::RateLimiterInterface *) const { - using SessionPtr = std::conditional_t; - LoggerPtr log = getLogger("AWSClient"); auto uri = request.GetUri().GetURIString(); @@ -396,40 +384,17 @@ void PocoHTTPClient::makeRequestInternalImpl( for (unsigned int attempt = 0; attempt <= s3_max_redirects; ++attempt) { Poco::URI target_uri(uri); - SessionPtr session; - if (!proxy_configuration.host.empty()) - { - if (enable_s3_requests_logging) - LOG_TEST(log, "Due to reverse proxy host name ({}) won't be resolved on ClickHouse side", uri); - /// Reverse proxy can replace host header with resolved ip address instead of host name. - /// This can lead to request signature difference on S3 side. - if constexpr (pooled) - session = makePooledHTTPSession( - target_uri, - getTimeouts(method, first_attempt, /*first_byte*/ true), - http_connection_pool_size, - wait_on_pool_size_limit, - proxy_configuration); - else - session = makeHTTPSession( - target_uri, - getTimeouts(method, first_attempt, /*first_byte*/ true), - proxy_configuration); - } - else - { - if constexpr (pooled) - session = makePooledHTTPSession( - target_uri, - getTimeouts(method, first_attempt, /*first_byte*/ true), - http_connection_pool_size, - wait_on_pool_size_limit); - else - session = makeHTTPSession( - target_uri, - getTimeouts(method, first_attempt, /*first_byte*/ true)); - } + if (enable_s3_requests_logging && !proxy_configuration.isEmpty()) + LOG_TEST(log, "Due to reverse proxy host name ({}) won't be resolved on ClickHouse side", uri); + + auto group = for_disk_s3 ? HTTPConnectionGroupType::DISK : HTTPConnectionGroupType::STORAGE; + + auto session = makeHTTPSession( + group, + target_uri, + getTimeouts(method, first_attempt, /*first_byte*/ true), + proxy_configuration); /// In case of error this address will be written to logs request.SetResolvedRemoteHost(session->getResolvedAddress()); @@ -612,10 +577,6 @@ void PocoHTTPClient::makeRequestInternalImpl( response->SetClientErrorMessage(getCurrentExceptionMessage(false)); addMetric(request, S3MetricType::Errors); - - /// Probably this is socket timeout or something more or less related to DNS - /// Let's just remove this host from DNS cache to be more safe - DNSResolver::instance().removeHostFromCache(Poco::URI(uri).getHost()); } } diff --git a/src/IO/S3/PocoHTTPClient.h b/src/IO/S3/PocoHTTPClient.h index 5178d75e7b6..a93a4dfbaf7 100644 --- a/src/IO/S3/PocoHTTPClient.h +++ b/src/IO/S3/PocoHTTPClient.h @@ -13,7 +13,7 @@ #include #include #include -#include +#include #include #include @@ -49,12 +49,7 @@ struct PocoHTTPClientConfiguration : public Aws::Client::ClientConfiguration ThrottlerPtr put_request_throttler; HTTPHeaderEntries extra_headers; - /// Not a client parameter in terms of HTTP and we won't send it to the server. Used internally to determine when connection have to be re-established. - uint32_t http_keep_alive_timeout_ms = 0; - /// Zero means pooling will not be used. - size_t http_connection_pool_size = 0; /// See PoolBase::BehaviourOnLimit - bool wait_on_pool_size_limit = true; bool s3_use_adaptive_timeouts = true; std::function error_report; @@ -98,12 +93,6 @@ public: ); } - void SetResponseBody(Aws::IStream & incoming_stream, PooledHTTPSessionPtr & session_) /// NOLINT - { - body_stream = Aws::Utils::Stream::ResponseStream( - Aws::New>("http result streambuf", session_, incoming_stream.rdbuf())); - } - void SetResponseBody(std::string & response_body) /// NOLINT { auto stream = Aws::New("http result buf", response_body); // STYLE_CHECK_ALLOW_STD_STRING_STREAM @@ -163,7 +152,6 @@ private: EnumSize, }; - template void makeRequestInternalImpl( Aws::Http::HttpRequest & request, const DB::ProxyConfiguration & proxy_configuration, @@ -196,9 +184,6 @@ protected: ThrottlerPtr put_request_throttler; const HTTPHeaderEntries extra_headers; - - size_t http_connection_pool_size = 0; - bool wait_on_pool_size_limit = true; }; } diff --git a/src/IO/SeekableReadBuffer.h b/src/IO/SeekableReadBuffer.h index c002d30e633..798833e1a9b 100644 --- a/src/IO/SeekableReadBuffer.h +++ b/src/IO/SeekableReadBuffer.h @@ -82,7 +82,7 @@ public: /// (e.g. next() or supportsReadAt()). /// * Performance: there's no buffering. Each readBigAt() call typically translates into actual /// IO operation (e.g. HTTP request). Don't use it for small adjacent reads. - virtual size_t readBigAt(char * /*to*/, size_t /*n*/, size_t /*offset*/, const std::function & /*progress_callback*/ = nullptr) + virtual size_t readBigAt(char * /*to*/, size_t /*n*/, size_t /*offset*/, const std::function & /*progress_callback*/ = nullptr) const { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Method readBigAt() not implemented"); } /// Checks if readBigAt() is allowed. May be slow, may throw (e.g. it may do an HTTP request or an fstat). diff --git a/src/IO/S3/SessionAwareIOStream.h b/src/IO/SessionAwareIOStream.h similarity index 97% rename from src/IO/S3/SessionAwareIOStream.h rename to src/IO/SessionAwareIOStream.h index babe52545d1..2380bd0fd60 100644 --- a/src/IO/S3/SessionAwareIOStream.h +++ b/src/IO/SessionAwareIOStream.h @@ -3,7 +3,7 @@ #include -namespace DB::S3 +namespace DB { /** * Wrapper of IOStream to store response stream and corresponding HTTP session. diff --git a/src/IO/WriteBufferFromHTTP.cpp b/src/IO/WriteBufferFromHTTP.cpp index 8ddcbc03b84..d54e1685017 100644 --- a/src/IO/WriteBufferFromHTTP.cpp +++ b/src/IO/WriteBufferFromHTTP.cpp @@ -7,6 +7,7 @@ namespace DB { WriteBufferFromHTTP::WriteBufferFromHTTP( + const HTTPConnectionGroupType & connection_group, const Poco::URI & uri, const std::string & method, const std::string & content_type, @@ -14,9 +15,10 @@ WriteBufferFromHTTP::WriteBufferFromHTTP( const HTTPHeaderEntries & additional_headers, const ConnectionTimeouts & timeouts, size_t buffer_size_, - ProxyConfiguration proxy_configuration) + ProxyConfiguration proxy_configuration +) : WriteBufferFromOStream(buffer_size_) - , session{makeHTTPSession(uri, timeouts, proxy_configuration)} + , session{makeHTTPSession(connection_group, uri, timeouts, proxy_configuration)} , request{method, uri.getPathAndQuery(), Poco::Net::HTTPRequest::HTTP_1_1} { request.setHost(uri.getHost()); diff --git a/src/IO/WriteBufferFromHTTP.h b/src/IO/WriteBufferFromHTTP.h index f1e1e2a9e91..09fd55ec290 100644 --- a/src/IO/WriteBufferFromHTTP.h +++ b/src/IO/WriteBufferFromHTTP.h @@ -19,7 +19,8 @@ namespace DB class WriteBufferFromHTTP : public WriteBufferFromOStream { public: - explicit WriteBufferFromHTTP(const Poco::URI & uri, + explicit WriteBufferFromHTTP(const HTTPConnectionGroupType & connection_group, + const Poco::URI & uri, const std::string & method = Poco::Net::HTTPRequest::HTTP_POST, // POST or PUT only const std::string & content_type = "", const std::string & content_encoding = "", diff --git a/src/IO/copyData.cpp b/src/IO/copyData.cpp index 07222a930b5..d2c7200c350 100644 --- a/src/IO/copyData.cpp +++ b/src/IO/copyData.cpp @@ -35,7 +35,7 @@ void copyDataImpl(ReadBuffer & from, WriteBuffer & to, bool check_bytes, size_t } if (check_bytes && bytes > 0) - throw Exception(ErrorCodes::ATTEMPT_TO_READ_AFTER_EOF, "Attempt to read after EOF."); + throw Exception(ErrorCodes::ATTEMPT_TO_READ_AFTER_EOF, "Attempt to read after EOF, left to copy {} bytes.", bytes); } void copyDataImpl(ReadBuffer & from, WriteBuffer & to, bool check_bytes, size_t bytes, std::function cancellation_hook, ThrottlerPtr throttler) diff --git a/src/Interpreters/ActionsDAG.cpp b/src/Interpreters/ActionsDAG.cpp index 7240679abb7..37ef217cb6d 100644 --- a/src/Interpreters/ActionsDAG.cpp +++ b/src/Interpreters/ActionsDAG.cpp @@ -64,6 +64,37 @@ std::pair getFunctionArguments(const ActionsDAG::N return { std::move(arguments), all_const }; } +bool isConstantFromScalarSubquery(const ActionsDAG::Node * node) +{ + std::stack stack; + stack.push(node); + while (!stack.empty()) + { + const auto * arg = stack.top(); + stack.pop(); + + if (arg->column && isColumnConst(*arg->column)) + continue; + + while (arg->type == ActionsDAG::ActionType::ALIAS) + arg = arg->children.at(0); + + if (arg->type != ActionsDAG::ActionType::FUNCTION) + return false; + + if (arg->function_base->getName() == "__scalarSubqueryResult") + continue; + + if (arg->children.empty() || !arg->function_base->isSuitableForConstantFolding()) + return false; + + for (const auto * child : arg->children) + stack.push(child); + } + + return true; +} + } void ActionsDAG::Node::toTree(JSONBuilder::JSONMap & map) const @@ -196,6 +227,19 @@ const ActionsDAG::Node & ActionsDAG::addFunction( { auto [arguments, all_const] = getFunctionArguments(children); + auto constant_args = function->getArgumentsThatAreAlwaysConstant(); + for (size_t pos : constant_args) + { + if (pos >= children.size()) + continue; + + if (arguments[pos].column && isColumnConst(*arguments[pos].column)) + continue; + + if (isConstantFromScalarSubquery(children[pos])) + arguments[pos].column = arguments[pos].type->createColumnConstWithDefaultValue(0); + } + auto function_base = function->build(arguments); return addFunctionImpl( function_base, @@ -1318,7 +1362,7 @@ ActionsDAGPtr ActionsDAG::makeConvertingActions( size_t num_result_columns = result.size(); if (mode == MatchColumnsMode::Position && num_input_columns != num_result_columns) - throw Exception(ErrorCodes::NUMBER_OF_COLUMNS_DOESNT_MATCH, "Number of columns doesn't match"); + throw Exception(ErrorCodes::NUMBER_OF_COLUMNS_DOESNT_MATCH, "Number of columns doesn't match (source: {} and result: {})", num_input_columns, num_result_columns); if (add_casted_columns && mode != MatchColumnsMode::Name) throw Exception(ErrorCodes::LOGICAL_ERROR, "Converting with add_casted_columns supported only for MatchColumnsMode::Name"); diff --git a/src/Interpreters/AsynchronousInsertLog.cpp b/src/Interpreters/AsynchronousInsertLog.cpp index 5d851f6b47d..0fc39c77fb4 100644 --- a/src/Interpreters/AsynchronousInsertLog.cpp +++ b/src/Interpreters/AsynchronousInsertLog.cpp @@ -33,26 +33,26 @@ ColumnsDescription AsynchronousInsertLogElement::getColumnsDescription() }); return ColumnsDescription{ - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, + {"hostname", std::make_shared(std::make_shared()), "Hostname of the server executing the query."}, + {"event_date", std::make_shared(), "The date when the async insert happened."}, + {"event_time", std::make_shared(), "The date and time when the async insert finished execution."}, + {"event_time_microseconds", std::make_shared(6), "The date and time when the async insert finished execution with microseconds precision."}, - {"query", std::make_shared()}, - {"database", std::make_shared(std::make_shared())}, - {"table", std::make_shared(std::make_shared())}, - {"format", std::make_shared(std::make_shared())}, - {"query_id", std::make_shared()}, - {"bytes", std::make_shared()}, - {"rows", std::make_shared()}, - {"exception", std::make_shared()}, - {"status", type_status}, - {"data_kind", type_data_kind}, + {"query", std::make_shared(), "Query string."}, + {"database", std::make_shared(std::make_shared()), "The name of the database the table is in."}, + {"table", std::make_shared(std::make_shared()), "Table name."}, + {"format", std::make_shared(std::make_shared()), "Format name."}, + {"query_id", std::make_shared(), "ID of the initial query."}, + {"bytes", std::make_shared(), "Number of inserted bytes."}, + {"rows", std::make_shared(), "Number of inserted rows."}, + {"exception", std::make_shared(), "Exception message."}, + {"status", type_status, "Status of the view. Values: 'Ok' = 1 — Successful insert, 'ParsingError' = 2 — Exception when parsing the data, 'FlushError' = 3 — Exception when flushing the data"}, + {"data_kind", type_data_kind, "The status of the data. Value: 'Parsed' and 'Preprocessed'."}, - {"flush_time", std::make_shared()}, - {"flush_time_microseconds", std::make_shared(6)}, - {"flush_query_id", std::make_shared()}, - {"timeout_milliseconds", std::make_shared()}, + {"flush_time", std::make_shared(), "The date and time when the flush happened."}, + {"flush_time_microseconds", std::make_shared(6), "The date and time when the flush happened with microseconds precision."}, + {"flush_query_id", std::make_shared(), "ID of the flush query."}, + {"timeout_milliseconds", std::make_shared(), "The adaptive timeout calculated for this entry."}, }; } diff --git a/src/Interpreters/AsynchronousInsertQueue.cpp b/src/Interpreters/AsynchronousInsertQueue.cpp index c7a39ad610b..7d56dbabe3c 100644 --- a/src/Interpreters/AsynchronousInsertQueue.cpp +++ b/src/Interpreters/AsynchronousInsertQueue.cpp @@ -214,44 +214,57 @@ AsynchronousInsertQueue::AsynchronousInsertQueue(ContextPtr context_, size_t poo AsynchronousInsertQueue::~AsynchronousInsertQueue() { - LOG_TRACE(log, "Shutting down the asynchronous insertion queue"); - shutdown = true; - - for (size_t i = 0; i < pool_size; ++i) + try { - auto & shard = queue_shards[i]; + LOG_TRACE(log, "Shutting down the asynchronous insertion queue"); + shutdown = true; - shard.are_tasks_available.notify_one(); - assert(dump_by_first_update_threads[i].joinable()); - dump_by_first_update_threads[i].join(); - - if (flush_on_shutdown) - { - for (auto & [_, elem] : shard.queue) - scheduleDataProcessingJob(elem.key, std::move(elem.data), getContext(), i); - } - else + for (size_t i = 0; i < pool_size; ++i) { + auto & shard = queue_shards[i]; - for (auto & [_, elem] : shard.queue) - for (const auto & entry : elem.data->entries) - entry->finish(std::make_exception_ptr(Exception( - ErrorCodes::TIMEOUT_EXCEEDED, "Wait for async insert timeout exceeded)"))); + shard.are_tasks_available.notify_one(); + assert(dump_by_first_update_threads[i].joinable()); + dump_by_first_update_threads[i].join(); + + if (flush_on_shutdown) + { + for (auto & [_, elem] : shard.queue) + scheduleDataProcessingJob(elem.key, std::move(elem.data), getContext(), i); + } + else + { + for (auto & [_, elem] : shard.queue) + for (const auto & entry : elem.data->entries) + entry->finish( + std::make_exception_ptr(Exception(ErrorCodes::TIMEOUT_EXCEEDED, "Wait for async insert timeout exceeded)"))); + } } + + pool.wait(); + LOG_TRACE(log, "Asynchronous insertion queue finished"); + } + catch (...) + { + tryLogCurrentException(log); + pool.wait(); } - - pool.wait(); - LOG_TRACE(log, "Asynchronous insertion queue finished"); } void AsynchronousInsertQueue::scheduleDataProcessingJob( const InsertQuery & key, InsertDataPtr data, ContextPtr global_context, size_t shard_num) { + /// Intuitively it seems reasonable to process first inserted blocks first. + /// We add new chunks in the end of entries list, so they are automatically ordered by creation time + chassert(!data->entries.empty()); + const auto priority = Priority{data->entries.front()->create_time.time_since_epoch().count()}; + /// Wrap 'unique_ptr' with 'shared_ptr' to make this /// lambda copyable and allow to save it to the thread pool. pool.scheduleOrThrowOnError( [this, key, global_context, shard_num, my_data = std::make_shared(std::move(data))]() mutable - { processData(key, std::move(*my_data), std::move(global_context), flush_time_history_per_queue_shard[shard_num]); }); + { processData(key, std::move(*my_data), std::move(global_context), flush_time_history_per_queue_shard[shard_num]); }, + priority); } void AsynchronousInsertQueue::preprocessInsertQuery(const ASTPtr & query, const ContextPtr & query_context) @@ -375,6 +388,7 @@ AsynchronousInsertQueue::pushDataChunk(ASTPtr query, DataChunk chunk, ContextPtr assert(data); auto size_in_bytes = data->size_in_bytes; data->size_in_bytes += entry_data_size; + /// We rely on the fact that entries are being added to the list in order of creation time in `scheduleDataProcessingJob()` data->entries.emplace_back(entry); insert_future = entry->getFuture(); diff --git a/src/Interpreters/BackupLog.cpp b/src/Interpreters/BackupLog.cpp index d5b69bc0728..af6c7cf6234 100644 --- a/src/Interpreters/BackupLog.cpp +++ b/src/Interpreters/BackupLog.cpp @@ -22,24 +22,24 @@ ColumnsDescription BackupLogElement::getColumnsDescription() { return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, - {"id", std::make_shared()}, - {"name", std::make_shared()}, - {"base_backup_name", std::make_shared()}, - {"query_id", std::make_shared()}, - {"status", std::make_shared(getBackupStatusEnumValues())}, - {"error", std::make_shared()}, - {"start_time", std::make_shared()}, - {"end_time", std::make_shared()}, - {"num_files", std::make_shared()}, - {"total_size", std::make_shared()}, - {"num_entries", std::make_shared()}, - {"uncompressed_size", std::make_shared()}, - {"compressed_size", std::make_shared()}, - {"files_read", std::make_shared()}, - {"bytes_read", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "Hostname of the server executing the query."}, + {"event_date", std::make_shared(), "Date of the entry."}, + {"event_time_microseconds", std::make_shared(6), "Time of the entry with microseconds precision."}, + {"id", std::make_shared(), "Identifier of the backup or restore operation."}, + {"name", std::make_shared(), "Name of the backup storage (the contents of the FROM or TO clause)."}, + {"base_backup_name", std::make_shared(), "The name of base backup in case incremental one."}, + {"query_id", std::make_shared(), "The ID of a query associated with a backup operation."}, + {"status", std::make_shared(getBackupStatusEnumValues()), "Operation status."}, + {"error", std::make_shared(), "Error message of the failed operation (empty string for successful operations)."}, + {"start_time", std::make_shared(), "Start time of the operation."}, + {"end_time", std::make_shared(), "End time of the operation."}, + {"num_files", std::make_shared(), "Number of files stored in the backup."}, + {"total_size", std::make_shared(), "Total size of files stored in the backup."}, + {"num_entries", std::make_shared(), "Number of entries in the backup, i.e. the number of files inside the folder if the backup is stored as a folder, or the number of files inside the archive if the backup is stored as an archive. It is not the same as num_files if it's an incremental backup or if it contains empty files or duplicates. The following is always true: num_entries <= num_files."}, + {"uncompressed_size", std::make_shared(), "Uncompressed size of the backup."}, + {"compressed_size", std::make_shared(), "Compressed size of the backup. If the backup is not stored as an archive it equals to uncompressed_size."}, + {"files_read", std::make_shared(), "Number of files read during the restore operation."}, + {"bytes_read", std::make_shared(), "Total size of files read during the restore operation."}, }; } diff --git a/src/Interpreters/BlobStorageLog.cpp b/src/Interpreters/BlobStorageLog.cpp index 520405374ca..f9d5b0d6790 100644 --- a/src/Interpreters/BlobStorageLog.cpp +++ b/src/Interpreters/BlobStorageLog.cpp @@ -26,23 +26,23 @@ ColumnsDescription BlobStorageLogElement::getColumnsDescription() return ColumnsDescription { - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, + {"event_date", std::make_shared(), "Date of the event."}, + {"event_time", std::make_shared(), "Time of the event."}, + {"event_time_microseconds", std::make_shared(6), "Time of the event with microseconds precision."}, - {"event_type", event_enum_type}, + {"event_type", event_enum_type, "Type of the event. Possible values: 'Upload', 'Delete', 'MultiPartUploadCreate', 'MultiPartUploadWrite', 'MultiPartUploadComplete', 'MultiPartUploadAbort'"}, - {"query_id", std::make_shared()}, - {"thread_id", std::make_shared()}, - {"thread_name", std::make_shared()}, + {"query_id", std::make_shared(), "Identifier of the query associated with the event, if any."}, + {"thread_id", std::make_shared(), "Identifier of the thread performing the operation."}, + {"thread_name", std::make_shared(), "Name of the thread performing the operation."}, - {"disk_name", std::make_shared(std::make_shared())}, - {"bucket", std::make_shared()}, - {"remote_path", std::make_shared()}, - {"local_path", std::make_shared()}, - {"data_size", std::make_shared()}, + {"disk_name", std::make_shared(std::make_shared()), "Name of the associated disk."}, + {"bucket", std::make_shared(), "Name of the bucket."}, + {"remote_path", std::make_shared(), "Path to the remote resource."}, + {"local_path", std::make_shared(), "Path to the metadata file on the local system, which references the remote resource."}, + {"data_size", std::make_shared(), "Size of the data involved in the upload event."}, - {"error", std::make_shared()}, + {"error", std::make_shared(), "Error message associated with the event, if any."}, }; } diff --git a/src/Interpreters/Cache/FileCache.cpp b/src/Interpreters/Cache/FileCache.cpp index d242544f787..9c705ddc27c 100644 --- a/src/Interpreters/Cache/FileCache.cpp +++ b/src/Interpreters/Cache/FileCache.cpp @@ -188,6 +188,11 @@ CacheGuard::Lock FileCache::lockCache() const return cache_guard.lock(); } +CacheGuard::Lock FileCache::tryLockCache() const +{ + return cache_guard.tryLock(); +} + FileSegments FileCache::getImpl(const LockedKey & locked_key, const FileSegment::Range & range, size_t file_segments_limit) const { /// Given range = [left, right] and non-overlapping ordered set of file segments, diff --git a/src/Interpreters/Cache/FileCache.h b/src/Interpreters/Cache/FileCache.h index 2de2f347999..5b665ad0271 100644 --- a/src/Interpreters/Cache/FileCache.h +++ b/src/Interpreters/Cache/FileCache.h @@ -173,6 +173,7 @@ public: void deactivateBackgroundOperations(); CacheGuard::Lock lockCache() const; + CacheGuard::Lock tryLockCache() const; std::vector sync(); diff --git a/src/Interpreters/Cache/FileSegment.cpp b/src/Interpreters/Cache/FileSegment.cpp index 8bd89465917..e58402dac03 100644 --- a/src/Interpreters/Cache/FileSegment.cpp +++ b/src/Interpreters/Cache/FileSegment.cpp @@ -23,6 +23,7 @@ namespace ProfileEvents extern const Event FileSegmentWriteMicroseconds; extern const Event FileSegmentUseMicroseconds; extern const Event FileSegmentHolderCompleteMicroseconds; + extern const Event FileSegmentFailToIncreasePriority; extern const Event FilesystemCacheHoldFileSegments; extern const Event FilesystemCacheUnusedHoldFileSegments; } @@ -965,8 +966,10 @@ void FileSegment::increasePriority() auto it = getQueueIterator(); if (it) { - auto cache_lock = cache->lockCache(); - hits_count = it->increasePriority(cache_lock); + if (auto cache_lock = cache->tryLockCache()) + hits_count = it->increasePriority(cache_lock); + else + ProfileEvents::increment(ProfileEvents::FileSegmentFailToIncreasePriority); } } diff --git a/src/Interpreters/Cache/Guards.h b/src/Interpreters/Cache/Guards.h index 09586b55c61..5729620d82f 100644 --- a/src/Interpreters/Cache/Guards.h +++ b/src/Interpreters/Cache/Guards.h @@ -65,10 +65,12 @@ struct CacheGuard : private boost::noncopyable /// so, we wouldn't be able to pass CacheGuard::Lock to a function which accepts KeyGuard::Lock, for example struct Lock : public std::unique_lock { - explicit Lock(std::mutex & mutex_) : std::unique_lock(mutex_) {} + using Base = std::unique_lock; + using Base::Base; }; Lock lock() { return Lock(mutex); } + Lock tryLock() { return Lock(mutex, std::try_to_lock); } std::mutex mutex; }; diff --git a/src/Interpreters/Context.cpp b/src/Interpreters/Context.cpp index cdeaa46cff2..d658fbe9920 100644 --- a/src/Interpreters/Context.cpp +++ b/src/Interpreters/Context.cpp @@ -380,8 +380,6 @@ struct ContextSharedPart : boost::noncopyable OrdinaryBackgroundExecutorPtr moves_executor TSA_GUARDED_BY(background_executors_mutex); OrdinaryBackgroundExecutorPtr fetch_executor TSA_GUARDED_BY(background_executors_mutex); OrdinaryBackgroundExecutorPtr common_executor TSA_GUARDED_BY(background_executors_mutex); - /// The global pool of HTTP sessions for background fetches. - PooledSessionFactoryPtr fetches_session_factory TSA_GUARDED_BY(background_executors_mutex); RemoteHostFilter remote_host_filter; /// Allowed URL from config.xml HTTPHeaderFilter http_header_filter; /// Forbidden HTTP headers from config.xml @@ -3270,7 +3268,7 @@ bool checkZooKeeperConfigIsLocal(const Poco::Util::AbstractConfiguration & confi if (startsWith(key, "node")) { String host = config.getString(config_name + "." + key + ".host"); - if (isLocalAddress(DNSResolver::instance().resolveHost(host))) + if (isLocalAddress(DNSResolver::instance().resolveHostAllInOriginOrder(host).front())) return true; } } @@ -5039,11 +5037,6 @@ void Context::initializeBackgroundExecutorsIfNeeded() ); LOG_INFO(shared->log, "Initialized background executor for move operations with num_threads={}, num_tasks={}", background_move_pool_size, background_move_pool_size); - auto timeouts = ConnectionTimeouts::getFetchPartHTTPTimeouts(getServerSettings(), getSettingsRef()); - /// The number of background fetches is limited by the number of threads in the background thread pool. - /// It doesn't make any sense to limit the number of connections per host any further. - shared->fetches_session_factory = std::make_shared(timeouts, background_fetches_pool_size); - shared->fetch_executor = std::make_shared ( "Fetch", @@ -5097,12 +5090,6 @@ OrdinaryBackgroundExecutorPtr Context::getCommonExecutor() const return shared->common_executor; } -PooledSessionFactoryPtr Context::getCommonFetchesSessionFactory() const -{ - SharedLockGuard lock(shared->background_executors_mutex); - return shared->fetches_session_factory; -} - IAsynchronousReader & Context::getThreadPoolReader(FilesystemReaderType type) const { callOnce(shared->readers_initialized, [&] { diff --git a/src/Interpreters/Context.h b/src/Interpreters/Context.h index b2310eaa85d..c8aa3604a6f 100644 --- a/src/Interpreters/Context.h +++ b/src/Interpreters/Context.h @@ -206,9 +206,6 @@ using TemporaryDataOnDiskScopePtr = std::shared_ptr; class PreparedSetsCache; using PreparedSetsCachePtr = std::shared_ptr; -class PooledSessionFactory; -using PooledSessionFactoryPtr = std::shared_ptr; - class SessionTracker; struct ServerSettings; @@ -1226,7 +1223,6 @@ public: OrdinaryBackgroundExecutorPtr getMovesExecutor() const; OrdinaryBackgroundExecutorPtr getFetchesExecutor() const; OrdinaryBackgroundExecutorPtr getCommonExecutor() const; - PooledSessionFactoryPtr getCommonFetchesSessionFactory() const; IAsynchronousReader & getThreadPoolReader(FilesystemReaderType type) const; #if USE_LIBURING diff --git a/src/Interpreters/CrashLog.cpp b/src/Interpreters/CrashLog.cpp index 4fb81e4bcf7..410ea922429 100644 --- a/src/Interpreters/CrashLog.cpp +++ b/src/Interpreters/CrashLog.cpp @@ -23,18 +23,18 @@ ColumnsDescription CrashLogElement::getColumnsDescription() { return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"timestamp_ns", std::make_shared()}, - {"signal", std::make_shared()}, - {"thread_id", std::make_shared()}, - {"query_id", std::make_shared()}, - {"trace", std::make_shared(std::make_shared())}, - {"trace_full", std::make_shared(std::make_shared())}, - {"version", std::make_shared()}, - {"revision", std::make_shared()}, - {"build_id", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "The hostname where the crash occurred."}, + {"event_date", std::make_shared(), "The date of the crash."}, + {"event_time", std::make_shared(), "The time of the crash."}, + {"timestamp_ns", std::make_shared(), "Timestamp of the event with nanoseconds."}, + {"signal", std::make_shared(), "Signal number."}, + {"thread_id", std::make_shared(), "Thread ID."}, + {"query_id", std::make_shared(), "Query ID."}, + {"trace", std::make_shared(std::make_shared()), "Stack trace at the moment of crash. Each element is a virtual memory address inside ClickHouse server process."}, + {"trace_full", std::make_shared(std::make_shared()), "Stack trace at the moment of crash. Each element contains a called method inside ClickHouse server process."}, + {"version", std::make_shared(), "ClickHouse server version."}, + {"revision", std::make_shared(), "ClickHouse server revision."}, + {"build_id", std::make_shared(), "BuildID that is generated by compiler."}, }; } diff --git a/src/Interpreters/DDLTask.cpp b/src/Interpreters/DDLTask.cpp index 6a8f82914bf..fe2baea6b4e 100644 --- a/src/Interpreters/DDLTask.cpp +++ b/src/Interpreters/DDLTask.cpp @@ -44,6 +44,11 @@ bool HostID::isLocalAddress(UInt16 clickhouse_port) const { return DB::isLocalAddress(DNSResolver::instance().resolveAddress(host_name, port), clickhouse_port); } + catch (const DB::NetException &) + { + /// Avoid "Host not found" exceptions + return false; + } catch (const Poco::Net::NetException &) { /// Avoid "Host not found" exceptions diff --git a/src/Interpreters/ExecuteScalarSubqueriesVisitor.cpp b/src/Interpreters/ExecuteScalarSubqueriesVisitor.cpp index 77a022e066b..0cf138c14f6 100644 --- a/src/Interpreters/ExecuteScalarSubqueriesVisitor.cpp +++ b/src/Interpreters/ExecuteScalarSubqueriesVisitor.cpp @@ -281,7 +281,7 @@ void ExecuteScalarSubqueriesMatcher::visit(const ASTSubquery & subquery, ASTPtr if (data.only_analyze) { ast->as()->alias.clear(); - auto func = makeASTFunction("identity", std::move(ast)); + auto func = makeASTFunction("__scalarSubqueryResult", std::move(ast)); func->alias = subquery_alias; func->prefer_alias_to_column_name = prefer_alias_to_column_name; ast = std::move(func); diff --git a/src/Interpreters/FilesystemCacheLog.cpp b/src/Interpreters/FilesystemCacheLog.cpp index ccfee49a66f..80fe1c3a8ef 100644 --- a/src/Interpreters/FilesystemCacheLog.cpp +++ b/src/Interpreters/FilesystemCacheLog.cpp @@ -38,20 +38,20 @@ ColumnsDescription FilesystemCacheLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"query_id", std::make_shared()}, - {"source_file_path", std::make_shared()}, - {"file_segment_range", std::make_shared(types)}, - {"total_requested_range", std::make_shared(types)}, - {"key", std::make_shared()}, - {"offset", std::make_shared()}, - {"size", std::make_shared()}, - {"read_type", std::make_shared()}, - {"read_from_cache_attempted", std::make_shared()}, - {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared())}, - {"read_buffer_id", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "Hostname"}, + {"event_date", std::make_shared(), "Event date"}, + {"event_time", std::make_shared(), "Event time"}, + {"query_id", std::make_shared(), "Id of the query"}, + {"source_file_path", std::make_shared(), "File segment path on filesystem"}, + {"file_segment_range", std::make_shared(types), "File segment range"}, + {"total_requested_range", std::make_shared(types), "Full read range"}, + {"key", std::make_shared(), "File segment key"}, + {"offset", std::make_shared(), "File segment offset"}, + {"size", std::make_shared(), "Read size"}, + {"read_type", std::make_shared(), "Read type: READ_FROM_CACHE, READ_FROM_FS_AND_DOWNLOADED_TO_CACHE, READ_FROM_FS_BYPASSING_CACHE"}, + {"read_from_cache_attempted", std::make_shared(), "Whether reading from cache was attempted"}, + {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared()), "Profile events collected while reading this file segment"}, + {"read_buffer_id", std::make_shared(), "Internal implementation read buffer id"}, }; } diff --git a/src/Interpreters/GlobalSubqueriesVisitor.h b/src/Interpreters/GlobalSubqueriesVisitor.h index c53e54573c5..64b6eb5dce9 100644 --- a/src/Interpreters/GlobalSubqueriesVisitor.h +++ b/src/Interpreters/GlobalSubqueriesVisitor.h @@ -216,7 +216,7 @@ private: if (enable_parallel_processing_of_joins) { /// We don't enable parallel replicas for IN (subquery) - if (ast->as()) + if (!settings.parallel_replicas_allow_in_with_subquery && ast->as()) { if (settings.allow_experimental_parallel_reading_from_replicas == 1) { diff --git a/src/Interpreters/HashJoin.cpp b/src/Interpreters/HashJoin.cpp index 77452936fbd..ddd65f95627 100644 --- a/src/Interpreters/HashJoin.cpp +++ b/src/Interpreters/HashJoin.cpp @@ -1019,6 +1019,7 @@ struct JoinOnKeyColumns bool isRowFiltered(size_t i) const { return join_mask_column.isRowFiltered(i); } }; +template class AddedColumns { public: @@ -1034,6 +1035,12 @@ public: } }; + struct LazyOutput + { + PaddedPODArray blocks; + PaddedPODArray row_nums; + }; + AddedColumns( const Block & left_block, const Block & block_with_columns_to_add, @@ -1050,6 +1057,13 @@ public: if (is_asof_join) ++num_columns_to_add; + if constexpr (lazy) + { + has_columns_to_add = num_columns_to_add > 0; + lazy_output.blocks.reserve(rows_to_add); + lazy_output.row_nums.reserve(rows_to_add); + } + columns.reserve(num_columns_to_add); type_name.reserve(num_columns_to_add); right_indexes.reserve(num_columns_to_add); @@ -1089,81 +1103,18 @@ public: size_t size() const { return columns.size(); } + void buildOutput(); + ColumnWithTypeAndName moveColumn(size_t i) { return ColumnWithTypeAndName(std::move(columns[i]), type_name[i].type, type_name[i].qualified_name); } + void appendFromBlock(const Block & block, size_t row_num, bool has_default); - template - void appendFromBlock(const Block & block, size_t row_num) - { - if constexpr (has_defaults) - applyLazyDefaults(); + void appendDefaultRow(); -#ifndef NDEBUG - for (size_t j = 0; j < right_indexes.size(); ++j) - { - const auto * column_from_block = block.getByPosition(right_indexes[j]).column.get(); - const auto * dest_column = columns[j].get(); - if (auto * nullable_col = nullable_column_ptrs[j]) - { - if (!is_join_get) - throw Exception(ErrorCodes::LOGICAL_ERROR, - "Columns {} and {} can have different nullability only in joinGetOrNull", - dest_column->getName(), column_from_block->getName()); - dest_column = nullable_col->getNestedColumnPtr().get(); - } - /** Using dest_column->structureEquals(*column_from_block) will not work for low cardinality columns, - * because dictionaries can be different, while calling insertFrom on them is safe, for example: - * ColumnLowCardinality(size = 0, UInt8(size = 0), ColumnUnique(size = 1, String(size = 1))) - * and - * ColumnLowCardinality(size = 0, UInt16(size = 0), ColumnUnique(size = 1, String(size = 1))) - */ - if (typeid(*dest_column) != typeid(*column_from_block)) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Columns {} and {} have different types {} and {}", - dest_column->getName(), column_from_block->getName(), - demangle(typeid(*dest_column).name()), demangle(typeid(*column_from_block).name())); - } -#endif - - if (is_join_get) - { - size_t right_indexes_size = right_indexes.size(); - for (size_t j = 0; j < right_indexes_size; ++j) - { - const auto & column_from_block = block.getByPosition(right_indexes[j]); - if (auto * nullable_col = nullable_column_ptrs[j]) - nullable_col->insertFromNotNullable(*column_from_block.column, row_num); - else - columns[j]->insertFrom(*column_from_block.column, row_num); - } - } - else - { - size_t right_indexes_size = right_indexes.size(); - for (size_t j = 0; j < right_indexes_size; ++j) - { - const auto & column_from_block = block.getByPosition(right_indexes[j]); - columns[j]->insertFrom(*column_from_block.column, row_num); - } - } - } - - void appendDefaultRow() - { - ++lazy_defaults_count; - } - - void applyLazyDefaults() - { - if (lazy_defaults_count) - { - for (size_t j = 0, size = right_indexes.size(); j < size; ++j) - JoinCommon::addDefaultValues(*columns[j], type_name[j].type, lazy_defaults_count); - lazy_defaults_count = 0; - } - } + void applyLazyDefaults(); const IColumn & leftAsofKey() const { return *left_asof_key; } @@ -1192,16 +1143,50 @@ public: } private: - std::vector type_name; - MutableColumns columns; - std::vector nullable_column_ptrs; + void checkBlock(const Block & block) + { + for (size_t j = 0; j < right_indexes.size(); ++j) + { + const auto * column_from_block = block.getByPosition(right_indexes[j]).column.get(); + const auto * dest_column = columns[j].get(); + if (auto * nullable_col = nullable_column_ptrs[j]) + { + if (!is_join_get) + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Columns {} and {} can have different nullability only in joinGetOrNull", + dest_column->getName(), column_from_block->getName()); + dest_column = nullable_col->getNestedColumnPtr().get(); + } + /** Using dest_column->structureEquals(*column_from_block) will not work for low cardinality columns, + * because dictionaries can be different, while calling insertFrom on them is safe, for example: + * ColumnLowCardinality(size = 0, UInt8(size = 0), ColumnUnique(size = 1, String(size = 1))) + * and + * ColumnLowCardinality(size = 0, UInt16(size = 0), ColumnUnique(size = 1, String(size = 1))) + */ + if (typeid(*dest_column) != typeid(*column_from_block)) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Columns {} and {} have different types {} and {}", + dest_column->getName(), column_from_block->getName(), + demangle(typeid(*dest_column).name()), demangle(typeid(*column_from_block).name())); + } + } + + MutableColumns columns; + bool is_join_get; std::vector right_indexes; + std::vector type_name; + std::vector nullable_column_ptrs; size_t lazy_defaults_count = 0; + + /// for lazy + // The default row is represented by an empty RowRef, so that fixed-size blocks can be generated sequentially, + // default_count cannot represent the position of the row + LazyOutput lazy_output; + bool has_columns_to_add; + /// for ASOF const IColumn * left_asof_key = nullptr; - bool is_join_get; void addColumn(const ColumnWithTypeAndName & src_column, const std::string & qualified_name) { @@ -1210,6 +1195,126 @@ private: type_name.emplace_back(src_column.type, src_column.name, qualified_name); } }; +template<> void AddedColumns::buildOutput() +{ +} + +template<> +void AddedColumns::buildOutput() +{ + for (size_t i = 0; i < this->size(); ++i) + { + auto& col = columns[i]; + size_t default_count = 0; + auto apply_default = [&]() + { + if (default_count > 0) + { + JoinCommon::addDefaultValues(*col, type_name[i].type, default_count); + default_count = 0; + } + }; + + for (size_t j = 0; j < lazy_output.blocks.size(); ++j) + { + if (!lazy_output.blocks[j]) + { + default_count ++; + continue; + } + apply_default(); + const auto & column_from_block = reinterpret_cast(lazy_output.blocks[j])->getByPosition(right_indexes[i]); + /// If it's joinGetOrNull, we need to wrap not-nullable columns in StorageJoin. + if (is_join_get) + { + if (auto * nullable_col = typeid_cast(col.get()); + nullable_col && !column_from_block.column->isNullable()) + { + nullable_col->insertFromNotNullable(*column_from_block.column, lazy_output.row_nums[j]); + continue; + } + } + col->insertFrom(*column_from_block.column, lazy_output.row_nums[j]); + } + apply_default(); + } +} + +template<> +void AddedColumns::applyLazyDefaults() +{ + if (lazy_defaults_count) + { + for (size_t j = 0, size = right_indexes.size(); j < size; ++j) + JoinCommon::addDefaultValues(*columns[j], type_name[j].type, lazy_defaults_count); + lazy_defaults_count = 0; + } +} + +template<> +void AddedColumns::applyLazyDefaults() +{ +} + +template <> +void AddedColumns::appendFromBlock(const Block & block, size_t row_num,const bool has_defaults) +{ + if (has_defaults) + applyLazyDefaults(); + +#ifndef NDEBUG + checkBlock(block); +#endif + if (is_join_get) + { + size_t right_indexes_size = right_indexes.size(); + for (size_t j = 0; j < right_indexes_size; ++j) + { + const auto & column_from_block = block.getByPosition(right_indexes[j]); + if (auto * nullable_col = nullable_column_ptrs[j]) + nullable_col->insertFromNotNullable(*column_from_block.column, row_num); + else + columns[j]->insertFrom(*column_from_block.column, row_num); + } + } + else + { + size_t right_indexes_size = right_indexes.size(); + for (size_t j = 0; j < right_indexes_size; ++j) + { + const auto & column_from_block = block.getByPosition(right_indexes[j]); + columns[j]->insertFrom(*column_from_block.column, row_num); + } + } +} + +template <> +void AddedColumns::appendFromBlock(const Block & block, size_t row_num, bool) +{ +#ifndef NDEBUG + checkBlock(block); +#endif + if (has_columns_to_add) + { + lazy_output.blocks.emplace_back(reinterpret_cast(&block)); + lazy_output.row_nums.emplace_back(static_cast(row_num)); + } +} +template<> +void AddedColumns::appendDefaultRow() +{ + ++lazy_defaults_count; +} + +template<> +void AddedColumns::appendDefaultRow() +{ + if (has_columns_to_add) + { + lazy_output.blocks.emplace_back(0); + lazy_output.row_nums.emplace_back(0); + } +} template struct JoinFeatures @@ -1308,7 +1413,7 @@ public: } }; -template +template void addFoundRowAll( const typename Map::mapped_type & mapped, AddedColumns & added, @@ -1327,7 +1432,7 @@ void addFoundRowAll( { if (!known_rows.isKnown(std::make_pair(it->block, it->row_num))) { - added.appendFromBlock(*it->block, it->row_num); + added.appendFromBlock(*it->block, it->row_num, false); ++current_offset; if (!new_known_rows_ptr) { @@ -1351,13 +1456,13 @@ void addFoundRowAll( { for (auto it = mapped.begin(); it.ok(); ++it) { - added.appendFromBlock(*it->block, it->row_num); + added.appendFromBlock(*it->block, it->row_num, false); ++current_offset; } } } -template +template void addNotFoundRow(AddedColumns & added [[maybe_unused]], IColumn::Offset & current_offset [[maybe_unused]]) { if constexpr (add_missing) @@ -1377,7 +1482,7 @@ void setUsed(IColumn::Filter & filter [[maybe_unused]], size_t pos [[maybe_unuse /// Joins right table columns which indexes are present in right_indexes using specified map. /// Makes filter (1 if row presented in right table) and returns offsets to replicate (for ALL JOINS). -template +template NO_INLINE size_t joinRightColumns( std::vector && key_getter_vector, const std::vector & mapv, @@ -1440,7 +1545,7 @@ NO_INLINE size_t joinRightColumns( else used_flags.template setUsed(find_result); - added_columns.appendFromBlock(*row_ref.block, row_ref.row_num); + added_columns.appendFromBlock(*row_ref.block, row_ref.row_num, join_features.add_missing); } else addNotFoundRow(added_columns, current_offset); @@ -1471,7 +1576,7 @@ NO_INLINE size_t joinRightColumns( if (used_once) { setUsed(added_columns.filter, i); - added_columns.appendFromBlock(*mapped.block, mapped.row_num); + added_columns.appendFromBlock(*mapped.block, mapped.row_num, join_features.add_missing); } break; @@ -1489,7 +1594,7 @@ NO_INLINE size_t joinRightColumns( { setUsed(added_columns.filter, i); used_flags.template setUsed(find_result); - added_columns.appendFromBlock(*mapped.block, mapped.row_num); + added_columns.appendFromBlock(*mapped.block, mapped.row_num, join_features.add_missing); if (join_features.is_any_or_semi_join) { @@ -1516,7 +1621,7 @@ NO_INLINE size_t joinRightColumns( return i; } -template +template size_t joinRightColumnsSwitchMultipleDisjuncts( std::vector && key_getter_vector, const std::vector & mapv, @@ -1528,7 +1633,7 @@ size_t joinRightColumnsSwitchMultipleDisjuncts( : joinRightColumns(std::forward>(key_getter_vector), mapv, added_columns, used_flags); } -template +template size_t joinRightColumnsSwitchNullability( std::vector && key_getter_vector, const std::vector & mapv, @@ -1545,7 +1650,7 @@ size_t joinRightColumnsSwitchNullability( } } -template +template size_t switchJoinRightColumns( const std::vector & mapv, AddedColumns & added_columns, @@ -1680,14 +1785,9 @@ Block HashJoin::joinBlockImpl( * but they will not be used at this stage of joining (and will be in `AdderNonJoined`), and they need to be skipped. * For ASOF, the last column is used as the ASOF column */ - AddedColumns added_columns( - block, - block_with_columns_to_add, - savedBlockSample(), - *this, - std::move(join_on_keys), - join_features.is_asof_join, - is_join_get); + AddedColumns added_columns( + block, block_with_columns_to_add, savedBlockSample(), *this, std::move(join_on_keys), join_features.is_asof_join, is_join_get); + bool has_required_right_keys = (required_right_keys.columns() != 0); added_columns.need_filter = join_features.need_filter || has_required_right_keys; @@ -1702,6 +1802,7 @@ Block HashJoin::joinBlockImpl( added_columns.join_on_keys.clear(); Block remaining_block = sliceBlock(block, num_joined); + added_columns.buildOutput(); for (size_t i = 0; i < added_columns.size(); ++i) block.insert(added_columns.moveColumn(i)); diff --git a/src/Interpreters/InterpreterCreateQuery.cpp b/src/Interpreters/InterpreterCreateQuery.cpp index 0047ea1bc78..d05d8b8deb1 100644 --- a/src/Interpreters/InterpreterCreateQuery.cpp +++ b/src/Interpreters/InterpreterCreateQuery.cpp @@ -38,7 +38,6 @@ #include #include #include -#include #include #include @@ -812,24 +811,7 @@ InterpreterCreateQuery::TableProperties InterpreterCreateQuery::getTableProperti } else { - /** To get valid sample block we need to prepare query without only_analyze, because we need to execute scalar - * subqueries. Otherwise functions that expect only constant arguments will throw error during query analysis, - * because the result of scalar subquery is not a constant. - * - * Example: - * CREATE MATERIALIZED VIEW test_mv ENGINE=MergeTree ORDER BY arr - * AS - * WITH (SELECT '\d[a-z]') AS constant_value - * SELECT extractAll(concat(toString(number), 'a'), assumeNotNull(constant_value)) AS arr - * FROM test_table; - * - * For new analyzer this issue does not exists because we always execute scalar subqueries. - * We can improve this in new analyzer, and execute scalar subqueries only in contexts when we expect constant - * for example: LIMIT, OFFSET, functions parameters, functions constant only arguments. - */ - - InterpreterSelectWithUnionQuery interpreter(create.select->clone(), getContext(), SelectQueryOptions()); - as_select_sample = interpreter.getSampleBlock(); + as_select_sample = InterpreterSelectWithUnionQuery::getSampleBlock(create.select->clone(), getContext()); } properties.columns = ColumnsDescription(as_select_sample.getNamesAndTypesList()); @@ -894,24 +876,6 @@ void InterpreterCreateQuery::validateTableStructure(const ASTCreateQuery & creat throw Exception(ErrorCodes::DUPLICATE_COLUMN, "Column {} already exists", backQuoteIfNeed(column.name)); } - /// Check if _row_exists for lightweight delete column in column_lists for merge tree family. - if (create.storage && create.storage->engine && endsWith(create.storage->engine->name, "MergeTree")) - { - auto search = all_columns.find(LightweightDeleteDescription::FILTER_COLUMN.name); - if (search != all_columns.end()) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Cannot create table with column '{}' for *MergeTree engines because it " - "is reserved for lightweight delete feature", - LightweightDeleteDescription::FILTER_COLUMN.name); - - auto search_block_number = all_columns.find(BlockNumberColumn::name); - if (search_block_number != all_columns.end()) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Cannot create table with column '{}' for *MergeTree engines because it " - "is reserved for storing block number", - BlockNumberColumn::name); - } - const auto & settings = getContext()->getSettingsRef(); /// If it's not attach and not materialized view to existing table, @@ -924,9 +888,23 @@ void InterpreterCreateQuery::validateTableStructure(const ASTCreateQuery & creat } } +void validateVirtualColumns(const IStorage & storage) +{ + auto virtual_columns = storage.getVirtualsPtr(); + for (const auto & storage_column : storage.getInMemoryMetadataPtr()->getColumns()) + { + if (virtual_columns->tryGet(storage_column.name, VirtualsKind::Persistent)) + { + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Cannot create table with column '{}' for {} engines because it is reserved for persistent virtual column", + storage_column.name, storage.getName()); + } + } +} + namespace { - void checkTemporaryTableEngineName(const String& name) + void checkTemporaryTableEngineName(const String & name) { if (name.starts_with("Replicated") || name.starts_with("Shared") || name == "KeeperMap") throw Exception(ErrorCodes::INCORRECT_QUERY, "Temporary tables cannot be created with Replicated, Shared or KeeperMap table engines"); @@ -1246,7 +1224,7 @@ BlockIO InterpreterCreateQuery::createTable(ASTCreateQuery & create) { input_block = InterpreterSelectWithUnionQuery(create.select->clone(), getContext(), - {}).getSampleBlock(); + SelectQueryOptions().analyze()).getSampleBlock(); } Block output_block = to_table->getInMemoryMetadataPtr()->getSampleBlock(); @@ -1509,6 +1487,16 @@ bool InterpreterCreateQuery::doCreateTable(ASTCreateQuery & create, addColumnsDescriptionToCreateQueryIfNecessary(query_ptr->as(), res); } + validateVirtualColumns(*res); + + if (!res->supportsDynamicSubcolumns() && hasDynamicSubcolumns(res->getInMemoryMetadataPtr()->getColumns())) + { + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Cannot create table with column of type Object, " + "because storage {} doesn't support dynamic subcolumns", + res->getName()); + } + if (!create.attach && getContext()->getSettingsRef().database_replicated_allow_only_replicated_engine) { bool is_replicated_storage = typeid_cast(res.get()) != nullptr; @@ -1558,14 +1546,6 @@ bool InterpreterCreateQuery::doCreateTable(ASTCreateQuery & create, /// we can safely destroy the object without a call to "shutdown", because there is guarantee /// that no background threads/similar resources remain after exception from "startup". - if (!res->supportsDynamicSubcolumns() && hasDynamicSubcolumns(res->getInMemoryMetadataPtr()->getColumns())) - { - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Cannot create table with column of type Object, " - "because storage {} doesn't support dynamic subcolumns", - res->getName()); - } - res->startup(); return true; } diff --git a/src/Interpreters/InterpreterDeleteQuery.cpp b/src/Interpreters/InterpreterDeleteQuery.cpp index 5c13a1145d1..97ae9649ae8 100644 --- a/src/Interpreters/InterpreterDeleteQuery.cpp +++ b/src/Interpreters/InterpreterDeleteQuery.cpp @@ -15,7 +15,6 @@ #include #include #include -#include namespace DB diff --git a/src/Interpreters/InterpreterDescribeQuery.cpp b/src/Interpreters/InterpreterDescribeQuery.cpp index 1aab72afcc1..04d44e34fff 100644 --- a/src/Interpreters/InterpreterDescribeQuery.cpp +++ b/src/Interpreters/InterpreterDescribeQuery.cpp @@ -123,28 +123,29 @@ BlockIO InterpreterDescribeQuery::execute() void InterpreterDescribeQuery::fillColumnsFromSubquery(const ASTTableExpression & table_expression) { - NamesAndTypesList names_and_types; + Block sample_block; auto select_query = table_expression.subquery->children.at(0); auto current_context = getContext(); if (settings.allow_experimental_analyzer) { SelectQueryOptions select_query_options; - names_and_types = InterpreterSelectQueryAnalyzer(select_query, current_context, select_query_options).getSampleBlock().getNamesAndTypesList(); + sample_block = InterpreterSelectQueryAnalyzer(select_query, current_context, select_query_options).getSampleBlock(); } else { - names_and_types = InterpreterSelectWithUnionQuery::getSampleBlock(select_query, current_context).getNamesAndTypesList(); + sample_block = InterpreterSelectWithUnionQuery::getSampleBlock(select_query, current_context); } - for (auto && [name, type] : names_and_types) - columns.emplace_back(std::move(name), std::move(type)); + for (auto && column : sample_block) + columns.emplace_back(std::move(column.name), std::move(column.type)); } void InterpreterDescribeQuery::fillColumnsFromTableFunction(const ASTTableExpression & table_expression) { auto current_context = getContext(); TableFunctionPtr table_function_ptr = TableFunctionFactory::instance().get(table_expression.table_function, current_context); + auto column_descriptions = table_function_ptr->getActualTableStructure(getContext(), /*is_insert_query*/ true); for (const auto & column : column_descriptions) columns.emplace_back(column); @@ -154,14 +155,16 @@ void InterpreterDescribeQuery::fillColumnsFromTableFunction(const ASTTableExpres auto table = table_function_ptr->execute(table_expression.table_function, getContext(), table_function_ptr->getName()); if (table) { - for (const auto & column : table->getVirtuals()) + auto virtuals = table->getVirtualsPtr(); + for (const auto & column : *virtuals) { if (!column_descriptions.has(column.name)) - virtual_columns.emplace_back(column.name, column.type); + virtual_columns.push_back(column); } } } } + void InterpreterDescribeQuery::fillColumnsFromTable(const ASTTableExpression & table_expression) { auto table_id = getContext()->resolveStorageID(table_expression.database_and_table_name); @@ -176,10 +179,11 @@ void InterpreterDescribeQuery::fillColumnsFromTable(const ASTTableExpression & t if (settings.describe_include_virtual_columns) { - for (const auto & column : table->getVirtuals()) + auto virtuals = table->getVirtualsPtr(); + for (const auto & column : *virtuals) { if (!column_descriptions.has(column.name)) - virtual_columns.emplace_back(column.name, column.type); + virtual_columns.push_back(column); } } diff --git a/src/Interpreters/InterpreterInsertQuery.cpp b/src/Interpreters/InterpreterInsertQuery.cpp index e27a8bd414b..70f9e0c51da 100644 --- a/src/Interpreters/InterpreterInsertQuery.cpp +++ b/src/Interpreters/InterpreterInsertQuery.cpp @@ -136,7 +136,7 @@ Block InterpreterInsertQuery::getSampleBlock( if (auto * window_view = dynamic_cast(table.get())) return window_view->getInputHeader(); else if (no_destination) - return metadata_snapshot->getSampleBlockWithVirtuals(table->getVirtuals()); + return metadata_snapshot->getSampleBlockWithVirtuals(table->getVirtualsList()); else return metadata_snapshot->getSampleBlockNonMaterialized(); } @@ -513,6 +513,16 @@ BlockIO InterpreterInsertQuery::execute() const bool resize_to_max_insert_threads = !table->isView() && views.empty(); pre_streams_size = resize_to_max_insert_threads ? settings.max_insert_threads : std::min(settings.max_insert_threads, pipeline.getNumStreams()); + + /// Deduplication when passing insert_deduplication_token breaks if using more than one thread + if (!settings.insert_deduplication_token.toString().empty()) + { + LOG_DEBUG( + getLogger("InsertQuery"), + "Insert-select query using insert_deduplication_token, setting streams to 1 to avoid deduplication issues"); + pre_streams_size = 1; + } + if (table->supportsParallelInsert()) sink_streams_size = pre_streams_size; } diff --git a/src/Interpreters/InterpreterSelectQuery.cpp b/src/Interpreters/InterpreterSelectQuery.cpp index eaa256a16b0..bcedba7346d 100644 --- a/src/Interpreters/InterpreterSelectQuery.cpp +++ b/src/Interpreters/InterpreterSelectQuery.cpp @@ -2111,8 +2111,7 @@ void InterpreterSelectQuery::applyFiltersToPrewhereInAnalysis(ExpressionAnalysis if (does_storage_support_prewhere && shouldMoveToPrewhere()) { /// Execute row level filter in prewhere as a part of "move to prewhere" optimization. - analysis.prewhere_info - = std::make_shared(std::move(analysis.filter_info->actions), std::move(analysis.filter_info->column_name)); + analysis.prewhere_info = std::make_shared(analysis.filter_info->actions, analysis.filter_info->column_name); analysis.prewhere_info->prewhere_actions->projectInput(false); analysis.prewhere_info->remove_prewhere_column = analysis.filter_info->do_remove_column; analysis.prewhere_info->need_filter = true; @@ -2122,8 +2121,8 @@ void InterpreterSelectQuery::applyFiltersToPrewhereInAnalysis(ExpressionAnalysis else { /// Add row level security actions to prewhere. - analysis.prewhere_info->row_level_filter = std::move(analysis.filter_info->actions); - analysis.prewhere_info->row_level_column_name = std::move(analysis.filter_info->column_name); + analysis.prewhere_info->row_level_filter = analysis.filter_info->actions; + analysis.prewhere_info->row_level_column_name = analysis.filter_info->column_name; analysis.prewhere_info->row_level_filter->projectInput(false); analysis.filter_info = nullptr; } diff --git a/src/Interpreters/InterpreterSelectQueryAnalyzer.cpp b/src/Interpreters/InterpreterSelectQueryAnalyzer.cpp index 4897101d80b..922f4a99b4a 100644 --- a/src/Interpreters/InterpreterSelectQueryAnalyzer.cpp +++ b/src/Interpreters/InterpreterSelectQueryAnalyzer.cpp @@ -103,7 +103,7 @@ QueryTreeNodePtr buildQueryTreeAndRunPasses(const ASTPtr & query, auto query_tree = buildQueryTree(query, context); QueryTreePassManager query_tree_pass_manager(context); - addQueryTreePasses(query_tree_pass_manager); + addQueryTreePasses(query_tree_pass_manager, select_query_options.only_analyze); /// We should not apply any query tree level optimizations on shards /// because it can lead to a changed header. diff --git a/src/Interpreters/InterpreterShowColumnsQuery.cpp b/src/Interpreters/InterpreterShowColumnsQuery.cpp index 149ba6d7575..f32ebceaa63 100644 --- a/src/Interpreters/InterpreterShowColumnsQuery.cpp +++ b/src/Interpreters/InterpreterShowColumnsQuery.cpp @@ -107,7 +107,7 @@ SELECT '' AS extra )"; // TODO Interpret query.extended. It is supposed to show internal/virtual columns. Need to fetch virtual column names, see - // IStorage::getVirtuals(). We can't easily do that via SQL. + // IStorage::getVirtualsList(). We can't easily do that via SQL. if (query.full) { diff --git a/src/Interpreters/InterpreterSystemQuery.cpp b/src/Interpreters/InterpreterSystemQuery.cpp index 4bb47a8c9e3..026e0c166b4 100644 --- a/src/Interpreters/InterpreterSystemQuery.cpp +++ b/src/Interpreters/InterpreterSystemQuery.cpp @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -58,6 +59,7 @@ #include #include #include +#include #include #include #include @@ -332,10 +334,17 @@ BlockIO InterpreterSystemQuery::execute() { getContext()->checkAccess(AccessType::SYSTEM_DROP_DNS_CACHE); DNSResolver::instance().dropCache(); + HostResolversPool::instance().dropCache(); /// Reinitialize clusters to update their resolved_addresses system_context->reloadClusterConfig(); break; } + case Type::DROP_CONNECTIONS_CACHE: + { + getContext()->checkAccess(AccessType::SYSTEM_DROP_CONNECTIONS_CACHE); + HTTPConnectionPools::instance().dropCache(); + break; + } case Type::DROP_MARK_CACHE: getContext()->checkAccess(AccessType::SYSTEM_DROP_MARK_CACHE); system_context->clearMarkCache(); @@ -1200,6 +1209,7 @@ AccessRightsElements InterpreterSystemQuery::getRequiredAccessForDDLOnCluster() break; } case Type::DROP_DNS_CACHE: + case Type::DROP_CONNECTIONS_CACHE: case Type::DROP_MARK_CACHE: case Type::DROP_MMAP_CACHE: case Type::DROP_QUERY_CACHE: diff --git a/src/Interpreters/JoinedTables.cpp b/src/Interpreters/JoinedTables.cpp index 9be8bf178a1..49693332280 100644 --- a/src/Interpreters/JoinedTables.cpp +++ b/src/Interpreters/JoinedTables.cpp @@ -272,7 +272,7 @@ void JoinedTables::makeFakeTable(StoragePtr storage, const StorageMetadataPtr & auto & table = tables_with_columns.back(); table.addHiddenColumns(storage_columns.getMaterialized()); table.addHiddenColumns(storage_columns.getAliases()); - table.addHiddenColumns(storage->getVirtuals()); + table.addHiddenColumns(storage->getVirtualsList()); } else tables_with_columns.emplace_back(DatabaseAndTableWithAlias{}, source_header.getNamesAndTypesList()); diff --git a/src/Interpreters/MutationsInterpreter.cpp b/src/Interpreters/MutationsInterpreter.cpp index a3d1b84fdc1..3b1a499255b 100644 --- a/src/Interpreters/MutationsInterpreter.cpp +++ b/src/Interpreters/MutationsInterpreter.cpp @@ -7,7 +7,7 @@ #include #include #include -#include +#include #include #include #include @@ -31,7 +31,6 @@ #include #include #include -#include #include #include #include @@ -265,7 +264,7 @@ MutationCommand createCommandToApplyDeletedMask(const MutationCommand & command) alter_command->partition = alter_command->children.emplace_back(command.partition).get(); auto row_exists_predicate = makeASTFunction("equals", - std::make_shared(LightweightDeleteDescription::FILTER_COLUMN.name), + std::make_shared(RowExistsColumn::name), std::make_shared(Field(0))); if (command.predicate) @@ -350,7 +349,8 @@ bool MutationsInterpreter::Source::isCompactPart() const static Names getAvailableColumnsWithVirtuals(StorageMetadataPtr metadata_snapshot, const IStorage & storage) { auto all_columns = metadata_snapshot->getColumns().getNamesOfPhysical(); - for (const auto & column : storage.getVirtuals()) + auto virtuals = storage.getVirtualsPtr(); + for (const auto & column : *virtuals) all_columns.push_back(column.name); return all_columns; } @@ -435,60 +435,54 @@ static NameSet getKeyColumns(const MutationsInterpreter::Source & source, const static void validateUpdateColumns( const MutationsInterpreter::Source & source, - const StorageMetadataPtr & metadata_snapshot, const NameSet & updated_columns, - const std::unordered_map & column_to_affected_materialized) + const StorageMetadataPtr & metadata_snapshot, + const NameSet & updated_columns, + const std::unordered_map & column_to_affected_materialized, + const ContextPtr & context) { + auto storage_snapshot = source.getStorageSnapshot(metadata_snapshot, context); NameSet key_columns = getKeyColumns(source, metadata_snapshot); - for (const String & column_name : updated_columns) + const auto & storage_columns = storage_snapshot->metadata->getColumns(); + const auto & virtual_columns = *storage_snapshot->virtual_columns; + + for (const auto & column_name : updated_columns) { - auto found = false; - for (const auto & col : metadata_snapshot->getColumns().getOrdinary()) - { - if (col.name == column_name) - { - found = true; - break; - } - } - - /// Allow to override value of lightweight delete filter virtual column - if (!found && column_name == LightweightDeleteDescription::FILTER_COLUMN.name) - { - if (!source.supportsLightweightDelete()) - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Lightweight delete is not supported for table"); - found = true; - } - - /// Dont allow to override value of block number virtual column - if (!found && column_name == BlockNumberColumn::name) - { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Update is not supported for virtual column {} ", backQuote(column_name)); - } - - if (!found) - { - for (const auto & col : metadata_snapshot->getColumns().getMaterialized()) - { - if (col.name == column_name) - throw Exception(ErrorCodes::CANNOT_UPDATE_COLUMN, "Cannot UPDATE materialized column {}", backQuote(column_name)); - } - - throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "There is no column {} in table", backQuote(column_name)); - } - if (key_columns.contains(column_name)) throw Exception(ErrorCodes::CANNOT_UPDATE_COLUMN, "Cannot UPDATE key column {}", backQuote(column_name)); + if (storage_columns.tryGetColumn(GetColumnsOptions::Materialized, column_name)) + throw Exception(ErrorCodes::CANNOT_UPDATE_COLUMN, "Cannot UPDATE materialized column {}", backQuote(column_name)); + auto materialized_it = column_to_affected_materialized.find(column_name); if (materialized_it != column_to_affected_materialized.end()) { - for (const String & materialized : materialized_it->second) + for (const auto & materialized : materialized_it->second) { if (key_columns.contains(materialized)) + { throw Exception(ErrorCodes::CANNOT_UPDATE_COLUMN, "Updated column {} affects MATERIALIZED column {}, which is a key column. " "Cannot UPDATE it.", backQuote(column_name), backQuote(materialized)); + } + } + } + + if (!storage_columns.tryGetColumn(GetColumnsOptions::Ordinary, column_name)) + { + /// Allow to override value of lightweight delete filter virtual column + if (column_name == RowExistsColumn::name) + { + if (!source.supportsLightweightDelete()) + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Lightweight delete is not supported for table"); + } + else if (virtual_columns.tryGet(column_name)) + { + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Update is not supported for virtual column {} ", backQuote(column_name)); + } + else + { + throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "There is no column {} in table", backQuote(column_name)); } } } @@ -546,8 +540,8 @@ void MutationsInterpreter::prepare(bool dry_run) /// Add _row_exists column if it is physically present in the part if (source.hasLightweightDeleteMask()) { - all_columns.push_back({LightweightDeleteDescription::FILTER_COLUMN}); - available_columns_set.insert(LightweightDeleteDescription::FILTER_COLUMN.name); + all_columns.emplace_back(RowExistsColumn::name, RowExistsColumn::type); + available_columns_set.insert(RowExistsColumn::name); } NameSet updated_columns; @@ -563,9 +557,7 @@ void MutationsInterpreter::prepare(bool dry_run) for (const auto & [name, _] : command.column_to_update_expression) { - if (!available_columns_set.contains(name) - && name != LightweightDeleteDescription::FILTER_COLUMN.name - && name != BlockNumberColumn::name) + if (!available_columns_set.contains(name) && name != RowExistsColumn::name) throw Exception(ErrorCodes::THERE_IS_NO_COLUMN, "Column {} is updated but not requested to read", name); @@ -590,7 +582,7 @@ void MutationsInterpreter::prepare(bool dry_run) } } - validateUpdateColumns(source, metadata_snapshot, updated_columns, column_to_affected_materialized); + validateUpdateColumns(source, metadata_snapshot, updated_columns, column_to_affected_materialized, context); } StorageInMemoryMetadata::HasDependencyCallback has_dependency = @@ -666,15 +658,11 @@ void MutationsInterpreter::prepare(bool dry_run) { type = physical_column->type; } - else if (column_name == LightweightDeleteDescription::FILTER_COLUMN.name) + else if (column_name == RowExistsColumn::name) { - type = LightweightDeleteDescription::FILTER_COLUMN.type; + type = RowExistsColumn::type; deleted_mask_updated = true; } - else if (column_name == BlockNumberColumn::name) - { - type = BlockNumberColumn::type; - } else { throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown column {}", column_name); @@ -1028,7 +1016,7 @@ void MutationsInterpreter::prepareMutationStages(std::vector & prepared_s /// Add _row_exists column if it is present in the part if (source.hasLightweightDeleteMask() || deleted_mask_updated) - all_columns.push_back(LightweightDeleteDescription::FILTER_COLUMN); + all_columns.emplace_back(RowExistsColumn::name, RowExistsColumn::type); bool has_filters = false; /// Next, for each stage calculate columns changed by this and previous stages. @@ -1038,7 +1026,7 @@ void MutationsInterpreter::prepareMutationStages(std::vector & prepared_s { for (const auto & column : all_columns) { - if (column.name == LightweightDeleteDescription::FILTER_COLUMN.name && !deleted_mask_updated) + if (column.name == RowExistsColumn::name && !deleted_mask_updated) continue; prepared_stages[i].output_columns.insert(column.name); @@ -1057,7 +1045,7 @@ void MutationsInterpreter::prepareMutationStages(std::vector & prepared_s /// and so it is not in the list of AllPhysical columns. for (const auto & [column_name, _] : prepared_stages[i].column_to_updated) { - if (column_name == LightweightDeleteDescription::FILTER_COLUMN.name && has_filters && !deleted_mask_updated) + if (column_name == RowExistsColumn::name && has_filters && !deleted_mask_updated) continue; prepared_stages[i].output_columns.insert(column_name); @@ -1148,93 +1136,6 @@ void MutationsInterpreter::prepareMutationStages(std::vector & prepared_s } } -/// This structure re-implements adding virtual columns while reading from MergeTree part. -/// It would be good to unify it with IMergeTreeSelectAlgorithm. -struct VirtualColumns -{ - struct ColumnAndPosition - { - ColumnWithTypeAndName column; - size_t position; - }; - - using Columns = std::vector; - - Columns virtuals; - Names columns_to_read; - - VirtualColumns(Names required_columns, const MergeTreeData::DataPartPtr & part) : columns_to_read(std::move(required_columns)) - { - for (size_t i = 0; i < columns_to_read.size(); ++i) - { - if (columns_to_read[i] == LightweightDeleteDescription::FILTER_COLUMN.name) - { - if (!part->getColumns().contains(LightweightDeleteDescription::FILTER_COLUMN.name)) - { - ColumnWithTypeAndName mask_column; - mask_column.type = LightweightDeleteDescription::FILTER_COLUMN.type; - mask_column.column = mask_column.type->createColumnConst(0, 1); - mask_column.name = std::move(columns_to_read[i]); - - virtuals.emplace_back(ColumnAndPosition{.column = std::move(mask_column), .position = i}); - } - } - else if (columns_to_read[i] == "_partition_id") - { - ColumnWithTypeAndName column; - column.type = std::make_shared(); - column.column = column.type->createColumnConst(0, part->info.partition_id); - column.name = std::move(columns_to_read[i]); - - virtuals.emplace_back(ColumnAndPosition{.column = std::move(column), .position = i}); - } - else if (columns_to_read[i] == BlockNumberColumn::name) - { - if (!part->getColumns().contains(BlockNumberColumn::name)) - { - ColumnWithTypeAndName block_number_column; - block_number_column.type = BlockNumberColumn::type; - block_number_column.column = block_number_column.type->createColumnConst(0, part->info.min_block); - block_number_column.name = std::move(columns_to_read[i]); - - virtuals.emplace_back(ColumnAndPosition{.column = std::move(block_number_column), .position = i}); - } - } - } - - if (!virtuals.empty()) - { - Names columns_no_virtuals; - columns_no_virtuals.reserve(columns_to_read.size()); - size_t next_virtual = 0; - for (size_t i = 0; i < columns_to_read.size(); ++i) - { - if (next_virtual < virtuals.size() && i == virtuals[next_virtual].position) - ++next_virtual; - else - columns_no_virtuals.emplace_back(std::move(columns_to_read[i])); - } - - columns_to_read.swap(columns_no_virtuals); - } - } - - void addVirtuals(QueryPlan & plan) - { - auto dag = std::make_unique(plan.getCurrentDataStream().header.getColumnsWithTypeAndName()); - - for (auto & column : virtuals) - { - const auto & adding_const = dag->addColumn(std::move(column.column)); - auto & outputs = dag->getOutputs(); - outputs.insert(outputs.begin() + column.position, &adding_const); - } - - auto step = std::make_unique(plan.getCurrentDataStream(), std::move(dag)); - plan.addStep(std::move(step)); - } -}; - void MutationsInterpreter::Source::read( Stage & first_stage, QueryPlan & plan, @@ -1277,16 +1178,12 @@ void MutationsInterpreter::Source::read( filter = ActionsDAG::buildFilterActionsDAG(nodes); } - VirtualColumns virtual_columns(std::move(required_columns), part); - createReadFromPartStep( MergeTreeSequentialSourceType::Mutation, - plan, *data, storage_snapshot, part, - std::move(virtual_columns.columns_to_read), + plan, *data, storage_snapshot, + part, required_columns, apply_deleted_mask_, filter, context_, getLogger("MutationsInterpreter")); - - virtual_columns.addVirtuals(plan); } else { diff --git a/src/Interpreters/OpenTelemetrySpanLog.cpp b/src/Interpreters/OpenTelemetrySpanLog.cpp index fffc1e50da0..aa11749f8a6 100644 --- a/src/Interpreters/OpenTelemetrySpanLog.cpp +++ b/src/Interpreters/OpenTelemetrySpanLog.cpp @@ -32,12 +32,17 @@ ColumnsDescription OpenTelemetrySpanLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", low_cardinality_string}, - {"trace_id", std::make_shared()}, - {"span_id", std::make_shared()}, - {"parent_span_id", std::make_shared()}, - {"operation_name", low_cardinality_string}, - {"kind", std::move(span_kind_type)}, + {"hostname", low_cardinality_string, "The hostname where this span was captured."}, + {"trace_id", std::make_shared(), "ID of the trace for executed query."}, + {"span_id", std::make_shared(), "ID of the trace span."}, + {"parent_span_id", std::make_shared(), "ID of the parent trace span."}, + {"operation_name", low_cardinality_string, "The name of the operation."}, + {"kind", std::move(span_kind_type), "The SpanKind of the span. " + "INTERNAL — Indicates that the span represents an internal operation within an application. " + "SERVER — Indicates that the span covers server-side handling of a synchronous RPC or other remote request. " + "CLIENT — Indicates that the span describes a request to some remote service. " + "PRODUCER — Indicates that the span describes the initiators of an asynchronous request. This parent span will often end before the corresponding child CONSUMER span, possibly even before the child span starts. " + "CONSUMER - Indicates that the span describes a child of an asynchronous PRODUCER request."}, // DateTime64 is really unwieldy -- there is no "normal" way to convert // it to an UInt64 count of microseconds, except: // 1) reinterpretAsUInt64(reinterpretAsFixedString(date)), which just @@ -48,10 +53,10 @@ ColumnsDescription OpenTelemetrySpanLogElement::getColumnsDescription() // Also subtraction of two DateTime64 points doesn't work, so you can't // get duration. // It is much less hassle to just use UInt64 of microseconds. - {"start_time_us", std::make_shared()}, - {"finish_time_us", std::make_shared()}, - {"finish_date", std::make_shared()}, - {"attribute", std::make_shared(low_cardinality_string, std::make_shared())}, + {"start_time_us", std::make_shared(), "The start time of the trace span (in microseconds)."}, + {"finish_time_us", std::make_shared(), "The finish time of the trace span (in microseconds)."}, + {"finish_date", std::make_shared(), "The finish date of the trace span."}, + {"attribute", std::make_shared(low_cardinality_string, std::make_shared()), "Attribute depending on the trace span. They are filled in according to the recommendations in the OpenTelemetry standard."}, }; } diff --git a/src/Interpreters/OptimizeIfWithConstantConditionVisitor.cpp b/src/Interpreters/OptimizeIfWithConstantConditionVisitor.cpp index 13b6311a877..f3504f3f403 100644 --- a/src/Interpreters/OptimizeIfWithConstantConditionVisitor.cpp +++ b/src/Interpreters/OptimizeIfWithConstantConditionVisitor.cpp @@ -53,7 +53,7 @@ static bool tryExtractConstValueFromCondition(const ASTPtr & condition, bool & v } } } - else if (function->name == "toUInt8" || function->name == "toInt8" || function->name == "identity") + else if (function->name == "toUInt8" || function->name == "toInt8" || function->name == "identity" || function->name == "__scalarSubqueryResult") { if (const auto * expr_list = function->arguments->as()) { diff --git a/src/Interpreters/PartLog.cpp b/src/Interpreters/PartLog.cpp index a7f20a06785..66f933f1afa 100644 --- a/src/Interpreters/PartLog.cpp +++ b/src/Interpreters/PartLog.cpp @@ -123,7 +123,7 @@ ColumnsDescription PartLogElement::getColumnsDescription() {"table_uuid", std::make_shared(), "UUID of the table the data part belongs to."}, {"part_name", std::make_shared(), "Name of the data part."}, {"partition_id", std::make_shared(), "ID of the partition that the data part was inserted to. The column takes the `all` value if the partitioning is by `tuple()`."}, - {"partition", std::make_shared()}, + {"partition", std::make_shared(), "The partition name."}, {"part_type", std::make_shared(), "The type of the part. Possible values: Wide and Compact."}, {"disk_name", std::make_shared(), "The disk name data part lies on."}, {"path_on_disk", std::make_shared(), "Absolute path to the folder with data part files."}, diff --git a/src/Interpreters/ProcessorsProfileLog.cpp b/src/Interpreters/ProcessorsProfileLog.cpp index 088d193257c..015b4abc712 100644 --- a/src/Interpreters/ProcessorsProfileLog.cpp +++ b/src/Interpreters/ProcessorsProfileLog.cpp @@ -21,26 +21,26 @@ ColumnsDescription ProcessorProfileLogElement::getColumnsDescription() { return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, + {"hostname", std::make_shared(std::make_shared()), "Hostname of the server executing the query."}, + {"event_date", std::make_shared(), "The date when the event happened."}, + {"event_time", std::make_shared(), "The date and time when the event happened."}, + {"event_time_microseconds", std::make_shared(6), "The date and time with microseconds precision when the event happened."}, - {"id", std::make_shared()}, - {"parent_ids", std::make_shared(std::make_shared())}, - {"plan_step", std::make_shared()}, - {"plan_group", std::make_shared()}, + {"id", std::make_shared(), "ID of processor."}, + {"parent_ids", std::make_shared(std::make_shared()), "Parent processors IDs."}, + {"plan_step", std::make_shared(), "ID of the query plan step which created this processor. The value is zero if the processor was not added from any step."}, + {"plan_group", std::make_shared(), "Group of the processor if it was created by query plan step. A group is a logical partitioning of processors added from the same query plan step. Group is used only for beautifying the result of EXPLAIN PIPELINE result."}, - {"initial_query_id", std::make_shared()}, - {"query_id", std::make_shared()}, - {"name", std::make_shared(std::make_shared())}, - {"elapsed_us", std::make_shared()}, - {"input_wait_elapsed_us", std::make_shared()}, - {"output_wait_elapsed_us", std::make_shared()}, - {"input_rows", std::make_shared()}, - {"input_bytes", std::make_shared()}, - {"output_rows", std::make_shared()}, - {"output_bytes", std::make_shared()}, + {"initial_query_id", std::make_shared(), "ID of the initial query (for distributed query execution)."}, + {"query_id", std::make_shared(), "ID of the query."}, + {"name", std::make_shared(std::make_shared()), "Name of the processor."}, + {"elapsed_us", std::make_shared(), "Number of microseconds this processor was executed."}, + {"input_wait_elapsed_us", std::make_shared(), "Number of microseconds this processor was waiting for data (from other processor)."}, + {"output_wait_elapsed_us", std::make_shared(), "Number of microseconds this processor was waiting because output port was full."}, + {"input_rows", std::make_shared(), "The number of rows consumed by processor."}, + {"input_bytes", std::make_shared(), "The number of bytes consumed by processor."}, + {"output_rows", std::make_shared(), "The number of rows generated by processor."}, + {"output_bytes", std::make_shared(), "The number of bytes generated by processor."}, }; } diff --git a/src/Interpreters/QueryLog.cpp b/src/Interpreters/QueryLog.cpp index ad6e344655b..92f8ddae141 100644 --- a/src/Interpreters/QueryLog.cpp +++ b/src/Interpreters/QueryLog.cpp @@ -134,13 +134,13 @@ ColumnsDescription QueryLogElement::getColumnsDescription() {"used_storages", array_low_cardinality_string, "Canonical names of storages, which were used during query execution."}, {"used_table_functions", array_low_cardinality_string, "Canonical names of table functions, which were used during query execution."}, - {"used_row_policies", array_low_cardinality_string}, + {"used_row_policies", array_low_cardinality_string, "The list of row policies names that were used during query execution."}, - {"transaction_id", getTransactionIDDataType()}, + {"transaction_id", getTransactionIDDataType(), "The identifier of the transaction in scope of which this query was executed."}, {"query_cache_usage", std::move(query_cache_usage_datatype), "Usage of the query cache during query execution. Values: 'Unknown' = Status unknown, 'None' = The query result was neither written into nor read from the query cache, 'Write' = The query result was written into the query cache, 'Read' = The query result was read from the query cache."}, - {"asynchronous_read_counters", std::make_shared(low_cardinality_string, std::make_shared())}, + {"asynchronous_read_counters", std::make_shared(low_cardinality_string, std::make_shared()), "Metrics for asynchronous reading."}, }; } diff --git a/src/Interpreters/QueryThreadLog.cpp b/src/Interpreters/QueryThreadLog.cpp index d153e30a4ce..f50458745b9 100644 --- a/src/Interpreters/QueryThreadLog.cpp +++ b/src/Interpreters/QueryThreadLog.cpp @@ -28,58 +28,58 @@ ColumnsDescription QueryThreadLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", low_cardinality_string}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, - {"query_start_time", std::make_shared()}, - {"query_start_time_microseconds", std::make_shared(6)}, - {"query_duration_ms", std::make_shared()}, + {"hostname", low_cardinality_string, "Hostname of the server executing the query."}, + {"event_date", std::make_shared(), "The date when the thread has finished execution of the query."}, + {"event_time", std::make_shared(), "The date and time when the thread has finished execution of the query."}, + {"event_time_microseconds", std::make_shared(6), "The date and time when the thread has finished execution of the query with microseconds precision."}, + {"query_start_time", std::make_shared(), "Start time of query execution."}, + {"query_start_time_microseconds", std::make_shared(6), "Start time of query execution with microsecond precision."}, + {"query_duration_ms", std::make_shared(), "Duration of query execution."}, - {"read_rows", std::make_shared()}, - {"read_bytes", std::make_shared()}, - {"written_rows", std::make_shared()}, - {"written_bytes", std::make_shared()}, - {"memory_usage", std::make_shared()}, - {"peak_memory_usage", std::make_shared()}, + {"read_rows", std::make_shared(), "Number of read rows."}, + {"read_bytes", std::make_shared(), "Number of read bytes."}, + {"written_rows", std::make_shared(), "For INSERT queries, the number of written rows. For other queries, the column value is 0."}, + {"written_bytes", std::make_shared(), "For INSERT queries, the number of written bytes. For other queries, the column value is 0."}, + {"memory_usage", std::make_shared(), "The difference between the amount of allocated and freed memory in context of this thread."}, + {"peak_memory_usage", std::make_shared(), "The maximum difference between the amount of allocated and freed memory in context of this thread."}, - {"thread_name", low_cardinality_string}, - {"thread_id", std::make_shared()}, - {"master_thread_id", std::make_shared()}, - {"current_database", low_cardinality_string}, - {"query", std::make_shared()}, - {"normalized_query_hash", std::make_shared()}, + {"thread_name", low_cardinality_string, "Name of the thread."}, + {"thread_id", std::make_shared(), "Internal thread ID."}, + {"master_thread_id", std::make_shared(), "OS initial ID of initial thread."}, + {"current_database", low_cardinality_string, "Name of the current database."}, + {"query", std::make_shared(), "Query string."}, + {"normalized_query_hash", std::make_shared(), "The hash of normalized query - with wiped constanstans, etc."}, - {"is_initial_query", std::make_shared()}, - {"user", low_cardinality_string}, - {"query_id", std::make_shared()}, - {"address", DataTypeFactory::instance().get("IPv6")}, - {"port", std::make_shared()}, - {"initial_user", low_cardinality_string}, - {"initial_query_id", std::make_shared()}, - {"initial_address", DataTypeFactory::instance().get("IPv6")}, - {"initial_port", std::make_shared()}, - {"initial_query_start_time", std::make_shared()}, - {"initial_query_start_time_microseconds", std::make_shared(6)}, - {"interface", std::make_shared()}, - {"is_secure", std::make_shared()}, - {"os_user", low_cardinality_string}, - {"client_hostname", low_cardinality_string}, - {"client_name", low_cardinality_string}, - {"client_revision", std::make_shared()}, - {"client_version_major", std::make_shared()}, - {"client_version_minor", std::make_shared()}, - {"client_version_patch", std::make_shared()}, - {"http_method", std::make_shared()}, - {"http_user_agent", low_cardinality_string}, - {"http_referer", std::make_shared()}, - {"forwarded_for", std::make_shared()}, - {"quota_key", std::make_shared()}, - {"distributed_depth", std::make_shared()}, + {"is_initial_query", std::make_shared(), "Query type. Possible values: 1 — Query was initiated by the client, 0 — Query was initiated by another query for distributed query execution."}, + {"user", low_cardinality_string, "Name of the user who initiated the current query."}, + {"query_id", std::make_shared(), "ID of the query."}, + {"address", DataTypeFactory::instance().get("IPv6"), "IP address that was used to make the query."}, + {"port", std::make_shared(), "The client port that was used to make the query."}, + {"initial_user", low_cardinality_string, "Name of the user who ran the initial query (for distributed query execution)."}, + {"initial_query_id", std::make_shared(), "ID of the initial query (for distributed query execution)."}, + {"initial_address", DataTypeFactory::instance().get("IPv6"), "IP address that the parent query was launched from."}, + {"initial_port", std::make_shared(), "The client port that was used to make the parent query."}, + {"initial_query_start_time", std::make_shared(), "Start time of the initial query execution."}, + {"initial_query_start_time_microseconds", std::make_shared(6), "Start time of the initial query execution "}, + {"interface", std::make_shared(), "Interface that the query was initiated from. Possible values: 1 — TCP, 2 — HTTP."}, + {"is_secure", std::make_shared(), "The flag which shows whether the connection was secure."}, + {"os_user", low_cardinality_string, "OSs username who runs clickhouse-client."}, + {"client_hostname", low_cardinality_string, "Hostname of the client machine where the clickhouse-client or another TCP client is run."}, + {"client_name", low_cardinality_string, "The clickhouse-client or another TCP client name."}, + {"client_revision", std::make_shared(), "Revision of the clickhouse-client or another TCP client."}, + {"client_version_major", std::make_shared(), "Major version of the clickhouse-client or another TCP client."}, + {"client_version_minor", std::make_shared(), "Minor version of the clickhouse-client or another TCP client."}, + {"client_version_patch", std::make_shared(), "Patch component of the clickhouse-client or another TCP client version."}, + {"http_method", std::make_shared(), "HTTP method that initiated the query. Possible values: 0 — The query was launched from the TCP interface, 1 — GET method was used., 2 — POST method was used."}, + {"http_user_agent", low_cardinality_string, "The UserAgent header passed in the HTTP request."}, + {"http_referer", std::make_shared(), "HTTP header `Referer` passed in the HTTP query (contains an absolute or partial address of the page making the query)."}, + {"forwarded_for", std::make_shared(), "HTTP header `X-Forwarded-For` passed in the HTTP query."}, + {"quota_key", std::make_shared(), "The 'quota key' specified in the quotas setting."}, + {"distributed_depth", std::make_shared(), "How many times a query was forwarded between servers."}, - {"revision", std::make_shared()}, + {"revision", std::make_shared(), "ClickHouse revision."}, - {"ProfileEvents", std::make_shared(low_cardinality_string, std::make_shared())}, + {"ProfileEvents", std::make_shared(low_cardinality_string, std::make_shared()), "ProfileEvents that measure different metrics for this thread. The description of them could be found in the table system.events."}, }; } diff --git a/src/Interpreters/QueryViewsLog.cpp b/src/Interpreters/QueryViewsLog.cpp index c426f2d3cf0..a5441363340 100644 --- a/src/Interpreters/QueryViewsLog.cpp +++ b/src/Interpreters/QueryViewsLog.cpp @@ -35,30 +35,34 @@ ColumnsDescription QueryViewsLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, - {"view_duration_ms", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "Hostname of the server executing the query."}, + {"event_date", std::make_shared(), "The date when the last event of the view happened."}, + {"event_time", std::make_shared(), "The date and time when the view finished execution."}, + {"event_time_microseconds", std::make_shared(6), "The date and time when the view finished execution with microseconds precision."}, + {"view_duration_ms", std::make_shared(), "Duration of view execution (sum of its stages) in milliseconds."}, - {"initial_query_id", std::make_shared()}, - {"view_name", std::make_shared()}, - {"view_uuid", std::make_shared()}, - {"view_type", std::move(view_type_datatype)}, - {"view_query", std::make_shared()}, - {"view_target", std::make_shared()}, + {"initial_query_id", std::make_shared(), "ID of the initial query (for distributed query execution)."}, + {"view_name", std::make_shared(), "Name of the view."}, + {"view_uuid", std::make_shared(), "UUID of the view."}, + {"view_type", std::move(view_type_datatype), "Type of the view. Values: 'Default' = 1 — Default views. Should not appear in this log, 'Materialized' = 2 — Materialized views, 'Live' = 3 — Live views."}, + {"view_query", std::make_shared(), "The query executed by the view."}, + {"view_target", std::make_shared(), "The name of the view target table."}, - {"read_rows", std::make_shared()}, - {"read_bytes", std::make_shared()}, - {"written_rows", std::make_shared()}, - {"written_bytes", std::make_shared()}, - {"peak_memory_usage", std::make_shared()}, - {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared())}, + {"read_rows", std::make_shared(), "Number of read rows."}, + {"read_bytes", std::make_shared(), "Number of read bytes."}, + {"written_rows", std::make_shared(), "Number of written rows."}, + {"written_bytes", std::make_shared(), "Number of written bytes."}, + {"peak_memory_usage", std::make_shared(), "The maximum difference between the amount of allocated and freed memory in context of this view."}, + {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared()), "ProfileEvents that measure different metrics. The description of them could be found in the table system.events."}, - {"status", std::move(view_status_datatype)}, - {"exception_code", std::make_shared()}, - {"exception", std::make_shared()}, - {"stack_trace", std::make_shared()} + {"status", std::move(view_status_datatype), "Status of the view. Values: " + "'QueryStart' = 1 — Successful start the view execution. Should not appear, " + "'QueryFinish' = 2 — Successful end of the view execution, " + "'ExceptionBeforeStart' = 3 — Exception before the start of the view execution., " + "'ExceptionWhileProcessing' = 4 — Exception during the view execution."}, + {"exception_code", std::make_shared(), "Code of an exception."}, + {"exception", std::make_shared(), "Exception message."}, + {"stack_trace", std::make_shared(), "Stack trace. An empty string, if the query was completed successfully."} }; } diff --git a/src/Interpreters/S3QueueLog.cpp b/src/Interpreters/S3QueueLog.cpp index 3ed58de0f87..ba990a8ac25 100644 --- a/src/Interpreters/S3QueueLog.cpp +++ b/src/Interpreters/S3QueueLog.cpp @@ -25,19 +25,19 @@ ColumnsDescription S3QueueLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"database", std::make_shared()}, - {"table", std::make_shared()}, - {"uuid", std::make_shared()}, - {"file_name", std::make_shared()}, - {"rows_processed", std::make_shared()}, - {"status", status_datatype}, - {"processing_start_time", std::make_shared(std::make_shared())}, - {"processing_end_time", std::make_shared(std::make_shared())}, - {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared())}, - {"exception", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "Hostname"}, + {"event_date", std::make_shared(), "Event date of writing this log row"}, + {"event_time", std::make_shared(), "Event time of writing this log row"}, + {"database", std::make_shared(), "The name of a database where current S3Queue table lives."}, + {"table", std::make_shared(), "The name of S3Queue table."}, + {"uuid", std::make_shared(), "The UUID of S3Queue table"}, + {"file_name", std::make_shared(), "File name of the processing file"}, + {"rows_processed", std::make_shared(), "Number of processed rows"}, + {"status", status_datatype, "Status of the processing file"}, + {"processing_start_time", std::make_shared(std::make_shared()), "Time of the start of processing the file"}, + {"processing_end_time", std::make_shared(std::make_shared()), "Time of the end of processing the file"}, + {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared()), "Profile events collected while loading this file"}, + {"exception", std::make_shared(), "Exception message if happened"}, }; } diff --git a/src/Interpreters/SessionLog.cpp b/src/Interpreters/SessionLog.cpp index a5bc5012292..dc0ac963d0b 100644 --- a/src/Interpreters/SessionLog.cpp +++ b/src/Interpreters/SessionLog.cpp @@ -121,33 +121,36 @@ ColumnsDescription SessionLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", lc_string_datatype}, - {"type", std::move(event_type)}, - {"auth_id", std::make_shared()}, - {"session_id", std::make_shared()}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, + {"hostname", lc_string_datatype, "Hostname of the server executing the query."}, + {"type", std::move(event_type), "Login/logout result. Possible values: " + "LoginFailure — Login error. " + "LoginSuccess — Successful login. " + "Logout — Logout from the system."}, + {"auth_id", std::make_shared(), "Authentication ID, which is a UUID that is automatically generated each time user logins."}, + {"session_id", std::make_shared(), "Session ID that is passed by client via HTTP interface."}, + {"event_date", std::make_shared(), "Login/logout date."}, + {"event_time", std::make_shared(), "Login/logout time."}, + {"event_time_microseconds", std::make_shared(6), "Login/logout starting time with microseconds precision."}, - {"user", std::make_shared(std::make_shared())}, - {"auth_type", std::make_shared(std::move(identified_with_column))}, + {"user", std::make_shared(std::make_shared()), "User name."}, + {"auth_type", std::make_shared(std::move(identified_with_column)), "The authentication type."}, - {"profiles", std::make_shared(lc_string_datatype)}, - {"roles", std::make_shared(lc_string_datatype)}, - {"settings", std::move(settings_type_column)}, + {"profiles", std::make_shared(lc_string_datatype), "The list of profiles set for all roles and/or users."}, + {"roles", std::make_shared(lc_string_datatype), "The list of roles to which the profile is applied."}, + {"settings", std::move(settings_type_column), "Settings that were changed when the client logged in/out."}, - {"client_address", DataTypeFactory::instance().get("IPv6")}, - {"client_port", std::make_shared()}, - {"interface", std::move(interface_type_column)}, + {"client_address", DataTypeFactory::instance().get("IPv6"), "The IP address that was used to log in/out."}, + {"client_port", std::make_shared(), "The client port that was used to log in/out."}, + {"interface", std::move(interface_type_column), "The interface from which the login was initiated."}, - {"client_hostname", std::make_shared()}, - {"client_name", std::make_shared()}, - {"client_revision", std::make_shared()}, - {"client_version_major", std::make_shared()}, - {"client_version_minor", std::make_shared()}, - {"client_version_patch", std::make_shared()}, + {"client_hostname", std::make_shared(), "The hostname of the client machine where the clickhouse-client or another TCP client is run."}, + {"client_name", std::make_shared(), "The clickhouse-client or another TCP client name."}, + {"client_revision", std::make_shared(), "Revision of the clickhouse-client or another TCP client."}, + {"client_version_major", std::make_shared(), "The major version of the clickhouse-client or another TCP client."}, + {"client_version_minor", std::make_shared(), "The minor version of the clickhouse-client or another TCP client."}, + {"client_version_patch", std::make_shared(), "Patch component of the clickhouse-client or another TCP client version."}, - {"failure_reason", std::make_shared()}, + {"failure_reason", std::make_shared(), "The exception message containing the reason for the login/logout failure."}, }; } diff --git a/src/Interpreters/TraceLog.cpp b/src/Interpreters/TraceLog.cpp index 26adb0cfc3f..01bedf34f15 100644 --- a/src/Interpreters/TraceLog.cpp +++ b/src/Interpreters/TraceLog.cpp @@ -29,20 +29,27 @@ ColumnsDescription TraceLogElement::getColumnsDescription() { return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared()}, - {"event_time_microseconds", std::make_shared(6)}, - {"timestamp_ns", std::make_shared()}, - {"revision", std::make_shared()}, - {"trace_type", std::make_shared(trace_values)}, - {"thread_id", std::make_shared()}, - {"query_id", std::make_shared()}, - {"trace", std::make_shared(std::make_shared())}, - {"size", std::make_shared()}, - {"ptr", std::make_shared()}, - {"event", std::make_shared(std::make_shared())}, - {"increment", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "Hostname of the server executing the query."}, + {"event_date", std::make_shared(), "Date of sampling moment."}, + {"event_time", std::make_shared(), "Timestamp of the sampling moment."}, + {"event_time_microseconds", std::make_shared(6), "Timestamp of the sampling moment with microseconds precision."}, + {"timestamp_ns", std::make_shared(), "Timestamp of the sampling moment in nanoseconds."}, + {"revision", std::make_shared(), "ClickHouse server build revision."}, + {"trace_type", std::make_shared(trace_values), "Trace type: " + "`Real` represents collecting stack traces by wall-clock time. " + "`CPU` represents collecting stack traces by CPU time. " + "`Memory` represents collecting allocations and deallocations when memory allocation exceeds the subsequent watermark. " + "`MemorySample` represents collecting random allocations and deallocations. " + "`MemoryPeak` represents collecting updates of peak memory usage. " + "`ProfileEvent` represents collecting of increments of profile events." + }, + {"thread_id", std::make_shared(), "Thread identifier."}, + {"query_id", std::make_shared(), "Query identifier that can be used to get details about a query that was running from the query_log system table."}, + {"trace", std::make_shared(std::make_shared()), "Stack trace at the moment of sampling. Each element is a virtual memory address inside ClickHouse server process."}, + {"size", std::make_shared(), "For trace types Memory, MemorySample or MemoryPeak is the amount of memory allocated, for other trace types is 0."}, + {"ptr", std::make_shared(), "The address of the allocated chunk."}, + {"event", std::make_shared(std::make_shared()), "For trace type ProfileEvent is the name of updated profile event, for other trace types is an empty string."}, + {"increment", std::make_shared(), "For trace type ProfileEvent is the amount of increment of profile event, for other trace types is 0."}, }; } diff --git a/src/Interpreters/TransactionsInfoLog.cpp b/src/Interpreters/TransactionsInfoLog.cpp index 4a413439671..d13b31518d2 100644 --- a/src/Interpreters/TransactionsInfoLog.cpp +++ b/src/Interpreters/TransactionsInfoLog.cpp @@ -34,22 +34,22 @@ ColumnsDescription TransactionsInfoLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"type", std::move(type_enum)}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared(6)}, - {"thread_id", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "The hostname where transaction was executed."}, + {"type", std::move(type_enum), "The type of the transaction. Possible values: Begin, Commit, Rollback, AddPart, LockPart, UnlockPart."}, + {"event_date", std::make_shared(), "Date of the entry."}, + {"event_time", std::make_shared(6), "Time of the entry"}, + {"thread_id", std::make_shared(), "The identifier of a thread."}, /// which thread? - {"query_id", std::make_shared()}, - {"tid", getTransactionIDDataType()}, - {"tid_hash", std::make_shared()}, + {"query_id", std::make_shared(), "The ID of a query executed in a scope of transaction."}, + {"tid", getTransactionIDDataType(), "The identifier of a transaction."}, + {"tid_hash", std::make_shared(), "The hash of the identifier."}, - {"csn", std::make_shared()}, + {"csn", std::make_shared(), "The Commit Sequence Number"}, - {"database", std::make_shared()}, - {"table", std::make_shared()}, - {"uuid", std::make_shared()}, - {"part", std::make_shared()}, + {"database", std::make_shared(), "The name of the database the transaction was executed against."}, + {"table", std::make_shared(), "The name of the table the transaction was executed against."}, + {"uuid", std::make_shared(), "The uuid of the table the transaction was executed against."}, + {"part", std::make_shared(), "The name of the part participated in the transaction."}, // ? }; } diff --git a/src/Interpreters/TreeRewriter.cpp b/src/Interpreters/TreeRewriter.cpp index 960fad5dec6..5588fc55a64 100644 --- a/src/Interpreters/TreeRewriter.cpp +++ b/src/Interpreters/TreeRewriter.cpp @@ -56,6 +56,7 @@ #include #include #include +#include #include #include @@ -990,8 +991,7 @@ void TreeRewriterResult::collectSourceColumns(bool add_special) { auto options = GetColumnsOptions(add_special ? GetColumnsOptions::All : GetColumnsOptions::AllPhysical); options.withExtendedObjects(); - if (storage->supportsSubcolumns()) - options.withSubcolumns(); + options.withSubcolumns(storage->supportsSubcolumns()); auto columns_from_storage = storage_snapshot->getColumns(options); @@ -1001,8 +1001,7 @@ void TreeRewriterResult::collectSourceColumns(bool add_special) source_columns.insert(source_columns.end(), columns_from_storage.begin(), columns_from_storage.end()); auto metadata_snapshot = storage->getInMemoryMetadataPtr(); - auto metadata_column_descriptions = metadata_snapshot->getColumns(); - source_columns_ordinary = metadata_column_descriptions.getOrdinary(); + source_columns_ordinary = metadata_snapshot->getColumns().getOrdinary(); } source_columns_set = removeDuplicateColumns(source_columns); @@ -1109,16 +1108,16 @@ bool TreeRewriterResult::collectUsedColumns(const ASTPtr & query, bool is_select const auto & partition_desc = storage_snapshot->metadata->getPartitionKey(); if (partition_desc.expression) { - auto partition_source_columns = partition_desc.expression->getRequiredColumns(); - partition_source_columns.push_back("_part"); - partition_source_columns.push_back("_partition_id"); - partition_source_columns.push_back("_part_uuid"); - partition_source_columns.push_back("_partition_value"); + auto partition_columns = partition_desc.expression->getRequiredColumns(); + NameSet partition_columns_set(partition_columns.begin(), partition_columns.end()); + + const auto & parititon_virtuals = MergeTreeData::virtuals_useful_for_filter; + partition_columns_set.insert(parititon_virtuals.begin(), parititon_virtuals.end()); + optimize_trivial_count = true; for (const auto & required_column : required) { - if (std::find(partition_source_columns.begin(), partition_source_columns.end(), required_column) - == partition_source_columns.end()) + if (!partition_columns_set.contains(required_column)) { optimize_trivial_count = false; break; @@ -1129,7 +1128,7 @@ bool TreeRewriterResult::collectUsedColumns(const ASTPtr & query, bool is_select NameSet unknown_required_source_columns = required; - for (NamesAndTypesList::iterator it = source_columns.begin(); it != source_columns.end();) + for (auto it = source_columns.begin(); it != source_columns.end();) { const String & column_name = it->name; unknown_required_source_columns.erase(column_name); @@ -1143,32 +1142,23 @@ bool TreeRewriterResult::collectUsedColumns(const ASTPtr & query, bool is_select has_virtual_shard_num = false; /// If there are virtual columns among the unknown columns. Remove them from the list of unknown and add /// in columns list, so that when further processing they are also considered. - if (storage) + if (storage_snapshot) { - const auto storage_virtuals = storage->getVirtuals(); + const auto & virtuals = storage_snapshot->virtual_columns; for (auto it = unknown_required_source_columns.begin(); it != unknown_required_source_columns.end();) { - auto column = storage_virtuals.tryGetByName(*it); - if (column) + if (auto column = virtuals->tryGet(*it)) { source_columns.push_back(*column); it = unknown_required_source_columns.erase(it); } else - ++it; - } - - if (is_remote_storage) - { - for (const auto & name_type : storage_virtuals) { - if (name_type.name == "_shard_num" && storage->isVirtualColumn("_shard_num", storage_snapshot->getMetadataForQuery())) - { - has_virtual_shard_num = true; - break; - } + ++it; } } + + has_virtual_shard_num = is_remote_storage && storage->isVirtualColumn("_shard_num", storage_snapshot->getMetadataForQuery()) && virtuals->has("_shard_num"); } /// Collect missed object subcolumns diff --git a/src/Interpreters/ZooKeeperLog.cpp b/src/Interpreters/ZooKeeperLog.cpp index 9cc31edfe56..6f6d4568064 100644 --- a/src/Interpreters/ZooKeeperLog.cpp +++ b/src/Interpreters/ZooKeeperLog.cpp @@ -122,49 +122,49 @@ ColumnsDescription ZooKeeperLogElement::getColumnsDescription() return ColumnsDescription { - {"hostname", std::make_shared(std::make_shared())}, - {"type", std::move(type_enum)}, - {"event_date", std::make_shared()}, - {"event_time", std::make_shared(6)}, - {"thread_id", std::make_shared()}, - {"query_id", std::make_shared()}, - {"address", DataTypeFactory::instance().get("IPv6")}, - {"port", std::make_shared()}, - {"session_id", std::make_shared()}, - {"duration_ms", std::make_shared()}, + {"hostname", std::make_shared(std::make_shared()), "Hostname of the server executing the query."}, + {"type", std::move(type_enum), "Event type in the ZooKeeper client. Can have one of the following values: Request — The request has been sent, Response — The response was received, Finalize — The connection is lost, no response was received."}, + {"event_date", std::make_shared(), "The date when the event happened."}, + {"event_time", std::make_shared(6), "The date and time when the event happened."}, + {"thread_id", std::make_shared(), "The ID of the thread executed this request."}, + {"query_id", std::make_shared(), "The ID of a query in scope of which this request was executed."}, + {"address", DataTypeFactory::instance().get("IPv6"), "IP address of ZooKeeper server that was used to make the request."}, + {"port", std::make_shared(), "The port of ZooKeeper server that was used to make the request."}, + {"session_id", std::make_shared(), "The session ID that the ZooKeeper server sets for each connection."}, + {"duration_ms", std::make_shared(), "The time taken by ZooKeeper to execute the request."}, - {"xid", std::make_shared()}, - {"has_watch", std::make_shared()}, - {"op_num", op_num_enum}, - {"path", std::make_shared()}, + {"xid", std::make_shared(), "The ID of the request within the session. This is usually a sequential request number. It is the same for the request row and the paired response/finalize row."}, + {"has_watch", std::make_shared(), "The request whether the watch has been set."}, + {"op_num", op_num_enum, "The type of request or response."}, + {"path", std::make_shared(), "The path to the ZooKeeper node specified in the request, or an empty string if the request not requires specifying a path."}, - {"data", std::make_shared()}, + {"data", std::make_shared(), "The data written to the ZooKeeper node (for the SET and CREATE requests — what the request wanted to write, for the response to the GET request — what was read) or an empty string."}, - {"is_ephemeral", std::make_shared()}, - {"is_sequential", std::make_shared()}, + {"is_ephemeral", std::make_shared(), "Is the ZooKeeper node being created as an ephemeral."}, + {"is_sequential", std::make_shared(), "Is the ZooKeeper node being created as an sequential."}, - {"version", std::make_shared(std::make_shared())}, + {"version", std::make_shared(std::make_shared()), "The version of the ZooKeeper node that the request expects when executing. This is supported for CHECK, SET, REMOVE requests (is relevant -1 if the request does not check the version or NULL for other requests that do not support version checking)."}, - {"requests_size", std::make_shared()}, - {"request_idx", std::make_shared()}, + {"requests_size", std::make_shared(), "The number of requests included in the multi request (this is a special request that consists of several consecutive ordinary requests and executes them atomically). All requests included in multi request will have the same xid."}, + {"request_idx", std::make_shared(), "The number of the request included in multi request (for multi request — 0, then in order from 1)."}, - {"zxid", std::make_shared()}, - {"error", std::make_shared(error_enum)}, + {"zxid", std::make_shared(), "ZooKeeper transaction ID. The serial number issued by the ZooKeeper server in response to a successfully executed request (0 if the request was not executed/returned an error/the client does not know whether the request was executed)."}, + {"error", std::make_shared(error_enum), "Error code. Can have many values, here are just some of them: ZOK — The request was executed successfully, ZCONNECTIONLOSS — The connection was lost, ZOPERATIONTIMEOUT — The request execution timeout has expired, ZSESSIONEXPIRED — The session has expired, NULL — The request is completed."}, - {"watch_type", std::make_shared(watch_type_enum)}, - {"watch_state", std::make_shared(watch_state_enum)}, + {"watch_type", std::make_shared(watch_type_enum), "The type of the watch event (for responses with op_num = Watch), for the remaining responses: NULL."}, + {"watch_state", std::make_shared(watch_state_enum), "The status of the watch event (for responses with op_num = Watch), for the remaining responses: NULL."}, - {"path_created", std::make_shared()}, + {"path_created", std::make_shared(), "The path to the created ZooKeeper node (for responses to the CREATE request), may differ from the path if the node is created as a sequential."}, - {"stat_czxid", std::make_shared()}, - {"stat_mzxid", std::make_shared()}, - {"stat_pzxid", std::make_shared()}, - {"stat_version", std::make_shared()}, - {"stat_cversion", std::make_shared()}, - {"stat_dataLength", std::make_shared()}, - {"stat_numChildren", std::make_shared()}, + {"stat_czxid", std::make_shared(), "The zxid of the change that caused this ZooKeeper node to be created."}, + {"stat_mzxid", std::make_shared(), "The zxid of the change that last modified this ZooKeeper node."}, + {"stat_pzxid", std::make_shared(), "The transaction ID of the change that last modified children of this ZooKeeper node."}, + {"stat_version", std::make_shared(), "The number of changes to the data of this ZooKeeper node."}, + {"stat_cversion", std::make_shared(), "The number of changes to the children of this ZooKeeper node."}, + {"stat_dataLength", std::make_shared(), "The length of the data field of this ZooKeeper node."}, + {"stat_numChildren", std::make_shared(), "The number of children of this ZooKeeper node."}, - {"children", std::make_shared(std::make_shared())}, + {"children", std::make_shared(std::make_shared()), "The list of child ZooKeeper nodes (for responses to LIST request)."}, }; } diff --git a/src/Interpreters/executeDDLQueryOnCluster.cpp b/src/Interpreters/executeDDLQueryOnCluster.cpp index 6122ec6180a..df8236c11f4 100644 --- a/src/Interpreters/executeDDLQueryOnCluster.cpp +++ b/src/Interpreters/executeDDLQueryOnCluster.cpp @@ -252,7 +252,8 @@ BlockIO getDistributedDDLStatus(const String & node_path, const DDLLogEntry & en auto source = std::make_shared(node_path, entry, context, hosts_to_wait); io.pipeline = QueryPipeline(std::move(source)); - if (context->getSettingsRef().distributed_ddl_output_mode == DistributedDDLOutputMode::NONE) + if (context->getSettingsRef().distributed_ddl_output_mode == DistributedDDLOutputMode::NONE || + context->getSettingsRef().distributed_ddl_output_mode == DistributedDDLOutputMode::NONE_ONLY_ACTIVE) io.pipeline.complete(std::make_shared(io.pipeline.getHeader())); return io; @@ -264,7 +265,9 @@ Block DDLQueryStatusSource::getSampleBlock(ContextPtr context_, bool hosts_to_wa auto maybe_make_nullable = [&](const DataTypePtr & type) -> DataTypePtr { - if (output_mode == DistributedDDLOutputMode::THROW || output_mode == DistributedDDLOutputMode::NONE) + if (output_mode == DistributedDDLOutputMode::THROW || + output_mode == DistributedDDLOutputMode::NONE || + output_mode == DistributedDDLOutputMode::NONE_ONLY_ACTIVE) return type; return std::make_shared(type); }; @@ -313,14 +316,15 @@ DDLQueryStatusSource::DDLQueryStatusSource( { auto output_mode = context->getSettingsRef().distributed_ddl_output_mode; throw_on_timeout = output_mode == DistributedDDLOutputMode::THROW || output_mode == DistributedDDLOutputMode::THROW_ONLY_ACTIVE - || output_mode == DistributedDDLOutputMode::NONE; + || output_mode == DistributedDDLOutputMode::NONE || output_mode == DistributedDDLOutputMode::NONE_ONLY_ACTIVE; if (hosts_to_wait) { waiting_hosts = NameSet(hosts_to_wait->begin(), hosts_to_wait->end()); is_replicated_database = true; only_running_hosts = output_mode == DistributedDDLOutputMode::THROW_ONLY_ACTIVE || - output_mode == DistributedDDLOutputMode::NULL_STATUS_ON_TIMEOUT_ONLY_ACTIVE; + output_mode == DistributedDDLOutputMode::NULL_STATUS_ON_TIMEOUT_ONLY_ACTIVE || + output_mode == DistributedDDLOutputMode::NONE_ONLY_ACTIVE; } else { @@ -442,14 +446,16 @@ Chunk DDLQueryStatusSource::generate() size_t num_unfinished_hosts = waiting_hosts.size() - num_hosts_finished; size_t num_active_hosts = current_active_hosts.size(); - constexpr auto msg_format = "Watching task {} is executing longer than distributed_ddl_task_timeout (={}) seconds. " - "There are {} unfinished hosts ({} of them are currently executing the task), " - "they are going to execute the query in background"; + constexpr auto msg_format = "Distributed DDL task {} is not finished on {} of {} hosts " + "({} of them are currently executing the task, {} are inactive). " + "They are going to execute the query in background. Was waiting for {} seconds{}"; + if (throw_on_timeout) { if (!first_exception) first_exception = std::make_unique(Exception(ErrorCodes::TIMEOUT_EXCEEDED, - msg_format, node_path, timeout_seconds, num_unfinished_hosts, num_active_hosts)); + msg_format, node_path, num_unfinished_hosts, waiting_hosts.size(), num_active_hosts, offline_hosts.size(), + watch.elapsedSeconds(), stop_waiting_offline_hosts ? "" : ", which is longer than distributed_ddl_task_timeout")); /// For Replicated database print a list of unfinished hosts as well. Will return empty block on next iteration. if (is_replicated_database) @@ -457,7 +463,8 @@ Chunk DDLQueryStatusSource::generate() return {}; } - LOG_INFO(log, msg_format, node_path, timeout_seconds, num_unfinished_hosts, num_active_hosts); + LOG_INFO(log, msg_format, node_path, num_unfinished_hosts, waiting_hosts.size(), num_active_hosts, offline_hosts.size(), + watch.elapsedSeconds(), stop_waiting_offline_hosts ? "" : "which is longer than distributed_ddl_task_timeout"); return generateChunkWithUnfinishedHosts(); } diff --git a/src/Interpreters/executeQuery.cpp b/src/Interpreters/executeQuery.cpp index 53fd5088372..88021038ebb 100644 --- a/src/Interpreters/executeQuery.cpp +++ b/src/Interpreters/executeQuery.cpp @@ -104,6 +104,8 @@ namespace ErrorCodes extern const int QUERY_WAS_CANCELLED; extern const int INCORRECT_DATA; extern const int SYNTAX_ERROR; + extern const int SUPPORT_IS_DISABLED; + extern const int INCORRECT_QUERY; } namespace FailPoints @@ -651,6 +653,36 @@ static void setQuerySpecificSettings(ASTPtr & ast, ContextMutablePtr context) } } +void validateAnalyzerSettings(ASTPtr ast, bool context_value) +{ + if (ast->as()) + return; + + bool top_level = context_value; + + std::vector nodes_to_process{ ast }; + while (!nodes_to_process.empty()) + { + auto node = nodes_to_process.back(); + nodes_to_process.pop_back(); + + if (auto * set_query = node->as()) + { + if (auto * value = set_query->changes.tryGet("allow_experimental_analyzer")) + { + if (top_level != value->safeGet()) + throw Exception(ErrorCodes::INCORRECT_QUERY, "Setting 'allow_experimental_analyzer' is changed in the subquery. Top level value: {}", top_level); + } + } + + for (auto child : node->children) + { + if (child) + nodes_to_process.push_back(std::move(child)); + } + } +} + static std::tuple executeQueryImpl( const char * begin, const char * end, @@ -861,6 +893,7 @@ static std::tuple executeQueryImpl( /// Interpret SETTINGS clauses as early as possible (before invoking the corresponding interpreter), /// to allow settings to take effect. InterpreterSetQuery::applySettingsFromQuery(ast, context); + validateAnalyzerSettings(ast, context->getSettingsRef().allow_experimental_analyzer); if (auto * insert_query = ast->as()) insert_query->tail = istr; @@ -991,6 +1024,21 @@ static std::tuple executeQueryImpl( if (settings.implicit_transaction && settings.throw_on_unsupported_query_inside_transaction) throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Async inserts with 'implicit_transaction' are not supported"); + /// Let's agree on terminology and say that a mini-INSERT is an asynchronous INSERT + /// which typically contains not a lot of data inside and a big-INSERT in an INSERT + /// which was formed by concatenating several mini-INSERTs together. + /// In case when the client had to retry some mini-INSERTs then they will be properly deduplicated + /// by the source tables. This functionality is controlled by a setting `async_insert_deduplicate`. + /// But then they will be glued together into a block and pushed through a chain of Materialized Views if any. + /// The process of forming such blocks is not deteministic so each time we retry mini-INSERTs the resulting + /// block may be concatenated differently. + /// That's why deduplication in dependent Materialized Views doesn't make sense in presence of async INSERTs. + if (settings.throw_if_deduplication_in_dependent_materialized_views_enabled_with_async_insert && + settings.deduplicate_blocks_in_dependent_materialized_views) + throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, + "Deduplication is dependent materialized view cannot work together with async inserts. "\ + "Please disable eiher `deduplicate_blocks_in_dependent_materialized_views` or `async_insert` setting."); + quota = context->getQuota(); if (quota) { diff --git a/src/Interpreters/fuzzers/execute_query_fuzzer.cpp b/src/Interpreters/fuzzers/execute_query_fuzzer.cpp index 6f84a60f2af..edff202d547 100644 --- a/src/Interpreters/fuzzers/execute_query_fuzzer.cpp +++ b/src/Interpreters/fuzzers/execute_query_fuzzer.cpp @@ -3,7 +3,7 @@ #include #include "Processors/Executors/PullingPipelineExecutor.h" -#include +#include #include #include #include diff --git a/src/Interpreters/getHeaderForProcessingStage.cpp b/src/Interpreters/getHeaderForProcessingStage.cpp index 21739298036..06c5d424d2f 100644 --- a/src/Interpreters/getHeaderForProcessingStage.cpp +++ b/src/Interpreters/getHeaderForProcessingStage.cpp @@ -121,7 +121,12 @@ Block getHeaderForProcessingStage( auto & table_expression_data = query_info.planner_context->getTableExpressionDataOrThrow(left_table_expression); const auto & query_context = query_info.planner_context->getQueryContext(); - auto columns = table_expression_data.getColumns(); + + NamesAndTypes columns; + const auto & column_name_to_column = table_expression_data.getColumnNameToColumn(); + for (const auto & column_name : table_expression_data.getSelectedColumnsNames()) + columns.push_back(column_name_to_column.at(column_name)); + auto new_query_node = buildSubqueryToReadColumnsFromTableExpression(columns, left_table_expression, query_context); query = new_query_node->toAST(); } @@ -137,8 +142,9 @@ Block getHeaderForProcessingStage( if (context->getSettingsRef().allow_experimental_analyzer) { - auto storage = std::make_shared( - storage_snapshot->storage.getStorageID(), storage_snapshot->metadata->getColumns(), storage_snapshot); + auto storage = std::make_shared(storage_snapshot->storage.getStorageID(), + storage_snapshot->getAllColumnsDescription(), + storage_snapshot); InterpreterSelectQueryAnalyzer interpreter(query, context, storage, SelectQueryOptions(processed_stage).analyze()); result = interpreter.getSampleBlock(); } diff --git a/src/Interpreters/getTableExpressions.cpp b/src/Interpreters/getTableExpressions.cpp index 70e38526648..2853be4c05e 100644 --- a/src/Interpreters/getTableExpressions.cpp +++ b/src/Interpreters/getTableExpressions.cpp @@ -99,7 +99,7 @@ static NamesAndTypesList getColumnsFromTableExpression( names_and_type_list = columns.getOrdinary(); materialized = columns.getMaterialized(); aliases = columns.getAliases(); - virtuals = function_storage->getVirtuals(); + virtuals = function_storage->getVirtualsList(); } else if (table_expression.database_and_table_name) { @@ -110,7 +110,7 @@ static NamesAndTypesList getColumnsFromTableExpression( names_and_type_list = columns.getOrdinary(); materialized = columns.getMaterialized(); aliases = columns.getAliases(); - virtuals = table->getVirtuals(); + virtuals = table->getVirtualsList(); } return names_and_type_list; diff --git a/src/Interpreters/inplaceBlockConversions.cpp b/src/Interpreters/inplaceBlockConversions.cpp index fd8f5b154c4..239cce5b427 100644 --- a/src/Interpreters/inplaceBlockConversions.cpp +++ b/src/Interpreters/inplaceBlockConversions.cpp @@ -20,7 +20,6 @@ #include #include #include -#include namespace DB @@ -280,7 +279,7 @@ void fillMissingColumns( const NamesAndTypesList & requested_columns, const NamesAndTypesList & available_columns, const NameSet & partially_read_columns, - StorageMetadataPtr metadata_snapshot, size_t block_number) + StorageMetadataPtr metadata_snapshot) { size_t num_columns = requested_columns.size(); if (num_columns != res_columns.size()) @@ -359,14 +358,9 @@ void fillMissingColumns( } else { - if (requested_column->name == BlockNumberColumn::name) - res_columns[i] = type->createColumnConst(num_rows, block_number)->convertToFullColumnIfConst(); - else - /// We must turn a constant column into a full column because the interpreter could infer - /// that it is constant everywhere but in some blocks (from other parts) it can be a full column. - res_columns[i] = type->createColumnConstWithDefaultValue(num_rows)->convertToFullColumnIfConst(); - - + /// We must turn a constant column into a full column because the interpreter could infer + /// that it is constant everywhere but in some blocks (from other parts) it can be a full column. + res_columns[i] = type->createColumnConstWithDefaultValue(num_rows)->convertToFullColumnIfConst(); } } } diff --git a/src/Interpreters/inplaceBlockConversions.h b/src/Interpreters/inplaceBlockConversions.h index 7a13a75ec8b..bea44bf6db9 100644 --- a/src/Interpreters/inplaceBlockConversions.h +++ b/src/Interpreters/inplaceBlockConversions.h @@ -46,6 +46,6 @@ void fillMissingColumns( const NamesAndTypesList & requested_columns, const NamesAndTypesList & available_columns, const NameSet & partially_read_columns, - StorageMetadataPtr metadata_snapshot, size_t block_number = 0); + StorageMetadataPtr metadata_snapshot); } diff --git a/src/Interpreters/processColumnTransformers.cpp b/src/Interpreters/processColumnTransformers.cpp index 2a704d4a937..5ef331eb119 100644 --- a/src/Interpreters/processColumnTransformers.cpp +++ b/src/Interpreters/processColumnTransformers.cpp @@ -32,7 +32,7 @@ ASTPtr processColumnTransformers( tables_with_columns[0].addHiddenColumns(columns.getMaterialized()); tables_with_columns[0].addHiddenColumns(columns.getAliases()); - tables_with_columns[0].addHiddenColumns(table->getVirtuals()); + tables_with_columns[0].addHiddenColumns(table->getVirtualsList()); NameSet source_columns_set; for (const auto & identifier : query_columns->children) diff --git a/src/Parsers/ASTAlterQuery.cpp b/src/Parsers/ASTAlterQuery.cpp index a93ad1d1746..f104e715452 100644 --- a/src/Parsers/ASTAlterQuery.cpp +++ b/src/Parsers/ASTAlterQuery.cpp @@ -634,6 +634,7 @@ void ASTAlterQuery::formatQueryImpl(const FormatSettings & settings, FormatState settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } else if (alter_object == AlterObjectType::DATABASE && database) diff --git a/src/Parsers/ASTCheckQuery.h b/src/Parsers/ASTCheckQuery.h index 5e9032e77fb..eca08b2b094 100644 --- a/src/Parsers/ASTCheckQuery.h +++ b/src/Parsers/ASTCheckQuery.h @@ -53,6 +53,7 @@ protected: settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTCreateIndexQuery.cpp b/src/Parsers/ASTCreateIndexQuery.cpp index 84493ff22f1..f6def3ed85c 100644 --- a/src/Parsers/ASTCreateIndexQuery.cpp +++ b/src/Parsers/ASTCreateIndexQuery.cpp @@ -52,6 +52,7 @@ void ASTCreateIndexQuery::formatQueryImpl(const FormatSettings & settings, Forma settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTCreateQuery.cpp b/src/Parsers/ASTCreateQuery.cpp index 1315ea5784c..de5eb40837f 100644 --- a/src/Parsers/ASTCreateQuery.cpp +++ b/src/Parsers/ASTCreateQuery.cpp @@ -337,6 +337,7 @@ void ASTCreateQuery::formatQueryImpl(const FormatSettings & settings, FormatStat settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); if (uuid != UUIDHelpers::Nil) @@ -370,6 +371,7 @@ void ASTCreateQuery::formatQueryImpl(const FormatSettings & settings, FormatStat settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); if (uuid != UUIDHelpers::Nil) diff --git a/src/Parsers/ASTDeleteQuery.cpp b/src/Parsers/ASTDeleteQuery.cpp index 1e8303dac62..67f3a85c9a5 100644 --- a/src/Parsers/ASTDeleteQuery.cpp +++ b/src/Parsers/ASTDeleteQuery.cpp @@ -40,6 +40,7 @@ void ASTDeleteQuery::formatQueryImpl(const FormatSettings & settings, FormatStat settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); formatOnCluster(settings); diff --git a/src/Parsers/ASTDropIndexQuery.cpp b/src/Parsers/ASTDropIndexQuery.cpp index 1a1481ea27f..1109f32f019 100644 --- a/src/Parsers/ASTDropIndexQuery.cpp +++ b/src/Parsers/ASTDropIndexQuery.cpp @@ -47,6 +47,7 @@ void ASTDropIndexQuery::formatQueryImpl(const FormatSettings & settings, FormatS settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTDropQuery.cpp b/src/Parsers/ASTDropQuery.cpp index f41f28f745d..ca47ceccb85 100644 --- a/src/Parsers/ASTDropQuery.cpp +++ b/src/Parsers/ASTDropQuery.cpp @@ -76,6 +76,7 @@ void ASTDropQuery::formatQueryImpl(const FormatSettings & settings, FormatState settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTFunction.cpp b/src/Parsers/ASTFunction.cpp index 69a2e7c9ada..07eea86ef81 100644 --- a/src/Parsers/ASTFunction.cpp +++ b/src/Parsers/ASTFunction.cpp @@ -18,6 +18,7 @@ #include #include #include +#include #include #include @@ -36,508 +37,6 @@ namespace ErrorCodes } -namespace -{ - /// Finds arguments of a specified function which should not be displayed for most users for security reasons. - /// That involves passwords and secret keys. - class FunctionSecretArgumentsFinder - { - public: - explicit FunctionSecretArgumentsFinder(const ASTFunction & function_) : function(function_) - { - if (!function.arguments) - return; - - const auto * expr_list = function.arguments->as(); - if (!expr_list) - return; - - arguments = &expr_list->children; - switch (function.kind) - { - case ASTFunction::Kind::ORDINARY_FUNCTION: findOrdinaryFunctionSecretArguments(); break; - case ASTFunction::Kind::WINDOW_FUNCTION: break; - case ASTFunction::Kind::LAMBDA_FUNCTION: break; - case ASTFunction::Kind::TABLE_ENGINE: findTableEngineSecretArguments(); break; - case ASTFunction::Kind::DATABASE_ENGINE: findDatabaseEngineSecretArguments(); break; - case ASTFunction::Kind::BACKUP_NAME: findBackupNameSecretArguments(); break; - } - } - - struct Result - { - /// Result constructed by default means no arguments will be hidden. - size_t start = static_cast(-1); - size_t count = 0; /// Mostly it's either 0 or 1. There are only a few cases where `count` can be greater than 1 (e.g. see `encrypt`). - /// In all known cases secret arguments are consecutive - bool are_named = false; /// Arguments like `password = 'password'` are considered as named arguments. - /// E.g. "headers" in `url('..', headers('foo' = '[HIDDEN]'))` - std::vector nested_maps; - - bool hasSecrets() const - { - return count != 0 || !nested_maps.empty(); - } - }; - - Result getResult() const { return result; } - - private: - const ASTFunction & function; - const ASTs * arguments = nullptr; - Result result; - - void markSecretArgument(size_t index, bool argument_is_named = false) - { - if (index >= arguments->size()) - return; - if (!result.count) - { - result.start = index; - result.are_named = argument_is_named; - } - chassert(index >= result.start); /// We always check arguments consecutively - result.count = index + 1 - result.start; - if (!argument_is_named) - result.are_named = false; - } - - void findOrdinaryFunctionSecretArguments() - { - if ((function.name == "mysql") || (function.name == "postgresql") || (function.name == "mongodb")) - { - /// mysql('host:port', 'database', 'table', 'user', 'password', ...) - /// postgresql('host:port', 'database', 'table', 'user', 'password', ...) - /// mongodb('host:port', 'database', 'collection', 'user', 'password', ...) - findMySQLFunctionSecretArguments(); - } - else if ((function.name == "s3") || (function.name == "cosn") || (function.name == "oss") || - (function.name == "deltaLake") || (function.name == "hudi") || (function.name == "iceberg")) - { - /// s3('url', 'aws_access_key_id', 'aws_secret_access_key', ...) - findS3FunctionSecretArguments(/* is_cluster_function= */ false); - } - else if (function.name == "s3Cluster") - { - /// s3Cluster('cluster_name', 'url', 'aws_access_key_id', 'aws_secret_access_key', ...) - findS3FunctionSecretArguments(/* is_cluster_function= */ true); - } - else if ((function.name == "remote") || (function.name == "remoteSecure")) - { - /// remote('addresses_expr', 'db', 'table', 'user', 'password', ...) - findRemoteFunctionSecretArguments(); - } - else if ((function.name == "encrypt") || (function.name == "decrypt") || - (function.name == "aes_encrypt_mysql") || (function.name == "aes_decrypt_mysql") || - (function.name == "tryDecrypt")) - { - /// encrypt('mode', 'plaintext', 'key' [, iv, aad]) - findEncryptionFunctionSecretArguments(); - } - else if (function.name == "url") - { - findURLSecretArguments(); - } - } - - void findMySQLFunctionSecretArguments() - { - if (isNamedCollectionName(0)) - { - /// mysql(named_collection, ..., password = 'password', ...) - findSecretNamedArgument("password", 1); - } - else - { - /// mysql('host:port', 'database', 'table', 'user', 'password', ...) - markSecretArgument(4); - } - } - - /// Returns the number of arguments excluding "headers" and "extra_credentials" (which should - /// always be at the end). Marks "headers" as secret, if found. - size_t excludeS3OrURLNestedMaps() - { - size_t count = arguments->size(); - while (count > 0) - { - const ASTFunction * f = arguments->at(count - 1)->as(); - if (!f) - break; - if (f->name == "headers") - result.nested_maps.push_back(f->name); - else if (f->name != "extra_credentials") - break; - count -= 1; - } - return count; - } - - void findS3FunctionSecretArguments(bool is_cluster_function) - { - /// s3Cluster('cluster_name', 'url', ...) has 'url' as its second argument. - size_t url_arg_idx = is_cluster_function ? 1 : 0; - - if (!is_cluster_function && isNamedCollectionName(0)) - { - /// s3(named_collection, ..., secret_access_key = 'secret_access_key', ...) - findSecretNamedArgument("secret_access_key", 1); - return; - } - - /// We should check other arguments first because we don't need to do any replacement in case of - /// s3('url', NOSIGN, 'format' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) - /// s3('url', 'format', 'structure' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) - size_t count = excludeS3OrURLNestedMaps(); - if ((url_arg_idx + 3 <= count) && (count <= url_arg_idx + 4)) - { - String second_arg; - if (tryGetStringFromArgument(url_arg_idx + 1, &second_arg)) - { - if (boost::iequals(second_arg, "NOSIGN")) - return; /// The argument after 'url' is "NOSIGN". - - if (second_arg == "auto" || KnownFormatNames::instance().exists(second_arg)) - return; /// The argument after 'url' is a format: s3('url', 'format', ...) - } - } - - /// We're going to replace 'aws_secret_access_key' with '[HIDDEN]' for the following signatures: - /// s3('url', 'aws_access_key_id', 'aws_secret_access_key', ...) - /// s3Cluster('cluster_name', 'url', 'aws_access_key_id', 'aws_secret_access_key', 'format', 'compression') - if (url_arg_idx + 2 < count) - markSecretArgument(url_arg_idx + 2); - } - - void findURLSecretArguments() - { - if (!isNamedCollectionName(0)) - excludeS3OrURLNestedMaps(); - } - - bool tryGetStringFromArgument(size_t arg_idx, String * res, bool allow_identifier = true) const - { - if (arg_idx >= arguments->size()) - return false; - - return tryGetStringFromArgument(*(*arguments)[arg_idx], res, allow_identifier); - } - - static bool tryGetStringFromArgument(const IAST & argument, String * res, bool allow_identifier = true) - { - if (const auto * literal = argument.as()) - { - if (literal->value.getType() != Field::Types::String) - return false; - if (res) - *res = literal->value.safeGet(); - return true; - } - - if (allow_identifier) - { - if (const auto * id = argument.as()) - { - if (res) - *res = id->name(); - return true; - } - } - - return false; - } - - void findRemoteFunctionSecretArguments() - { - if (isNamedCollectionName(0)) - { - /// remote(named_collection, ..., password = 'password', ...) - findSecretNamedArgument("password", 1); - return; - } - - /// We're going to replace 'password' with '[HIDDEN'] for the following signatures: - /// remote('addresses_expr', db.table, 'user' [, 'password'] [, sharding_key]) - /// remote('addresses_expr', 'db', 'table', 'user' [, 'password'] [, sharding_key]) - /// remote('addresses_expr', table_function(), 'user' [, 'password'] [, sharding_key]) - - /// But we should check the number of arguments first because we don't need to do any replacements in case of - /// remote('addresses_expr', db.table) - if (arguments->size() < 3) - return; - - size_t arg_num = 1; - - /// Skip 1 or 2 arguments with table_function() or db.table or 'db', 'table'. - const auto * table_function = (*arguments)[arg_num]->as(); - if (table_function && KnownTableFunctionNames::instance().exists(table_function->name)) - { - ++arg_num; - } - else - { - std::optional database; - std::optional qualified_table_name; - if (!tryGetDatabaseNameOrQualifiedTableName(arg_num, database, qualified_table_name)) - { - /// We couldn't evaluate the argument so we don't know whether it is 'db.table' or just 'db'. - /// Hence we can't figure out whether we should skip one argument 'user' or two arguments 'table', 'user' - /// before the argument 'password'. So it's safer to wipe two arguments just in case. - /// The last argument can be also a `sharding_key`, so we need to check that argument is a literal string - /// before wiping it (because the `password` argument is always a literal string). - if (tryGetStringFromArgument(arg_num + 2, nullptr, /* allow_identifier= */ false)) - { - /// Wipe either `password` or `user`. - markSecretArgument(arg_num + 2); - } - if (tryGetStringFromArgument(arg_num + 3, nullptr, /* allow_identifier= */ false)) - { - /// Wipe either `password` or `sharding_key`. - markSecretArgument(arg_num + 3); - } - return; - } - - /// Skip the current argument (which is either a database name or a qualified table name). - ++arg_num; - if (database) - { - /// Skip the 'table' argument if the previous argument was a database name. - ++arg_num; - } - } - - /// Skip username. - ++arg_num; - - /// Do our replacement: - /// remote('addresses_expr', db.table, 'user', 'password', ...) -> remote('addresses_expr', db.table, 'user', '[HIDDEN]', ...) - /// The last argument can be also a `sharding_key`, so we need to check that argument is a literal string - /// before wiping it (because the `password` argument is always a literal string). - bool can_be_password = tryGetStringFromArgument(arg_num, nullptr, /* allow_identifier= */ false); - if (can_be_password) - markSecretArgument(arg_num); - } - - /// Tries to get either a database name or a qualified table name from an argument. - /// Empty string is also allowed (it means the default database). - /// The function is used by findRemoteFunctionSecretArguments() to determine how many arguments to skip before a password. - bool tryGetDatabaseNameOrQualifiedTableName( - size_t arg_idx, - std::optional & res_database, - std::optional & res_qualified_table_name) const - { - res_database.reset(); - res_qualified_table_name.reset(); - - String str; - if (!tryGetStringFromArgument(arg_idx, &str, /* allow_identifier= */ true)) - return false; - - if (str.empty()) - { - res_database = ""; - return true; - } - - auto qualified_table_name = QualifiedTableName::tryParseFromString(str); - if (!qualified_table_name) - return false; - - if (qualified_table_name->database.empty()) - res_database = std::move(qualified_table_name->table); - else - res_qualified_table_name = std::move(qualified_table_name); - return true; - } - - void findEncryptionFunctionSecretArguments() - { - if (arguments->empty()) - return; - - /// We replace all arguments after 'mode' with '[HIDDEN]': - /// encrypt('mode', 'plaintext', 'key' [, iv, aad]) -> encrypt('mode', '[HIDDEN]') - result.start = 1; - result.count = arguments->size() - 1; - } - - void findTableEngineSecretArguments() - { - const String & engine_name = function.name; - if (engine_name == "ExternalDistributed") - { - /// ExternalDistributed('engine', 'host:port', 'database', 'table', 'user', 'password') - findExternalDistributedTableEngineSecretArguments(); - } - else if ((engine_name == "MySQL") || (engine_name == "PostgreSQL") || - (engine_name == "MaterializedPostgreSQL") || (engine_name == "MongoDB")) - { - /// MySQL('host:port', 'database', 'table', 'user', 'password', ...) - /// PostgreSQL('host:port', 'database', 'table', 'user', 'password', ...) - /// MaterializedPostgreSQL('host:port', 'database', 'table', 'user', 'password', ...) - /// MongoDB('host:port', 'database', 'collection', 'user', 'password', ...) - findMySQLFunctionSecretArguments(); - } - else if ((engine_name == "S3") || (engine_name == "COSN") || (engine_name == "OSS") || - (engine_name == "DeltaLake") || (engine_name == "Hudi") || (engine_name == "Iceberg") || (engine_name == "S3Queue")) - { - /// S3('url', ['aws_access_key_id', 'aws_secret_access_key',] ...) - findS3TableEngineSecretArguments(); - } - else if (engine_name == "URL") - { - findURLSecretArguments(); - } - } - - void findExternalDistributedTableEngineSecretArguments() - { - if (isNamedCollectionName(1)) - { - /// ExternalDistributed('engine', named_collection, ..., password = 'password', ...) - findSecretNamedArgument("password", 2); - } - else - { - /// ExternalDistributed('engine', 'host:port', 'database', 'table', 'user', 'password') - markSecretArgument(5); - } - } - - void findS3TableEngineSecretArguments() - { - if (isNamedCollectionName(0)) - { - /// S3(named_collection, ..., secret_access_key = 'secret_access_key') - findSecretNamedArgument("secret_access_key", 1); - return; - } - - /// We should check other arguments first because we don't need to do any replacement in case of - /// S3('url', NOSIGN, 'format' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) - /// S3('url', 'format', 'compression' [, extra_credentials(..)] [, headers(..)]) - size_t count = excludeS3OrURLNestedMaps(); - if ((3 <= count) && (count <= 4)) - { - String second_arg; - if (tryGetStringFromArgument(1, &second_arg)) - { - if (boost::iequals(second_arg, "NOSIGN")) - return; /// The argument after 'url' is "NOSIGN". - - if (count == 3) - { - if (second_arg == "auto" || KnownFormatNames::instance().exists(second_arg)) - return; /// The argument after 'url' is a format: S3('url', 'format', ...) - } - } - } - - /// We replace 'aws_secret_access_key' with '[HIDDEN]' for the following signatures: - /// S3('url', 'aws_access_key_id', 'aws_secret_access_key') - /// S3('url', 'aws_access_key_id', 'aws_secret_access_key', 'format') - /// S3('url', 'aws_access_key_id', 'aws_secret_access_key', 'format', 'compression') - if (2 < count) - markSecretArgument(2); - } - - void findDatabaseEngineSecretArguments() - { - const String & engine_name = function.name; - if ((engine_name == "MySQL") || (engine_name == "MaterializeMySQL") || - (engine_name == "MaterializedMySQL") || (engine_name == "PostgreSQL") || - (engine_name == "MaterializedPostgreSQL")) - { - /// MySQL('host:port', 'database', 'user', 'password') - /// PostgreSQL('host:port', 'database', 'user', 'password') - findMySQLDatabaseSecretArguments(); - } - else if (engine_name == "S3") - { - /// S3('url', 'access_key_id', 'secret_access_key') - findS3DatabaseSecretArguments(); - } - } - - void findMySQLDatabaseSecretArguments() - { - if (isNamedCollectionName(0)) - { - /// MySQL(named_collection, ..., password = 'password', ...) - findSecretNamedArgument("password", 1); - } - else - { - /// MySQL('host:port', 'database', 'user', 'password') - markSecretArgument(3); - } - } - - void findS3DatabaseSecretArguments() - { - if (isNamedCollectionName(0)) - { - /// S3(named_collection, ..., secret_access_key = 'password', ...) - findSecretNamedArgument("secret_access_key", 1); - } - else - { - /// S3('url', 'access_key_id', 'secret_access_key') - markSecretArgument(2); - } - } - - void findBackupNameSecretArguments() - { - const String & engine_name = function.name; - if (engine_name == "S3") - { - /// BACKUP ... TO S3(url, [aws_access_key_id, aws_secret_access_key]) - markSecretArgument(2); - } - } - - /// Whether a specified argument can be the name of a named collection? - bool isNamedCollectionName(size_t arg_idx) const - { - if (arguments->size() <= arg_idx) - return false; - - const auto * identifier = (*arguments)[arg_idx]->as(); - return identifier != nullptr; - } - - /// Looks for a secret argument with a specified name. This function looks for arguments in format `key=value` where the key is specified. - void findSecretNamedArgument(const std::string_view & key, size_t start = 0) - { - for (size_t i = start; i < arguments->size(); ++i) - { - const auto & argument = (*arguments)[i]; - const auto * equals_func = argument->as(); - if (!equals_func || (equals_func->name != "equals")) - continue; - - const auto * expr_list = equals_func->arguments->as(); - if (!expr_list) - continue; - - const auto & equal_args = expr_list->children; - if (equal_args.size() != 2) - continue; - - String found_key; - if (!tryGetStringFromArgument(*equal_args[0], &found_key)) - continue; - - if (found_key == key) - markSecretArgument(i, /* argument_is_named= */ true); - } - } - }; -} - - void ASTFunction::appendColumnNameImpl(WriteBuffer & ostr) const { /// These functions contain some unexpected ASTs in arguments (e.g. SETTINGS or even a SELECT query) @@ -1195,7 +694,7 @@ void ASTFunction::formatImplWithoutAlias(const FormatSettings & settings, Format FunctionSecretArgumentsFinder::Result secret_arguments; if (!settings.show_secrets) - secret_arguments = FunctionSecretArgumentsFinder{*this}.getResult(); + secret_arguments = FunctionSecretArgumentsFinderAST(*this).getResult(); for (size_t i = 0, size = arguments->children.size(); i < size; ++i) { @@ -1260,7 +759,7 @@ void ASTFunction::formatImplWithoutAlias(const FormatSettings & settings, Format bool ASTFunction::hasSecretParts() const { - return (FunctionSecretArgumentsFinder{*this}.getResult().hasSecrets()) || childrenHaveSecretParts(); + return (FunctionSecretArgumentsFinderAST(*this).getResult().hasSecrets()) || childrenHaveSecretParts(); } String getFunctionName(const IAST * ast) diff --git a/src/Parsers/ASTInsertQuery.cpp b/src/Parsers/ASTInsertQuery.cpp index eb4bcce6578..72a569fe047 100644 --- a/src/Parsers/ASTInsertQuery.cpp +++ b/src/Parsers/ASTInsertQuery.cpp @@ -74,6 +74,7 @@ void ASTInsertQuery::formatImpl(const FormatSettings & settings, FormatState & s settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTInsertQuery.h b/src/Parsers/ASTInsertQuery.h index 6a4ce078f79..b0f444ed755 100644 --- a/src/Parsers/ASTInsertQuery.h +++ b/src/Parsers/ASTInsertQuery.h @@ -59,11 +59,13 @@ public: if (database) { res->database = database->clone(); res->children.push_back(res->database); } if (table) { res->table = table->clone(); res->children.push_back(res->table); } if (columns) { res->columns = columns->clone(); res->children.push_back(res->columns); } - if (select) { res->select = select->clone(); res->children.push_back(res->select); } - if (watch) { res->watch = watch->clone(); res->children.push_back(res->watch); } if (table_function) { res->table_function = table_function->clone(); res->children.push_back(res->table_function); } if (partition_by) { res->partition_by = partition_by->clone(); res->children.push_back(res->partition_by); } if (settings_ast) { res->settings_ast = settings_ast->clone(); res->children.push_back(res->settings_ast); } + if (select) { res->select = select->clone(); res->children.push_back(res->select); } + if (watch) { res->watch = watch->clone(); res->children.push_back(res->watch); } + if (infile) { res->infile = infile->clone(); res->children.push_back(res->infile); } + if (compression) { res->compression = compression->clone(); res->children.push_back(res->compression); } return res; } diff --git a/src/Parsers/ASTOptimizeQuery.cpp b/src/Parsers/ASTOptimizeQuery.cpp index d9406a9eca0..397a37586fc 100644 --- a/src/Parsers/ASTOptimizeQuery.cpp +++ b/src/Parsers/ASTOptimizeQuery.cpp @@ -15,6 +15,7 @@ void ASTOptimizeQuery::formatQueryImpl(const FormatSettings & settings, FormatSt settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); formatOnCluster(settings); diff --git a/src/Parsers/ASTQueryWithOutput.cpp b/src/Parsers/ASTQueryWithOutput.cpp index 4bf1e6cb231..c57aa759969 100644 --- a/src/Parsers/ASTQueryWithOutput.cpp +++ b/src/Parsers/ASTQueryWithOutput.cpp @@ -23,6 +23,16 @@ void ASTQueryWithOutput::cloneOutputOptions(ASTQueryWithOutput & cloned) const cloned.settings_ast = settings_ast->clone(); cloned.children.push_back(cloned.settings_ast); } + if (compression) + { + cloned.compression = compression->clone(); + cloned.children.push_back(cloned.compression); + } + if (compression_level) + { + cloned.compression_level = compression_level->clone(); + cloned.children.push_back(cloned.compression_level); + } } void ASTQueryWithOutput::formatImpl(const FormatSettings & s, FormatState & state, FormatStateStacked frame) const @@ -64,9 +74,23 @@ bool ASTQueryWithOutput::resetOutputASTIfExist(IAST & ast) /// FIXME: try to prettify this cast using `as<>()` if (auto * ast_with_output = dynamic_cast(&ast)) { - ast_with_output->format.reset(); - ast_with_output->out_file.reset(); - ast_with_output->settings_ast.reset(); + auto remove_if_exists = [&](ASTPtr & p) + { + if (p) + { + if (auto * it = std::find(ast_with_output->children.begin(), ast_with_output->children.end(), p); + it != ast_with_output->children.end()) + ast_with_output->children.erase(it); + p.reset(); + } + }; + + remove_if_exists(ast_with_output->out_file); + remove_if_exists(ast_with_output->format); + remove_if_exists(ast_with_output->settings_ast); + remove_if_exists(ast_with_output->compression); + remove_if_exists(ast_with_output->compression_level); + return true; } diff --git a/src/Parsers/ASTQueryWithTableAndOutput.h b/src/Parsers/ASTQueryWithTableAndOutput.h index 1b8621fb63b..29dbd30e54b 100644 --- a/src/Parsers/ASTQueryWithTableAndOutput.h +++ b/src/Parsers/ASTQueryWithTableAndOutput.h @@ -1,5 +1,6 @@ #pragma once +#include #include #include #include @@ -61,6 +62,7 @@ protected: settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } }; diff --git a/src/Parsers/ASTRenameQuery.h b/src/Parsers/ASTRenameQuery.h index 901f0121a28..73d12be094a 100644 --- a/src/Parsers/ASTRenameQuery.h +++ b/src/Parsers/ASTRenameQuery.h @@ -127,6 +127,7 @@ protected: settings.ostr << '.'; } + chassert(it->from.table); it->from.table->formatImpl(settings, state, frame); settings.ostr << (settings.hilite ? hilite_keyword : "") << (exchange ? " AND " : " TO ") << (settings.hilite ? hilite_none : ""); @@ -137,6 +138,7 @@ protected: settings.ostr << '.'; } + chassert(it->to.table); it->to.table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTSystemQuery.cpp b/src/Parsers/ASTSystemQuery.cpp index 63311a70e42..effc7207793 100644 --- a/src/Parsers/ASTSystemQuery.cpp +++ b/src/Parsers/ASTSystemQuery.cpp @@ -114,6 +114,7 @@ void ASTSystemQuery::formatImpl(const FormatSettings & settings, FormatState & s settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); return settings.ostr; }; @@ -383,6 +384,7 @@ void ASTSystemQuery::formatImpl(const FormatSettings & settings, FormatState & s case Type::KILL: case Type::SHUTDOWN: case Type::DROP_DNS_CACHE: + case Type::DROP_CONNECTIONS_CACHE: case Type::DROP_MMAP_CACHE: case Type::DROP_QUERY_CACHE: case Type::DROP_MARK_CACHE: diff --git a/src/Parsers/ASTSystemQuery.h b/src/Parsers/ASTSystemQuery.h index 48be7f6b84f..70a9e27178d 100644 --- a/src/Parsers/ASTSystemQuery.h +++ b/src/Parsers/ASTSystemQuery.h @@ -22,6 +22,7 @@ public: KILL, SUSPEND, DROP_DNS_CACHE, + DROP_CONNECTIONS_CACHE, DROP_MARK_CACHE, DROP_UNCOMPRESSED_CACHE, DROP_INDEX_MARK_CACHE, diff --git a/src/Parsers/ASTUndropQuery.cpp b/src/Parsers/ASTUndropQuery.cpp index ab96ca4711e..7212e264c0e 100644 --- a/src/Parsers/ASTUndropQuery.cpp +++ b/src/Parsers/ASTUndropQuery.cpp @@ -36,6 +36,7 @@ void ASTUndropQuery::formatQueryImpl(const FormatSettings & settings, FormatStat settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); } diff --git a/src/Parsers/ASTWatchQuery.h b/src/Parsers/ASTWatchQuery.h index 80b675f513f..a5b76c07605 100644 --- a/src/Parsers/ASTWatchQuery.h +++ b/src/Parsers/ASTWatchQuery.h @@ -52,6 +52,7 @@ protected: settings.ostr << '.'; } + chassert(table); table->formatImpl(settings, state, frame); if (is_watch_events) diff --git a/src/Parsers/FunctionSecretArgumentsFinder.h b/src/Parsers/FunctionSecretArgumentsFinder.h new file mode 100644 index 00000000000..002ad94f6ea --- /dev/null +++ b/src/Parsers/FunctionSecretArgumentsFinder.h @@ -0,0 +1,28 @@ +#pragma once + +#include + +namespace DB +{ + +class FunctionSecretArgumentsFinder +{ +public: + struct Result + { + /// Result constructed by default means no arguments will be hidden. + size_t start = static_cast(-1); + size_t count = 0; /// Mostly it's either 0 or 1. There are only a few cases where `count` can be greater than 1 (e.g. see `encrypt`). + /// In all known cases secret arguments are consecutive + bool are_named = false; /// Arguments like `password = 'password'` are considered as named arguments. + /// E.g. "headers" in `url('..', headers('foo' = '[HIDDEN]'))` + std::vector nested_maps; + + bool hasSecrets() const + { + return count != 0 || !nested_maps.empty(); + } + }; +}; + +} diff --git a/src/Parsers/FunctionSecretArgumentsFinderAST.h b/src/Parsers/FunctionSecretArgumentsFinderAST.h new file mode 100644 index 00000000000..348b2ca9e3a --- /dev/null +++ b/src/Parsers/FunctionSecretArgumentsFinderAST.h @@ -0,0 +1,499 @@ +#pragma once + +#include +#include +#include +#include +#include +#include + +#include + + +namespace DB +{ + + +/// Finds arguments of a specified function which should not be displayed for most users for security reasons. +/// That involves passwords and secret keys. +class FunctionSecretArgumentsFinderAST +{ +public: + explicit FunctionSecretArgumentsFinderAST(const ASTFunction & function_) : function(function_) + { + if (!function.arguments) + return; + + const auto * expr_list = function.arguments->as(); + if (!expr_list) + return; + + arguments = &expr_list->children; + switch (function.kind) + { + case ASTFunction::Kind::ORDINARY_FUNCTION: findOrdinaryFunctionSecretArguments(); break; + case ASTFunction::Kind::WINDOW_FUNCTION: break; + case ASTFunction::Kind::LAMBDA_FUNCTION: break; + case ASTFunction::Kind::TABLE_ENGINE: findTableEngineSecretArguments(); break; + case ASTFunction::Kind::DATABASE_ENGINE: findDatabaseEngineSecretArguments(); break; + case ASTFunction::Kind::BACKUP_NAME: findBackupNameSecretArguments(); break; + } + } + + FunctionSecretArgumentsFinder::Result getResult() const { return result; } + +private: + const ASTFunction & function; + const ASTs * arguments = nullptr; + FunctionSecretArgumentsFinder::Result result; + + void markSecretArgument(size_t index, bool argument_is_named = false) + { + if (index >= arguments->size()) + return; + if (!result.count) + { + result.start = index; + result.are_named = argument_is_named; + } + chassert(index >= result.start); /// We always check arguments consecutively + result.count = index + 1 - result.start; + if (!argument_is_named) + result.are_named = false; + } + + void findOrdinaryFunctionSecretArguments() + { + if ((function.name == "mysql") || (function.name == "postgresql") || (function.name == "mongodb")) + { + /// mysql('host:port', 'database', 'table', 'user', 'password', ...) + /// postgresql('host:port', 'database', 'table', 'user', 'password', ...) + /// mongodb('host:port', 'database', 'collection', 'user', 'password', ...) + findMySQLFunctionSecretArguments(); + } + else if ((function.name == "s3") || (function.name == "cosn") || (function.name == "oss") || + (function.name == "deltaLake") || (function.name == "hudi") || (function.name == "iceberg")) + { + /// s3('url', 'aws_access_key_id', 'aws_secret_access_key', ...) + findS3FunctionSecretArguments(/* is_cluster_function= */ false); + } + else if (function.name == "s3Cluster") + { + /// s3Cluster('cluster_name', 'url', 'aws_access_key_id', 'aws_secret_access_key', ...) + findS3FunctionSecretArguments(/* is_cluster_function= */ true); + } + else if ((function.name == "remote") || (function.name == "remoteSecure")) + { + /// remote('addresses_expr', 'db', 'table', 'user', 'password', ...) + findRemoteFunctionSecretArguments(); + } + else if ((function.name == "encrypt") || (function.name == "decrypt") || + (function.name == "aes_encrypt_mysql") || (function.name == "aes_decrypt_mysql") || + (function.name == "tryDecrypt")) + { + /// encrypt('mode', 'plaintext', 'key' [, iv, aad]) + findEncryptionFunctionSecretArguments(); + } + else if (function.name == "url") + { + findURLSecretArguments(); + } + } + + void findMySQLFunctionSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// mysql(named_collection, ..., password = 'password', ...) + findSecretNamedArgument("password", 1); + } + else + { + /// mysql('host:port', 'database', 'table', 'user', 'password', ...) + markSecretArgument(4); + } + } + + /// Returns the number of arguments excluding "headers" and "extra_credentials" (which should + /// always be at the end). Marks "headers" as secret, if found. + size_t excludeS3OrURLNestedMaps() + { + size_t count = arguments->size(); + while (count > 0) + { + const ASTFunction * f = arguments->at(count - 1)->as(); + if (!f) + break; + if (f->name == "headers") + result.nested_maps.push_back(f->name); + else if (f->name != "extra_credentials") + break; + count -= 1; + } + return count; + } + + void findS3FunctionSecretArguments(bool is_cluster_function) + { + /// s3Cluster('cluster_name', 'url', ...) has 'url' as its second argument. + size_t url_arg_idx = is_cluster_function ? 1 : 0; + + if (!is_cluster_function && isNamedCollectionName(0)) + { + /// s3(named_collection, ..., secret_access_key = 'secret_access_key', ...) + findSecretNamedArgument("secret_access_key", 1); + return; + } + + /// We should check other arguments first because we don't need to do any replacement in case of + /// s3('url', NOSIGN, 'format' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) + /// s3('url', 'format', 'structure' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) + size_t count = excludeS3OrURLNestedMaps(); + if ((url_arg_idx + 3 <= count) && (count <= url_arg_idx + 4)) + { + String second_arg; + if (tryGetStringFromArgument(url_arg_idx + 1, &second_arg)) + { + if (boost::iequals(second_arg, "NOSIGN")) + return; /// The argument after 'url' is "NOSIGN". + + if (second_arg == "auto" || KnownFormatNames::instance().exists(second_arg)) + return; /// The argument after 'url' is a format: s3('url', 'format', ...) + } + } + + /// We're going to replace 'aws_secret_access_key' with '[HIDDEN]' for the following signatures: + /// s3('url', 'aws_access_key_id', 'aws_secret_access_key', ...) + /// s3Cluster('cluster_name', 'url', 'aws_access_key_id', 'aws_secret_access_key', 'format', 'compression') + if (url_arg_idx + 2 < count) + markSecretArgument(url_arg_idx + 2); + } + + void findURLSecretArguments() + { + if (!isNamedCollectionName(0)) + excludeS3OrURLNestedMaps(); + } + + bool tryGetStringFromArgument(size_t arg_idx, String * res, bool allow_identifier = true) const + { + if (arg_idx >= arguments->size()) + return false; + + return tryGetStringFromArgument(*(*arguments)[arg_idx], res, allow_identifier); + } + + static bool tryGetStringFromArgument(const IAST & argument, String * res, bool allow_identifier = true) + { + if (const auto * literal = argument.as()) + { + if (literal->value.getType() != Field::Types::String) + return false; + if (res) + *res = literal->value.safeGet(); + return true; + } + + if (allow_identifier) + { + if (const auto * id = argument.as()) + { + if (res) + *res = id->name(); + return true; + } + } + + return false; + } + + void findRemoteFunctionSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// remote(named_collection, ..., password = 'password', ...) + findSecretNamedArgument("password", 1); + return; + } + + /// We're going to replace 'password' with '[HIDDEN'] for the following signatures: + /// remote('addresses_expr', db.table, 'user' [, 'password'] [, sharding_key]) + /// remote('addresses_expr', 'db', 'table', 'user' [, 'password'] [, sharding_key]) + /// remote('addresses_expr', table_function(), 'user' [, 'password'] [, sharding_key]) + + /// But we should check the number of arguments first because we don't need to do any replacements in case of + /// remote('addresses_expr', db.table) + if (arguments->size() < 3) + return; + + size_t arg_num = 1; + + /// Skip 1 or 2 arguments with table_function() or db.table or 'db', 'table'. + const auto * table_function = (*arguments)[arg_num]->as(); + if (table_function && KnownTableFunctionNames::instance().exists(table_function->name)) + { + ++arg_num; + } + else + { + std::optional database; + std::optional qualified_table_name; + if (!tryGetDatabaseNameOrQualifiedTableName(arg_num, database, qualified_table_name)) + { + /// We couldn't evaluate the argument so we don't know whether it is 'db.table' or just 'db'. + /// Hence we can't figure out whether we should skip one argument 'user' or two arguments 'table', 'user' + /// before the argument 'password'. So it's safer to wipe two arguments just in case. + /// The last argument can be also a `sharding_key`, so we need to check that argument is a literal string + /// before wiping it (because the `password` argument is always a literal string). + if (tryGetStringFromArgument(arg_num + 2, nullptr, /* allow_identifier= */ false)) + { + /// Wipe either `password` or `user`. + markSecretArgument(arg_num + 2); + } + if (tryGetStringFromArgument(arg_num + 3, nullptr, /* allow_identifier= */ false)) + { + /// Wipe either `password` or `sharding_key`. + markSecretArgument(arg_num + 3); + } + return; + } + + /// Skip the current argument (which is either a database name or a qualified table name). + ++arg_num; + if (database) + { + /// Skip the 'table' argument if the previous argument was a database name. + ++arg_num; + } + } + + /// Skip username. + ++arg_num; + + /// Do our replacement: + /// remote('addresses_expr', db.table, 'user', 'password', ...) -> remote('addresses_expr', db.table, 'user', '[HIDDEN]', ...) + /// The last argument can be also a `sharding_key`, so we need to check that argument is a literal string + /// before wiping it (because the `password` argument is always a literal string). + bool can_be_password = tryGetStringFromArgument(arg_num, nullptr, /* allow_identifier= */ false); + if (can_be_password) + markSecretArgument(arg_num); + } + + /// Tries to get either a database name or a qualified table name from an argument. + /// Empty string is also allowed (it means the default database). + /// The function is used by findRemoteFunctionSecretArguments() to determine how many arguments to skip before a password. + bool tryGetDatabaseNameOrQualifiedTableName( + size_t arg_idx, + std::optional & res_database, + std::optional & res_qualified_table_name) const + { + res_database.reset(); + res_qualified_table_name.reset(); + + String str; + if (!tryGetStringFromArgument(arg_idx, &str, /* allow_identifier= */ true)) + return false; + + if (str.empty()) + { + res_database = ""; + return true; + } + + auto qualified_table_name = QualifiedTableName::tryParseFromString(str); + if (!qualified_table_name) + return false; + + if (qualified_table_name->database.empty()) + res_database = std::move(qualified_table_name->table); + else + res_qualified_table_name = std::move(qualified_table_name); + return true; + } + + void findEncryptionFunctionSecretArguments() + { + if (arguments->empty()) + return; + + /// We replace all arguments after 'mode' with '[HIDDEN]': + /// encrypt('mode', 'plaintext', 'key' [, iv, aad]) -> encrypt('mode', '[HIDDEN]') + result.start = 1; + result.count = arguments->size() - 1; + } + + void findTableEngineSecretArguments() + { + const String & engine_name = function.name; + if (engine_name == "ExternalDistributed") + { + /// ExternalDistributed('engine', 'host:port', 'database', 'table', 'user', 'password') + findExternalDistributedTableEngineSecretArguments(); + } + else if ((engine_name == "MySQL") || (engine_name == "PostgreSQL") || + (engine_name == "MaterializedPostgreSQL") || (engine_name == "MongoDB")) + { + /// MySQL('host:port', 'database', 'table', 'user', 'password', ...) + /// PostgreSQL('host:port', 'database', 'table', 'user', 'password', ...) + /// MaterializedPostgreSQL('host:port', 'database', 'table', 'user', 'password', ...) + /// MongoDB('host:port', 'database', 'collection', 'user', 'password', ...) + findMySQLFunctionSecretArguments(); + } + else if ((engine_name == "S3") || (engine_name == "COSN") || (engine_name == "OSS") || + (engine_name == "DeltaLake") || (engine_name == "Hudi") || (engine_name == "Iceberg") || (engine_name == "S3Queue")) + { + /// S3('url', ['aws_access_key_id', 'aws_secret_access_key',] ...) + findS3TableEngineSecretArguments(); + } + else if (engine_name == "URL") + { + findURLSecretArguments(); + } + } + + void findExternalDistributedTableEngineSecretArguments() + { + if (isNamedCollectionName(1)) + { + /// ExternalDistributed('engine', named_collection, ..., password = 'password', ...) + findSecretNamedArgument("password", 2); + } + else + { + /// ExternalDistributed('engine', 'host:port', 'database', 'table', 'user', 'password') + markSecretArgument(5); + } + } + + void findS3TableEngineSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// S3(named_collection, ..., secret_access_key = 'secret_access_key') + findSecretNamedArgument("secret_access_key", 1); + return; + } + + /// We should check other arguments first because we don't need to do any replacement in case of + /// S3('url', NOSIGN, 'format' [, 'compression'] [, extra_credentials(..)] [, headers(..)]) + /// S3('url', 'format', 'compression' [, extra_credentials(..)] [, headers(..)]) + size_t count = excludeS3OrURLNestedMaps(); + if ((3 <= count) && (count <= 4)) + { + String second_arg; + if (tryGetStringFromArgument(1, &second_arg)) + { + if (boost::iequals(second_arg, "NOSIGN")) + return; /// The argument after 'url' is "NOSIGN". + + if (count == 3) + { + if (second_arg == "auto" || KnownFormatNames::instance().exists(second_arg)) + return; /// The argument after 'url' is a format: S3('url', 'format', ...) + } + } + } + + /// We replace 'aws_secret_access_key' with '[HIDDEN]' for the following signatures: + /// S3('url', 'aws_access_key_id', 'aws_secret_access_key') + /// S3('url', 'aws_access_key_id', 'aws_secret_access_key', 'format') + /// S3('url', 'aws_access_key_id', 'aws_secret_access_key', 'format', 'compression') + if (2 < count) + markSecretArgument(2); + } + + void findDatabaseEngineSecretArguments() + { + const String & engine_name = function.name; + if ((engine_name == "MySQL") || (engine_name == "MaterializeMySQL") || + (engine_name == "MaterializedMySQL") || (engine_name == "PostgreSQL") || + (engine_name == "MaterializedPostgreSQL")) + { + /// MySQL('host:port', 'database', 'user', 'password') + /// PostgreSQL('host:port', 'database', 'user', 'password') + findMySQLDatabaseSecretArguments(); + } + else if (engine_name == "S3") + { + /// S3('url', 'access_key_id', 'secret_access_key') + findS3DatabaseSecretArguments(); + } + } + + void findMySQLDatabaseSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// MySQL(named_collection, ..., password = 'password', ...) + findSecretNamedArgument("password", 1); + } + else + { + /// MySQL('host:port', 'database', 'user', 'password') + markSecretArgument(3); + } + } + + void findS3DatabaseSecretArguments() + { + if (isNamedCollectionName(0)) + { + /// S3(named_collection, ..., secret_access_key = 'password', ...) + findSecretNamedArgument("secret_access_key", 1); + } + else + { + /// S3('url', 'access_key_id', 'secret_access_key') + markSecretArgument(2); + } + } + + void findBackupNameSecretArguments() + { + const String & engine_name = function.name; + if (engine_name == "S3") + { + /// BACKUP ... TO S3(url, [aws_access_key_id, aws_secret_access_key]) + markSecretArgument(2); + } + } + + /// Whether a specified argument can be the name of a named collection? + bool isNamedCollectionName(size_t arg_idx) const + { + if (arguments->size() <= arg_idx) + return false; + + const auto * identifier = (*arguments)[arg_idx]->as(); + return identifier != nullptr; + } + + /// Looks for a secret argument with a specified name. This function looks for arguments in format `key=value` where the key is specified. + void findSecretNamedArgument(const std::string_view & key, size_t start = 0) + { + for (size_t i = start; i < arguments->size(); ++i) + { + const auto & argument = (*arguments)[i]; + const auto * equals_func = argument->as(); + if (!equals_func || (equals_func->name != "equals")) + continue; + + const auto * expr_list = equals_func->arguments->as(); + if (!expr_list) + continue; + + const auto & equal_args = expr_list->children; + if (equal_args.size() != 2) + continue; + + String found_key; + if (!tryGetStringFromArgument(*equal_args[0], &found_key)) + continue; + + if (found_key == key) + markSecretArgument(i, /* argument_is_named= */ true); + } + } +}; + +} diff --git a/src/Parsers/Kusto/ParserKQLQuery.cpp b/src/Parsers/Kusto/ParserKQLQuery.cpp index 29b26b4e3fe..a54a2b0eda9 100644 --- a/src/Parsers/Kusto/ParserKQLQuery.cpp +++ b/src/Parsers/Kusto/ParserKQLQuery.cpp @@ -416,8 +416,9 @@ bool ParserKQLQuery::parseImpl(Pos & pos, ASTPtr & node, Expected & expected) ParserToken s_dash(TokenType::Minus); if (s_dash.ignore(pos, expected)) { - String tmp_op(op_pos_begin->begin, pos->end); - kql_operator = tmp_op; + if (!isValidKQLPos(pos)) + return false; + kql_operator = String(op_pos_begin->begin, pos->end); } else --pos; diff --git a/src/Parsers/ParserQueryWithOutput.cpp b/src/Parsers/ParserQueryWithOutput.cpp index 7a627ae5f6a..4a0b928608b 100644 --- a/src/Parsers/ParserQueryWithOutput.cpp +++ b/src/Parsers/ParserQueryWithOutput.cpp @@ -136,6 +136,7 @@ bool ParserQueryWithOutput::parseImpl(Pos & pos, ASTPtr & node, Expected & expec ParserStringLiteral compression; if (!compression.parse(pos, query_with_output.compression, expected)) return false; + query_with_output.children.push_back(query_with_output.compression); ParserKeyword s_compression_level("LEVEL"); if (s_compression_level.ignore(pos, expected)) @@ -143,6 +144,7 @@ bool ParserQueryWithOutput::parseImpl(Pos & pos, ASTPtr & node, Expected & expec ParserNumber compression_level; if (!compression_level.parse(pos, query_with_output.compression_level, expected)) return false; + query_with_output.children.push_back(query_with_output.compression_level); } } diff --git a/src/Planner/CollectTableExpressionData.cpp b/src/Planner/CollectTableExpressionData.cpp index 78a7c7074c3..385381f1355 100644 --- a/src/Planner/CollectTableExpressionData.cpp +++ b/src/Planner/CollectTableExpressionData.cpp @@ -29,34 +29,13 @@ namespace class CollectSourceColumnsVisitor : public InDepthQueryTreeVisitor { public: - explicit CollectSourceColumnsVisitor(PlannerContext & planner_context_) + explicit CollectSourceColumnsVisitor(PlannerContextPtr & planner_context_, bool keep_alias_columns_ = true) : planner_context(planner_context_) + , keep_alias_columns(keep_alias_columns_) {} void visitImpl(QueryTreeNodePtr & node) { - /// Special case for USING clause which contains references to ALIAS columns. - /// We can not modify such ColumnNode. - if (auto * join_node = node->as()) - { - if (!join_node->isUsingJoinExpression()) - return; - - auto & using_list = join_node->getJoinExpression()->as(); - for (auto & using_element : using_list) - { - auto & column_node = using_element->as(); - /// This list contains column nodes from left and right tables. - auto & columns_from_subtrees = column_node.getExpressionOrThrow()->as().getNodes(); - - /// Visit left table column node. - visitUsingColumn(columns_from_subtrees[0]); - /// Visit right table column node. - visitUsingColumn(columns_from_subtrees[1]); - } - return; - } - auto * column_node = node->as(); if (!column_node) return; @@ -72,22 +51,55 @@ public: /// JOIN using expression if (column_node->hasExpression() && column_source_node_type == QueryTreeNodeType::JOIN) - return; - - auto & table_expression_data = planner_context.getOrCreateTableExpressionData(column_source_node); - - if (column_node->hasExpression() && column_source_node_type != QueryTreeNodeType::ARRAY_JOIN) { - /// Replace ALIAS column with expression + auto & columns_from_subtrees = column_node->getExpression()->as().getNodes(); + if (columns_from_subtrees.size() != 2) + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Expected two columns in JOIN using expression for column {}", column_node->dumpTree()); + + visit(columns_from_subtrees[0]); + visit(columns_from_subtrees[1]); + return; + } + + auto & table_expression_data = planner_context->getOrCreateTableExpressionData(column_source_node); + + if (isAliasColumn(node)) + { + /// Column is an ALIAS column with expression bool column_already_exists = table_expression_data.hasColumn(column_node->getColumnName()); if (!column_already_exists) { - auto column_identifier = planner_context.getGlobalPlannerContext()->createColumnIdentifier(node); - table_expression_data.addAliasColumnName(column_node->getColumnName(), column_identifier); + CollectSourceColumnsVisitor visitor_for_alias_column(planner_context); + /// While we are processing expression of ALIAS columns we should not add source columns to selected. + /// See also comment for `select_added_columns` + visitor_for_alias_column.select_added_columns = false; + visitor_for_alias_column.keep_alias_columns = keep_alias_columns; + visitor_for_alias_column.visit(column_node->getExpression()); + + if (!keep_alias_columns) + { + /// For PREWHERE we can just replace ALIAS column with it's expression, + /// because ActionsDAG for PREWHERE applied right on top of table expression + /// and cannot affect subqueries or other table expressions. + node = column_node->getExpression(); + return; + } + + auto column_identifier = planner_context->getGlobalPlannerContext()->createColumnIdentifier(node); + + ActionsDAGPtr alias_column_actions_dag = std::make_shared(); + PlannerActionsVisitor actions_visitor(planner_context, false); + auto outputs = actions_visitor.visit(alias_column_actions_dag, column_node->getExpression()); + if (outputs.size() != 1) + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Expected single output in actions dag for alias column {}. Actual {}", column_node->dumpTree(), outputs.size()); + const auto & column_name = column_node->getColumnName(); + const auto & alias_node = alias_column_actions_dag->addAlias(*outputs[0], column_name); + alias_column_actions_dag->addOrReplaceInOutputs(alias_node); + table_expression_data.addAliasColumn(column_node->getColumn(), column_identifier, alias_column_actions_dag, select_added_columns); } - node = column_node->getExpression(); - visitImpl(node); return; } @@ -102,45 +114,58 @@ public: bool column_already_exists = table_expression_data.hasColumn(column_node->getColumnName()); if (column_already_exists) + { + /// Column may be added when we collected data for ALIAS column + /// But now we see it directly in the query, so make sure it's marked as selected + if (select_added_columns) + table_expression_data.markSelectedColumn(column_node->getColumnName()); return; + } - auto column_identifier = planner_context.getGlobalPlannerContext()->createColumnIdentifier(node); - table_expression_data.addColumn(column_node->getColumn(), column_identifier); + auto column_identifier = planner_context->getGlobalPlannerContext()->createColumnIdentifier(node); + table_expression_data.addColumn(column_node->getColumn(), column_identifier, select_added_columns); } - static bool needChildVisit(const QueryTreeNodePtr & parent, const QueryTreeNodePtr & child_node) + static bool isAliasColumn(const QueryTreeNodePtr & node) + { + const auto * column_node = node->as(); + if (!column_node || !column_node->hasExpression()) + return false; + const auto & column_source = column_node->getColumnSourceOrNull(); + if (!column_source) + return false; + return column_source->getNodeType() != QueryTreeNodeType::JOIN && + column_source->getNodeType() != QueryTreeNodeType::ARRAY_JOIN; + } + + static bool needChildVisit(const QueryTreeNodePtr & parent_node, const QueryTreeNodePtr & child_node) { - if (auto * join_node = parent->as()) - { - if (join_node->getJoinExpression() == child_node && join_node->isUsingJoinExpression()) - return false; - } auto child_node_type = child_node->getNodeType(); - return !(child_node_type == QueryTreeNodeType::QUERY || child_node_type == QueryTreeNodeType::UNION); + return !(child_node_type == QueryTreeNodeType::QUERY || + child_node_type == QueryTreeNodeType::UNION || + isAliasColumn(parent_node)); + } + + void setKeepAliasColumns(bool keep_alias_columns_) + { + keep_alias_columns = keep_alias_columns_; } private: + PlannerContextPtr & planner_context; - void visitUsingColumn(QueryTreeNodePtr & node) - { - auto & column_node = node->as(); - if (column_node.hasExpression()) - { - auto & table_expression_data = planner_context.getOrCreateTableExpressionData(column_node.getColumnSource()); - bool column_already_exists = table_expression_data.hasColumn(column_node.getColumnName()); - if (column_already_exists) - return; + /// Replace ALIAS columns with their expressions or register them in table expression data. + /// Usually we can replace them when we build some "local" actions DAG + /// (for example Row Policy or PREWHERE) that is applied on top of the table expression. + /// In other cases, we keep ALIAS columns as ColumnNode with an expression child node, + /// and handle them in the Planner by inserting ActionsDAG to compute them after reading from storage. + bool keep_alias_columns = true; - auto column_identifier = planner_context.getGlobalPlannerContext()->createColumnIdentifier(node); - table_expression_data.addAliasColumnName(column_node.getColumnName(), column_identifier); - - visitImpl(column_node.getExpressionOrThrow()); - } - else - visitImpl(node); - } - - PlannerContext & planner_context; + /// Flag `select_added_columns` indicates if we should mark column as explicitly selected. + /// For example, for table with columns (a Int32, b ALIAS a+1) and query SELECT b FROM table + /// Column `b` is selected explicitly by user, but not `a` (that is also read though). + /// Distinguishing such columns is important for checking access rights for ALIAS columns. + bool select_added_columns = true; }; class CollectPrewhereTableExpressionVisitor : public ConstInDepthQueryTreeVisitor @@ -274,7 +299,7 @@ void collectTableExpressionData(QueryTreeNodePtr & query_node, PlannerContextPtr } } - CollectSourceColumnsVisitor collect_source_columns_visitor(*planner_context); + CollectSourceColumnsVisitor collect_source_columns_visitor(planner_context); for (auto & node : query_node_typed.getChildren()) { if (!node || node == query_node_typed.getPrewhere()) @@ -300,21 +325,26 @@ void collectTableExpressionData(QueryTreeNodePtr & query_node, PlannerContextPtr } auto & table_expression_data = planner_context->getOrCreateTableExpressionData(prewhere_table_expression); - const auto & column_names = table_expression_data.getColumnNames(); - NameSet required_column_names_without_prewhere(column_names.begin(), column_names.end()); + const auto & read_column_names = table_expression_data.getColumnNames(); + NameSet required_column_names_without_prewhere(read_column_names.begin(), read_column_names.end()); + const auto & selected_column_names = table_expression_data.getSelectedColumnsNames(); + required_column_names_without_prewhere.insert(selected_column_names.begin(), selected_column_names.end()); + collect_source_columns_visitor.setKeepAliasColumns(false); collect_source_columns_visitor.visit(query_node_typed.getPrewhere()); auto prewhere_actions_dag = std::make_shared(); + QueryTreeNodePtr query_tree_node = query_node_typed.getPrewhere(); + PlannerActionsVisitor visitor(planner_context, false /*use_column_identifier_as_action_node_name*/); - auto expression_nodes = visitor.visit(prewhere_actions_dag, query_node_typed.getPrewhere()); + auto expression_nodes = visitor.visit(prewhere_actions_dag, query_tree_node); if (expression_nodes.size() != 1) throw Exception(ErrorCodes::ILLEGAL_PREWHERE, "Invalid PREWHERE. Expected single boolean expression. In query {}", query_node->formatASTForErrorMessage()); - prewhere_actions_dag->getOutputs().push_back(expression_nodes[0]); + prewhere_actions_dag->getOutputs().push_back(expression_nodes.back()); for (const auto & prewhere_input_node : prewhere_actions_dag->getInputs()) if (required_column_names_without_prewhere.contains(prewhere_input_node->result_name)) @@ -324,9 +354,9 @@ void collectTableExpressionData(QueryTreeNodePtr & query_node, PlannerContextPtr } } -void collectSourceColumns(QueryTreeNodePtr & expression_node, PlannerContextPtr & planner_context) +void collectSourceColumns(QueryTreeNodePtr & expression_node, PlannerContextPtr & planner_context, bool keep_alias_columns) { - CollectSourceColumnsVisitor collect_source_columns_visitor(*planner_context); + CollectSourceColumnsVisitor collect_source_columns_visitor(planner_context, keep_alias_columns); collect_source_columns_visitor.visit(expression_node); } diff --git a/src/Planner/CollectTableExpressionData.h b/src/Planner/CollectTableExpressionData.h index ed3f0ff7a47..b0cebc15682 100644 --- a/src/Planner/CollectTableExpressionData.h +++ b/src/Planner/CollectTableExpressionData.h @@ -19,6 +19,6 @@ void collectTableExpressionData(QueryTreeNodePtr & query_node, PlannerContextPtr * * ALIAS table column nodes are registered in table expression data and replaced in query tree with inner alias expression. */ -void collectSourceColumns(QueryTreeNodePtr & expression_node, PlannerContextPtr & planner_context); +void collectSourceColumns(QueryTreeNodePtr & expression_node, PlannerContextPtr & planner_context, bool keep_alias_columns = true); } diff --git a/src/Planner/Planner.cpp b/src/Planner/Planner.cpp index efd8f4faa42..bc1fb30781d 100644 --- a/src/Planner/Planner.cpp +++ b/src/Planner/Planner.cpp @@ -1181,7 +1181,7 @@ PlannerContextPtr buildPlannerContext(const QueryTreeNodePtr & query_tree_node, if (select_query_options.is_subquery) updateContextForSubqueryExecution(mutable_context); - return std::make_shared(mutable_context, std::move(global_planner_context)); + return std::make_shared(mutable_context, std::move(global_planner_context), select_query_options); } Planner::Planner(const QueryTreeNodePtr & query_tree_, @@ -1373,7 +1373,7 @@ void Planner::buildPlanForQueryNode() const auto & settings = query_context->getSettingsRef(); if (query_context->canUseTaskBasedParallelReplicas()) { - if (planner_context->getPreparedSets().hasSubqueries()) + if (!settings.parallel_replicas_allow_in_with_subquery && planner_context->getPreparedSets().hasSubqueries()) { if (settings.allow_experimental_parallel_reading_from_replicas >= 2) throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "IN with subquery is not supported with parallel replicas"); diff --git a/src/Planner/PlannerActionsVisitor.cpp b/src/Planner/PlannerActionsVisitor.cpp index 511e9396a35..af23e684f23 100644 --- a/src/Planner/PlannerActionsVisitor.cpp +++ b/src/Planner/PlannerActionsVisitor.cpp @@ -44,6 +44,27 @@ namespace ErrorCodes namespace { +/* Calculates Action node name for ConstantNode. + * + * If converting to AST will add a '_CAST' function call, + * the result action name will also include it. + */ +String calculateActionNodeNameWithCastIfNeeded(const ConstantNode & constant_node) +{ + WriteBufferFromOwnString buffer; + if (constant_node.requiresCastCall()) + buffer << "_CAST("; + + buffer << calculateConstantActionNodeName(constant_node.getValue(), constant_node.getResultType()); + + if (constant_node.requiresCastCall()) + { + buffer << ", '" << constant_node.getResultType()->getName() << "'_String)"; + } + + return buffer.str(); +} + class ActionNodeNameHelper { public: @@ -88,7 +109,49 @@ public: case QueryTreeNodeType::CONSTANT: { const auto & constant_node = node->as(); - result = calculateConstantActionNodeName(constant_node.getValue(), constant_node.getResultType()); + /* To ensure that headers match during distributed query we need to simulate action node naming on + * secondary servers. If we don't do that headers will mismatch due to constant folding. + * + * +--------+ + * -----------------| Server |---------------- + * / +--------+ \ + * / \ + * v v + * +-----------+ +-----------+ + * | Initiator | ------ | Secondary |------ + * +-----------+ / +-----------+ \ + * | / \ + * | / \ + * v / \ + * +---------------+ v v + * | Wrap in _CAST | +----------------------------+ +----------------------+ + * | if needed | | Constant folded from _CAST | | Constant folded from | + * +---------------+ +----------------------------+ | another expression | + * | +----------------------+ + * v | + * +----------------------------+ v + * | Name ConstantNode the same | +--------------------------+ + * | as on initiator server | | Generate action name for | + * | (wrap in _CAST if needed) | | original expression | + * +----------------------------+ +--------------------------+ + */ + if (planner_context.isASTLevelOptimizationAllowed()) + { + result = calculateActionNodeNameWithCastIfNeeded(constant_node); + } + else + { + // Need to check if constant folded from QueryNode until https://github.com/ClickHouse/ClickHouse/issues/60847 is fixed. + if (constant_node.hasSourceExpression() && constant_node.getSourceExpression()->getNodeType() != QueryTreeNodeType::QUERY) + { + if (constant_node.receivedFromInitiatorServer()) + result = calculateActionNodeNameWithCastIfNeeded(constant_node); + else + result = calculateActionNodeName(constant_node.getSourceExpression()); + } + else + result = calculateConstantActionNodeName(constant_node.getValue(), constant_node.getResultType()); + } break; } case QueryTreeNodeType::FUNCTION: @@ -451,6 +514,7 @@ private: std::unordered_map node_to_node_name; const PlannerContextPtr planner_context; ActionNodeNameHelper action_node_name_helper; + bool use_column_identifier_as_action_node_name; }; PlannerActionsVisitorImpl::PlannerActionsVisitorImpl(ActionsDAGPtr actions_dag, @@ -458,6 +522,7 @@ PlannerActionsVisitorImpl::PlannerActionsVisitorImpl(ActionsDAGPtr actions_dag, bool use_column_identifier_as_action_node_name_) : planner_context(planner_context_) , action_node_name_helper(node_to_node_name, *planner_context, use_column_identifier_as_action_node_name_) + , use_column_identifier_as_action_node_name(use_column_identifier_as_action_node_name_) { actions_stack.emplace_back(std::move(actions_dag), nullptr); } @@ -503,7 +568,8 @@ PlannerActionsVisitorImpl::NodeNameAndNodeMinLevel PlannerActionsVisitorImpl::vi { auto column_node_name = action_node_name_helper.calculateActionNodeName(node); const auto & column_node = node->as(); - + if (column_node.hasExpression() && !use_column_identifier_as_action_node_name) + return visitImpl(column_node.getExpression()); Int64 actions_stack_size = static_cast(actions_stack.size() - 1); for (Int64 i = actions_stack_size; i >= 0; --i) { @@ -527,7 +593,52 @@ PlannerActionsVisitorImpl::NodeNameAndNodeMinLevel PlannerActionsVisitorImpl::vi const auto & constant_literal = constant_node.getValue(); const auto & constant_type = constant_node.getResultType(); - auto constant_node_name = calculateConstantActionNodeName(constant_literal, constant_type); + auto constant_node_name = [&]() + { + /* To ensure that headers match during distributed query we need to simulate action node naming on + * secondary servers. If we don't do that headers will mismatch due to constant folding. + * + * +--------+ + * -----------------| Server |---------------- + * / +--------+ \ + * / \ + * v v + * +-----------+ +-----------+ + * | Initiator | ------ | Secondary |------ + * +-----------+ / +-----------+ \ + * | / \ + * | / \ + * v / \ + * +---------------+ v v + * | Wrap in _CAST | +----------------------------+ +----------------------+ + * | if needed | | Constant folded from _CAST | | Constant folded from | + * +---------------+ +----------------------------+ | another expression | + * | +----------------------+ + * v | + * +----------------------------+ v + * | Name ConstantNode the same | +--------------------------+ + * | as on initiator server | | Generate action name for | + * | (wrap in _CAST if needed) | | original expression | + * +----------------------------+ +--------------------------+ + */ + if (planner_context->isASTLevelOptimizationAllowed()) + { + return calculateActionNodeNameWithCastIfNeeded(constant_node); + } + else + { + // Need to check if constant folded from QueryNode until https://github.com/ClickHouse/ClickHouse/issues/60847 is fixed. + if (constant_node.hasSourceExpression() && constant_node.getSourceExpression()->getNodeType() != QueryTreeNodeType::QUERY) + { + if (constant_node.receivedFromInitiatorServer()) + return calculateActionNodeNameWithCastIfNeeded(constant_node); + else + return action_node_name_helper.calculateActionNodeName(constant_node.getSourceExpression()); + } + else + return calculateConstantActionNodeName(constant_literal, constant_type); + } + }(); ColumnWithTypeAndName column; column.name = constant_node_name; diff --git a/src/Planner/PlannerContext.cpp b/src/Planner/PlannerContext.cpp index 422c8c1d01f..f33255f0a44 100644 --- a/src/Planner/PlannerContext.cpp +++ b/src/Planner/PlannerContext.cpp @@ -3,6 +3,7 @@ #include #include #include +#include namespace DB { @@ -41,9 +42,10 @@ bool GlobalPlannerContext::hasColumnIdentifier(const ColumnIdentifier & column_i return column_identifiers.contains(column_identifier); } -PlannerContext::PlannerContext(ContextMutablePtr query_context_, GlobalPlannerContextPtr global_planner_context_) +PlannerContext::PlannerContext(ContextMutablePtr query_context_, GlobalPlannerContextPtr global_planner_context_, const SelectQueryOptions & select_query_options_) : query_context(std::move(query_context_)) , global_planner_context(std::move(global_planner_context_)) + , is_ast_level_optimization_allowed(!(query_context->getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY || select_query_options_.ignore_ast_optimizations)) {} TableExpressionData & PlannerContext::getOrCreateTableExpressionData(const QueryTreeNodePtr & table_expression_node) diff --git a/src/Planner/PlannerContext.h b/src/Planner/PlannerContext.h index 853f55a5f00..4d9ba037cac 100644 --- a/src/Planner/PlannerContext.h +++ b/src/Planner/PlannerContext.h @@ -10,6 +10,7 @@ #include #include +#include namespace DB { @@ -78,7 +79,7 @@ class PlannerContext { public: /// Create planner context with query context and global planner context - PlannerContext(ContextMutablePtr query_context_, GlobalPlannerContextPtr global_planner_context_); + PlannerContext(ContextMutablePtr query_context_, GlobalPlannerContextPtr global_planner_context_, const SelectQueryOptions & select_query_options_); /// Get planner context query context ContextPtr getQueryContext() const @@ -165,6 +166,12 @@ public: static SetKey createSetKey(const DataTypePtr & left_operand_type, const QueryTreeNodePtr & set_source_node); PreparedSets & getPreparedSets() { return prepared_sets; } + + /// Returns false if any of following conditions met: + /// 1. Query is executed on a follower node. + /// 2. ignore_ast_optimizations is set. + bool isASTLevelOptimizationAllowed() const { return is_ast_level_optimization_allowed; } + private: /// Query context ContextMutablePtr query_context; @@ -172,6 +179,8 @@ private: /// Global planner context GlobalPlannerContextPtr global_planner_context; + bool is_ast_level_optimization_allowed; + /// Column node to column identifier std::unordered_map column_node_to_column_identifier; diff --git a/src/Planner/PlannerJoinTree.cpp b/src/Planner/PlannerJoinTree.cpp index 59da88f4e45..7b3fb0c5c91 100644 --- a/src/Planner/PlannerJoinTree.cpp +++ b/src/Planner/PlannerJoinTree.cpp @@ -86,7 +86,7 @@ namespace /// Check if current user has privileges to SELECT columns from table /// Throws an exception if access to any column from `column_names` is not granted /// If `column_names` is empty, check access to any columns and return names of accessible columns -NameSet checkAccessRights(const TableNode & table_node, Names & column_names, const ContextPtr & query_context) +NameSet checkAccessRights(const TableNode & table_node, const Names & column_names, const ContextPtr & query_context) { /// StorageDummy is created on preliminary stage, ignore access check for it. if (typeid_cast(table_node.getStorage().get())) @@ -353,9 +353,7 @@ void prepareBuildQueryPlanForTableExpression(const QueryTreeNodePtr & table_expr NameSet columns_names_allowed_to_select; if (table_node) { - auto column_names_with_aliases = columns_names; - const auto & alias_columns_names = table_expression_data.getAliasColumnsNames(); - column_names_with_aliases.insert(column_names_with_aliases.end(), alias_columns_names.begin(), alias_columns_names.end()); + const auto & column_names_with_aliases = table_expression_data.getSelectedColumnsNames(); columns_names_allowed_to_select = checkAccessRights(*table_node, column_names_with_aliases, query_context); } @@ -864,6 +862,28 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres max_block_size, max_streams); + const auto & alias_column_expressions = table_expression_data.getAliasColumnExpressions(); + if (!alias_column_expressions.empty() && query_plan.isInitialized() && from_stage == QueryProcessingStage::FetchColumns) + { + ActionsDAGPtr merged_alias_columns_actions_dag = std::make_shared(query_plan.getCurrentDataStream().header.getColumnsWithTypeAndName()); + ActionsDAG::NodeRawConstPtrs action_dag_outputs = merged_alias_columns_actions_dag->getInputs(); + + for (const auto & [column_name, alias_column_actions_dag] : alias_column_expressions) + { + const auto & current_outputs = alias_column_actions_dag->getOutputs(); + action_dag_outputs.insert(action_dag_outputs.end(), current_outputs.begin(), current_outputs.end()); + merged_alias_columns_actions_dag->mergeNodes(std::move(*alias_column_actions_dag)); + } + + for (const auto * output_node : action_dag_outputs) + merged_alias_columns_actions_dag->addOrReplaceInOutputs(*output_node); + merged_alias_columns_actions_dag->removeUnusedActions(false); + + auto alias_column_step = std::make_unique(query_plan.getCurrentDataStream(), std::move(merged_alias_columns_actions_dag)); + alias_column_step->setStepDescription("Compute alias columns"); + query_plan.addStep(std::move(alias_column_step)); + } + for (const auto & filter_info_and_description : where_filters) { const auto & [filter_info, description] = filter_info_and_description; @@ -907,7 +927,8 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres else { /// Create step which reads from empty source if storage has no data. - auto source_header = storage_snapshot->getSampleBlockForColumns(table_expression_data.getColumnNames()); + const auto & column_names = table_expression_data.getSelectedColumnsNames(); + auto source_header = storage_snapshot->getSampleBlockForColumns(column_names); Pipe pipe(std::make_shared(source_header)); auto read_from_pipe = std::make_unique(std::move(pipe)); read_from_pipe->setStepDescription("Read from NullSource"); @@ -1024,57 +1045,6 @@ void joinCastPlanColumnsToNullable(QueryPlan & plan_to_add_cast, PlannerContextP plan_to_add_cast.addStep(std::move(cast_join_columns_step)); } -/// Actions to calculate table columns that have a functional representation (ALIASes and subcolumns) -/// and used in USING clause of JOIN expression. -struct UsingAliasKeyActions -{ - UsingAliasKeyActions( - const ColumnsWithTypeAndName & left_plan_output_columns, - const ColumnsWithTypeAndName & right_plan_output_columns - ) - : left_alias_columns_keys(std::make_shared(left_plan_output_columns)) - , right_alias_columns_keys(std::make_shared(right_plan_output_columns)) - {} - - void addLeftColumn(QueryTreeNodePtr & node, const ColumnsWithTypeAndName & plan_output_columns, const PlannerContextPtr & planner_context) - { - addColumnImpl(left_alias_columns_keys, node, plan_output_columns, planner_context); - } - - void addRightColumn(QueryTreeNodePtr & node, const ColumnsWithTypeAndName & plan_output_columns, const PlannerContextPtr & planner_context) - { - addColumnImpl(right_alias_columns_keys, node, plan_output_columns, planner_context); - } - - ActionsDAGPtr getLeftActions() - { - left_alias_columns_keys->projectInput(); - return std::move(left_alias_columns_keys); - } - - ActionsDAGPtr getRightActions() - { - right_alias_columns_keys->projectInput(); - return std::move(right_alias_columns_keys); - } - -private: - void addColumnImpl(ActionsDAGPtr & alias_columns_keys, QueryTreeNodePtr & node, const ColumnsWithTypeAndName & plan_output_columns, const PlannerContextPtr & planner_context) - { - auto & column_node = node->as(); - if (column_node.hasExpression()) - { - auto dag = buildActionsDAGFromExpressionNode(column_node.getExpressionOrThrow(), plan_output_columns, planner_context); - const auto & left_inner_column_identifier = planner_context->getColumnNodeIdentifierOrThrow(node); - dag->addOrReplaceInOutputs(dag->addAlias(*dag->getOutputs().front(), left_inner_column_identifier)); - alias_columns_keys->mergeInplace(std::move(*dag)); - } - } - - ActionsDAGPtr left_alias_columns_keys; - ActionsDAGPtr right_alias_columns_keys; -}; - JoinTreeQueryPlan buildQueryPlanForJoinNode(const QueryTreeNodePtr & join_table_expression, JoinTreeQueryPlan left_join_tree_query_plan, JoinTreeQueryPlan right_join_tree_query_plan, @@ -1143,8 +1113,6 @@ JoinTreeQueryPlan buildQueryPlanForJoinNode(const QueryTreeNodePtr & join_table_ if (join_node.isUsingJoinExpression()) { - UsingAliasKeyActions using_alias_key_actions{left_plan_output_columns, right_plan_output_columns}; - auto & join_node_using_columns_list = join_node.getJoinExpression()->as(); for (auto & join_node_using_node : join_node_using_columns_list.getNodes()) { @@ -1154,13 +1122,9 @@ JoinTreeQueryPlan buildQueryPlanForJoinNode(const QueryTreeNodePtr & join_table_ auto & left_inner_column_node = inner_columns_list.getNodes().at(0); auto & left_inner_column = left_inner_column_node->as(); - using_alias_key_actions.addLeftColumn(left_inner_column_node, left_plan_output_columns, planner_context); - auto & right_inner_column_node = inner_columns_list.getNodes().at(1); auto & right_inner_column = right_inner_column_node->as(); - using_alias_key_actions.addRightColumn(right_inner_column_node, right_plan_output_columns, planner_context); - const auto & join_node_using_column_node_type = join_node_using_column_node.getColumnType(); if (!left_inner_column.getColumnType()->equals(*join_node_using_column_node_type)) { @@ -1174,14 +1138,6 @@ JoinTreeQueryPlan buildQueryPlanForJoinNode(const QueryTreeNodePtr & join_table_ right_plan_column_name_to_cast_type.emplace(right_inner_column_identifier, join_node_using_column_node_type); } } - - auto left_alias_columns_keys_step = std::make_unique(left_plan.getCurrentDataStream(), using_alias_key_actions.getLeftActions()); - left_alias_columns_keys_step->setStepDescription("Actions for left table alias column keys"); - left_plan.addStep(std::move(left_alias_columns_keys_step)); - - auto right_alias_columns_keys_step = std::make_unique(right_plan.getCurrentDataStream(), using_alias_key_actions.getRightActions()); - right_alias_columns_keys_step->setStepDescription("Actions for right table alias column keys"); - right_plan.addStep(std::move(right_alias_columns_keys_step)); } auto join_cast_plan_output_nodes = [&](QueryPlan & plan_to_add_cast, std::unordered_map & plan_column_name_to_cast_type) diff --git a/src/Planner/TableExpressionData.h b/src/Planner/TableExpressionData.h index 20c4f05ea7e..9ab7a8e64fe 100644 --- a/src/Planner/TableExpressionData.h +++ b/src/Planner/TableExpressionData.h @@ -55,7 +55,7 @@ public: /// Return true if column with name exists, false otherwise bool hasColumn(const std::string & column_name) const { - return alias_columns_names.contains(column_name) || column_name_to_column.contains(column_name); + return column_name_to_column.contains(column_name); } /** Add column in table expression data. @@ -63,37 +63,40 @@ public: * * Logical error exception is thrown if column already exists. */ - void addColumn(const NameAndTypePair & column, const ColumnIdentifier & column_identifier) + void addColumn(const NameAndTypePair & column, const ColumnIdentifier & column_identifier, bool is_selected_column = true) { if (hasColumn(column.name)) throw Exception(ErrorCodes::LOGICAL_ERROR, "Column with name {} already exists", column.name); - addColumnImpl(column, column_identifier); + column_names.push_back(column.name); + addColumnImpl(column, column_identifier, is_selected_column); } - /** Add column if it does not exists in table expression data. - * Column identifier must be created using global planner context. - */ - void addColumnIfNotExists(const NameAndTypePair & column, const ColumnIdentifier & column_identifier) + /// Add alias column + void addAliasColumn(const NameAndTypePair & column, const ColumnIdentifier & column_identifier, ActionsDAGPtr actions_dag, bool is_selected_column = true) { - if (hasColumn(column.name)) - return; - - addColumnImpl(column, column_identifier); + alias_column_expressions.emplace(column.name, std::move(actions_dag)); + addColumnImpl(column, column_identifier, is_selected_column); } - /// Add alias column name - void addAliasColumnName(const std::string & column_name, const ColumnIdentifier & column_identifier) + /// Mark existing column as selected + void markSelectedColumn(const std::string & column_name) { - alias_columns_names.insert(column_name); - - column_name_to_column_identifier.emplace(column_name, column_identifier); + auto [_, inserted] = selected_column_names_set.emplace(column_name); + if (inserted) + selected_column_names.push_back(column_name); } - /// Get alias columns names - const NameSet & getAliasColumnsNames() const + /// Get columns that are requested from table expression, including ALIAS columns + const Names & getSelectedColumnsNames() const { - return alias_columns_names; + return selected_column_names; + } + + /// Get ALIAS columns names mapped to expressions + const std::unordered_map & getAliasColumnExpressions() const + { + return alias_column_expressions; } /// Get column name to column map @@ -102,7 +105,7 @@ public: return column_name_to_column; } - /// Get column names + /// Get column names that are read from table expression const Names & getColumnNames() const { return column_names; @@ -119,23 +122,6 @@ public: return result; } - ColumnIdentifiers getColumnIdentifiers() const - { - ColumnIdentifiers result; - result.reserve(column_identifier_to_column_name.size()); - - for (const auto & [column_identifier, _] : column_identifier_to_column_name) - result.push_back(column_identifier); - - return result; - } - - /// Get column name to column identifier map - const ColumnNameToColumnIdentifier & getColumnNameToIdentifier() const - { - return column_name_to_column_identifier; - } - /// Get column identifier to column name map const ColumnNameToColumnIdentifier & getColumnIdentifierToColumnName() const { @@ -159,18 +145,6 @@ public: return it->second; } - /** Get column for column name. - * Null is returned if there are no column for column name. - */ - const NameAndTypePair * getColumnOrNull(const std::string & column_name) const - { - auto it = column_name_to_column.find(column_name); - if (it == column_name_to_column.end()) - return nullptr; - - return &it->second; - } - /** Get column identifier for column name. * Exception is thrown if there are no column identifier for column name. */ @@ -200,24 +174,6 @@ public: return &it->second; } - /** Get column name for column identifier. - * Exception is thrown if there are no column name for column identifier. - */ - const std::string & getColumnNameOrThrow(const ColumnIdentifier & column_identifier) const - { - auto it = column_identifier_to_column_name.find(column_identifier); - if (it == column_identifier_to_column_name.end()) - { - auto column_identifiers = getColumnIdentifiers(); - throw Exception(ErrorCodes::LOGICAL_ERROR, - "Column name for column identifier {} does not exists. There are only column identifiers: {}", - column_identifier, - fmt::join(column_identifiers.begin(), column_identifiers.end(), ", ")); - } - - return it->second; - } - /** Get column name for column identifier. * Null is returned if there are no column name for column identifier. */ @@ -296,23 +252,36 @@ public: } private: - void addColumnImpl(const NameAndTypePair & column, const ColumnIdentifier & column_identifier) + void addColumnImpl(const NameAndTypePair & column, const ColumnIdentifier & column_identifier, bool add_to_selected_columns) { - column_names.push_back(column.name); + if (add_to_selected_columns) + markSelectedColumn(column.name); + column_name_to_column.emplace(column.name, column); column_name_to_column_identifier.emplace(column.name, column_identifier); column_identifier_to_column_name.emplace(column_identifier, column.name); } - /// Valid for table, table function, array join, query, union nodes + /// Set of columns that are physically read from table expression + /// In case of ALIAS columns it contains source column names that are used to calculate alias + /// This source column may be not used by user Names column_names; + /// Set of columns that are SELECTed from table expression + /// It may contain ALIAS columns. + /// Mainly it's used to determine access to which columns to check + /// For example user may have an access to column `a ALIAS x + y` but not to `x` and `y` + /// In that case we can read `x` and `y` and calculate `a`, but not return `x` and `y` to user + Names selected_column_names; + /// To deduplicate columns in `selected_column_names` + NameSet selected_column_names_set; + + /// Expression to calculate ALIAS columns + std::unordered_map alias_column_expressions; + /// Valid for table, table function, array join, query, union nodes ColumnNameToColumn column_name_to_column; - /// Valid only for table node - NameSet alias_columns_names; - /// Valid for table, table function, array join, query, union nodes ColumnNameToColumnIdentifier column_name_to_column_identifier; diff --git a/src/Planner/Utils.cpp b/src/Planner/Utils.cpp index 5f5875b8019..bd0b831ee58 100644 --- a/src/Planner/Utils.cpp +++ b/src/Planner/Utils.cpp @@ -469,12 +469,19 @@ FilterDAGInfo buildFilterInfo(ASTPtr filter_expression, NameSet table_expression_required_names_without_filter) { const auto & query_context = planner_context->getQueryContext(); - auto filter_query_tree = buildQueryTree(filter_expression, query_context); QueryAnalysisPass query_analysis_pass(table_expression); query_analysis_pass.run(filter_query_tree, query_context); + return buildFilterInfo(std::move(filter_query_tree), table_expression, planner_context, std::move(table_expression_required_names_without_filter)); +} + +FilterDAGInfo buildFilterInfo(QueryTreeNodePtr filter_query_tree, + const QueryTreeNodePtr & table_expression, + PlannerContextPtr & planner_context, + NameSet table_expression_required_names_without_filter) +{ if (table_expression_required_names_without_filter.empty()) { auto & table_expression_data = planner_context->getTableExpressionDataOrThrow(table_expression); @@ -482,7 +489,7 @@ FilterDAGInfo buildFilterInfo(ASTPtr filter_expression, table_expression_required_names_without_filter.insert(table_expression_names.begin(), table_expression_names.end()); } - collectSourceColumns(filter_query_tree, planner_context); + collectSourceColumns(filter_query_tree, planner_context, false /*keep_alias_columns*/); collectSets(filter_query_tree, *planner_context); auto filter_actions_dag = std::make_shared(); diff --git a/src/Planner/Utils.h b/src/Planner/Utils.h index 3060b1c2711..bf45770552b 100644 --- a/src/Planner/Utils.h +++ b/src/Planner/Utils.h @@ -89,6 +89,11 @@ FilterDAGInfo buildFilterInfo(ASTPtr filter_expression, PlannerContextPtr & planner_context, NameSet table_expression_required_names_without_filter = {}); +FilterDAGInfo buildFilterInfo(QueryTreeNodePtr filter_query_tree, + const QueryTreeNodePtr & table_expression, + PlannerContextPtr & planner_context, + NameSet table_expression_required_names_without_filter = {}); + ASTPtr parseAdditionalResultFilter(const Settings & settings); } diff --git a/src/Planner/findParallelReplicasQuery.cpp b/src/Planner/findParallelReplicasQuery.cpp index 95afcb605b3..ef640bcd42d 100644 --- a/src/Planner/findParallelReplicasQuery.cpp +++ b/src/Planner/findParallelReplicasQuery.cpp @@ -2,6 +2,7 @@ #include #include #include +#include #include #include #include @@ -156,7 +157,8 @@ QueryTreeNodePtr replaceTablesWithDummyTables(const QueryTreeNodePtr & query, co /// Otherwise we can execute current query up to WithMergableStage only. const QueryNode * findQueryForParallelReplicas( std::stack stack, - const std::unordered_map & mapping) + const std::unordered_map & mapping, + const Settings & settings) { const QueryPlan::Node * prev_checked_node = nullptr; const QueryNode * res = nullptr; @@ -192,7 +194,11 @@ const QueryNode * findQueryForParallelReplicas( { const auto * expression = typeid_cast(step); const auto * filter = typeid_cast(step); - if (!expression && !filter) + + const auto * creating_sets = typeid_cast(step); + bool allowed_creating_sets = settings.parallel_replicas_allow_in_with_subquery && creating_sets; + + if (!expression && !filter && !allowed_creating_sets) can_distribute_full_node = false; next_node_to_check = children.front(); @@ -274,7 +280,7 @@ const QueryNode * findQueryForParallelReplicas(const QueryTreeNodePtr & query_tr /// So that we build a list of candidates again, and call findQueryForParallelReplicas for it. auto new_stack = getSupportingParallelReplicasQuery(updated_query_tree.get()); const auto & mapping = planner.getQueryNodeToPlanStepMapping(); - const auto * res = findQueryForParallelReplicas(new_stack, mapping); + const auto * res = findQueryForParallelReplicas(new_stack, mapping, context->getSettingsRef()); /// Now, return a query from initial stack. if (res) diff --git a/src/Processors/Executors/PipelineExecutor.cpp b/src/Processors/Executors/PipelineExecutor.cpp index a06bacd7d3b..5b5880759e6 100644 --- a/src/Processors/Executors/PipelineExecutor.cpp +++ b/src/Processors/Executors/PipelineExecutor.cpp @@ -391,7 +391,9 @@ void PipelineExecutor::executeImpl(size_t num_threads, bool concurrency_control) SCOPE_EXIT_SAFE( if (!finished_flag) { - finish(); + /// If finished_flag is not set, there was an exception. + /// Cancel execution in this case. + cancel(); if (pool) pool->wait(); } diff --git a/src/Processors/Formats/Impl/AvroRowInputFormat.cpp b/src/Processors/Formats/Impl/AvroRowInputFormat.cpp index 8ef2cda5587..2ed55cca30c 100644 --- a/src/Processors/Formats/Impl/AvroRowInputFormat.cpp +++ b/src/Processors/Formats/Impl/AvroRowInputFormat.cpp @@ -1016,7 +1016,7 @@ private: http_basic_credentials.authenticate(request); } - auto session = makePooledHTTPSession(url, timeouts, 1); + auto session = makeHTTPSession(HTTPConnectionGroupType::HTTP, url, timeouts); session->sendRequest(request); Poco::Net::HTTPResponse response; @@ -1025,8 +1025,6 @@ private: Poco::JSON::Parser parser; auto json_body = parser.parse(*response_body).extract(); - /// Response was fully read. - markSessionForReuse(session); auto schema = json_body->getValue("schema"); LOG_TRACE((getLogger("AvroConfluentRowInputFormat")), "Successfully fetched schema id = {}\n{}", id, schema); diff --git a/src/Processors/Formats/Impl/PrometheusTextOutputFormat.cpp b/src/Processors/Formats/Impl/PrometheusTextOutputFormat.cpp index c2ee5923c01..3578401a0f8 100644 --- a/src/Processors/Formats/Impl/PrometheusTextOutputFormat.cpp +++ b/src/Processors/Formats/Impl/PrometheusTextOutputFormat.cpp @@ -12,6 +12,7 @@ #include #include +#include "DataTypes/IDataType.h" #include #include @@ -35,9 +36,12 @@ namespace ErrorCodes extern const int BAD_ARGUMENTS; } +namespace +{ + constexpr auto FORMAT_NAME = "Prometheus"; -static bool isDataTypeMapString(const DataTypePtr & type) +bool isDataTypeMapString(const DataTypePtr & type) { if (!isMap(type)) return false; @@ -45,8 +49,8 @@ static bool isDataTypeMapString(const DataTypePtr & type) return isStringOrFixedString(type_map->getKeyType()) && isStringOrFixedString(type_map->getValueType()); } -template -static void getColumnPos(const Block & header, const String & col_name, Pred pred, ResType & res) +template +void getColumnPos(const Block & header, const String & col_name, bool (*pred)(const DataTypePtr &), ResType & res) { static_assert(std::is_same_v || std::is_same_v>, "Illegal ResType"); @@ -71,7 +75,7 @@ static void getColumnPos(const Block & header, const String & col_name, Pred pre } } -static Float64 tryParseFloat(const String & s) +Float64 tryParseFloat(const String & s) { Float64 t = 0; ReadBufferFromString buf(s); @@ -79,6 +83,8 @@ static Float64 tryParseFloat(const String & s) return t; } +} + PrometheusTextOutputFormat::PrometheusTextOutputFormat( WriteBuffer & out_, const Block & header_, @@ -89,12 +95,12 @@ PrometheusTextOutputFormat::PrometheusTextOutputFormat( { const Block & header = getPort(PortKind::Main).getHeader(); - getColumnPos(header, "name", isStringOrFixedString, pos.name); - getColumnPos(header, "value", isNumber, pos.value); + getColumnPos(header, "name", isStringOrFixedString, pos.name); + getColumnPos(header, "value", isNumber, pos.value); - getColumnPos(header, "help", isStringOrFixedString, pos.help); - getColumnPos(header, "type", isStringOrFixedString, pos.type); - getColumnPos(header, "timestamp", isNumber, pos.timestamp); + getColumnPos(header, "help", isStringOrFixedString, pos.help); + getColumnPos(header, "type", isStringOrFixedString, pos.type); + getColumnPos(header, "timestamp", isNumber, pos.timestamp); getColumnPos(header, "labels", isDataTypeMapString, pos.labels); } diff --git a/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp b/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp index 845cf561968..28160b18269 100644 --- a/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp +++ b/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp @@ -12,7 +12,7 @@ #include #include #include -#include +#include namespace DB diff --git a/src/Processors/QueryPlan/Optimizations/optimizeUseAggregateProjection.cpp b/src/Processors/QueryPlan/Optimizations/optimizeUseAggregateProjection.cpp index 28f0f979499..91f4213ff43 100644 --- a/src/Processors/QueryPlan/Optimizations/optimizeUseAggregateProjection.cpp +++ b/src/Processors/QueryPlan/Optimizations/optimizeUseAggregateProjection.cpp @@ -431,7 +431,7 @@ AggregateProjectionCandidates getAggregateProjectionCandidates( { const auto & keys = aggregating.getParams().keys; const auto & aggregates = aggregating.getParams().aggregates; - Block key_virtual_columns = reading.getMergeTreeData().getSampleBlockWithVirtualColumns(); + Block key_virtual_columns = reading.getMergeTreeData().getHeaderWithVirtualsForFilter(); AggregateProjectionCandidates candidates; diff --git a/src/Processors/QueryPlan/Optimizations/optimizeUseNormalProjection.cpp b/src/Processors/QueryPlan/Optimizations/optimizeUseNormalProjection.cpp index 73caf70627f..cac172a856f 100644 --- a/src/Processors/QueryPlan/Optimizations/optimizeUseNormalProjection.cpp +++ b/src/Processors/QueryPlan/Optimizations/optimizeUseNormalProjection.cpp @@ -135,7 +135,7 @@ bool optimizeUseNormalProjections(Stack & stack, QueryPlan::Nodes & nodes) std::list candidates; NormalProjectionCandidate * best_candidate = nullptr; - const Names & required_columns = reading->getRealColumnNames(); + const Names & required_columns = reading->getAllColumnNames(); const auto & parts = reading->getParts(); const auto & alter_conversions = reading->getAlterConvertionsForParts(); const auto & query_info = reading->getQueryInfo(); diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.cpp b/src/Processors/QueryPlan/ReadFromMergeTree.cpp index c7b9eb72d4d..21e3cfcceab 100644 --- a/src/Processors/QueryPlan/ReadFromMergeTree.cpp +++ b/src/Processors/QueryPlan/ReadFromMergeTree.cpp @@ -43,6 +43,7 @@ #include #include #include +#include #include #include @@ -260,30 +261,24 @@ void ReadFromMergeTree::AnalysisResult::checkLimits(const Settings & settings, c ReadFromMergeTree::ReadFromMergeTree( MergeTreeData::DataPartsVector parts_, std::vector alter_conversions_, - const Names & column_names_, - Names real_column_names_, - Names virt_column_names_, + Names all_column_names_, const MergeTreeData & data_, const SelectQueryInfo & query_info_, const StorageSnapshotPtr & storage_snapshot_, const ContextPtr & context_, size_t max_block_size_, size_t num_streams_, - bool sample_factor_column_queried_, std::shared_ptr max_block_numbers_to_read_, LoggerPtr log_, AnalysisResultPtr analyzed_result_ptr_, bool enable_parallel_reading) : SourceStepWithFilter(DataStream{.header = MergeTreeSelectProcessor::transformHeader( - storage_snapshot_->getSampleBlockForColumns(real_column_names_), - query_info_.prewhere_info, - data_.getPartitionValueType(), - virt_column_names_)}, column_names_, query_info_, storage_snapshot_, context_) + storage_snapshot_->getSampleBlockForColumns(all_column_names_), + query_info_.prewhere_info)}, all_column_names_, query_info_, storage_snapshot_, context_) , reader_settings(getMergeTreeReaderSettings(context_, query_info_)) , prepared_parts(std::move(parts_)) , alter_conversions_for_parts(std::move(alter_conversions_)) - , real_column_names(std::move(real_column_names_)) - , virt_column_names(std::move(virt_column_names_)) + , all_column_names(std::move(all_column_names_)) , data(data_) , actions_settings(ExpressionActionsSettings::fromContext(context_)) , metadata_for_reading(storage_snapshot->getMetadataForQuery()) @@ -292,20 +287,11 @@ ReadFromMergeTree::ReadFromMergeTree( .preferred_block_size_bytes = context->getSettingsRef().preferred_block_size_bytes, .preferred_max_column_in_block_size_bytes = context->getSettingsRef().preferred_max_column_in_block_size_bytes} , requested_num_streams(num_streams_) - , sample_factor_column_queried(sample_factor_column_queried_) , max_block_numbers_to_read(std::move(max_block_numbers_to_read_)) , log(std::move(log_)) , analyzed_result_ptr(analyzed_result_ptr_) , is_parallel_reading_from_replicas(enable_parallel_reading) { - if (sample_factor_column_queried) - { - /// Only _sample_factor virtual column is added by ReadFromMergeTree - /// Other virtual columns are added by MergeTreeSelectProcessor. - auto type = std::make_shared(); - output_stream->header.insert({type->createColumn(), type, "_sample_factor"}); - } - if (is_parallel_reading_from_replicas) { all_ranges_callback = context->getMergeTreeAllRangesCallback(); @@ -367,12 +353,12 @@ Pipe ReadFromMergeTree::readFromPoolParallelReplicas( auto pool = std::make_shared( std::move(extension), std::move(parts_with_range), + shared_virtual_fields, storage_snapshot, prewhere_info, actions_settings, reader_settings, required_columns, - virt_column_names, pool_settings, context); @@ -386,8 +372,8 @@ Pipe ReadFromMergeTree::readFromPoolParallelReplicas( auto algorithm = std::make_unique(i); auto processor = std::make_unique( - pool, std::move(algorithm), data, prewhere_info, - actions_settings, block_size_copy, reader_settings, virt_column_names); + pool, std::move(algorithm), storage_snapshot, prewhere_info, + actions_settings, block_size_copy, reader_settings); auto source = std::make_shared(std::move(processor)); pipes.emplace_back(std::move(source)); @@ -448,12 +434,12 @@ Pipe ReadFromMergeTree::readFromPool( { pool = std::make_shared( std::move(parts_with_range), + shared_virtual_fields, storage_snapshot, prewhere_info, actions_settings, reader_settings, required_columns, - virt_column_names, pool_settings, context); } @@ -461,12 +447,12 @@ Pipe ReadFromMergeTree::readFromPool( { pool = std::make_shared( std::move(parts_with_range), + shared_virtual_fields, storage_snapshot, prewhere_info, actions_settings, reader_settings, required_columns, - virt_column_names, pool_settings, context); } @@ -485,8 +471,8 @@ Pipe ReadFromMergeTree::readFromPool( auto algorithm = std::make_unique(i); auto processor = std::make_unique( - pool, std::move(algorithm), data, prewhere_info, - actions_settings, block_size_copy, reader_settings, virt_column_names); + pool, std::move(algorithm), storage_snapshot, prewhere_info, + actions_settings, block_size_copy, reader_settings); auto source = std::make_shared(std::move(processor)); @@ -537,12 +523,12 @@ Pipe ReadFromMergeTree::readInOrder( std::move(extension), mode, parts_with_ranges, + shared_virtual_fields, storage_snapshot, prewhere_info, actions_settings, reader_settings, required_columns, - virt_column_names, pool_settings, context); } @@ -552,12 +538,12 @@ Pipe ReadFromMergeTree::readInOrder( has_limit_below_one_block, read_type, parts_with_ranges, + shared_virtual_fields, storage_snapshot, prewhere_info, actions_settings, reader_settings, required_columns, - virt_column_names, pool_settings, context); } @@ -591,8 +577,8 @@ Pipe ReadFromMergeTree::readInOrder( algorithm = std::make_unique(i); auto processor = std::make_unique( - pool, std::move(algorithm), data, prewhere_info, - actions_settings, block_size, reader_settings, virt_column_names); + pool, std::move(algorithm), storage_snapshot, prewhere_info, + actions_settings, block_size, reader_settings); processor->addPartLevelToChunk(isQueryWithFinal()); @@ -1301,8 +1287,7 @@ ReadFromMergeTree::AnalysisResultPtr ReadFromMergeTree::selectRangesToRead( requested_num_streams, max_block_numbers_to_read, data, - real_column_names, - sample_factor_column_queried, + all_column_names, log, indexes); } @@ -1421,6 +1406,20 @@ static void buildIndexes( } } + // move minmax indices to first positions, so they will be applied first as cheapest ones + std::stable_sort(begin(skip_indexes.useful_indices), end(skip_indexes.useful_indices), [](const auto & l, const auto & r) + { + const bool l_min_max = (typeid_cast(l.index.get())); + const bool r_min_max = (typeid_cast(r.index.get())); + if (l_min_max == r_min_max) + return false; + + if (l_min_max) + return true; // left is min max but right is not + + return false; // right is min max but left is not + }); + indexes->skip_indexes = std::move(skip_indexes); } @@ -1434,8 +1433,13 @@ void ReadFromMergeTree::applyFilters(ActionDAGNodes added_filter_nodes) if (query_info.planner_context) { const auto & table_expression_data = query_info.planner_context->getTableExpressionDataOrThrow(query_info.table_expression); + const auto & alias_column_expressions = table_expression_data.getAliasColumnExpressions(); for (const auto & [column_identifier, column_name] : table_expression_data.getColumnIdentifierToColumnName()) { + /// ALIAS columns cannot be used in the filter expression without being calculated in ActionsDAG, + /// so they should not be added to the input nodes. + if (alias_column_expressions.contains(column_name)) + continue; const auto & column = table_expression_data.getColumnOrThrow(column_name); node_name_to_input_node_column.emplace(column_identifier, ColumnWithTypeAndName(column.type, column_name)); } @@ -1469,8 +1473,7 @@ ReadFromMergeTree::AnalysisResultPtr ReadFromMergeTree::selectRangesToRead( size_t num_streams, std::shared_ptr max_block_numbers_to_read, const MergeTreeData & data, - const Names & real_column_names, - bool sample_factor_column_queried, + const Names & all_column_names, LoggerPtr log, std::optional & indexes) { @@ -1483,8 +1486,7 @@ ReadFromMergeTree::AnalysisResultPtr ReadFromMergeTree::selectRangesToRead( num_streams, max_block_numbers_to_read, data, - real_column_names, - sample_factor_column_queried, + all_column_names, log, indexes); } @@ -1498,8 +1500,7 @@ ReadFromMergeTree::AnalysisResultPtr ReadFromMergeTree::selectRangesToReadImpl( size_t num_streams, std::shared_ptr max_block_numbers_to_read, const MergeTreeData & data, - const Names & real_column_names, - bool sample_factor_column_queried, + const Names & all_column_names, LoggerPtr log, std::optional & indexes) { @@ -1508,7 +1509,7 @@ ReadFromMergeTree::AnalysisResultPtr ReadFromMergeTree::selectRangesToReadImpl( size_t total_parts = parts.size(); - result.column_names_to_read = real_column_names; + result.column_names_to_read = all_column_names; /// If there are only virtual columns in the query, you must request at least one non-virtual one. if (result.column_names_to_read.empty()) @@ -1567,7 +1568,6 @@ ReadFromMergeTree::AnalysisResultPtr ReadFromMergeTree::selectRangesToReadImpl( data, metadata_snapshot, context_, - sample_factor_column_queried, log); if (result.sampling.read_nothing) @@ -1684,10 +1684,8 @@ void ReadFromMergeTree::updatePrewhereInfo(const PrewhereInfoPtr & prewhere_info prewhere_info = prewhere_info_value; output_stream = DataStream{.header = MergeTreeSelectProcessor::transformHeader( - storage_snapshot->getSampleBlockForColumns(real_column_names), - prewhere_info_value, - data.getPartitionValueType(), - virt_column_names)}; + storage_snapshot->getSampleBlockForColumns(all_column_names), + prewhere_info_value)}; updateSortDescriptionForOutputStream( *output_stream, @@ -1881,6 +1879,7 @@ void ReadFromMergeTree::initializePipeline(QueryPipelineBuilder & pipeline, cons storage_snapshot->data = std::make_unique(); result.checkLimits(context->getSettingsRef(), query_info); + shared_virtual_fields.emplace("_sample_factor", result.sampling.used_sample_factor); LOG_DEBUG( log, @@ -1965,18 +1964,6 @@ void ReadFromMergeTree::initializePipeline(QueryPipelineBuilder & pipeline, cons result_projection = ActionsDAG::merge(std::move(*result_projection), std::move(*actions)); }; - /// By the way, if a distributed query or query to a Merge table is made, then the `_sample_factor` column can have different values. - if (sample_factor_column_queried) - { - ColumnWithTypeAndName column; - column.name = "_sample_factor"; - column.type = std::make_shared(); - column.column = column.type->createColumnConst(0, Field(result.sampling.used_sample_factor)); - - auto adding_column = ActionsDAG::makeAddingColumnActions(std::move(column)); - append_actions(std::move(adding_column)); - } - if (result_projection) cur_header = result_projection->updateHeader(cur_header); diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.h b/src/Processors/QueryPlan/ReadFromMergeTree.h index 8efe5ed7b13..5ed742a9bfd 100644 --- a/src/Processors/QueryPlan/ReadFromMergeTree.h +++ b/src/Processors/QueryPlan/ReadFromMergeTree.h @@ -110,16 +110,13 @@ public: ReadFromMergeTree( MergeTreeData::DataPartsVector parts_, std::vector alter_conversions_, - const Names & column_names_, - Names real_column_names_, - Names virt_column_names_, + Names all_column_names_, const MergeTreeData & data_, const SelectQueryInfo & query_info_, const StorageSnapshotPtr & storage_snapshot, const ContextPtr & context_, size_t max_block_size_, size_t num_streams_, - bool sample_factor_column_queried_, std::shared_ptr max_block_numbers_to_read_, LoggerPtr log_, AnalysisResultPtr analyzed_result_ptr_, @@ -136,8 +133,7 @@ public: void describeActions(JSONBuilder::JSONMap & map) const override; void describeIndexes(JSONBuilder::JSONMap & map) const override; - const Names & getRealColumnNames() const { return real_column_names; } - const Names & getVirtualColumnNames() const { return virt_column_names; } + const Names & getAllColumnNames() const { return all_column_names; } StorageID getStorageID() const { return data.getStorageID(); } UInt64 getSelectedParts() const { return selected_parts; } @@ -164,8 +160,7 @@ public: size_t num_streams, std::shared_ptr max_block_numbers_to_read, const MergeTreeData & data, - const Names & real_column_names, - bool sample_factor_column_queried, + const Names & all_column_names, LoggerPtr log, std::optional & indexes); @@ -209,8 +204,7 @@ private: size_t num_streams, std::shared_ptr max_block_numbers_to_read, const MergeTreeData & data, - const Names & real_column_names, - bool sample_factor_column_queried, + const Names & all_column_names, LoggerPtr log, std::optional & indexes); @@ -227,8 +221,7 @@ private: MergeTreeData::DataPartsVector prepared_parts; std::vector alter_conversions_for_parts; - Names real_column_names; - Names virt_column_names; + Names all_column_names; const MergeTreeData & data; ExpressionActionsSettings actions_settings; @@ -239,7 +232,6 @@ private: size_t requested_num_streams; size_t output_streams_limit = 0; - const bool sample_factor_column_queried; /// Used for aggregation optimization (see DB::QueryPlanOptimizations::tryAggregateEachPartitionIndependently). bool output_each_partition_through_separate_port = false; @@ -280,7 +272,9 @@ private: RangesInDataParts && parts, size_t num_streams, const Names & origin_column_names, const Names & column_names, ActionsDAGPtr & out_projection); ReadFromMergeTree::AnalysisResult getAnalysisResult() const; + AnalysisResultPtr analyzed_result_ptr; + VirtualFields shared_virtual_fields; bool is_parallel_reading_from_replicas; std::optional all_ranges_callback; diff --git a/src/Processors/QueryPlan/ReadFromSystemNumbersStep.cpp b/src/Processors/QueryPlan/ReadFromSystemNumbersStep.cpp index c72c63d09c4..a294683c640 100644 --- a/src/Processors/QueryPlan/ReadFromSystemNumbersStep.cpp +++ b/src/Processors/QueryPlan/ReadFromSystemNumbersStep.cpp @@ -26,9 +26,11 @@ namespace class NumbersSource : public ISource { public: - NumbersSource(UInt64 block_size_, UInt64 offset_, UInt64 step_) + NumbersSource(UInt64 block_size_, UInt64 offset_, std::optional limit_, UInt64 step_) : ISource(createHeader()), block_size(block_size_), next(offset_), step(step_) { + if (limit_.has_value()) + end = limit_.value() + offset_; } String getName() const override { return "Numbers"; } @@ -38,24 +40,32 @@ public: protected: Chunk generate() override { - auto column = ColumnUInt64::create(block_size); + UInt64 real_block_size = block_size; + if (end.has_value()) + { + if (end.value() <= next) + return {}; + real_block_size = std::min(block_size, end.value() - next); + } + auto column = ColumnUInt64::create(real_block_size); ColumnUInt64::Container & vec = column->getData(); UInt64 curr = next; /// The local variable for some reason works faster (>20%) than member of class. UInt64 * pos = vec.data(); /// This also accelerates the code. - UInt64 * end = &vec[block_size]; - iota(pos, static_cast(end - pos), curr); + UInt64 * end_ = &vec[real_block_size]; + iota(pos, static_cast(end_ - pos), curr); next += step; progress(column->size(), column->byteSize()); - return {Columns{std::move(column)}, block_size}; + return {Columns{std::move(column)}, real_block_size}; } private: UInt64 block_size; UInt64 next; + std::optional end; /// not included UInt64 step; }; @@ -478,7 +488,7 @@ Pipe ReadFromSystemNumbersStep::makePipe() for (size_t i = 0; i < num_streams; ++i) { auto source - = std::make_shared(max_block_size, numbers_storage.offset + i * max_block_size, num_streams * max_block_size); + = std::make_shared(max_block_size, numbers_storage.offset + i * max_block_size, numbers_storage.limit, num_streams * max_block_size); if (numbers_storage.limit && i == 0) { diff --git a/src/Processors/Sources/RemoteSource.cpp b/src/Processors/Sources/RemoteSource.cpp index 33bd824dd5c..3d7dd3f76b8 100644 --- a/src/Processors/Sources/RemoteSource.cpp +++ b/src/Processors/Sources/RemoteSource.cpp @@ -71,19 +71,36 @@ ISource::Status RemoteSource::prepare() if (is_async_state) return Status::Async; + if (executor_finished) + return Status::Finished; + Status status = ISource::prepare(); /// To avoid resetting the connection (because of "unfinished" query) in the /// RemoteQueryExecutor it should be finished explicitly. if (status == Status::Finished) { - query_executor->finish(); is_async_state = false; - return status; + need_drain = true; + return Status::Ready; } return status; } +void RemoteSource::work() +{ + /// Connection drain is a heavy operation that may take a long time. + /// Therefore we move connection drain from prepare() to work(), and drain multiple connections in parallel. + /// See issue: https://github.com/ClickHouse/ClickHouse/issues/60844 + if (need_drain) + { + query_executor->finish(); + executor_finished = true; + return; + } + ISource::work(); +} + std::optional RemoteSource::tryGenerate() { /// onCancel() will do the cancel if the query was sent. diff --git a/src/Processors/Sources/RemoteSource.h b/src/Processors/Sources/RemoteSource.h index 4e179371cf5..052567bc261 100644 --- a/src/Processors/Sources/RemoteSource.h +++ b/src/Processors/Sources/RemoteSource.h @@ -22,6 +22,7 @@ public: ~RemoteSource() override; Status prepare() override; + void work() override; String getName() const override { return "Remote"; } void setRowsBeforeLimitCounter(RowsBeforeLimitCounterPtr counter) override { rows_before_limit.swap(counter); } @@ -39,6 +40,8 @@ protected: private: bool was_query_sent = false; + bool need_drain = false; + bool executor_finished = false; bool add_aggregation_info = false; RemoteQueryExecutorPtr query_executor; RowsBeforeLimitCounterPtr rows_before_limit; diff --git a/src/Processors/Transforms/FilterTransform.cpp b/src/Processors/Transforms/FilterTransform.cpp index ea4dcff4808..0f2509c7510 100644 --- a/src/Processors/Transforms/FilterTransform.cpp +++ b/src/Processors/Transforms/FilterTransform.cpp @@ -265,6 +265,7 @@ void FilterTransform::doTransform(Chunk & chunk) { size_t num_rows_before_filtration = chunk.getNumRows(); auto columns = chunk.detachColumns(); + DataTypes types; auto select_final_indices_info = getSelectByFinalIndices(chunk); { @@ -275,6 +276,7 @@ void FilterTransform::doTransform(Chunk & chunk) expression->execute(block, num_rows_before_filtration); columns = block.getColumns(); + types = block.getDataTypes(); } if (constant_filter_description.always_true || on_totals) @@ -323,14 +325,22 @@ void FilterTransform::doTransform(Chunk & chunk) * or calculate number of set bytes in the filter. */ size_t first_non_constant_column = num_columns; + size_t min_size_in_memory = std::numeric_limits::max(); for (size_t i = 0; i < num_columns; ++i) { - if (i != filter_column_position && !isColumnConst(*columns[i])) + DataTypePtr type_not_null = removeNullableOrLowCardinalityNullable(types[i]); + if (i != filter_column_position && !isColumnConst(*columns[i]) && type_not_null->isValueRepresentedByNumber()) { - first_non_constant_column = i; + size_t size_in_memory = type_not_null->getSizeOfValueInMemory() + (isNullableOrLowCardinalityNullable(types[i]) ? 1 : 0); + if (size_in_memory < min_size_in_memory) + { + min_size_in_memory = size_in_memory; + first_non_constant_column = i; + } break; } } + (void)min_size_in_memory; /// Suppress error of clang-analyzer-deadcode.DeadStores size_t num_filtered_rows = 0; if (first_non_constant_column != num_columns) diff --git a/src/Processors/Transforms/buildPushingToViewsChain.cpp b/src/Processors/Transforms/buildPushingToViewsChain.cpp index dd07d043599..aa5a1c0cc1a 100644 --- a/src/Processors/Transforms/buildPushingToViewsChain.cpp +++ b/src/Processors/Transforms/buildPushingToViewsChain.cpp @@ -451,7 +451,7 @@ Chain buildPushingToViewsChain( /// If we don't write directly to the destination /// then expect that we're inserting with precalculated virtual columns - auto storage_header = no_destination ? metadata_snapshot->getSampleBlockWithVirtuals(storage->getVirtuals()) + auto storage_header = no_destination ? metadata_snapshot->getSampleBlockWithVirtuals(storage->getVirtualsList()) : metadata_snapshot->getSampleBlock(); /** TODO This is a very important line. At any insertion into the table one of chains should own lock. @@ -597,7 +597,7 @@ static QueryPipeline process(Block block, ViewRuntimeData & view, const ViewsDat views_data.source_storage_id, views_data.source_metadata_snapshot->getColumns(), std::move(block), - views_data.source_storage->getVirtuals())); + *views_data.source_storage->getVirtualsPtr())); QueryPipelineBuilder pipeline; diff --git a/src/Server/TCPHandler.cpp b/src/Server/TCPHandler.cpp index d883029408c..9637e5bf9ec 100644 --- a/src/Server/TCPHandler.cpp +++ b/src/Server/TCPHandler.cpp @@ -933,9 +933,26 @@ void TCPHandler::processInsertQuery() if (insert_queue && async_insert_enabled && !insert_query.select) { + /// Let's agree on terminology and say that a mini-INSERT is an asynchronous INSERT + /// which typically contains not a lot of data inside and a big-INSERT in an INSERT + /// which was formed by concatenating several mini-INSERTs together. + /// In case when the client had to retry some mini-INSERTs then they will be properly deduplicated + /// by the source tables. This functionality is controlled by a setting `async_insert_deduplicate`. + /// But then they will be glued together into a block and pushed through a chain of Materialized Views if any. + /// The process of forming such blocks is not deteministic so each time we retry mini-INSERTs the resulting + /// block may be concatenated differently. + /// That's why deduplication in dependent Materialized Views doesn't make sense in presence of async INSERTs. + if (settings.throw_if_deduplication_in_dependent_materialized_views_enabled_with_async_insert && + settings.deduplicate_blocks_in_dependent_materialized_views) + throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, + "Deduplication is dependent materialized view cannot work together with async inserts. "\ + "Please disable eiher `deduplicate_blocks_in_dependent_materialized_views` or `async_insert` setting."); + auto result = processAsyncInsertQuery(*insert_queue); if (result.status == AsynchronousInsertQueue::PushResult::OK) { + /// Reset pipeline because it may hold write lock for some storages. + state.io.pipeline.reset(); if (settings.wait_for_async_insert) { size_t timeout_ms = settings.wait_for_async_insert_timeout.totalMilliseconds(); @@ -968,7 +985,7 @@ void TCPHandler::processInsertQuery() else { PushingPipelineExecutor executor(state.io.pipeline); - run_executor(executor, processed_block); + run_executor(executor, std::move(processed_block)); } sendInsertProfileEvents(); diff --git a/src/Storages/AlterCommands.cpp b/src/Storages/AlterCommands.cpp index b09200f06ff..eae5e1a8a47 100644 --- a/src/Storages/AlterCommands.cpp +++ b/src/Storages/AlterCommands.cpp @@ -31,8 +31,6 @@ #include #include #include -#include -#include #include #include #include @@ -965,8 +963,7 @@ bool AlterCommand::isRequireMutationStage(const StorageInMemoryMetadata & metada /// Drop alias is metadata alter, in other case mutation is required. if (type == DROP_COLUMN) - return metadata.columns.hasColumnOrNested(GetColumnsOptions::AllPhysical, column_name) || - column_name == LightweightDeleteDescription::FILTER_COLUMN.name || column_name == BlockNumberColumn::name; + return metadata.columns.hasColumnOrNested(GetColumnsOptions::AllPhysical, column_name); if (type != MODIFY_COLUMN || data_type == nullptr) return false; @@ -1256,7 +1253,9 @@ void AlterCommands::prepare(const StorageInMemoryMetadata & metadata) void AlterCommands::validate(const StoragePtr & table, ContextPtr context) const { - const StorageInMemoryMetadata & metadata = table->getInMemoryMetadata(); + const auto & metadata = table->getInMemoryMetadata(); + auto virtuals = table->getVirtualsPtr(); + auto all_columns = metadata.columns; /// Default expression for all added/modified columns ASTPtr default_expr_list = std::make_shared(); @@ -1292,16 +1291,20 @@ void AlterCommands::validate(const StoragePtr & table, ContextPtr context) const if (command.data_type->hasDynamicSubcolumns()) throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "Adding a new column of a type which has dynamic subcolumns to an existing table is not allowed. It has known bugs"); - if (column_name == LightweightDeleteDescription::FILTER_COLUMN.name && std::dynamic_pointer_cast(table)) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot add column {}: " - "this column name is reserved for lightweight delete feature", backQuote(column_name)); - - if (column_name == BlockNumberColumn::name && std::dynamic_pointer_cast(table)) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot add column {}: " - "this column name is reserved for _block_number persisting feature", backQuote(column_name)); + if (virtuals->tryGet(column_name, VirtualsKind::Persistent)) + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Cannot add column {}: this column name is reserved for persistent virtual column", backQuote(column_name)); if (command.codec) - CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST(command.codec, command.data_type, !context->getSettingsRef().allow_suspicious_codecs, context->getSettingsRef().allow_experimental_codecs, context->getSettingsRef().enable_deflate_qpl_codec, context->getSettingsRef().enable_zstd_qat_codec); + { + const auto & settings = context->getSettingsRef(); + CompressionCodecFactory::instance().validateCodecAndGetPreprocessedAST( + command.codec, command.data_type, + !settings.allow_suspicious_codecs, + settings.allow_experimental_codecs, + settings.enable_deflate_qpl_codec, + settings.enable_zstd_qat_codec); + } all_columns.add(ColumnDescription(column_name, command.data_type)); } @@ -1415,9 +1418,7 @@ void AlterCommands::validate(const StoragePtr & table, ContextPtr context) const } else if (command.type == AlterCommand::DROP_COLUMN) { - if (all_columns.has(command.column_name) || - all_columns.hasNested(command.column_name) || - (command.clear && column_name == LightweightDeleteDescription::FILTER_COLUMN.name)) + if (all_columns.has(command.column_name) || all_columns.hasNested(command.column_name)) { if (!command.clear) /// CLEAR column is Ok even if there are dependencies. { @@ -1501,16 +1502,12 @@ void AlterCommands::validate(const StoragePtr & table, ContextPtr context) const } if (all_columns.has(command.rename_to)) - throw Exception(ErrorCodes::DUPLICATE_COLUMN, "Cannot rename to {}: " - "column with this name already exists", backQuote(command.rename_to)); + throw Exception(ErrorCodes::DUPLICATE_COLUMN, + "Cannot rename to {}: column with this name already exists", backQuote(command.rename_to)); - if (command.rename_to == LightweightDeleteDescription::FILTER_COLUMN.name && std::dynamic_pointer_cast(table)) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot rename to {}: " - "this column name is reserved for lightweight delete feature", backQuote(command.rename_to)); - - if (command.rename_to == BlockNumberColumn::name && std::dynamic_pointer_cast(table)) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot rename to {}: " - "this column name is reserved for _block_number persisting feature", backQuote(command.rename_to)); + if (virtuals->tryGet(command.rename_to, VirtualsKind::Persistent)) + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Cannot rename to {}: this column name is reserved for persistent virtual column", backQuote(command.rename_to)); if (modified_columns.contains(column_name)) throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot rename and modify the same column {} " diff --git a/src/Storages/BlockNumberColumn.cpp b/src/Storages/BlockNumberColumn.cpp deleted file mode 100644 index 8c9e1fd902a..00000000000 --- a/src/Storages/BlockNumberColumn.cpp +++ /dev/null @@ -1,23 +0,0 @@ -#include -#include - -namespace DB -{ - -CompressionCodecPtr getCompressionCodecDelta(UInt8 delta_bytes_size); - -CompressionCodecPtr getCompressionCodecForBlockNumberColumn() -{ - std::vector codecs; - codecs.reserve(2); - auto data_bytes_size = BlockNumberColumn::type->getSizeOfValueInMemory(); - codecs.emplace_back(getCompressionCodecDelta(data_bytes_size)); - codecs.emplace_back(CompressionCodecFactory::instance().get("LZ4", {})); - return std::make_shared(codecs); -} - -const String BlockNumberColumn::name = "_block_number"; -const DataTypePtr BlockNumberColumn::type = std::make_shared(); -const CompressionCodecPtr BlockNumberColumn::compression_codec = getCompressionCodecForBlockNumberColumn(); - -} diff --git a/src/Storages/BlockNumberColumn.h b/src/Storages/BlockNumberColumn.h deleted file mode 100644 index fffa68bfd49..00000000000 --- a/src/Storages/BlockNumberColumn.h +++ /dev/null @@ -1,16 +0,0 @@ -#pragma once -#include -#include -#include - -namespace DB -{ - -struct BlockNumberColumn -{ - static const String name; - static const DataTypePtr type; - static const CompressionCodecPtr compression_codec; -}; - -} diff --git a/src/Storages/ColumnsDescription.cpp b/src/Storages/ColumnsDescription.cpp index d6a241da032..e08dac3a332 100644 --- a/src/Storages/ColumnsDescription.cpp +++ b/src/Storages/ColumnsDescription.cpp @@ -31,15 +31,11 @@ #include #include #include -#include namespace DB { -CompressionCodecPtr getCompressionCodecDelta(UInt8 delta_bytes_size); - - namespace ErrorCodes { extern const int NO_SUCH_COLUMN_IN_TABLE; @@ -482,6 +478,10 @@ NamesAndTypesList ColumnsDescription::get(const GetColumnsOptions & options) con NamesAndTypesList res; switch (options.kind) { + case GetColumnsOptions::None: + { + break; + } case GetColumnsOptions::All: { res = getAll(); @@ -559,6 +559,12 @@ const ColumnDescription & ColumnsDescription::get(const String & column_name) co return *it; } +const ColumnDescription * ColumnsDescription::tryGet(const String & column_name) const +{ + auto it = columns.get<1>().find(column_name); + return it == columns.get<1>().end() ? nullptr : &(*it); +} + static GetColumnsOptions::Kind defaultKindToGetKind(ColumnDefaultKind kind) { switch (kind) @@ -572,7 +578,8 @@ static GetColumnsOptions::Kind defaultKindToGetKind(ColumnDefaultKind kind) case ColumnDefaultKind::Ephemeral: return GetColumnsOptions::Ephemeral; } - UNREACHABLE(); + + return GetColumnsOptions::None; } NamesAndTypesList ColumnsDescription::getByNames(const GetColumnsOptions & options, const Names & names) const @@ -784,33 +791,6 @@ bool ColumnsDescription::hasCompressionCodec(const String & column_name) const return it != columns.get<1>().end() && it->codec != nullptr; } -CompressionCodecPtr ColumnsDescription::getCodecOrDefault(const String & column_name, CompressionCodecPtr default_codec) const -{ - const auto it = columns.get<1>().find(column_name); - - if (it == columns.get<1>().end() || !it->codec) - return default_codec; - - return CompressionCodecFactory::instance().get(it->codec, it->type, default_codec); -} - -CompressionCodecPtr ColumnsDescription::getCodecOrDefault(const String & column_name) const -{ - assert (column_name != BlockNumberColumn::name); - return getCodecOrDefault(column_name, CompressionCodecFactory::instance().getDefaultCodec()); -} - -ASTPtr ColumnsDescription::getCodecDescOrDefault(const String & column_name, CompressionCodecPtr default_codec) const -{ - assert (column_name != BlockNumberColumn::name); - const auto it = columns.get<1>().find(column_name); - - if (it == columns.get<1>().end() || !it->codec) - return default_codec->getFullCodecDesc(); - - return it->codec; -} - ColumnsDescription::ColumnTTLs ColumnsDescription::getColumnTTLs() const { ColumnTTLs ret; diff --git a/src/Storages/ColumnsDescription.h b/src/Storages/ColumnsDescription.h index 5ba655ee10d..82e55e29073 100644 --- a/src/Storages/ColumnsDescription.h +++ b/src/Storages/ColumnsDescription.h @@ -29,10 +29,19 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } +enum class VirtualsKind : UInt8 +{ + None = 0, + Ephemeral = 1, + Persistent = 2, + All = Ephemeral | Persistent, +}; + struct GetColumnsOptions { enum Kind : UInt8 { + None = 0, Ordinary = 1, Materialized = 2, Aliases = 4, @@ -51,9 +60,9 @@ struct GetColumnsOptions return *this; } - GetColumnsOptions & withVirtuals(bool value = true) + GetColumnsOptions & withVirtuals(VirtualsKind value = VirtualsKind::All) { - with_virtuals = value; + virtuals_kind = value; return *this; } @@ -63,17 +72,11 @@ struct GetColumnsOptions return *this; } - GetColumnsOptions & withSystemColumns(bool value = true) - { - with_system_columns = value; - return *this; - } - Kind kind; + VirtualsKind virtuals_kind = VirtualsKind::None; + bool with_subcolumns = false; - bool with_virtuals = false; bool with_extended_objects = false; - bool with_system_columns = false; }; /// Description of a single table column (in CREATE TABLE for example). @@ -160,6 +163,7 @@ public: bool hasNested(const String & column_name) const; bool hasSubcolumn(const String & column_name) const; const ColumnDescription & get(const String & column_name) const; + const ColumnDescription * tryGet(const String & column_name) const; template void modify(const String & column_name, F && f) @@ -213,9 +217,6 @@ public: /// Does column has non default specified compression codec bool hasCompressionCodec(const String & column_name) const; - CompressionCodecPtr getCodecOrDefault(const String & column_name, CompressionCodecPtr default_codec) const; - CompressionCodecPtr getCodecOrDefault(const String & column_name) const; - ASTPtr getCodecDescOrDefault(const String & column_name, CompressionCodecPtr default_codec) const; String toString() const; static ColumnsDescription parse(const String & str); @@ -269,4 +270,5 @@ private: /// don't have strange constructions in default expression like SELECT query or /// arrayJoin function. Block validateColumnsDefaultsAndGetSampleBlock(ASTPtr default_expr_list, const NamesAndTypesList & all_columns, ContextPtr context); + } diff --git a/src/Storages/FileLog/FileLogSource.cpp b/src/Storages/FileLog/FileLogSource.cpp index b1192af4ced..eb3ff0436a5 100644 --- a/src/Storages/FileLog/FileLogSource.cpp +++ b/src/Storages/FileLog/FileLogSource.cpp @@ -31,7 +31,7 @@ FileLogSource::FileLogSource( , max_streams_number(max_streams_number_) , handle_error_mode(handle_error_mode_) , non_virtual_header(storage_snapshot->metadata->getSampleBlockNonMaterialized()) - , virtual_header(storage_snapshot->getSampleBlockForColumns(storage.getVirtuals().getNames())) + , virtual_header(storage_snapshot->virtual_columns->getSampleBlock()) { consumer = std::make_unique(storage, max_block_size, poll_time_out, context, stream_number_, max_streams_number_); diff --git a/src/Storages/FileLog/StorageFileLog.cpp b/src/Storages/FileLog/StorageFileLog.cpp index b0bac081d38..a5f2331a068 100644 --- a/src/Storages/FileLog/StorageFileLog.cpp +++ b/src/Storages/FileLog/StorageFileLog.cpp @@ -147,6 +147,7 @@ StorageFileLog::StorageFileLog( storage_metadata.setColumns(columns_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals(filelog_settings->handle_error_mode)); if (!fileOrSymlinkPathStartsWith(path, getContext()->getUserFilesPath())) { @@ -203,6 +204,22 @@ StorageFileLog::StorageFileLog( } } +VirtualColumnsDescription StorageFileLog::createVirtuals(StreamingHandleErrorMode handle_error_mode) +{ + VirtualColumnsDescription desc; + + desc.addEphemeral("_filename", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_offset", std::make_shared(), ""); + + if (handle_error_mode == StreamingHandleErrorMode::STREAM) + { + desc.addEphemeral("_raw_record", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_error", std::make_shared(std::make_shared()), ""); + } + + return desc; +} + void StorageFileLog::loadMetaFiles(bool attach) { /// Attach table @@ -1009,19 +1026,4 @@ bool StorageFileLog::updateFileInfos() return events.empty() || file_infos.file_names.empty(); } -NamesAndTypesList StorageFileLog::getVirtuals() const -{ - auto virtuals = NamesAndTypesList{ - {"_filename", std::make_shared(std::make_shared())}, - {"_offset", std::make_shared()}}; - - if (filelog_settings->handle_error_mode == StreamingHandleErrorMode::STREAM) - { - virtuals.push_back({"_raw_record", std::make_shared(std::make_shared())}); - virtuals.push_back({"_error", std::make_shared(std::make_shared())}); - } - - return virtuals; -} - } diff --git a/src/Storages/FileLog/StorageFileLog.h b/src/Storages/FileLog/StorageFileLog.h index bdcf85ad6e5..91d58540c94 100644 --- a/src/Storages/FileLog/StorageFileLog.h +++ b/src/Storages/FileLog/StorageFileLog.h @@ -102,8 +102,6 @@ public: String getFullMetaPath(const String & file_name) const { return std::filesystem::path(metadata_base_path) / file_name; } String getFullDataPath(const String & file_name) const { return std::filesystem::path(root_data_path) / file_name; } - NamesAndTypesList getVirtuals() const override; - static UInt64 getInode(const String & file_name); void openFilesAndSetPos(); @@ -212,6 +210,8 @@ private: UInt64 inode = 0; }; ReadMetadataResult readMetadata(const String & filename) const; + + static VirtualColumnsDescription createVirtuals(StreamingHandleErrorMode handle_error_mode); }; } diff --git a/src/Storages/HDFS/StorageHDFS.cpp b/src/Storages/HDFS/StorageHDFS.cpp index cd935fa3100..c574f57fc6a 100644 --- a/src/Storages/HDFS/StorageHDFS.cpp +++ b/src/Storages/HDFS/StorageHDFS.cpp @@ -241,8 +241,7 @@ StorageHDFS::StorageHDFS( storage_metadata.setConstraints(constraints_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } namespace @@ -975,7 +974,7 @@ void StorageHDFS::read( size_t max_block_size, size_t num_streams) { - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(context_), virtual_columns); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(context_)); bool need_only_count = (query_info.optimize_trivial_count || read_from_format_info.requested_columns.empty()) && context_->getSettingsRef().optimize_count_from_files; @@ -1011,7 +1010,7 @@ void ReadFromHDFS::createIterator(const ActionsDAG::Node * predicate) else if (storage->is_path_with_globs) { /// Iterate through disclosed globs and make a source for each file - auto glob_iterator = std::make_shared(storage->uris[0], predicate, storage->virtual_columns, context); + auto glob_iterator = std::make_shared(storage->uris[0], predicate, storage->getVirtualsList(), context); iterator_wrapper = std::make_shared([glob_iterator]() { return glob_iterator->next(); @@ -1019,7 +1018,7 @@ void ReadFromHDFS::createIterator(const ActionsDAG::Node * predicate) } else { - auto uris_iterator = std::make_shared(storage->uris, predicate, storage->virtual_columns, context); + auto uris_iterator = std::make_shared(storage->uris, predicate, storage->getVirtualsList(), context); iterator_wrapper = std::make_shared([uris_iterator]() { return uris_iterator->next(); @@ -1179,16 +1178,6 @@ void registerStorageHDFS(StorageFactory & factory) }); } -NamesAndTypesList StorageHDFS::getVirtuals() const -{ - return virtual_columns; -} - -Names StorageHDFS::getVirtualColumnNames() -{ - return VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage({}).getNames(); -} - SchemaCache & StorageHDFS::getSchemaCache(const ContextPtr & ctx) { static SchemaCache schema_cache(ctx->getConfigRef().getUInt("schema_inference_cache_max_elements_for_hdfs", DEFAULT_SCHEMA_CACHE_ELEMENTS)); diff --git a/src/Storages/HDFS/StorageHDFS.h b/src/Storages/HDFS/StorageHDFS.h index b36ff7ea37e..47e5addccb4 100644 --- a/src/Storages/HDFS/StorageHDFS.h +++ b/src/Storages/HDFS/StorageHDFS.h @@ -69,9 +69,6 @@ public: ContextPtr local_context, TableExclusiveLockHolder &) override; - NamesAndTypesList getVirtuals() const override; - static Names getVirtualColumnNames(); - bool supportsPartitionBy() const override { return true; } /// Check if the format is column-oriented. @@ -114,7 +111,6 @@ private: const bool distributed_processing; ASTPtr partition_by; bool is_path_with_globs; - NamesAndTypesList virtual_columns; LoggerPtr log = getLogger("StorageHDFS"); }; diff --git a/src/Storages/HDFS/StorageHDFSCluster.cpp b/src/Storages/HDFS/StorageHDFSCluster.cpp index 714d6391543..bde8b84e349 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.cpp +++ b/src/Storages/HDFS/StorageHDFSCluster.cpp @@ -72,8 +72,7 @@ StorageHDFSCluster::StorageHDFSCluster( storage_metadata.setConstraints(constraints_); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } void StorageHDFSCluster::updateQueryToSendIfNeeded(DB::ASTPtr & query, const DB::StorageSnapshotPtr & storage_snapshot, const DB::ContextPtr & context) @@ -89,18 +88,11 @@ void StorageHDFSCluster::updateQueryToSendIfNeeded(DB::ASTPtr & query, const DB: RemoteQueryExecutor::Extension StorageHDFSCluster::getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const { - auto iterator = std::make_shared(uri, predicate, virtual_columns, context); + auto iterator = std::make_shared(uri, predicate, getVirtualsList(), context); auto callback = std::make_shared>([iter = std::move(iterator)]() mutable -> String { return iter->next().path; }); return RemoteQueryExecutor::Extension{.task_iterator = std::move(callback)}; } -NamesAndTypesList StorageHDFSCluster::getVirtuals() const -{ - return NamesAndTypesList{ - {"_path", std::make_shared(std::make_shared())}, - {"_file", std::make_shared(std::make_shared())}}; -} - } #endif diff --git a/src/Storages/HDFS/StorageHDFSCluster.h b/src/Storages/HDFS/StorageHDFSCluster.h index 40884f98984..26ebc8601ee 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.h +++ b/src/Storages/HDFS/StorageHDFSCluster.h @@ -32,8 +32,6 @@ public: std::string getName() const override { return "HDFSCluster"; } - NamesAndTypesList getVirtuals() const override; - RemoteQueryExecutor::Extension getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const override; bool supportsSubcolumns() const override { return true; } @@ -45,7 +43,6 @@ private: String uri; String format_name; - NamesAndTypesList virtual_columns; }; diff --git a/src/Storages/Hive/StorageHive.cpp b/src/Storages/Hive/StorageHive.cpp index 183a4532281..88ab8e15e76 100644 --- a/src/Storages/Hive/StorageHive.cpp +++ b/src/Storages/Hive/StorageHive.cpp @@ -45,6 +45,7 @@ #include #include #include +#include namespace CurrentMetrics { @@ -444,6 +445,7 @@ StorageHive::StorageHive( storage_metadata.partition_key = KeyDescription::getKeyFromAST(partition_by_ast, storage_metadata.columns, getContext()); setInMemoryMetadata(storage_metadata); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } void StorageHive::lazyInitialize() @@ -1020,13 +1022,6 @@ SinkToStoragePtr StorageHive::write(const ASTPtr & /*query*/, const StorageMetad throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Method write is not implemented for StorageHive"); } -NamesAndTypesList StorageHive::getVirtuals() const -{ - return NamesAndTypesList{ - {"_path", std::make_shared(std::make_shared())}, - {"_file", std::make_shared(std::make_shared())}}; -} - std::optional StorageHive::totalRows(const Settings & settings) const { /// query_info is not used when prune_level == PruneLevel::None diff --git a/src/Storages/Hive/StorageHive.h b/src/Storages/Hive/StorageHive.h index 07440097f7a..67ef153af0e 100644 --- a/src/Storages/Hive/StorageHive.h +++ b/src/Storages/Hive/StorageHive.h @@ -54,8 +54,6 @@ public: SinkToStoragePtr write(const ASTPtr & /*query*/, const StorageMetadataPtr & metadata_snapshot, ContextPtr /*context*/, bool async_insert) override; - NamesAndTypesList getVirtuals() const override; - bool supportsSubsetOfColumns() const; std::optional totalRows(const Settings & settings) const override; diff --git a/src/Storages/IStorage.cpp b/src/Storages/IStorage.cpp index 85ef6a0bb35..9852220241f 100644 --- a/src/Storages/IStorage.cpp +++ b/src/Storages/IStorage.cpp @@ -27,10 +27,17 @@ namespace ErrorCodes extern const int CANNOT_RESTORE_TABLE; } +IStorage::IStorage(StorageID storage_id_) + : storage_id(std::move(storage_id_)) + , metadata(std::make_unique()) + , virtuals(std::make_unique()) +{ +} + bool IStorage::isVirtualColumn(const String & column_name, const StorageMetadataPtr & metadata_snapshot) const { /// Virtual column maybe overridden by real column - return !metadata_snapshot->getColumns().has(column_name) && getVirtuals().contains(column_name); + return !metadata_snapshot->getColumns().has(column_name) && virtuals.get()->has(column_name); } RWLockImpl::LockHolder IStorage::tryLockTimed( @@ -237,11 +244,6 @@ void IStorage::renameInMemory(const StorageID & new_table_id) storage_id = new_table_id; } -NamesAndTypesList IStorage::getVirtuals() const -{ - return {}; -} - Names IStorage::getAllRegisteredNames() const { Names result; diff --git a/src/Storages/IStorage.h b/src/Storages/IStorage.h index 4fa6bfdd617..1108eafc6b6 100644 --- a/src/Storages/IStorage.h +++ b/src/Storages/IStorage.h @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -98,9 +99,7 @@ class IStorage : public std::enable_shared_from_this, public TypePromo public: IStorage() = delete; /// Storage metadata can be set separately in setInMemoryMetadata method - explicit IStorage(StorageID storage_id_) - : storage_id(std::move(storage_id_)) - , metadata(std::make_unique()) {} + explicit IStorage(StorageID storage_id_); IStorage(const IStorage &) = delete; IStorage & operator=(const IStorage &) = delete; @@ -215,6 +214,10 @@ public: metadata.set(std::make_unique(metadata_)); } + void setVirtuals(VirtualColumnsDescription virtuals_) + { + virtuals.set(std::make_unique(std::move(virtuals_))); + } /// Return list of virtual columns (like _part, _table, etc). In the vast /// majority of cases virtual columns are static constant part of Storage @@ -226,7 +229,9 @@ public: /// virtual column will be overridden and inaccessible. /// /// By default return empty list of columns. - virtual NamesAndTypesList getVirtuals() const; + VirtualsDescriptionPtr getVirtualsPtr() const { return virtuals.get(); } + NamesAndTypesList getVirtualsList() const { return virtuals.get()->getNamesAndTypesList(); } + Block getVirtualsHeader() const { return virtuals.get()->getSampleBlock(); } Names getAllRegisteredNames() const override; @@ -263,15 +268,16 @@ public: virtual bool supportsTrivialCountOptimization() const { return false; } private: - StorageID storage_id; mutable std::mutex id_mutex; - /// Multiversion storage metadata. Allows to read/write storage metadata - /// without locks. + /// Multiversion storage metadata. Allows to read/write storage metadata without locks. MultiVersionStorageMetadataPtr metadata; + /// Description of virtual columns. Optional, may be set in constructor. + MultiVersionVirtualsDescriptionPtr virtuals; + protected: RWLockImpl::LockHolder tryLockTimed( const RWLock & rwlock, RWLockImpl::Type type, const String & query_id, const std::chrono::milliseconds & acquire_timeout) const; diff --git a/src/Storages/IndicesDescription.h b/src/Storages/IndicesDescription.h index e56642b8c76..21ba5fb632e 100644 --- a/src/Storages/IndicesDescription.h +++ b/src/Storages/IndicesDescription.h @@ -2,7 +2,6 @@ #include -#include #include #include #include diff --git a/src/Storages/Kafka/KafkaSource.cpp b/src/Storages/Kafka/KafkaSource.cpp index dc62c13f633..9c68107872e 100644 --- a/src/Storages/Kafka/KafkaSource.cpp +++ b/src/Storages/Kafka/KafkaSource.cpp @@ -45,7 +45,7 @@ KafkaSource::KafkaSource( , max_block_size(max_block_size_) , commit_in_suffix(commit_in_suffix_) , non_virtual_header(storage_snapshot->metadata->getSampleBlockNonMaterialized()) - , virtual_header(storage_snapshot->getSampleBlockForColumns(storage.getVirtualColumnNames())) + , virtual_header(storage.getVirtualsHeader()) , handle_error_mode(storage.getStreamingHandleErrorMode()) { } diff --git a/src/Storages/Kafka/StorageKafka.cpp b/src/Storages/Kafka/StorageKafka.cpp index 638f5fe2ef6..e41488189e9 100644 --- a/src/Storages/Kafka/StorageKafka.cpp +++ b/src/Storages/Kafka/StorageKafka.cpp @@ -363,6 +363,8 @@ StorageKafka::StorageKafka( StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(columns_); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals(kafka_settings->kafka_handle_error_mode)); + auto task_count = thread_per_consumer ? num_consumers : 1; for (size_t i = 0; i < task_count; ++i) { @@ -384,6 +386,28 @@ StorageKafka::StorageKafka( }); } +VirtualColumnsDescription StorageKafka::createVirtuals(StreamingHandleErrorMode handle_error_mode) +{ + VirtualColumnsDescription desc; + + desc.addEphemeral("_topic", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_key", std::make_shared(), ""); + desc.addEphemeral("_offset", std::make_shared(), ""); + desc.addEphemeral("_partition", std::make_shared(), ""); + desc.addEphemeral("_timestamp", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_timestamp_ms", std::make_shared(std::make_shared(3)), ""); + desc.addEphemeral("_headers.name", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_headers.value", std::make_shared(std::make_shared()), ""); + + if (handle_error_mode == StreamingHandleErrorMode::STREAM) + { + desc.addEphemeral("_raw_message", std::make_shared(), ""); + desc.addEphemeral("_error", std::make_shared(), ""); + } + + return desc; +} + SettingsChanges StorageKafka::createSettingsAdjustments() { SettingsChanges result; @@ -1194,43 +1218,4 @@ void registerStorageKafka(StorageFactory & factory) factory.registerStorage("Kafka", creator_fn, StorageFactory::StorageFeatures{ .supports_settings = true, }); } -NamesAndTypesList StorageKafka::getVirtuals() const -{ - auto result = NamesAndTypesList{ - {"_topic", std::make_shared(std::make_shared())}, - {"_key", std::make_shared()}, - {"_offset", std::make_shared()}, - {"_partition", std::make_shared()}, - {"_timestamp", std::make_shared(std::make_shared())}, - {"_timestamp_ms", std::make_shared(std::make_shared(3))}, - {"_headers.name", std::make_shared(std::make_shared())}, - {"_headers.value", std::make_shared(std::make_shared())}}; - if (kafka_settings->kafka_handle_error_mode == StreamingHandleErrorMode::STREAM) - { - result.push_back({"_raw_message", std::make_shared()}); - result.push_back({"_error", std::make_shared()}); - } - return result; -} - -Names StorageKafka::getVirtualColumnNames() const -{ - auto result = Names { - "_topic", - "_key", - "_offset", - "_partition", - "_timestamp", - "_timestamp_ms", - "_headers.name", - "_headers.value", - }; - if (kafka_settings->kafka_handle_error_mode == StreamingHandleErrorMode::STREAM) - { - result.push_back({"_raw_message"}); - result.push_back({"_error"}); - } - return result; -} - } diff --git a/src/Storages/Kafka/StorageKafka.h b/src/Storages/Kafka/StorageKafka.h index d5e319b8974..3b20e6b23f6 100644 --- a/src/Storages/Kafka/StorageKafka.h +++ b/src/Storages/Kafka/StorageKafka.h @@ -74,8 +74,6 @@ public: const auto & getFormatName() const { return format_name; } - NamesAndTypesList getVirtuals() const override; - Names getVirtualColumnNames() const; StreamingHandleErrorMode getStreamingHandleErrorMode() const { return kafka_settings->kafka_handle_error_mode; } struct SafeConsumers @@ -158,6 +156,8 @@ private: bool checkDependencies(const StorageID & table_id); void cleanConsumers(); + + static VirtualColumnsDescription createVirtuals(StreamingHandleErrorMode handle_error_mode); }; } diff --git a/src/Storages/LightweightDeleteDescription.cpp b/src/Storages/LightweightDeleteDescription.cpp deleted file mode 100644 index ae5e68da9c2..00000000000 --- a/src/Storages/LightweightDeleteDescription.cpp +++ /dev/null @@ -1,9 +0,0 @@ -#include -#include - -namespace DB -{ - -const NameAndTypePair LightweightDeleteDescription::FILTER_COLUMN {"_row_exists", std::make_shared()}; - -} diff --git a/src/Storages/LightweightDeleteDescription.h b/src/Storages/LightweightDeleteDescription.h deleted file mode 100644 index 45bde59ea71..00000000000 --- a/src/Storages/LightweightDeleteDescription.h +++ /dev/null @@ -1,13 +0,0 @@ -#pragma once -#include -#include "Storages/TTLDescription.h" - -namespace DB -{ - -struct LightweightDeleteDescription -{ - static const NameAndTypePair FILTER_COLUMN; -}; - -} diff --git a/src/Storages/LiveView/StorageLiveView.cpp b/src/Storages/LiveView/StorageLiveView.cpp index 958e0a326cf..c3aacfd67d3 100644 --- a/src/Storages/LiveView/StorageLiveView.cpp +++ b/src/Storages/LiveView/StorageLiveView.cpp @@ -218,6 +218,10 @@ StorageLiveView::StorageLiveView( setInMemoryMetadata(storage_metadata); + VirtualColumnsDescription virtuals; + virtuals.addEphemeral("_version", std::make_shared(), ""); + setVirtuals(std::move(virtuals)); + if (!query.select) throw Exception(ErrorCodes::INCORRECT_QUERY, "SELECT query is not specified for {}", getName()); @@ -236,13 +240,6 @@ StorageLiveView::~StorageLiveView() shutdown(false); } -NamesAndTypesList StorageLiveView::getVirtuals() const -{ - return NamesAndTypesList{ - NameAndTypePair("_version", std::make_shared()) - }; -} - void StorageLiveView::checkTableCanBeDropped([[ maybe_unused ]] ContextPtr query_context) const { auto table_id = getStorageID(); diff --git a/src/Storages/LiveView/StorageLiveView.h b/src/Storages/LiveView/StorageLiveView.h index bf6b13fc837..91daac32c7b 100644 --- a/src/Storages/LiveView/StorageLiveView.h +++ b/src/Storages/LiveView/StorageLiveView.h @@ -74,8 +74,6 @@ public: bool supportsFinal() const override { return true; } - NamesAndTypesList getVirtuals() const override; - void checkTableCanBeDropped([[ maybe_unused ]] ContextPtr query_context) const override; void drop() override; diff --git a/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp b/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp index c6407a99a4e..ff9941ee808 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp @@ -623,6 +623,15 @@ void DataPartStorageOnDiskBase::remove( } } + if (!disk->exists(from)) + { + LOG_ERROR(log, "Directory {} (part to remove) doesn't exist or one of nested files has gone. Most likely this is due to manual removing. This should be discouraged. Ignoring.", fullPath(disk, from)); + /// We will never touch this part again, so unlocking it from zero-copy + if (!can_remove_description) + can_remove_description.emplace(can_remove_callback()); + return; + } + try { disk->moveDirectory(from, to); diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index 168c5f729ce..05e1129f9dc 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -526,14 +526,12 @@ std::pair Fetcher::fetchSelected creds.setPassword(password); } - std::unique_ptr in = std::make_unique( - uri, - Poco::Net::HTTPRequest::HTTP_POST, - nullptr, - creds, - DBMS_DEFAULT_BUFFER_SIZE, - 0, /* no redirects */ - context->getCommonFetchesSessionFactory()); + auto in = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::HTTP) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(timeouts) + .withDelayInit(false) + .create(creds); int server_protocol_version = parse(in->getResponseCookie("server_protocol_version", "0")); String remote_fs_metadata = parse(in->getResponseCookie("remote_fs_metadata", "")); @@ -557,11 +555,13 @@ std::pair Fetcher::fetchSelected if (server_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_SIZE) { readBinary(sum_files_size, *in); + if (server_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_SIZE_AND_TTL_INFOS) { IMergeTreeDataPart::TTLInfos ttl_infos; String ttl_infos_string; readBinary(ttl_infos_string, *in); + ReadBufferFromString ttl_infos_buffer(ttl_infos_string); assertString("ttl format version: 1\n", ttl_infos_buffer); ttl_infos.read(ttl_infos_buffer); @@ -609,6 +609,7 @@ std::pair Fetcher::fetchSelected } UInt64 revision = parse(in->getResponseCookie("disk_revision", "0")); + if (revision) disk->syncRevision(revision); @@ -743,7 +744,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( const UUID & part_uuid, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, size_t projections, bool is_projection, ThrottlerPtr throttler) @@ -799,7 +800,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( void Fetcher::downloadBaseOrProjectionPartToDisk( const String & replica_path, const MutableDataPartStoragePtr & data_part_storage, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, OutputBufferGetter output_buffer_getter, MergeTreeData::DataPart::Checksums & checksums, ThrottlerPtr throttler, @@ -807,6 +808,8 @@ void Fetcher::downloadBaseOrProjectionPartToDisk( { size_t files; readBinary(files, in); + LOG_DEBUG(log, "Downloading files {}", files); + std::vector> written_files; @@ -872,7 +875,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDisk( const String & tmp_prefix, DiskPtr disk, bool to_remote_disk, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, OutputBufferGetter output_buffer_getter, size_t projections, ThrottlerPtr throttler, diff --git a/src/Storages/MergeTree/DataPartsExchange.h b/src/Storages/MergeTree/DataPartsExchange.h index 8c15dc3cfdb..45a6cf83872 100644 --- a/src/Storages/MergeTree/DataPartsExchange.h +++ b/src/Storages/MergeTree/DataPartsExchange.h @@ -20,7 +20,7 @@ namespace DB { class StorageReplicatedMergeTree; -class PooledReadWriteBufferFromHTTP; +class ReadWriteBufferFromHTTP; namespace DataPartsExchange { @@ -94,7 +94,7 @@ private: void downloadBaseOrProjectionPartToDisk( const String & replica_path, const MutableDataPartStoragePtr & data_part_storage, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, OutputBufferGetter output_buffer_getter, MergeTreeData::DataPart::Checksums & checksums, ThrottlerPtr throttler, @@ -107,7 +107,7 @@ private: const String & tmp_prefix_, DiskPtr disk, bool to_remote_disk, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, OutputBufferGetter output_buffer_getter, size_t projections, ThrottlerPtr throttler, @@ -120,7 +120,7 @@ private: const UUID & part_uuid, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, size_t projections, bool is_projection, ThrottlerPtr throttler); @@ -131,7 +131,7 @@ private: bool to_detached, const String & tmp_prefix_, DiskPtr disk, - PooledReadWriteBufferFromHTTP & in, + ReadWriteBufferFromHTTP & in, size_t projections, MergeTreeData::DataPart::Checksums & checksums, ThrottlerPtr throttler); diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index a9bdceacef0..526ea542b0e 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -1451,6 +1451,11 @@ bool IMergeTreeDataPart::supportLightweightDeleteMutate() const parent_part == nullptr && projection_parts.empty(); } +bool IMergeTreeDataPart::hasLightweightDelete() const +{ + return columns.contains(RowExistsColumn::name); +} + void IMergeTreeDataPart::assertHasVersionMetadata(MergeTreeTransaction * txn) const { TransactionID expected_tid = txn ? txn->tid : Tx::PrehistoricTID; diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.h b/src/Storages/MergeTree/IMergeTreeDataPart.h index 91c559d30c8..fba1e6ddbb1 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.h +++ b/src/Storages/MergeTree/IMergeTreeDataPart.h @@ -1,12 +1,12 @@ #pragma once +#include #include #include #include #include #include #include -#include #include #include #include @@ -48,6 +48,8 @@ class MarkCache; class UncompressedCache; class MergeTreeTransaction; +struct MergeTreeReadTaskInfo; +using MergeTreeReadTaskInfoPtr = std::shared_ptr; enum class DataPartRemovalState { @@ -69,6 +71,7 @@ public: using Checksums = MergeTreeDataPartChecksums; using Checksum = MergeTreeDataPartChecksums::Checksum; using ValueSizeMap = std::map; + using VirtualFields = std::unordered_map; using MergeTreeReaderPtr = std::unique_ptr; using MergeTreeWriterPtr = std::unique_ptr; @@ -95,6 +98,7 @@ public: const NamesAndTypesList & columns_, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * uncompressed_cache, MarkCache * mark_cache, const AlterConversionsPtr & alter_conversions, @@ -493,7 +497,7 @@ public: bool supportLightweightDeleteMutate() const; /// True if here is lightweight deleted mask file in part. - bool hasLightweightDelete() const { return columns.contains(LightweightDeleteDescription::FILTER_COLUMN.name); } + bool hasLightweightDelete() const; void writeChecksums(const MergeTreeDataPartChecksums & checksums_, const WriteSettings & settings); diff --git a/src/Storages/MergeTree/IMergeTreeReader.cpp b/src/Storages/MergeTree/IMergeTreeReader.cpp index 63ed8021f58..4936f1d33c6 100644 --- a/src/Storages/MergeTree/IMergeTreeReader.cpp +++ b/src/Storages/MergeTree/IMergeTreeReader.cpp @@ -1,7 +1,8 @@ #include +#include +#include +#include #include -#include -#include #include #include #include @@ -19,12 +20,13 @@ namespace namespace ErrorCodes { extern const int LOGICAL_ERROR; + extern const int NOT_IMPLEMENTED; } - IMergeTreeReader::IMergeTreeReader( MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, const NamesAndTypesList & columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, UncompressedCache * uncompressed_cache_, MarkCache * mark_cache_, @@ -47,6 +49,7 @@ IMergeTreeReader::IMergeTreeReader( , part_columns(data_part_info_for_read->isWidePart() ? data_part_info_for_read->getColumnsDescriptionWithCollectedNested() : data_part_info_for_read->getColumnsDescription()) + , virtual_fields(virtual_fields_) { columns_to_read.reserve(requested_columns.size()); serializations.reserve(requested_columns.size()); @@ -63,7 +66,49 @@ const IMergeTreeReader::ValueSizeMap & IMergeTreeReader::getAvgValueSizeHints() return avg_value_size_hints; } -void IMergeTreeReader::fillMissingColumns(Columns & res_columns, bool & should_evaluate_missing_defaults, size_t num_rows, size_t block_number) const +void IMergeTreeReader::fillVirtualColumns(Columns & columns, size_t rows) const +{ + chassert(columns.size() == requested_columns.size()); + + const auto * loaded_part_info = typeid_cast(data_part_info_for_read.get()); + if (!loaded_part_info) + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Filling of virtual columns is supported only for LoadedMergeTreeDataPartInfoForReader"); + + const auto & data_part = loaded_part_info->getDataPart(); + const auto & storage_columns = storage_snapshot->getMetadataForQuery()->getColumns(); + const auto & virtual_columns = storage_snapshot->virtual_columns; + + auto it = requested_columns.begin(); + for (size_t pos = 0; pos < columns.size(); ++pos, ++it) + { + if (columns[pos] || storage_columns.has(it->name)) + continue; + + auto virtual_column = virtual_columns->tryGet(it->name); + if (!virtual_column) + continue; + + if (!it->type->equals(*virtual_column->type)) + { + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Data type for virtual column {} mismatched. Requested type: {}, virtual column type: {}", + it->name, it->type->getName(), virtual_column->type->getName()); + } + + if (it->name == "_part_offset") + throw Exception(ErrorCodes::LOGICAL_ERROR, "Virtual column {} must be filled by range reader", it->name); + + Field field; + if (auto field_it = virtual_fields.find(it->name); field_it != virtual_fields.end()) + field = field_it->second; + else + field = getFieldForConstVirtualColumn(it->name, *data_part); + + columns[pos] = virtual_column->type->createColumnConst(rows, field)->convertToFullColumnIfConst(); + } +} + +void IMergeTreeReader::fillMissingColumns(Columns & res_columns, bool & should_evaluate_missing_defaults, size_t num_rows) const { try { @@ -72,7 +117,7 @@ void IMergeTreeReader::fillMissingColumns(Columns & res_columns, bool & should_e res_columns, num_rows, Nested::convertToSubcolumns(requested_columns), Nested::convertToSubcolumns(available_columns), - partially_read_columns, storage_snapshot->metadata, block_number); + partially_read_columns, storage_snapshot->metadata); should_evaluate_missing_defaults = std::any_of( res_columns.begin(), res_columns.end(), [](const auto & column) { return column == nullptr; }); diff --git a/src/Storages/MergeTree/IMergeTreeReader.h b/src/Storages/MergeTree/IMergeTreeReader.h index 997be064f28..a5b84eba241 100644 --- a/src/Storages/MergeTree/IMergeTreeReader.h +++ b/src/Storages/MergeTree/IMergeTreeReader.h @@ -9,8 +9,6 @@ namespace DB { -class IDataType; - /// Reads the data between pairs of marks in the same part. When reading consecutive ranges, avoids unnecessary seeks. /// When ranges are almost consecutive, seeks are fast because they are performed inside the buffer. /// Avoids loading the marks file if it is not needed (e.g. when reading the whole part). @@ -18,11 +16,13 @@ class IMergeTreeReader : private boost::noncopyable { public: using ValueSizeMap = std::map; + using VirtualFields = std::unordered_map; using DeserializeBinaryBulkStateMap = std::map; IMergeTreeReader( MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, const NamesAndTypesList & columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, UncompressedCache * uncompressed_cache_, MarkCache * mark_cache_, @@ -42,10 +42,13 @@ public: const ValueSizeMap & getAvgValueSizeHints() const; + /// Add virtual columns that are not present in the block. + void fillVirtualColumns(Columns & columns, size_t rows) const; + /// Add columns from ordered_names that are not present in the block. /// Missing columns are added in the order specified by ordered_names. /// num_rows is needed in case if all res_columns are nullptr. - void fillMissingColumns(Columns & res_columns, bool & should_evaluate_missing_defaults, size_t num_rows, size_t block_number = 0) const; + void fillMissingColumns(Columns & res_columns, bool & should_evaluate_missing_defaults, size_t num_rows) const; /// Evaluate defaulted columns if necessary. void evaluateMissingDefaults(Block additional_columns, Columns & res_columns) const; @@ -113,6 +116,9 @@ private: /// Actual columns description in part. const ColumnsDescription & part_columns; + + /// Fields of virtual columns that were filled in previous stages. + VirtualFields virtual_fields; }; } diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp index df64ae33713..06b6927bc43 100644 --- a/src/Storages/MergeTree/MergeTask.cpp +++ b/src/Storages/MergeTree/MergeTask.cpp @@ -8,7 +8,6 @@ #include #include #include -#include #include #include @@ -297,7 +296,7 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() switch (global_ctx->chosen_merge_algorithm) { - case MergeAlgorithm::Horizontal : + case MergeAlgorithm::Horizontal: { global_ctx->merging_columns = global_ctx->storage_columns; global_ctx->merging_column_names = global_ctx->all_column_names; @@ -305,7 +304,7 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() global_ctx->gathering_column_names.clear(); break; } - case MergeAlgorithm::Vertical : + case MergeAlgorithm::Vertical: { ctx->rows_sources_uncompressed_write_buf = ctx->tmp_disk->createRawStream(); ctx->rows_sources_write_buf = std::make_unique(*ctx->rows_sources_uncompressed_write_buf); @@ -1075,14 +1074,18 @@ void MergeTask::ExecuteAndFinalizeHorizontalPart::createMergedStream() if (global_ctx->deduplicate) { - /// We don't want to deduplicate by block number column - /// so if deduplicate_by_columns is empty, add all columns except _block_number - if (supportsBlockNumberColumn(global_ctx) && global_ctx->deduplicate_by_columns.empty()) + const auto & virtuals = *global_ctx->data->getVirtualsPtr(); + + /// We don't want to deduplicate by virtual persistent column. + /// If deduplicate_by_columns is empty, add all columns except virtuals. + if (global_ctx->deduplicate_by_columns.empty()) { - for (const auto & col : global_ctx->merging_column_names) + for (const auto & column_name : global_ctx->merging_column_names) { - if (col != BlockNumberColumn::name) - global_ctx->deduplicate_by_columns.emplace_back(col); + if (virtuals.tryGet(column_name, VirtualsKind::Persistent)) + continue; + + global_ctx->deduplicate_by_columns.emplace_back(column_name); } } diff --git a/src/Storages/MergeTree/MergeTask.h b/src/Storages/MergeTree/MergeTask.h index 7fb4797e482..1f50e55f8a0 100644 --- a/src/Storages/MergeTree/MergeTask.h +++ b/src/Storages/MergeTree/MergeTask.h @@ -15,7 +15,7 @@ #include -#include +#include #include #include #include diff --git a/src/Storages/MergeTree/MergeTreeBlockReadUtils.cpp b/src/Storages/MergeTree/MergeTreeBlockReadUtils.cpp index f5f0fa6f726..67f5e7a53e8 100644 --- a/src/Storages/MergeTree/MergeTreeBlockReadUtils.cpp +++ b/src/Storages/MergeTree/MergeTreeBlockReadUtils.cpp @@ -6,6 +6,7 @@ #include #include #include +#include #include #include #include @@ -106,16 +107,14 @@ NameSet injectRequiredColumns( auto options = GetColumnsOptions(GetColumnsOptions::AllPhysical) .withExtendedObjects() - .withSystemColumns(); - - if (with_subcolumns) - options.withSubcolumns(); + .withVirtuals() + .withSubcolumns(with_subcolumns); for (size_t i = 0; i < columns.size(); ++i) { - /// We are going to fetch only physical columns and system columns + /// We are going to fetch physical columns and system columns first if (!storage_snapshot->tryGetColumn(options, columns[i])) - throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "There is no physical column or subcolumn {} in table", columns[i]); + throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "There is no column or subcolumn {} in table", columns[i]); have_at_least_one_physical_column |= injectRequiredColumnsRecursively( columns[i], storage_snapshot, alter_conversions, @@ -258,11 +257,10 @@ void MergeTreeBlockSizePredictor::update(const Block & sample_block, const Colum } -MergeTreeReadTask::Columns getReadTaskColumns( +MergeTreeReadTaskColumns getReadTaskColumns( const IMergeTreeDataPartInfoForReader & data_part_info_for_reader, const StorageSnapshotPtr & storage_snapshot, const Names & required_columns, - const Names & system_columns, const PrewhereInfoPtr & prewhere_info, const ExpressionActionsSettings & actions_settings, const MergeTreeReaderSettings & reader_settings, @@ -270,28 +268,30 @@ MergeTreeReadTask::Columns getReadTaskColumns( { Names column_to_read_after_prewhere = required_columns; - /// Read system columns such as lightweight delete mask "_row_exists" if it is persisted in the part - for (const auto & name : system_columns) - if (data_part_info_for_reader.getColumns().contains(name)) - column_to_read_after_prewhere.push_back(name); - /// Inject columns required for defaults evaluation injectRequiredColumns( data_part_info_for_reader, storage_snapshot, with_subcolumns, column_to_read_after_prewhere); - MergeTreeReadTask::Columns result; + MergeTreeReadTaskColumns result; auto options = GetColumnsOptions(GetColumnsOptions::All) .withExtendedObjects() - .withSystemColumns(); + .withVirtuals() + .withSubcolumns(with_subcolumns); - if (with_subcolumns) - options.withSubcolumns(); + static const NameSet columns_to_read_at_first_step = {"_part_offset"}; NameSet columns_from_previous_steps; auto add_step = [&](const PrewhereExprStep & step) { Names step_column_names; + if (columns_from_previous_steps.empty()) + { + for (const auto & required_column : required_columns) + if (columns_to_read_at_first_step.contains(required_column)) + step_column_names.push_back(required_column); + } + /// Computation results from previous steps might be used in the current step as well. In such a case these /// computed columns will be present in the current step inputs. They don't need to be read from the disk so /// exclude them from the list of columns to read. This filtering must be done before injecting required diff --git a/src/Storages/MergeTree/MergeTreeBlockReadUtils.h b/src/Storages/MergeTree/MergeTreeBlockReadUtils.h index 9417d47814a..b19c42c8db8 100644 --- a/src/Storages/MergeTree/MergeTreeBlockReadUtils.h +++ b/src/Storages/MergeTree/MergeTreeBlockReadUtils.h @@ -1,6 +1,5 @@ #pragma once -#include #include #include @@ -22,11 +21,10 @@ NameSet injectRequiredColumns( bool with_subcolumns, Names & columns); -MergeTreeReadTask::Columns getReadTaskColumns( +MergeTreeReadTaskColumns getReadTaskColumns( const IMergeTreeDataPartInfoForReader & data_part_info_for_reader, const StorageSnapshotPtr & storage_snapshot, const Names & required_columns, - const Names & system_columns, const PrewhereInfoPtr & prewhere_info, const ExpressionActionsSettings & actions_settings, const MergeTreeReaderSettings & reader_settings, diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 8aa188cfe5c..9bc360cbcac 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -67,7 +67,7 @@ #include #include #include -#include +#include #include #include #include @@ -430,6 +430,29 @@ MergeTreeData::MergeTreeData( }; } +VirtualColumnsDescription MergeTreeData::createVirtuals(const StorageInMemoryMetadata & metadata) +{ + VirtualColumnsDescription desc; + + desc.addEphemeral("_part", std::make_shared(std::make_shared()), "Name of part"); + desc.addEphemeral("_part_index", std::make_shared(), "Sequential index of the part in the query result"); + desc.addEphemeral("_part_uuid", std::make_shared(), "Unique part identifier (if enabled MergeTree setting assign_part_uuids)"); + desc.addEphemeral("_partition_id", std::make_shared(std::make_shared()), "Name of partition"); + desc.addEphemeral("_sample_factor", std::make_shared(), "Sample factor (from the query)"); + desc.addEphemeral("_part_offset", std::make_shared(), "Number of row in the part"); + + if (metadata.hasPartitionKey()) + { + auto partition_types = metadata.partition_key.sample_block.getDataTypes(); + desc.addEphemeral("_partition_value", std::make_shared(std::move(partition_types)), "Value (a tuple) of a PARTITION BY expression"); + } + + desc.addPersistent(RowExistsColumn::name, RowExistsColumn::type, nullptr, "Persisted mask created by lightweight delete that show whether row exists or is deleted"); + desc.addPersistent(BlockNumberColumn::name, BlockNumberColumn::type, BlockNumberColumn::codec, "Persisted original number of block that was assigned at insert"); + + return desc; +} + StoragePolicyPtr MergeTreeData::getStoragePolicy() const { auto settings = getSettings(); @@ -677,6 +700,7 @@ void MergeTreeData::setProperties( { checkProperties(new_metadata, old_metadata, attach, false, allow_nullable_key, local_context); setInMemoryMetadata(new_metadata); + setVirtuals(createVirtuals(new_metadata)); } namespace @@ -1002,73 +1026,38 @@ void MergeTreeData::MergingParams::check(const StorageInMemoryMetadata & metadat /// TODO Checks for Graphite mode. } +const Names MergeTreeData::virtuals_useful_for_filter = {"_part", "_partition_id", "_part_uuid", "_partition_value"}; -DataTypePtr MergeTreeData::getPartitionValueType() const +Block MergeTreeData::getHeaderWithVirtualsForFilter() const { - DataTypePtr partition_value_type; - auto partition_types = getInMemoryMetadataPtr()->partition_key.sample_block.getDataTypes(); - if (partition_types.empty()) - partition_value_type = std::make_shared(); - else - partition_value_type = std::make_shared(std::move(partition_types)); - return partition_value_type; + Block header; + auto virtuals_desc = getVirtualsPtr(); + for (const auto & name : virtuals_useful_for_filter) + if (auto column = virtuals_desc->tryGet(name)) + header.insert({column->type->createColumn(), column->type, name}); + return header; } - -Block MergeTreeData::getSampleBlockWithVirtualColumns() const +Block MergeTreeData::getBlockWithVirtualsForFilter(const MergeTreeData::DataPartsVector & parts, bool ignore_empty) const { - DataTypePtr partition_value_type = getPartitionValueType(); - return { - ColumnWithTypeAndName( - DataTypeLowCardinality{std::make_shared()}.createColumn(), - std::make_shared(std::make_shared()), - "_part"), - ColumnWithTypeAndName( - DataTypeLowCardinality{std::make_shared()}.createColumn(), - std::make_shared(std::make_shared()), - "_partition_id"), - ColumnWithTypeAndName(ColumnUUID::create(), std::make_shared(), "_part_uuid"), - ColumnWithTypeAndName(partition_value_type->createColumn(), partition_value_type, "_partition_value")}; -} + auto block = getHeaderWithVirtualsForFilter(); - -Block MergeTreeData::getBlockWithVirtualPartColumns(const MergeTreeData::DataPartsVector & parts, bool one_part, bool ignore_empty) const -{ - auto block = getSampleBlockWithVirtualColumns(); - MutableColumns columns = block.mutateColumns(); - - auto & part_column = columns[0]; - auto & partition_id_column = columns[1]; - auto & part_uuid_column = columns[2]; - auto & partition_value_column = columns[3]; - - bool has_partition_value = typeid_cast(partition_value_column.get()); for (const auto & part_or_projection : parts) { if (ignore_empty && part_or_projection->isEmpty()) continue; - const auto * part = part_or_projection->isProjectionPart() ? part_or_projection->getParentPart() : part_or_projection.get(); - part_column->insert(part->name); - partition_id_column->insert(part->info.partition_id); - part_uuid_column->insert(part->uuid); - Tuple tuple(part->partition.value.begin(), part->partition.value.end()); - if (has_partition_value) - partition_value_column->insert(tuple); - if (one_part) + const auto * part = part_or_projection->isProjectionPart() + ? part_or_projection->getParentPart() + : part_or_projection.get(); + + for (auto & column : block) { - part_column = ColumnConst::create(std::move(part_column), 1); - partition_id_column = ColumnConst::create(std::move(partition_id_column), 1); - part_uuid_column = ColumnConst::create(std::move(part_uuid_column), 1); - if (has_partition_value) - partition_value_column = ColumnConst::create(std::move(partition_value_column), 1); - break; + auto field = getFieldForConstVirtualColumn(column.name, *part); + column.column->assumeMutableRef().insert(field); } } - block.setColumns(std::move(columns)); - if (!has_partition_value) - block.erase("_partition_value"); return block; } @@ -1077,13 +1066,14 @@ std::optional MergeTreeData::totalRowsByPartitionPredicateImpl( const ActionsDAGPtr & filter_actions_dag, ContextPtr local_context, const DataPartsVector & parts) const { if (parts.empty()) - return 0u; + return 0; + auto metadata_snapshot = getInMemoryMetadataPtr(); - Block virtual_columns_block = getBlockWithVirtualPartColumns(parts, true /* one_part */); + auto virtual_columns_block = getBlockWithVirtualsForFilter({parts[0]}); auto filter_dag = VirtualColumnUtils::splitFilterDagForAllowedInputs(filter_actions_dag->getOutputs().at(0), nullptr); - // Generate valid expressions for filtering + /// Generate valid expressions for filtering bool valid = true; for (const auto * input : filter_dag->getInputs()) if (!virtual_columns_block.has(input->result_name)) @@ -1096,7 +1086,7 @@ std::optional MergeTreeData::totalRowsByPartitionPredicateImpl( std::unordered_set part_values; if (valid) { - virtual_columns_block = getBlockWithVirtualPartColumns(parts, false /* one_part */); + virtual_columns_block = getBlockWithVirtualsForFilter(parts); VirtualColumnUtils::filterBlockWithDAG(filter_dag, virtual_columns_block, local_context); part_values = VirtualColumnUtils::extractSingleValueFromBlock(virtual_columns_block, "_part"); if (part_values.empty()) @@ -3658,6 +3648,7 @@ void MergeTreeData::checkPartDynamicColumns(MutableDataPartPtr & part, DataParts { auto metadata_snapshot = getInMemoryMetadataPtr(); const auto & columns = metadata_snapshot->getColumns(); + auto virtuals = getVirtualsPtr(); if (!hasDynamicSubcolumns(columns)) return; @@ -3665,7 +3656,7 @@ void MergeTreeData::checkPartDynamicColumns(MutableDataPartPtr & part, DataParts const auto & part_columns = part->getColumns(); for (const auto & part_column : part_columns) { - if (part_column.name == LightweightDeleteDescription::FILTER_COLUMN.name || part_column.name == BlockNumberColumn::name) + if (virtuals->has(part_column.name)) continue; auto storage_column = columns.getPhysical(part_column.name); @@ -6669,14 +6660,6 @@ Block MergeTreeData::getMinMaxCountProjectionBlock( const auto & primary_key_max_column_name = metadata_snapshot->minmax_count_projection->primary_key_max_column_name; NameSet required_columns_set(required_columns.begin(), required_columns.end()); - if (required_columns_set.contains("_partition_value") && !typeid_cast(getPartitionValueType().get())) - { - throw Exception( - ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, - "Missing column `_partition_value` because there is no partition column in table {}", - getStorageID().getTableName()); - } - if (!primary_key_max_column_name.empty()) need_primary_key_max_column = required_columns_set.contains(primary_key_max_column_name); @@ -6702,11 +6685,11 @@ Block MergeTreeData::getMinMaxCountProjectionBlock( }; Block virtual_columns_block; - auto virtual_block = getSampleBlockWithVirtualColumns(); + auto virtual_block = getHeaderWithVirtualsForFilter(); bool has_virtual_column = std::any_of(required_columns.begin(), required_columns.end(), [&](const auto & name) { return virtual_block.has(name); }); if (has_virtual_column || filter_dag) { - virtual_columns_block = getBlockWithVirtualPartColumns(parts, false /* one_part */, true /* ignore_empty */); + virtual_columns_block = getBlockWithVirtualsForFilter(parts, /*ignore_empty=*/ true); if (virtual_columns_block.rows() == 0) return {}; } @@ -7952,21 +7935,6 @@ AlterConversionsPtr MergeTreeData::getAlterConversionsForPart(MergeTreeDataPartP return result; } -NamesAndTypesList MergeTreeData::getVirtuals() const -{ - return NamesAndTypesList{ - NameAndTypePair("_part", std::make_shared(std::make_shared())), - NameAndTypePair("_part_index", std::make_shared()), - NameAndTypePair("_part_uuid", std::make_shared()), - NameAndTypePair("_partition_id", std::make_shared(std::make_shared())), - NameAndTypePair("_partition_value", getPartitionValueType()), - NameAndTypePair("_sample_factor", std::make_shared()), - NameAndTypePair("_part_offset", std::make_shared()), - LightweightDeleteDescription::FILTER_COLUMN, - NameAndTypePair(BlockNumberColumn::name, BlockNumberColumn::type), - }; -} - size_t MergeTreeData::getTotalMergesWithTTLInMergeList() const { return getContext()->getMergeList().getMergesWithTTLCount(); diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 4475f2b6f12..dc84505f38f 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -444,8 +444,6 @@ public: bool supportsTrivialCountOptimization() const override { return !hasLightweightDeletedMask(); } - NamesAndTypesList getVirtuals() const override; - /// Snapshot for MergeTree contains the current set of data parts /// at the moment of the start of query. struct SnapshotData : public StorageSnapshot::Data @@ -988,15 +986,13 @@ public: void removeQueryId(const String & query_id) const; void removeQueryIdNoLock(const String & query_id) const TSA_REQUIRES(query_id_set_mutex); - /// Return the partition expression types as a Tuple type. Return DataTypeUInt8 if partition expression is empty. - DataTypePtr getPartitionValueType() const; + static const Names virtuals_useful_for_filter; /// Construct a sample block of virtual columns. - Block getSampleBlockWithVirtualColumns() const; + Block getHeaderWithVirtualsForFilter() const; /// Construct a block consisting only of possible virtual columns for part pruning. - /// If one_part is true, fill in at most one part. - Block getBlockWithVirtualPartColumns(const MergeTreeData::DataPartsVector & parts, bool one_part, bool ignore_empty = false) const; + Block getBlockWithVirtualsForFilter(const MergeTreeData::DataPartsVector & parts, bool ignore_empty = false) const; /// In merge tree we do inserts with several steps. One of them: /// X. write part to temporary directory with some temp name @@ -1087,6 +1083,8 @@ public: bool initializeDiskOnConfigChange(const std::set & /*new_added_disks*/) override; + static VirtualColumnsDescription createVirtuals(const StorageInMemoryMetadata & metadata); + protected: friend class IMergeTreeDataPart; friend class MergeTreeDataMergerMutator; diff --git a/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp index 0ecd7abe183..9f201ab3b81 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp @@ -5,7 +5,6 @@ #include #include #include -#include namespace DB @@ -33,6 +32,7 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartCompact::getReader( const NamesAndTypesList & columns_to_read, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * uncompressed_cache, MarkCache * mark_cache, const AlterConversionsPtr & alter_conversions, @@ -41,12 +41,21 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartCompact::getReader( const ReadBufferFromFileBase::ProfileCallback & profile_callback) const { auto read_info = std::make_shared(shared_from_this(), alter_conversions); - auto * load_marks_threadpool = reader_settings.read_settings.load_marks_asynchronously ? &read_info->getContext()->getLoadMarksThreadpool() : nullptr; + auto * load_marks_threadpool + = reader_settings.read_settings.load_marks_asynchronously ? &read_info->getContext()->getLoadMarksThreadpool() : nullptr; return std::make_unique( - read_info, columns_to_read, storage_snapshot, uncompressed_cache, - mark_cache, mark_ranges, reader_settings, load_marks_threadpool, - avg_value_size_hints, profile_callback); + read_info, + columns_to_read, + virtual_fields, + storage_snapshot, + uncompressed_cache, + mark_cache, + mark_ranges, + reader_settings, + load_marks_threadpool, + avg_value_size_hints, + profile_callback); } IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartCompact::getWriter( @@ -66,12 +75,6 @@ IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartCompact::getWriter( ordered_columns_list.sort([this](const auto & lhs, const auto & rhs) { return *getColumnPosition(lhs.name) < *getColumnPosition(rhs.name); }); - /// _block_number column is not added by user, but is persisted in a part after merge - /// If _block_number is not present in the parts to be merged, then it won't have a position - /// So check if its not present and add it at the end - if (columns_list.contains(BlockNumberColumn::name) && !ordered_columns_list.contains(BlockNumberColumn::name)) - ordered_columns_list.emplace_back(NameAndTypePair{BlockNumberColumn::name, BlockNumberColumn::type}); - return std::make_unique( shared_from_this(), ordered_columns_list, metadata_snapshot, indices_to_recalc, stats_to_recalc_, getMarksFileExtension(), diff --git a/src/Storages/MergeTree/MergeTreeDataPartCompact.h b/src/Storages/MergeTree/MergeTreeDataPartCompact.h index 35a358b3720..a97d15a08f3 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartCompact.h +++ b/src/Storages/MergeTree/MergeTreeDataPartCompact.h @@ -32,6 +32,7 @@ public: const NamesAndTypesList & columns, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * uncompressed_cache, MarkCache * mark_cache, const AlterConversionsPtr & alter_conversions, diff --git a/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp b/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp index 2f01dbfe04b..e023ae9be0e 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp @@ -33,6 +33,7 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartInMemory::getReader( const NamesAndTypesList & columns_to_read, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * /* uncompressed_cache */, MarkCache * /* mark_cache */, const AlterConversionsPtr & alter_conversions, @@ -44,7 +45,13 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartInMemory::getReader( auto ptr = std::static_pointer_cast(shared_from_this()); return std::make_unique( - read_info, ptr, columns_to_read, storage_snapshot, mark_ranges, reader_settings); + read_info, + ptr, + columns_to_read, + virtual_fields, + storage_snapshot, + mark_ranges, + reader_settings); } IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartInMemory::getWriter( diff --git a/src/Storages/MergeTree/MergeTreeDataPartInMemory.h b/src/Storages/MergeTree/MergeTreeDataPartInMemory.h index 27f8ba4bccb..90b4b0e3471 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartInMemory.h +++ b/src/Storages/MergeTree/MergeTreeDataPartInMemory.h @@ -21,6 +21,7 @@ public: const NamesAndTypesList & columns, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * uncompressed_cache, MarkCache * mark_cache, const AlterConversionsPtr & alter_conversions, diff --git a/src/Storages/MergeTree/MergeTreeDataPartWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWide.cpp index dc6c1f0019d..018b8a35534 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWide.cpp @@ -31,6 +31,7 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartWide::getReader( const NamesAndTypesList & columns_to_read, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * uncompressed_cache, MarkCache * mark_cache, const AlterConversionsPtr & alter_conversions, @@ -40,10 +41,16 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartWide::getReader( { auto read_info = std::make_shared(shared_from_this(), alter_conversions); return std::make_unique( - read_info, columns_to_read, - storage_snapshot, uncompressed_cache, - mark_cache, mark_ranges, reader_settings, - avg_value_size_hints, profile_callback); + read_info, + columns_to_read, + virtual_fields, + storage_snapshot, + uncompressed_cache, + mark_cache, + mark_ranges, + reader_settings, + avg_value_size_hints, + profile_callback); } IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartWide::getWriter( diff --git a/src/Storages/MergeTree/MergeTreeDataPartWide.h b/src/Storages/MergeTree/MergeTreeDataPartWide.h index 14147c4ad56..a8710dad679 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWide.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWide.h @@ -27,6 +27,7 @@ public: const NamesAndTypesList & columns, const StorageSnapshotPtr & storage_snapshot, const MarkRanges & mark_ranges, + const VirtualFields & virtual_fields, UncompressedCache * uncompressed_cache, MarkCache * mark_cache, const AlterConversionsPtr & alter_conversions, diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp index b05b4584259..1721fd15b8d 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp @@ -1,12 +1,9 @@ #include #include -#include namespace DB { - CompressionCodecPtr getCompressionCodecDelta(UInt8 delta_bytes_size); - namespace ErrorCodes { extern const int LOGICAL_ERROR; @@ -55,14 +52,10 @@ MergeTreeDataPartWriterCompact::MergeTreeDataPartWriterCompact( marks_source_hashing = std::make_unique(*marks_compressor); } - const auto & storage_columns = metadata_snapshot->getColumns(); + auto storage_snapshot = std::make_shared(data_part->storage, metadata_snapshot); for (const auto & column : columns_list) { - ASTPtr compression; - if (column.name == BlockNumberColumn::name) - compression = BlockNumberColumn::compression_codec->getFullCodecDesc(); - else - compression = storage_columns.getCodecDescOrDefault(column.name, default_codec); + auto compression = storage_snapshot->getCodecDescOrDefault(column.name, default_codec); addStreams(column, compression); } } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp index 9d373504473..d79590ded21 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp @@ -6,12 +6,10 @@ #include #include #include -#include #include namespace DB { - CompressionCodecPtr getCompressionCodecDelta(UInt8 delta_bytes_size); namespace ErrorCodes { @@ -91,15 +89,11 @@ MergeTreeDataPartWriterWide::MergeTreeDataPartWriterWide( indices_to_recalc_, stats_to_recalc_, marks_file_extension_, default_codec_, settings_, index_granularity_) { - const auto & columns = metadata_snapshot->getColumns(); - for (const auto & it : columns_list) + auto storage_snapshot = std::make_shared(data_part->storage, metadata_snapshot); + for (const auto & column : columns_list) { - ASTPtr compression; - if (it.name == BlockNumberColumn::name) - compression = BlockNumberColumn::compression_codec->getFullCodecDesc(); - else - compression = columns.getCodecDescOrDefault(it.name, default_codec); - addStreams(it, compression); + auto compression = storage_snapshot->getCodecDescOrDefault(column.name, default_codec); + addStreams(column, compression); } } diff --git a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp index 00aeac9cef4..428c8f92931 100644 --- a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp +++ b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp @@ -46,7 +46,6 @@ #include #include -#include #include namespace CurrentMetrics @@ -69,7 +68,6 @@ namespace ErrorCodes extern const int CANNOT_PARSE_TEXT; extern const int TOO_MANY_PARTITIONS; extern const int DUPLICATED_PART_UUIDS; - extern const int NO_SUCH_COLUMN_IN_TABLE; } @@ -166,7 +164,6 @@ MergeTreeDataSelectSamplingData MergeTreeDataSelectExecutor::getSampling( const MergeTreeData & data, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, - bool sample_factor_column_queried, LoggerPtr log) { const Settings & settings = context->getSettingsRef(); @@ -296,7 +293,7 @@ MergeTreeDataSelectSamplingData MergeTreeDataSelectExecutor::getSampling( if (sampling.use_sampling) { - if (sample_factor_column_queried && relative_sample_size != RelativeSize(0)) + if (relative_sample_size != RelativeSize(0)) sampling.used_sample_factor = 1.0 / boost::rational_cast(relative_sample_size); RelativeSize size_of_universum = 0; @@ -483,12 +480,13 @@ std::optional> MergeTreeDataSelectExecutor::filterPar { if (!filter_dag) return {}; - auto sample = data.getSampleBlockWithVirtualColumns(); + + auto sample = data.getHeaderWithVirtualsForFilter(); auto dag = VirtualColumnUtils::splitFilterDagForAllowedInputs(filter_dag->getOutputs().at(0), &sample); if (!dag) return {}; - auto virtual_columns_block = data.getBlockWithVirtualPartColumns(parts, false /* one_part */); + auto virtual_columns_block = data.getBlockWithVirtualsForFilter(parts); VirtualColumnUtils::filterBlockWithDAG(dag, virtual_columns_block, context); return VirtualColumnUtils::extractSingleValueFromBlock(virtual_columns_block, "_part"); } @@ -868,69 +866,6 @@ std::shared_ptr MergeTreeDataSelectExecutor::checkLimits( return nullptr; } -static void selectColumnNames( - const Names & column_names_to_return, - const MergeTreeData & data, - Names & real_column_names, - Names & virt_column_names, - bool & sample_factor_column_queried) -{ - sample_factor_column_queried = false; - - for (const String & name : column_names_to_return) - { - if (name == "_part") - { - virt_column_names.push_back(name); - } - else if (name == "_part_index") - { - virt_column_names.push_back(name); - } - else if (name == "_partition_id") - { - virt_column_names.push_back(name); - } - else if (name == "_part_offset") - { - virt_column_names.push_back(name); - } - else if (name == LightweightDeleteDescription::FILTER_COLUMN.name) - { - virt_column_names.push_back(name); - } - else if (name == BlockNumberColumn::name) - { - virt_column_names.push_back(name); - } - else if (name == "_part_uuid") - { - virt_column_names.push_back(name); - } - else if (name == "_partition_value") - { - if (!typeid_cast(data.getPartitionValueType().get())) - { - throw Exception( - ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, - "Missing column `_partition_value` because there is no partition column in table {}", - data.getStorageID().getTableName()); - } - - virt_column_names.push_back(name); - } - else if (name == "_sample_factor") - { - sample_factor_column_queried = true; - virt_column_names.push_back(name); - } - else - { - real_column_names.push_back(name); - } - } -} - ReadFromMergeTree::AnalysisResultPtr MergeTreeDataSelectExecutor::estimateNumMarksToRead( MergeTreeData::DataPartsVector parts, const Names & column_names_to_return, @@ -944,14 +879,6 @@ ReadFromMergeTree::AnalysisResultPtr MergeTreeDataSelectExecutor::estimateNumMar if (total_parts == 0) return std::make_shared(); - Names real_column_names; - Names virt_column_names; - /// If query contains restrictions on the virtual column `_part` or `_part_index`, select only parts suitable for it. - /// The virtual column `_sample_factor` (which is equal to 1 / used sample rate) can be requested in the query. - bool sample_factor_column_queried = false; - - selectColumnNames(column_names_to_return, data, real_column_names, virt_column_names, sample_factor_column_queried); - std::optional indexes; /// NOTE: We don't need alter_conversions because the returned analysis_result is only used for: /// 1. estimate the number of rows to read; 2. projection reading, which doesn't have alter_conversions. @@ -964,8 +891,7 @@ ReadFromMergeTree::AnalysisResultPtr MergeTreeDataSelectExecutor::estimateNumMar num_streams, max_block_numbers_to_read, data, - real_column_names, - sample_factor_column_queried, + column_names_to_return, log, indexes); } @@ -992,27 +918,16 @@ QueryPlanStepPtr MergeTreeDataSelectExecutor::readFromParts( else if (parts.empty()) return {}; - Names real_column_names; - Names virt_column_names; - /// If query contains restrictions on the virtual column `_part` or `_part_index`, select only parts suitable for it. - /// The virtual column `_sample_factor` (which is equal to 1 / used sample rate) can be requested in the query. - bool sample_factor_column_queried = false; - - selectColumnNames(column_names_to_return, data, real_column_names, virt_column_names, sample_factor_column_queried); - return std::make_unique( std::move(parts), std::move(alter_conversions), column_names_to_return, - real_column_names, - virt_column_names, data, query_info, storage_snapshot, context, max_block_size, num_streams, - sample_factor_column_queried, max_block_numbers_to_read, log, merge_tree_select_result_ptr, diff --git a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.h b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.h index d61c97ab73b..b1afd7e6668 100644 --- a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.h +++ b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.h @@ -213,7 +213,6 @@ public: const MergeTreeData & data, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, - bool sample_factor_column_queried, LoggerPtr log); /// Check query limits: max_partitions_to_read, max_concurrent_queries. diff --git a/src/Storages/MergeTree/MergeTreeIndexHypothesis.h b/src/Storages/MergeTree/MergeTreeIndexHypothesis.h index 2296e1b717d..130e708d76f 100644 --- a/src/Storages/MergeTree/MergeTreeIndexHypothesis.h +++ b/src/Storages/MergeTree/MergeTreeIndexHypothesis.h @@ -2,7 +2,6 @@ #include #include -#include namespace DB { diff --git a/src/Storages/MergeTree/MergeTreeIndexMinMax.h b/src/Storages/MergeTree/MergeTreeIndexMinMax.h index 1e2abe6983f..dca26fb7b28 100644 --- a/src/Storages/MergeTree/MergeTreeIndexMinMax.h +++ b/src/Storages/MergeTree/MergeTreeIndexMinMax.h @@ -4,8 +4,6 @@ #include #include -#include - namespace DB { diff --git a/src/Storages/MergeTree/MergeTreeIndexSet.h b/src/Storages/MergeTree/MergeTreeIndexSet.h index ea9f7ddef3d..7c66ba1a867 100644 --- a/src/Storages/MergeTree/MergeTreeIndexSet.h +++ b/src/Storages/MergeTree/MergeTreeIndexSet.h @@ -5,9 +5,6 @@ #include -#include -#include - namespace DB { diff --git a/src/Storages/MergeTree/MergeTreeIndices.h b/src/Storages/MergeTree/MergeTreeIndices.h index 4749470bedd..8fdadb4e5eb 100644 --- a/src/Storages/MergeTree/MergeTreeIndices.h +++ b/src/Storages/MergeTree/MergeTreeIndices.h @@ -1,12 +1,9 @@ #pragma once #include -#include #include #include #include -#include -#include #include #include #include diff --git a/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp b/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp index 8d8b0f1cc79..c19b4ddd8a2 100644 --- a/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp +++ b/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp @@ -9,6 +9,7 @@ #include #include #include +#include #include #include #include @@ -108,22 +109,22 @@ MergeTreeReadTask::Readers MergeTreePrefetchedReadPool::PrefetchedReaders::get() MergeTreePrefetchedReadPool::MergeTreePrefetchedReadPool( RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_) : MergeTreeReadPoolBase( std::move(parts_), + std::move(shared_virtual_fields_), storage_snapshot_, prewhere_info_, actions_settings_, reader_settings_, column_names_, - virtual_column_names_, settings_, context_) , WithContext(context_) @@ -375,7 +376,7 @@ void MergeTreePrefetchedReadPool::fillPerPartStatistics() update_stat_for_column(column.name); if (reader_settings.apply_deleted_mask && read_info.data_part->hasLightweightDelete()) - update_stat_for_column(LightweightDeleteDescription::FILTER_COLUMN.name); + update_stat_for_column(RowExistsColumn::name); for (const auto & pre_columns : read_info.task_columns.pre_columns) for (const auto & column : pre_columns) diff --git a/src/Storages/MergeTree/MergeTreePrefetchedReadPool.h b/src/Storages/MergeTree/MergeTreePrefetchedReadPool.h index 378034c5eae..0c8a6716d40 100644 --- a/src/Storages/MergeTree/MergeTreePrefetchedReadPool.h +++ b/src/Storages/MergeTree/MergeTreePrefetchedReadPool.h @@ -18,12 +18,12 @@ class MergeTreePrefetchedReadPool : public MergeTreeReadPoolBase, private WithCo public: MergeTreePrefetchedReadPool( RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_); @@ -67,7 +67,7 @@ private: struct ThreadTask { - using InfoPtr = MergeTreeReadTask::InfoPtr; + using InfoPtr = MergeTreeReadTaskInfoPtr; ThreadTask(InfoPtr read_info_, MarkRanges ranges_, Priority priority_) : read_info(std::move(read_info_)), ranges(std::move(ranges_)), priority(priority_) diff --git a/src/Storages/MergeTree/MergeTreeRangeReader.cpp b/src/Storages/MergeTree/MergeTreeRangeReader.cpp index 50d1216cdc2..6932762f58b 100644 --- a/src/Storages/MergeTree/MergeTreeRangeReader.cpp +++ b/src/Storages/MergeTree/MergeTreeRangeReader.cpp @@ -362,7 +362,7 @@ void MergeTreeRangeReader::ReadResult::shrink(Columns & old_columns, const NumRo } } -/// The main invariant of the data in the read result is that he number of rows is +/// The main invariant of the data in the read result is that the number of rows is /// either equal to total_rows_per_granule (if filter has not been applied) or to the number of /// 1s in the filter (if filter has been applied). void MergeTreeRangeReader::ReadResult::checkInternalConsistency() const @@ -803,8 +803,7 @@ MergeTreeRangeReader::MergeTreeRangeReader( IMergeTreeReader * merge_tree_reader_, MergeTreeRangeReader * prev_reader_, const PrewhereExprStep * prewhere_info_, - bool last_reader_in_chain_, - const Names & non_const_virtual_column_names_) + bool last_reader_in_chain_) : merge_tree_reader(merge_tree_reader_) , index_granularity(&(merge_tree_reader->data_part_info_for_read->getIndexGranularity())) , prev_reader(prev_reader_) @@ -821,21 +820,6 @@ MergeTreeRangeReader::MergeTreeRangeReader( result_sample_block.insert({name_and_type.type->createColumn(), name_and_type.type, name_and_type.name}); } - for (const auto & column_name : non_const_virtual_column_names_) - { - if (result_sample_block.has(column_name)) - continue; - - non_const_virtual_column_names.push_back(column_name); - - if (column_name == "_part_offset" && !prev_reader) - { - /// _part_offset column is filled by the first reader. - read_sample_block.insert(ColumnWithTypeAndName(ColumnUInt64::create(), std::make_shared(), column_name)); - result_sample_block.insert(ColumnWithTypeAndName(ColumnUInt64::create(), std::make_shared(), column_name)); - } - } - if (prewhere_info) { const auto & step = *prewhere_info; @@ -1001,6 +985,8 @@ MergeTreeRangeReader::ReadResult MergeTreeRangeReader::read(size_t max_rows, Mar if (num_read_rows == 0) num_read_rows = read_result.num_rows; + merge_tree_reader->fillVirtualColumns(columns, num_read_rows); + /// fillMissingColumns() must be called after reading but befoe any filterings because /// some columns (e.g. arrays) might be only partially filled and thus not be valid and /// fillMissingColumns() fixes this. @@ -1050,23 +1036,23 @@ MergeTreeRangeReader::ReadResult MergeTreeRangeReader::read(size_t max_rows, Mar return read_result; { - /// Physical columns go first and then some virtual columns follow - size_t physical_columns_count = merge_tree_reader->getColumns().size(); - Columns physical_columns(read_result.columns.begin(), read_result.columns.begin() + physical_columns_count); + size_t columns_count = merge_tree_reader->getColumns().size(); + Columns columns(read_result.columns.begin(), read_result.columns.begin() + columns_count); + merge_tree_reader->fillVirtualColumns(columns, read_result.num_rows); bool should_evaluate_missing_defaults; - merge_tree_reader->fillMissingColumns(physical_columns, should_evaluate_missing_defaults, read_result.num_rows); + merge_tree_reader->fillMissingColumns(columns, should_evaluate_missing_defaults, read_result.num_rows); /// If some columns absent in part, then evaluate default values if (should_evaluate_missing_defaults) - merge_tree_reader->evaluateMissingDefaults({}, physical_columns); + merge_tree_reader->evaluateMissingDefaults({}, columns); /// If result not empty, then apply on-fly alter conversions if any required if (!prewhere_info || prewhere_info->perform_alter_conversions) - merge_tree_reader->performRequiredConversions(physical_columns); + merge_tree_reader->performRequiredConversions(columns); - for (size_t i = 0; i < physical_columns.size(); ++i) - read_result.columns[i] = std::move(physical_columns[i]); + for (size_t i = 0; i < columns.size(); ++i) + read_result.columns[i] = std::move(columns[i]); } size_t total_bytes = 0; @@ -1158,12 +1144,17 @@ MergeTreeRangeReader::ReadResult MergeTreeRangeReader::startReadingChain(size_t result.adjustLastGranule(); if (read_sample_block.has("_part_offset")) - fillPartOffsetColumn(result, leading_begin_part_offset, leading_end_part_offset); + { + size_t pos = read_sample_block.getPositionByName("_part_offset"); + chassert(pos < result.columns.size()); + chassert(result.columns[pos] == nullptr); + result.columns[pos] = createPartOffsetColumn(result, leading_begin_part_offset, leading_end_part_offset); + } return result; } -void MergeTreeRangeReader::fillPartOffsetColumn(ReadResult & result, UInt64 leading_begin_part_offset, UInt64 leading_end_part_offset) +ColumnPtr MergeTreeRangeReader::createPartOffsetColumn(ReadResult & result, UInt64 leading_begin_part_offset, UInt64 leading_end_part_offset) { size_t num_rows = result.numReadRows(); @@ -1189,7 +1180,7 @@ void MergeTreeRangeReader::fillPartOffsetColumn(ReadResult & result, UInt64 lead *pos++ = start_part_offset++; } - result.columns.emplace_back(std::move(column)); + return column; } Columns MergeTreeRangeReader::continueReadingChain(const ReadResult & result, size_t & num_rows) @@ -1203,7 +1194,7 @@ Columns MergeTreeRangeReader::continueReadingChain(const ReadResult & result, si if (result.rows_per_granule.empty()) { - /// If zero rows were read on prev step, than there is no more rows to read. + /// If zero rows were read on prev step, there is no more rows to read. /// Last granule may have less rows than index_granularity, so finish reading manually. stream.finish(); return columns; diff --git a/src/Storages/MergeTree/MergeTreeRangeReader.h b/src/Storages/MergeTree/MergeTreeRangeReader.h index 79ed18f4d1f..688a6b0922b 100644 --- a/src/Storages/MergeTree/MergeTreeRangeReader.h +++ b/src/Storages/MergeTree/MergeTreeRangeReader.h @@ -101,8 +101,7 @@ public: IMergeTreeReader * merge_tree_reader_, MergeTreeRangeReader * prev_reader_, const PrewhereExprStep * prewhere_info_, - bool last_reader_in_chain_, - const Names & non_const_virtual_column_names); + bool last_reader_in_chain_); MergeTreeRangeReader() = default; @@ -309,7 +308,7 @@ private: ReadResult startReadingChain(size_t max_rows, MarkRanges & ranges); Columns continueReadingChain(const ReadResult & result, size_t & num_rows); void executePrewhereActionsAndFilterColumns(ReadResult & result) const; - void fillPartOffsetColumn(ReadResult & result, UInt64 leading_begin_part_offset, UInt64 leading_end_part_offset); + ColumnPtr createPartOffsetColumn(ReadResult & result, UInt64 leading_begin_part_offset, UInt64 leading_end_part_offset); IMergeTreeReader * merge_tree_reader = nullptr; const MergeTreeIndexGranularity * index_granularity = nullptr; @@ -323,7 +322,6 @@ private: bool last_reader_in_chain = false; bool is_initialized = false; - Names non_const_virtual_column_names; LoggerPtr log = getLogger("MergeTreeRangeReader"); }; diff --git a/src/Storages/MergeTree/MergeTreeReadPool.cpp b/src/Storages/MergeTree/MergeTreeReadPool.cpp index 68d57bf7b06..e525f7f5f65 100644 --- a/src/Storages/MergeTree/MergeTreeReadPool.cpp +++ b/src/Storages/MergeTree/MergeTreeReadPool.cpp @@ -35,22 +35,22 @@ size_t getApproxSizeOfPart(const IMergeTreeDataPart & part, const Names & column MergeTreeReadPool::MergeTreeReadPool( RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_) : MergeTreeReadPoolBase( std::move(parts_), + std::move(shared_virtual_fields_), storage_snapshot_, prewhere_info_, actions_settings_, reader_settings_, column_names_, - virtual_column_names_, settings_, context_) , min_marks_for_concurrent_read(pool_settings.min_marks_for_concurrent_read) diff --git a/src/Storages/MergeTree/MergeTreeReadPool.h b/src/Storages/MergeTree/MergeTreeReadPool.h index e45ccad912f..cb0e8a9657f 100644 --- a/src/Storages/MergeTree/MergeTreeReadPool.h +++ b/src/Storages/MergeTree/MergeTreeReadPool.h @@ -26,12 +26,12 @@ public: MergeTreeReadPool( RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_); diff --git a/src/Storages/MergeTree/MergeTreeReadPoolBase.cpp b/src/Storages/MergeTree/MergeTreeReadPoolBase.cpp index 94942c4aa0b..0cbb0a86b2f 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolBase.cpp +++ b/src/Storages/MergeTree/MergeTreeReadPoolBase.cpp @@ -8,21 +8,21 @@ namespace DB MergeTreeReadPoolBase::MergeTreeReadPoolBase( RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & pool_settings_, const ContextPtr & context_) : parts_ranges(std::move(parts_)) + , shared_virtual_fields(std::move(shared_virtual_fields_)) , storage_snapshot(storage_snapshot_) , prewhere_info(prewhere_info_) , actions_settings(actions_settings_) , reader_settings(reader_settings_) , column_names(column_names_) - , virtual_column_names(virtual_column_names_) , pool_settings(pool_settings_) , owned_mark_cache(context_->getGlobalContext()->getMarkCache()) , owned_uncompressed_cache(pool_settings_.use_uncompressed_cache ? context_->getGlobalContext()->getUncompressedCache() : nullptr) @@ -45,7 +45,7 @@ void MergeTreeReadPoolBase::fillPerPartInfos() assertSortedAndNonIntersecting(part_with_ranges.ranges); #endif - MergeTreeReadTask::Info read_task_info; + MergeTreeReadTaskInfo read_task_info; read_task_info.data_part = part_with_ranges.data_part; read_task_info.part_index_in_query = part_with_ranges.part_index_in_query; @@ -54,9 +54,16 @@ void MergeTreeReadPoolBase::fillPerPartInfos() LoadedMergeTreeDataPartInfoForReader part_info(part_with_ranges.data_part, part_with_ranges.alter_conversions); read_task_info.task_columns = getReadTaskColumns( - part_info, storage_snapshot, column_names, virtual_column_names, - prewhere_info, actions_settings, - reader_settings, /*with_subcolumns=*/ true); + part_info, + storage_snapshot, + column_names, + prewhere_info, + actions_settings, + reader_settings, + /*with_subcolumns=*/true); + + read_task_info.const_virtual_fields = shared_virtual_fields; + read_task_info.const_virtual_fields.emplace("_part_index", read_task_info.part_index_in_query); if (pool_settings.preferred_block_size_bytes > 0) { @@ -76,7 +83,7 @@ void MergeTreeReadPoolBase::fillPerPartInfos() } is_part_on_remote_disk.push_back(part_with_ranges.data_part->isStoredOnRemoteDisk()); - per_part_infos.push_back(std::make_shared(std::move(read_task_info))); + per_part_infos.push_back(std::make_shared(std::move(read_task_info))); } } @@ -98,7 +105,7 @@ std::vector MergeTreeReadPoolBase::getPerPartSumMarks() const } MergeTreeReadTaskPtr MergeTreeReadPoolBase::createTask( - MergeTreeReadTask::InfoPtr read_info, + MergeTreeReadTaskInfoPtr read_info, MarkRanges ranges, MergeTreeReadTask * previous_task) const { diff --git a/src/Storages/MergeTree/MergeTreeReadPoolBase.h b/src/Storages/MergeTree/MergeTreeReadPoolBase.h index 0081063cd37..1b5bfec5898 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolBase.h +++ b/src/Storages/MergeTree/MergeTreeReadPoolBase.h @@ -23,12 +23,12 @@ public: MergeTreeReadPoolBase( RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_); @@ -37,12 +37,12 @@ public: protected: /// Initialized in constructor const RangesInDataParts parts_ranges; + const VirtualFields shared_virtual_fields; const StorageSnapshotPtr storage_snapshot; const PrewhereInfoPtr prewhere_info; const ExpressionActionsSettings actions_settings; const MergeTreeReaderSettings reader_settings; const Names column_names; - const Names virtual_column_names; const PoolSettings pool_settings; const MarkCachePtr owned_mark_cache; const UncompressedCachePtr owned_uncompressed_cache; @@ -52,13 +52,13 @@ protected: std::vector getPerPartSumMarks() const; MergeTreeReadTaskPtr createTask( - MergeTreeReadTask::InfoPtr read_info, + MergeTreeReadTaskInfoPtr read_info, MarkRanges ranges, MergeTreeReadTask * previous_task) const; MergeTreeReadTask::Extras getExtras() const; - std::vector per_part_infos; + std::vector per_part_infos; std::vector is_part_on_remote_disk; ReadBufferFromFileBase::ProfileCallback profile_callback; diff --git a/src/Storages/MergeTree/MergeTreeReadPoolInOrder.cpp b/src/Storages/MergeTree/MergeTreeReadPoolInOrder.cpp index 1b621ad5055..4c0391ffa57 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolInOrder.cpp +++ b/src/Storages/MergeTree/MergeTreeReadPoolInOrder.cpp @@ -12,22 +12,22 @@ MergeTreeReadPoolInOrder::MergeTreeReadPoolInOrder( bool has_limit_below_one_block_, MergeTreeReadType read_type_, RangesInDataParts parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_) : MergeTreeReadPoolBase( std::move(parts_), + std::move(shared_virtual_fields_), storage_snapshot_, prewhere_info_, actions_settings_, reader_settings_, column_names_, - virtual_column_names_, settings_, context_) , has_limit_below_one_block(has_limit_below_one_block_) diff --git a/src/Storages/MergeTree/MergeTreeReadPoolInOrder.h b/src/Storages/MergeTree/MergeTreeReadPoolInOrder.h index d9cc1ba4984..9fedf396a6b 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolInOrder.h +++ b/src/Storages/MergeTree/MergeTreeReadPoolInOrder.h @@ -11,12 +11,12 @@ public: bool has_limit_below_one_block_, MergeTreeReadType read_type_, RangesInDataParts parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_); diff --git a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.cpp b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.cpp index 47436ed1407..38035d97f56 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.cpp +++ b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.cpp @@ -13,22 +13,22 @@ namespace ErrorCodes MergeTreeReadPoolParallelReplicas::MergeTreeReadPoolParallelReplicas( ParallelReadingExtension extension_, RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_) : MergeTreeReadPoolBase( std::move(parts_), + std::move(shared_virtual_fields_), storage_snapshot_, prewhere_info_, actions_settings_, reader_settings_, column_names_, - virtual_column_names_, settings_, context_) , extension(std::move(extension_)) diff --git a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.h b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.h index 6a548dffe37..ca159edb91c 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.h +++ b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicas.h @@ -11,12 +11,12 @@ public: MergeTreeReadPoolParallelReplicas( ParallelReadingExtension extension_, RangesInDataParts && parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_); diff --git a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.cpp b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.cpp index a822a517933..01c0a9f91be 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.cpp +++ b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.cpp @@ -12,22 +12,22 @@ MergeTreeReadPoolParallelReplicasInOrder::MergeTreeReadPoolParallelReplicasInOrd ParallelReadingExtension extension_, CoordinationMode mode_, RangesInDataParts parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_) : MergeTreeReadPoolBase( std::move(parts_), + std::move(shared_virtual_fields_), storage_snapshot_, prewhere_info_, actions_settings_, reader_settings_, column_names_, - virtual_column_names_, settings_, context_) , extension(std::move(extension_)) diff --git a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.h b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.h index 3e5f8f5dfba..4fe3f7a699c 100644 --- a/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.h +++ b/src/Storages/MergeTree/MergeTreeReadPoolParallelReplicasInOrder.h @@ -12,12 +12,12 @@ public: ParallelReadingExtension extension_, CoordinationMode mode_, RangesInDataParts parts_, + VirtualFields shared_virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReaderSettings & reader_settings_, const Names & column_names_, - const Names & virtual_column_names_, const PoolSettings & settings_, const ContextPtr & context_); diff --git a/src/Storages/MergeTree/MergeTreeReadTask.cpp b/src/Storages/MergeTree/MergeTreeReadTask.cpp index 41c7531b6a6..08b30e445e2 100644 --- a/src/Storages/MergeTree/MergeTreeReadTask.cpp +++ b/src/Storages/MergeTree/MergeTreeReadTask.cpp @@ -1,5 +1,6 @@ #include #include +#include #include namespace DB @@ -10,7 +11,7 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } -String MergeTreeReadTask::Columns::dump() const +String MergeTreeReadTaskColumns::dump() const { WriteBufferFromOwnString s; for (size_t i = 0; i < pre_columns.size(); ++i) @@ -22,7 +23,7 @@ String MergeTreeReadTask::Columns::dump() const } MergeTreeReadTask::MergeTreeReadTask( - InfoPtr info_, + MergeTreeReadTaskInfoPtr info_, Readers readers_, MarkRanges mark_ranges_, MergeTreeBlockSizePredictorPtr size_predictor_) @@ -34,23 +35,30 @@ MergeTreeReadTask::MergeTreeReadTask( } MergeTreeReadTask::Readers MergeTreeReadTask::createReaders( - const InfoPtr & read_info, const Extras & extras, const MarkRanges & ranges) + const MergeTreeReadTaskInfoPtr & read_info, const Extras & extras, const MarkRanges & ranges) { Readers new_readers; auto create_reader = [&](const NamesAndTypesList & columns_to_read) { return read_info->data_part->getReader( - columns_to_read, extras.storage_snapshot, ranges, - extras.uncompressed_cache, extras.mark_cache, - read_info->alter_conversions, extras.reader_settings, extras.value_size_map, extras.profile_callback); + columns_to_read, + extras.storage_snapshot, + ranges, + read_info->const_virtual_fields, + extras.uncompressed_cache, + extras.mark_cache, + read_info->alter_conversions, + extras.reader_settings, + extras.value_size_map, + extras.profile_callback); }; new_readers.main = create_reader(read_info->task_columns.columns); /// Add lightweight delete filtering step if (extras.reader_settings.apply_deleted_mask && read_info->data_part->hasLightweightDelete()) - new_readers.prewhere.push_back(create_reader({LightweightDeleteDescription::FILTER_COLUMN})); + new_readers.prewhere.push_back(create_reader({{RowExistsColumn::name, RowExistsColumn::type}})); for (const auto & pre_columns_per_step : read_info->task_columns.pre_columns) new_readers.prewhere.push_back(create_reader(pre_columns_per_step)); @@ -58,10 +66,8 @@ MergeTreeReadTask::Readers MergeTreeReadTask::createReaders( return new_readers; } -MergeTreeReadTask::RangeReaders MergeTreeReadTask::createRangeReaders( - const Readers & task_readers, - const PrewhereExprInfo & prewhere_actions, - const Names & non_const_virtual_column_names) +MergeTreeReadTask::RangeReaders +MergeTreeReadTask::createRangeReaders(const Readers & task_readers, const PrewhereExprInfo & prewhere_actions) { MergeTreeReadTask::RangeReaders new_range_readers; if (prewhere_actions.steps.size() != task_readers.prewhere.size()) @@ -77,10 +83,7 @@ MergeTreeReadTask::RangeReaders MergeTreeReadTask::createRangeReaders( { last_reader = task_readers.main->getColumns().empty() && (i + 1 == prewhere_actions.steps.size()); - MergeTreeRangeReader current_reader( - task_readers.prewhere[i].get(), - prev_reader, prewhere_actions.steps[i].get(), - last_reader, non_const_virtual_column_names); + MergeTreeRangeReader current_reader(task_readers.prewhere[i].get(), prev_reader, prewhere_actions.steps[i].get(), last_reader); new_range_readers.prewhere.push_back(std::move(current_reader)); prev_reader = &new_range_readers.prewhere.back(); @@ -88,11 +91,11 @@ MergeTreeReadTask::RangeReaders MergeTreeReadTask::createRangeReaders( if (!last_reader) { - new_range_readers.main = MergeTreeRangeReader(task_readers.main.get(), prev_reader, nullptr, true, non_const_virtual_column_names); + new_range_readers.main = MergeTreeRangeReader(task_readers.main.get(), prev_reader, nullptr, true); } else { - /// If all columns are read by prewhere range readers than move last prewhere range reader to main. + /// If all columns are read by prewhere range readers, move last prewhere range reader to main. new_range_readers.main = std::move(new_range_readers.prewhere.back()); new_range_readers.prewhere.pop_back(); } @@ -100,14 +103,12 @@ MergeTreeReadTask::RangeReaders MergeTreeReadTask::createRangeReaders( return new_range_readers; } -void MergeTreeReadTask::initializeRangeReaders( - const PrewhereExprInfo & prewhere_actions, - const Names & non_const_virtual_column_names) +void MergeTreeReadTask::initializeRangeReaders(const PrewhereExprInfo & prewhere_actions) { if (range_readers.main.isInitialized()) throw Exception(ErrorCodes::LOGICAL_ERROR, "Range reader is already initialized"); - range_readers = createRangeReaders(readers, prewhere_actions, non_const_virtual_column_names); + range_readers = createRangeReaders(readers, prewhere_actions); } UInt64 MergeTreeReadTask::estimateNumRows(const BlockSizeParams & params) const diff --git a/src/Storages/MergeTree/MergeTreeReadTask.h b/src/Storages/MergeTree/MergeTreeReadTask.h index 3fe79f1087c..c8bb501c0e8 100644 --- a/src/Storages/MergeTree/MergeTreeReadTask.h +++ b/src/Storages/MergeTree/MergeTreeReadTask.h @@ -20,6 +20,8 @@ using MergeTreeBlockSizePredictorPtr = std::shared_ptr; using MergeTreeReaderPtr = std::unique_ptr; +using VirtualFields = std::unordered_map; + enum class MergeTreeReadType { @@ -40,36 +42,38 @@ enum class MergeTreeReadType ParallelReplicas, }; +struct MergeTreeReadTaskColumns +{ + /// Column names to read during WHERE + NamesAndTypesList columns; + /// Column names to read during each PREWHERE step + std::vector pre_columns; + + String dump() const; +}; + +struct MergeTreeReadTaskInfo +{ + /// Data part which should be read while performing this task + DataPartPtr data_part; + /// For `part_index` virtual column + size_t part_index_in_query; + /// Alter converversionss that should be applied on-fly for part. + AlterConversionsPtr alter_conversions; + /// Column names to read during PREWHERE and WHERE + MergeTreeReadTaskColumns task_columns; + /// Shared initialized size predictor. It is copied for each new task. + MergeTreeBlockSizePredictorPtr shared_size_predictor; + /// TODO: comment + VirtualFields const_virtual_fields; +}; + +using MergeTreeReadTaskInfoPtr = std::shared_ptr; + /// A batch of work for MergeTreeSelectProcessor struct MergeTreeReadTask : private boost::noncopyable { public: - struct Columns - { - /// Column names to read during WHERE - NamesAndTypesList columns; - /// Column names to read during each PREWHERE step - std::vector pre_columns; - - String dump() const; - }; - - struct Info - { - /// Data part which should be read while performing this task - DataPartPtr data_part; - /// For virtual `part_index` virtual column - size_t part_index_in_query; - /// Alter converversionss that should be applied on-fly for part. - AlterConversionsPtr alter_conversions; - /// Column names to read during PREWHERE and WHERE - Columns task_columns; - /// Shared initialized size predictor. It is copied for each new task. - MergeTreeBlockSizePredictorPtr shared_size_predictor; - }; - - using InfoPtr = std::shared_ptr; - /// Extra params that required for creation of reader. struct Extras { @@ -115,27 +119,32 @@ public: size_t num_read_bytes = 0; }; - MergeTreeReadTask(InfoPtr info_, Readers readers_, MarkRanges mark_ranges_, MergeTreeBlockSizePredictorPtr size_predictor_); + MergeTreeReadTask( + MergeTreeReadTaskInfoPtr info_, + Readers readers_, + MarkRanges mark_ranges_, - void initializeRangeReaders(const PrewhereExprInfo & prewhere_actions, const Names & non_const_virtual_column_names); + MergeTreeBlockSizePredictorPtr size_predictor_); + + void initializeRangeReaders(const PrewhereExprInfo & prewhere_actions); BlockAndProgress read(const BlockSizeParams & params); bool isFinished() const { return mark_ranges.empty() && range_readers.main.isCurrentRangeFinished(); } - const Info & getInfo() const { return *info; } + const MergeTreeReadTaskInfo & getInfo() const { return *info; } const MergeTreeRangeReader & getMainRangeReader() const { return range_readers.main; } const IMergeTreeReader & getMainReader() const { return *readers.main; } Readers releaseReaders() { return std::move(readers); } - static Readers createReaders(const InfoPtr & read_info, const Extras & extras, const MarkRanges & ranges); - static RangeReaders createRangeReaders(const Readers & readers, const PrewhereExprInfo & prewhere_actions, const Names & non_const_virtual_column_names); + static Readers createReaders(const MergeTreeReadTaskInfoPtr & read_info, const Extras & extras, const MarkRanges & ranges); + static RangeReaders createRangeReaders(const Readers & readers, const PrewhereExprInfo & prewhere_actions); private: UInt64 estimateNumRows(const BlockSizeParams & params) const; /// Shared information required for reading. - InfoPtr info; + MergeTreeReadTaskInfoPtr info; /// Readers for data_part of this task. /// May be reused and released to the next task. diff --git a/src/Storages/MergeTree/MergeTreeReaderCompact.cpp b/src/Storages/MergeTree/MergeTreeReaderCompact.cpp index 02048009296..63824366722 100644 --- a/src/Storages/MergeTree/MergeTreeReaderCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeReaderCompact.cpp @@ -17,6 +17,7 @@ namespace ErrorCodes MergeTreeReaderCompact::MergeTreeReaderCompact( MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, NamesAndTypesList columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, UncompressedCache * uncompressed_cache_, MarkCache * mark_cache_, @@ -29,6 +30,7 @@ MergeTreeReaderCompact::MergeTreeReaderCompact( : IMergeTreeReader( data_part_info_for_read_, columns_, + virtual_fields_, storage_snapshot_, uncompressed_cache_, mark_cache_, diff --git a/src/Storages/MergeTree/MergeTreeReaderCompact.h b/src/Storages/MergeTree/MergeTreeReaderCompact.h index dace4ec468e..769e6a08be4 100644 --- a/src/Storages/MergeTree/MergeTreeReaderCompact.h +++ b/src/Storages/MergeTree/MergeTreeReaderCompact.h @@ -21,6 +21,7 @@ public: MergeTreeReaderCompact( MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, NamesAndTypesList columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, UncompressedCache * uncompressed_cache_, MarkCache * mark_cache_, diff --git a/src/Storages/MergeTree/MergeTreeReaderInMemory.cpp b/src/Storages/MergeTree/MergeTreeReaderInMemory.cpp index bacd86511f5..91fc8966a7a 100644 --- a/src/Storages/MergeTree/MergeTreeReaderInMemory.cpp +++ b/src/Storages/MergeTree/MergeTreeReaderInMemory.cpp @@ -19,12 +19,14 @@ MergeTreeReaderInMemory::MergeTreeReaderInMemory( MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, DataPartInMemoryPtr data_part_, NamesAndTypesList columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, MarkRanges mark_ranges_, MergeTreeReaderSettings settings_) : IMergeTreeReader( data_part_info_for_read_, columns_, + virtual_fields_, storage_snapshot_, nullptr, nullptr, diff --git a/src/Storages/MergeTree/MergeTreeReaderInMemory.h b/src/Storages/MergeTree/MergeTreeReaderInMemory.h index e26a98f0916..cc1e2e9e4e2 100644 --- a/src/Storages/MergeTree/MergeTreeReaderInMemory.h +++ b/src/Storages/MergeTree/MergeTreeReaderInMemory.h @@ -18,6 +18,7 @@ public: MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, DataPartInMemoryPtr data_part_, NamesAndTypesList columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, MarkRanges mark_ranges_, MergeTreeReaderSettings settings_); diff --git a/src/Storages/MergeTree/MergeTreeReaderStream.h b/src/Storages/MergeTree/MergeTreeReaderStream.h index baf8ec713f9..49ce3103434 100644 --- a/src/Storages/MergeTree/MergeTreeReaderStream.h +++ b/src/Storages/MergeTree/MergeTreeReaderStream.h @@ -1,5 +1,4 @@ #pragma once -#include #include #include #include diff --git a/src/Storages/MergeTree/MergeTreeReaderWide.cpp b/src/Storages/MergeTree/MergeTreeReaderWide.cpp index 640432ef755..d34a58a25b0 100644 --- a/src/Storages/MergeTree/MergeTreeReaderWide.cpp +++ b/src/Storages/MergeTree/MergeTreeReaderWide.cpp @@ -24,6 +24,7 @@ namespace MergeTreeReaderWide::MergeTreeReaderWide( MergeTreeDataPartInfoForReaderPtr data_part_info_, NamesAndTypesList columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, UncompressedCache * uncompressed_cache_, MarkCache * mark_cache_, @@ -35,6 +36,7 @@ MergeTreeReaderWide::MergeTreeReaderWide( : IMergeTreeReader( data_part_info_, columns_, + virtual_fields_, storage_snapshot_, uncompressed_cache_, mark_cache_, diff --git a/src/Storages/MergeTree/MergeTreeReaderWide.h b/src/Storages/MergeTree/MergeTreeReaderWide.h index 2a850cc2814..a9a5526dd65 100644 --- a/src/Storages/MergeTree/MergeTreeReaderWide.h +++ b/src/Storages/MergeTree/MergeTreeReaderWide.h @@ -17,6 +17,7 @@ public: MergeTreeReaderWide( MergeTreeDataPartInfoForReaderPtr data_part_info_for_read_, NamesAndTypesList columns_, + const VirtualFields & virtual_fields_, const StorageSnapshotPtr & storage_snapshot_, UncompressedCache * uncompressed_cache_, MarkCache * mark_cache_, diff --git a/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp b/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp index 64cec946991..fce733d47b7 100644 --- a/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp +++ b/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp @@ -12,7 +12,7 @@ #include #include #include -#include +#include #include namespace DB @@ -20,41 +20,26 @@ namespace DB namespace ErrorCodes { - extern const int LOGICAL_ERROR; extern const int QUERY_WAS_CANCELLED; } -static void injectNonConstVirtualColumns( - size_t rows, - Block & block, - const Names & virtual_columns, - MergeTreeReadTask * task = nullptr); - -static void injectPartConstVirtualColumns( - size_t rows, - Block & block, - MergeTreeReadTask * task, - const DataTypePtr & partition_value_type, - const Names & virtual_columns); - MergeTreeSelectProcessor::MergeTreeSelectProcessor( MergeTreeReadPoolPtr pool_, MergeTreeSelectAlgorithmPtr algorithm_, - const MergeTreeData & storage_, + const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReadTask::BlockSizeParams & block_size_params_, - const MergeTreeReaderSettings & reader_settings_, - const Names & virt_column_names_) + const MergeTreeReaderSettings & reader_settings_) : pool(std::move(pool_)) , algorithm(std::move(algorithm_)) + , storage_snapshot(storage_snapshot_) , prewhere_info(prewhere_info_) , actions_settings(actions_settings_) , prewhere_actions(getPrewhereActions(prewhere_info, actions_settings, reader_settings_.enable_multiple_prewhere_read_steps)) , reader_settings(reader_settings_) , block_size_params(block_size_params_) - , virt_column_names(virt_column_names_) - , partition_value_type(storage_.getPartitionValueType()) + , result_header(transformHeader(pool->getHeader(), prewhere_info)) { if (reader_settings.apply_deleted_mask) { @@ -62,7 +47,7 @@ MergeTreeSelectProcessor::MergeTreeSelectProcessor( { .type = PrewhereExprStep::Filter, .actions = nullptr, - .filter_column_name = LightweightDeleteDescription::FILTER_COLUMN.name, + .filter_column_name = RowExistsColumn::name, .remove_filter_column = true, .need_filter = true, .perform_alter_conversions = true, @@ -71,16 +56,6 @@ MergeTreeSelectProcessor::MergeTreeSelectProcessor( lightweight_delete_filter_step = std::make_shared(std::move(step)); } - header_without_const_virtual_columns = SourceStepWithFilter::applyPrewhereActions(pool->getHeader(), prewhere_info); - size_t non_const_columns_offset = header_without_const_virtual_columns.columns(); - injectNonConstVirtualColumns(0, header_without_const_virtual_columns, virt_column_names); - - for (size_t col_num = non_const_columns_offset; col_num < header_without_const_virtual_columns.columns(); ++col_num) - non_const_virtual_column_names.emplace_back(header_without_const_virtual_columns.getByPosition(col_num).name); - - result_header = header_without_const_virtual_columns; - injectPartConstVirtualColumns(0, result_header, nullptr, partition_value_type, virt_column_names); - if (!prewhere_actions.steps.empty()) LOG_TRACE(log, "PREWHERE condition was split into {} steps: {}", prewhere_actions.steps.size(), prewhere_actions.dumpConditions()); @@ -163,8 +138,6 @@ ChunkAndProgress MergeTreeSelectProcessor::read() if (res.row_count) { - injectVirtualColumns(res.block, res.row_count, task.get(), partition_value_type, virt_column_names); - /// Reorder the columns according to result_header Columns ordered_columns; ordered_columns.reserve(result_header.columns()); @@ -198,209 +171,12 @@ void MergeTreeSelectProcessor::initializeRangeReaders() for (const auto & step : prewhere_actions.steps) all_prewhere_actions.steps.push_back(step); - task->initializeRangeReaders(all_prewhere_actions, non_const_virtual_column_names); + task->initializeRangeReaders(all_prewhere_actions); } - -namespace +Block MergeTreeSelectProcessor::transformHeader(Block block, const PrewhereInfoPtr & prewhere_info) { - struct VirtualColumnsInserter - { - explicit VirtualColumnsInserter(Block & block_) : block(block_) {} - - bool columnExists(const String & name) const { return block.has(name); } - - void insertUInt8Column(const ColumnPtr & column, const String & name) - { - block.insert({column, std::make_shared(), name}); - } - - void insertUInt64Column(const ColumnPtr & column, const String & name) - { - block.insert({column, std::make_shared(), name}); - } - - void insertUUIDColumn(const ColumnPtr & column, const String & name) - { - block.insert({column, std::make_shared(), name}); - } - - void insertLowCardinalityColumn(const ColumnPtr & column, const String & name) - { - block.insert({column, std::make_shared(std::make_shared()), name}); - } - - void insertPartitionValueColumn( - size_t rows, const Row & partition_value, const DataTypePtr & partition_value_type, const String & name) - { - ColumnPtr column; - if (rows) - column = partition_value_type->createColumnConst(rows, Tuple(partition_value.begin(), partition_value.end())) - ->convertToFullColumnIfConst(); - else - column = partition_value_type->createColumn(); - - block.insert({column, partition_value_type, name}); - } - - Block & block; - }; -} - -/// Adds virtual columns that are not const for all rows -static void injectNonConstVirtualColumns( - size_t rows, - Block & block, - const Names & virtual_columns, - MergeTreeReadTask * task) -{ - VirtualColumnsInserter inserter(block); - for (const auto & virtual_column_name : virtual_columns) - { - if (virtual_column_name == "_part_offset") - { - if (!rows) - { - inserter.insertUInt64Column(DataTypeUInt64().createColumn(), virtual_column_name); - } - else - { - if (!inserter.columnExists(virtual_column_name)) - throw Exception(ErrorCodes::LOGICAL_ERROR, - "Column {} must have been filled part reader", - virtual_column_name); - } - } - - if (virtual_column_name == LightweightDeleteDescription::FILTER_COLUMN.name) - { - /// If _row_exists column isn't present in the part then fill it here with 1s - ColumnPtr column; - if (rows) - column = LightweightDeleteDescription::FILTER_COLUMN.type->createColumnConst(rows, 1)->convertToFullColumnIfConst(); - else - column = LightweightDeleteDescription::FILTER_COLUMN.type->createColumn(); - - inserter.insertUInt8Column(column, virtual_column_name); - } - - if (virtual_column_name == BlockNumberColumn::name) - { - ColumnPtr column; - if (rows) - { - size_t value = 0; - if (task) - { - value = task->getInfo().data_part ? task->getInfo().data_part->info.min_block : 0; - } - column = BlockNumberColumn::type->createColumnConst(rows, value)->convertToFullColumnIfConst(); - } - else - column = BlockNumberColumn::type->createColumn(); - - inserter.insertUInt64Column(column, virtual_column_name); - } - } -} - -/// Adds virtual columns that are const for the whole part -static void injectPartConstVirtualColumns( - size_t rows, - Block & block, - MergeTreeReadTask * task, - const DataTypePtr & partition_value_type, - const Names & virtual_columns) -{ - VirtualColumnsInserter inserter(block); - /// add virtual columns - /// Except _sample_factor, which is added from the outside. - if (!virtual_columns.empty()) - { - if (unlikely(rows && !task)) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot insert virtual columns to non-empty chunk without specified task."); - - const IMergeTreeDataPart * part = nullptr; - - if (rows) - { - part = task->getInfo().data_part.get(); - if (part->isProjectionPart()) - part = part->getParentPart(); - } - - for (const auto & virtual_column_name : virtual_columns) - { - if (virtual_column_name == "_part") - { - ColumnPtr column; - if (rows) - column = DataTypeLowCardinality{std::make_shared()} - .createColumnConst(rows, part->name) - ->convertToFullColumnIfConst(); - else - column = DataTypeLowCardinality{std::make_shared()}.createColumn(); - - inserter.insertLowCardinalityColumn(column, virtual_column_name); - } - else if (virtual_column_name == "_part_index") - { - ColumnPtr column; - if (rows) - column = DataTypeUInt64().createColumnConst(rows, task->getInfo().part_index_in_query)->convertToFullColumnIfConst(); - else - column = DataTypeUInt64().createColumn(); - - inserter.insertUInt64Column(column, virtual_column_name); - } - else if (virtual_column_name == "_part_uuid") - { - ColumnPtr column; - if (rows) - column = DataTypeUUID().createColumnConst(rows, part->uuid)->convertToFullColumnIfConst(); - else - column = DataTypeUUID().createColumn(); - - inserter.insertUUIDColumn(column, virtual_column_name); - } - else if (virtual_column_name == "_partition_id") - { - ColumnPtr column; - if (rows) - column = DataTypeLowCardinality{std::make_shared()} - .createColumnConst(rows, part->info.partition_id) - ->convertToFullColumnIfConst(); - else - column = DataTypeLowCardinality{std::make_shared()}.createColumn(); - - inserter.insertLowCardinalityColumn(column, virtual_column_name); - } - else if (virtual_column_name == "_partition_value") - { - if (rows) - inserter.insertPartitionValueColumn(rows, part->partition.value, partition_value_type, virtual_column_name); - else - inserter.insertPartitionValueColumn(rows, {}, partition_value_type, virtual_column_name); - } - } - } -} - -void MergeTreeSelectProcessor::injectVirtualColumns( - Block & block, size_t row_count, MergeTreeReadTask * task, const DataTypePtr & partition_value_type, const Names & virtual_columns) -{ - /// First add non-const columns that are filled by the range reader and then const columns that we will fill ourselves. - /// Note that the order is important: virtual columns filled by the range reader must go first - injectNonConstVirtualColumns(row_count, block, virtual_columns,task); - injectPartConstVirtualColumns(row_count, block, task, partition_value_type, virtual_columns); -} - -Block MergeTreeSelectProcessor::transformHeader( - Block block, const PrewhereInfoPtr & prewhere_info, const DataTypePtr & partition_value_type, const Names & virtual_columns) -{ - injectVirtualColumns(block, 0, nullptr, partition_value_type, virtual_columns); - auto transformed = SourceStepWithFilter::applyPrewhereActions(std::move(block), prewhere_info); - return transformed; + return SourceStepWithFilter::applyPrewhereActions(std::move(block), prewhere_info); } } diff --git a/src/Storages/MergeTree/MergeTreeSelectProcessor.h b/src/Storages/MergeTree/MergeTreeSelectProcessor.h index b1606f983a1..01bb3851e04 100644 --- a/src/Storages/MergeTree/MergeTreeSelectProcessor.h +++ b/src/Storages/MergeTree/MergeTreeSelectProcessor.h @@ -41,21 +41,15 @@ public: MergeTreeSelectProcessor( MergeTreeReadPoolPtr pool_, MergeTreeSelectAlgorithmPtr algorithm_, - const MergeTreeData & storage_, + const StorageSnapshotPtr & storage_snapshot_, const PrewhereInfoPtr & prewhere_info_, const ExpressionActionsSettings & actions_settings_, const MergeTreeReadTask::BlockSizeParams & block_size_params_, - const MergeTreeReaderSettings & reader_settings_, - const Names & virt_column_names_); + const MergeTreeReaderSettings & reader_settings_); String getName() const; - static Block transformHeader( - Block block, - const PrewhereInfoPtr & prewhere_info, - const DataTypePtr & partition_value_type, - const Names & virtual_columns); - + static Block transformHeader(Block block, const PrewhereInfoPtr & prewhere_info); Block getHeader() const { return result_header; } ChunkAndProgress read(); @@ -81,14 +75,12 @@ private: size_t num_read_bytes = 0; }; - /// Used for filling header with no rows as well as block with data - static void injectVirtualColumns(Block & block, size_t row_count, MergeTreeReadTask * task, const DataTypePtr & partition_value_type, const Names & virtual_columns); - /// Sets up range readers corresponding to data readers void initializeRangeReaders(); const MergeTreeReadPoolPtr pool; const MergeTreeSelectAlgorithmPtr algorithm; + const StorageSnapshotPtr storage_snapshot; const PrewhereInfoPtr prewhere_info; const ExpressionActionsSettings actions_settings; @@ -96,17 +88,11 @@ private: const MergeTreeReaderSettings reader_settings; const MergeTreeReadTask::BlockSizeParams block_size_params; - const Names virt_column_names; - const DataTypePtr partition_value_type; /// Current task to read from. MergeTreeReadTaskPtr task; /// This step is added when the part has lightweight delete mask PrewhereExprStepPtr lightweight_delete_filter_step; - /// These columns will be filled by the merge tree range reader - Names non_const_virtual_column_names; - /// This header is used for chunks from readFromPart(). - Block header_without_const_virtual_columns; /// A result of getHeader(). A chunk which this header is returned from read(). Block result_header; diff --git a/src/Storages/MergeTree/MergeTreeSequentialSource.cpp b/src/Storages/MergeTree/MergeTreeSequentialSource.cpp index d0fbc316024..e5545a92aea 100644 --- a/src/Storages/MergeTree/MergeTreeSequentialSource.cpp +++ b/src/Storages/MergeTree/MergeTreeSequentialSource.cpp @@ -2,6 +2,7 @@ #include #include #include +#include #include #include #include @@ -16,6 +17,7 @@ namespace DB { + namespace ErrorCodes { extern const int MEMORY_LIMIT_EXCEEDED; @@ -55,7 +57,6 @@ protected: Chunk generate() override; private: - const MergeTreeData & storage; StorageSnapshotPtr storage_snapshot; @@ -86,7 +87,6 @@ private: void finish(); }; - MergeTreeSequentialSource::MergeTreeSequentialSource( MergeTreeSequentialSourceType type, const MergeTreeData & storage_, @@ -136,10 +136,8 @@ MergeTreeSequentialSource::MergeTreeSequentialSource( { auto options = GetColumnsOptions(GetColumnsOptions::AllPhysical) .withExtendedObjects() - .withSystemColumns(); - - if (storage.supportsSubcolumns()) - options.withSubcolumns(); + .withVirtuals() + .withSubcolumns(storage.supportsSubcolumns()); columns_for_reader = storage_snapshot->getColumnsByNames(options, columns_to_read); } @@ -181,9 +179,37 @@ MergeTreeSequentialSource::MergeTreeSequentialSource( mark_ranges.emplace(MarkRanges{MarkRange(0, data_part->getMarksCount())}); reader = data_part->getReader( - columns_for_reader, storage_snapshot, - *mark_ranges, /* uncompressed_cache = */ nullptr, - mark_cache.get(), alter_conversions, reader_settings, {}, {}); + columns_for_reader, + storage_snapshot, + *mark_ranges, + /*virtual_fields=*/ {}, + /*uncompressed_cache=*/{}, + mark_cache.get(), + alter_conversions, + reader_settings, + {}, + {}); +} + +static void fillBlockNumberColumns( + Columns & res_columns, + const NamesAndTypesList & columns_list, + UInt64 block_number, + UInt64 num_rows) +{ + chassert(res_columns.size() == columns_list.size()); + + auto it = columns_list.begin(); + for (size_t i = 0; i < res_columns.size(); ++i, ++it) + { + if (res_columns[i]) + continue; + + if (it->name == BlockNumberColumn::name) + { + res_columns[i] = BlockNumberColumn::type->createColumnConst(num_rows, block_number)->convertToFullColumnIfConst(); + } + } } Chunk MergeTreeSequentialSource::generate() @@ -204,16 +230,17 @@ try if (rows_read) { + fillBlockNumberColumns(columns, sample, data_part->info.min_block, rows_read); + reader->fillVirtualColumns(columns, rows_read); + current_row += rows_read; current_mark += (rows_to_read == rows_read); bool should_evaluate_missing_defaults = false; - reader->fillMissingColumns(columns, should_evaluate_missing_defaults, rows_read, data_part->info.min_block); + reader->fillMissingColumns(columns, should_evaluate_missing_defaults, rows_read); if (should_evaluate_missing_defaults) - { reader->evaluateMissingDefaults({}, columns); - } reader->performRequiredConversions(columns); @@ -278,14 +305,13 @@ Pipe createMergeTreeSequentialSource( bool quiet, std::shared_ptr> filtered_rows_count) { - const auto & filter_column = LightweightDeleteDescription::FILTER_COLUMN; /// The part might have some rows masked by lightweight deletes const bool need_to_filter_deleted_rows = apply_deleted_mask && data_part->hasLightweightDelete(); - const bool has_filter_column = std::ranges::find(columns_to_read, filter_column.name) != columns_to_read.end(); + const bool has_filter_column = std::ranges::find(columns_to_read, RowExistsColumn::name) != columns_to_read.end(); if (need_to_filter_deleted_rows && !has_filter_column) - columns_to_read.emplace_back(filter_column.name); + columns_to_read.emplace_back(RowExistsColumn::name); auto column_part_source = std::make_shared(type, storage, storage_snapshot, data_part, columns_to_read, std::move(mark_ranges), @@ -299,7 +325,7 @@ Pipe createMergeTreeSequentialSource( pipe.addSimpleTransform([filtered_rows_count, has_filter_column](const Block & header) { return std::make_shared( - header, nullptr, filter_column.name, !has_filter_column, false, filtered_rows_count); + header, nullptr, RowExistsColumn::name, !has_filter_column, false, filtered_rows_count); }); } diff --git a/src/Storages/MergeTree/MergeTreeVirtualColumns.cpp b/src/Storages/MergeTree/MergeTreeVirtualColumns.cpp new file mode 100644 index 00000000000..b87dccc2b18 --- /dev/null +++ b/src/Storages/MergeTree/MergeTreeVirtualColumns.cpp @@ -0,0 +1,52 @@ +#include +#include +#include +#include +#include + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int NO_SUCH_COLUMN_IN_TABLE; +} + +static ASTPtr getCompressionCodecDeltaLZ4() +{ + return makeASTFunction("CODEC", + std::make_shared("Delta"), + std::make_shared("LZ4")); +} + +const String RowExistsColumn::name = "_row_exists"; +const DataTypePtr RowExistsColumn::type = std::make_shared(); + +const String BlockNumberColumn::name = "_block_number"; +const DataTypePtr BlockNumberColumn::type = std::make_shared(); +const ASTPtr BlockNumberColumn::codec = getCompressionCodecDeltaLZ4(); + +Field getFieldForConstVirtualColumn(const String & column_name, const IMergeTreeDataPart & part) +{ + if (column_name == RowExistsColumn::name) + return 1ULL; + + if (column_name == BlockNumberColumn::name) + return part.info.min_block; + + if (column_name == "_part") + return part.name; + + if (column_name == "_part_uuid") + return part.uuid; + + if (column_name == "_partition_id") + return part.info.partition_id; + + if (column_name == "_partition_value") + return Tuple(part.partition.value.begin(), part.partition.value.end()); + + throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "Unexpected const virtual column: {}", column_name); +} + +} diff --git a/src/Storages/MergeTree/MergeTreeVirtualColumns.h b/src/Storages/MergeTree/MergeTreeVirtualColumns.h new file mode 100644 index 00000000000..24721bf1ad1 --- /dev/null +++ b/src/Storages/MergeTree/MergeTreeVirtualColumns.h @@ -0,0 +1,26 @@ +#pragma once +#include +#include +#include + +namespace DB +{ + +class IMergeTreeDataPart; + +struct RowExistsColumn +{ + static const String name; + static const DataTypePtr type; +}; + +struct BlockNumberColumn +{ + static const String name; + static const DataTypePtr type; + static const ASTPtr codec; +}; + +Field getFieldForConstVirtualColumn(const String & column_name, const IMergeTreeDataPart & part); + +} diff --git a/src/Storages/MergeTree/MutateTask.cpp b/src/Storages/MergeTree/MutateTask.cpp index 1c33f018a5d..a5b8a2a2a6d 100644 --- a/src/Storages/MergeTree/MutateTask.cpp +++ b/src/Storages/MergeTree/MutateTask.cpp @@ -23,7 +23,7 @@ #include #include #include -#include +#include #include #include #include @@ -168,7 +168,7 @@ static void splitAndModifyMutationCommands( { if (!mutated_columns.contains(column.name)) { - if (!metadata_snapshot->getColumns().has(column.name) && !part->storage.getVirtuals().contains(column.name)) + if (!metadata_snapshot->getColumns().has(column.name) && !part->storage.getVirtualsPtr()->has(column.name)) { /// We cannot add the column because there's no such column in table. /// It's okay if the column was dropped. It may also absent in dropped_columns @@ -283,7 +283,6 @@ getColumnsForNewDataPart( ColumnsDescription part_columns(source_part->getColumns()); NamesAndTypesList system_columns; - const auto & deleted_mask_column = LightweightDeleteDescription::FILTER_COLUMN; bool supports_lightweight_deletes = source_part->supportLightweightDeleteMutate(); bool deleted_mask_updated = false; @@ -299,9 +298,9 @@ getColumnsForNewDataPart( { for (const auto & [column_name, _] : command.column_to_update_expression) { - if (column_name == deleted_mask_column.name + if (column_name == RowExistsColumn::name && supports_lightweight_deletes - && !storage_columns_set.contains(deleted_mask_column.name)) + && !storage_columns_set.contains(RowExistsColumn::name)) deleted_mask_updated = true; } } @@ -323,12 +322,12 @@ getColumnsForNewDataPart( } } - if (!storage_columns_set.contains(deleted_mask_column.name)) + if (!storage_columns_set.contains(RowExistsColumn::name)) { - if (deleted_mask_updated || (part_columns.has(deleted_mask_column.name) && !has_delete_command)) + if (deleted_mask_updated || (part_columns.has(RowExistsColumn::name) && !has_delete_command)) { - storage_columns.push_back(deleted_mask_column); - storage_columns_set.insert(deleted_mask_column.name); + storage_columns.emplace_back(RowExistsColumn::name, RowExistsColumn::type); + storage_columns_set.insert(RowExistsColumn::name); } } diff --git a/src/Storages/MergeTree/StorageFromMergeTreeDataPart.h b/src/Storages/MergeTree/StorageFromMergeTreeDataPart.h index bbb38346f38..ca8ed9abdb5 100644 --- a/src/Storages/MergeTree/StorageFromMergeTreeDataPart.h +++ b/src/Storages/MergeTree/StorageFromMergeTreeDataPart.h @@ -34,6 +34,7 @@ public: , partition_id(part_->info.partition_id) { setInMemoryMetadata(storage.getInMemoryMetadata()); + setVirtuals(*storage.getVirtualsPtr()); } /// Used in queries with projection. @@ -90,11 +91,6 @@ public: bool supportsSubcolumns() const override { return true; } - NamesAndTypesList getVirtuals() const override - { - return storage.getVirtuals(); - } - String getPartitionId() const { return partition_id; diff --git a/src/Storages/NATS/NATSSource.cpp b/src/Storages/NATS/NATSSource.cpp index 3fc01eacb22..54f479faacc 100644 --- a/src/Storages/NATS/NATSSource.cpp +++ b/src/Storages/NATS/NATSSource.cpp @@ -9,10 +9,10 @@ namespace DB { -static std::pair getHeaders(StorageNATS & storage, const StorageSnapshotPtr & storage_snapshot) +static std::pair getHeaders(const StorageSnapshotPtr & storage_snapshot) { auto non_virtual_header = storage_snapshot->metadata->getSampleBlockNonMaterialized(); - auto virtual_header = storage_snapshot->getSampleBlockForColumns(storage.getVirtuals().getNames()); + auto virtual_header = storage_snapshot->virtual_columns->getSampleBlock(); return {non_virtual_header, virtual_header}; } @@ -33,7 +33,7 @@ NATSSource::NATSSource( const Names & columns, size_t max_block_size_, StreamingHandleErrorMode handle_error_mode_) - : NATSSource(storage_, storage_snapshot_, getHeaders(storage_, storage_snapshot_), context_, columns, max_block_size_, handle_error_mode_) + : NATSSource(storage_, storage_snapshot_, getHeaders(storage_snapshot_), context_, columns, max_block_size_, handle_error_mode_) { } diff --git a/src/Storages/NATS/StorageNATS.cpp b/src/Storages/NATS/StorageNATS.cpp index c7a5d0b8d0a..0b88a9e8929 100644 --- a/src/Storages/NATS/StorageNATS.cpp +++ b/src/Storages/NATS/StorageNATS.cpp @@ -88,6 +88,7 @@ StorageNATS::StorageNATS( StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(columns_); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals(nats_settings->nats_handle_error_mode)); nats_context = addSettings(getContext()); nats_context->makeQueryContext(); @@ -131,6 +132,19 @@ StorageNATS::StorageNATS( connection_task->deactivate(); } +VirtualColumnsDescription StorageNATS::createVirtuals(StreamingHandleErrorMode handle_error_mode) +{ + VirtualColumnsDescription desc; + desc.addEphemeral("_subject", std::make_shared(), ""); + + if (handle_error_mode == StreamingHandleErrorMode::STREAM) + { + desc.addEphemeral("_raw_message", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_error", std::make_shared(std::make_shared()), ""); + } + + return desc; +} Names StorageNATS::parseList(const String & list, char delim) { @@ -746,20 +760,4 @@ void registerStorageNATS(StorageFactory & factory) factory.registerStorage("NATS", creator_fn, StorageFactory::StorageFeatures{ .supports_settings = true, }); } - -NamesAndTypesList StorageNATS::getVirtuals() const -{ - auto virtuals = NamesAndTypesList{ - {"_subject", std::make_shared()} - }; - - if (nats_settings->nats_handle_error_mode == StreamingHandleErrorMode::STREAM) - { - virtuals.push_back({"_raw_message", std::make_shared(std::make_shared())}); - virtuals.push_back({"_error", std::make_shared(std::make_shared())}); - } - - return virtuals; -} - } diff --git a/src/Storages/NATS/StorageNATS.h b/src/Storages/NATS/StorageNATS.h index 94f955ccdae..41d77acfde6 100644 --- a/src/Storages/NATS/StorageNATS.h +++ b/src/Storages/NATS/StorageNATS.h @@ -61,7 +61,6 @@ public: NATSConsumerPtr popConsumer(std::chrono::milliseconds timeout); const String & getFormatName() const { return format_name; } - NamesAndTypesList getVirtuals() const override; void incrementReader(); void decrementReader(); @@ -137,6 +136,7 @@ private: static Names parseList(const String & list, char delim); static String getTableBasedName(String name, const StorageID & table_id); + static VirtualColumnsDescription createVirtuals(StreamingHandleErrorMode handle_error_mode); ContextMutablePtr addSettings(ContextPtr context) const; size_t getMaxBlockSize() const; diff --git a/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.cpp b/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.cpp index 61d83750c31..64d329f74b2 100644 --- a/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.cpp +++ b/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.cpp @@ -72,6 +72,7 @@ StorageMaterializedPostgreSQL::StorageMaterializedPostgreSQL( throw Exception(ErrorCodes::BAD_ARGUMENTS, "Storage MaterializedPostgreSQL is allowed only for Atomic database"); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals()); replication_settings->materialized_postgresql_tables_list = remote_table_name_; @@ -127,8 +128,16 @@ StorageMaterializedPostgreSQL::StorageMaterializedPostgreSQL( , nested_table_id(nested_storage_->getStorageID()) { setInMemoryMetadata(nested_storage_->getInMemoryMetadata()); + setVirtuals(*nested_storage_->getVirtualsPtr()); } +VirtualColumnsDescription StorageMaterializedPostgreSQL::createVirtuals() +{ + VirtualColumnsDescription desc; + desc.addEphemeral("_sign", std::make_shared(), ""); + desc.addEphemeral("_version", std::make_shared(), ""); + return desc; +} /// A temporary clone table might be created for current table in order to update its schema and reload /// all data in the background while current table will still handle read requests. @@ -254,15 +263,6 @@ void StorageMaterializedPostgreSQL::dropInnerTableIfAny(bool sync, ContextPtr lo } -NamesAndTypesList StorageMaterializedPostgreSQL::getVirtuals() const -{ - return NamesAndTypesList{ - {"_sign", std::make_shared()}, - {"_version", std::make_shared()} - }; -} - - bool StorageMaterializedPostgreSQL::needRewriteQueryWithFinal(const Names & column_names) const { return needRewriteQueryWithFinalForStorage(column_names, getNested()); diff --git a/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.h b/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.h index 2d5c1efae7c..af2f13bb880 100644 --- a/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.h +++ b/src/Storages/PostgreSQL/StorageMaterializedPostgreSQL.h @@ -89,8 +89,6 @@ public: /// Used only for single MaterializedPostgreSQL storage. void dropInnerTableIfAny(bool sync, ContextPtr local_context) override; - NamesAndTypesList getVirtuals() const override; - bool needRewriteQueryWithFinal(const Names & column_names) const override; void read( @@ -138,6 +136,8 @@ private: static std::shared_ptr getMaterializedColumnsDeclaration( String name, String type, UInt64 default_value); + static VirtualColumnsDescription createVirtuals(); + ASTPtr getColumnDeclaration(const DataTypePtr & data_type) const; String getNestedTableName() const; diff --git a/src/Storages/RabbitMQ/RabbitMQSource.cpp b/src/Storages/RabbitMQ/RabbitMQSource.cpp index 72196e7dd3c..4dc257074f3 100644 --- a/src/Storages/RabbitMQ/RabbitMQSource.cpp +++ b/src/Storages/RabbitMQ/RabbitMQSource.cpp @@ -11,10 +11,10 @@ namespace DB { -static std::pair getHeaders(StorageRabbitMQ & storage_, const StorageSnapshotPtr & storage_snapshot) +static std::pair getHeaders(const StorageSnapshotPtr & storage_snapshot) { auto non_virtual_header = storage_snapshot->metadata->getSampleBlockNonMaterialized(); - auto virtual_header = storage_snapshot->getSampleBlockForColumns(storage_.getVirtuals().getNames()); + auto virtual_header = storage_snapshot->virtual_columns->getSampleBlock(); return {non_virtual_header, virtual_header}; } @@ -40,7 +40,7 @@ RabbitMQSource::RabbitMQSource( : RabbitMQSource( storage_, storage_snapshot_, - getHeaders(storage_, storage_snapshot_), + getHeaders(storage_snapshot_), context_, columns, max_block_size_, diff --git a/src/Storages/RabbitMQ/StorageRabbitMQ.cpp b/src/Storages/RabbitMQ/StorageRabbitMQ.cpp index e2ef23193c5..980fccd307e 100644 --- a/src/Storages/RabbitMQ/StorageRabbitMQ.cpp +++ b/src/Storages/RabbitMQ/StorageRabbitMQ.cpp @@ -136,6 +136,7 @@ StorageRabbitMQ::StorageRabbitMQ( StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(columns_); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals(rabbitmq_settings->rabbitmq_handle_error_mode)); rabbitmq_context = addSettings(getContext()); rabbitmq_context->makeQueryContext(); @@ -191,6 +192,26 @@ StorageRabbitMQ::StorageRabbitMQ( init_task->deactivate(); } +VirtualColumnsDescription StorageRabbitMQ::createVirtuals(StreamingHandleErrorMode handle_error_mode) +{ + VirtualColumnsDescription desc; + + desc.addEphemeral("_exchange_name", std::make_shared(), ""); + desc.addEphemeral("_channel_id", std::make_shared(), ""); + desc.addEphemeral("_delivery_tag", std::make_shared(), ""); + desc.addEphemeral("_redelivered", std::make_shared(), ""); + desc.addEphemeral("_message_id", std::make_shared(), ""); + desc.addEphemeral("_timestamp", std::make_shared(), ""); + + + if (handle_error_mode == StreamingHandleErrorMode::STREAM) + { + desc.addEphemeral("_raw_message", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_error", std::make_shared(std::make_shared()), ""); + } + + return desc; +} Names StorageRabbitMQ::parseSettings(String settings_list) { @@ -1213,25 +1234,4 @@ void registerStorageRabbitMQ(StorageFactory & factory) factory.registerStorage("RabbitMQ", creator_fn, StorageFactory::StorageFeatures{ .supports_settings = true, }); } - -NamesAndTypesList StorageRabbitMQ::getVirtuals() const -{ - auto virtuals = NamesAndTypesList{ - {"_exchange_name", std::make_shared()}, - {"_channel_id", std::make_shared()}, - {"_delivery_tag", std::make_shared()}, - {"_redelivered", std::make_shared()}, - {"_message_id", std::make_shared()}, - {"_timestamp", std::make_shared()} - }; - - if (rabbitmq_settings->rabbitmq_handle_error_mode == StreamingHandleErrorMode::STREAM) - { - virtuals.push_back({"_raw_message", std::make_shared(std::make_shared())}); - virtuals.push_back({"_error", std::make_shared(std::make_shared())}); - } - - return virtuals; -} - } diff --git a/src/Storages/RabbitMQ/StorageRabbitMQ.h b/src/Storages/RabbitMQ/StorageRabbitMQ.h index b3a0b53cde5..e14741d9636 100644 --- a/src/Storages/RabbitMQ/StorageRabbitMQ.h +++ b/src/Storages/RabbitMQ/StorageRabbitMQ.h @@ -68,7 +68,6 @@ public: RabbitMQConsumerPtr popConsumer(std::chrono::milliseconds timeout); const String & getFormatName() const { return format_name; } - NamesAndTypesList getVirtuals() const override; String getExchange() const { return exchange_name; } void unbindExchange(); @@ -191,6 +190,8 @@ private: bool tryStreamToViews(); bool hasDependencies(const StorageID & table_id); + static VirtualColumnsDescription createVirtuals(StreamingHandleErrorMode handle_error_mode); + static String getRandomName() { std::uniform_int_distribution distribution('a', 'z'); diff --git a/src/Storages/RocksDB/StorageSystemRocksDB.cpp b/src/Storages/RocksDB/StorageSystemRocksDB.cpp index d0533b5ba0c..eec2f53381f 100644 --- a/src/Storages/RocksDB/StorageSystemRocksDB.cpp +++ b/src/Storages/RocksDB/StorageSystemRocksDB.cpp @@ -39,7 +39,7 @@ ColumnsDescription StorageSystemRocksDB::getColumnsDescription() } -void StorageSystemRocksDB::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const +void StorageSystemRocksDB::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_databases = !access->isGranted(AccessType::SHOW_TABLES); @@ -87,7 +87,7 @@ void StorageSystemRocksDB::fillData(MutableColumns & res_columns, ContextPtr con { col_table_to_filter, std::make_shared(), "table" }, }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) return; diff --git a/src/Storages/RocksDB/StorageSystemRocksDB.h b/src/Storages/RocksDB/StorageSystemRocksDB.h index c1f10a7722d..ec351c75446 100644 --- a/src/Storages/RocksDB/StorageSystemRocksDB.h +++ b/src/Storages/RocksDB/StorageSystemRocksDB.h @@ -11,7 +11,7 @@ class Context; /** Implements the `rocksdb` system table, which expose various rocksdb metrics. */ -class StorageSystemRocksDB final : public IStorageSystemOneBlock +class StorageSystemRocksDB final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemRocksDB"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const override; }; } diff --git a/src/Storages/S3Queue/StorageS3Queue.cpp b/src/Storages/S3Queue/StorageS3Queue.cpp index 88504975490..6e7ac2b47b8 100644 --- a/src/Storages/S3Queue/StorageS3Queue.cpp +++ b/src/Storages/S3Queue/StorageS3Queue.cpp @@ -155,8 +155,7 @@ StorageS3Queue::StorageS3Queue( storage_metadata.setConstraints(constraints_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); LOG_INFO(log, "Using zookeeper path: {}", zk_path.string()); task = getContext()->getSchedulePool().createTask("S3QueueStreamingTask", [this] { threadFunc(); }); @@ -315,7 +314,7 @@ void StorageS3Queue::read( } auto this_ptr = std::static_pointer_cast(shared_from_this()); - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context), getVirtuals()); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context)); auto reading = std::make_unique( column_names, @@ -493,7 +492,7 @@ bool StorageS3Queue::streamToViews() auto block_io = interpreter.execute(); auto file_iterator = createFileIterator(s3queue_context, nullptr); - auto read_from_format_info = prepareReadingFromFormat(block_io.pipeline.getHeader().getNames(), storage_snapshot, supportsSubsetOfColumns(s3queue_context), getVirtuals()); + auto read_from_format_info = prepareReadingFromFormat(block_io.pipeline.getHeader().getNames(), storage_snapshot, supportsSubsetOfColumns(s3queue_context)); Pipes pipes; pipes.reserve(s3queue_settings->s3queue_processing_threads_num); @@ -602,8 +601,9 @@ void StorageS3Queue::checkTableStructure(const String & zookeeper_prefix, const std::shared_ptr StorageS3Queue::createFileIterator(ContextPtr local_context, const ActionsDAG::Node * predicate) { auto glob_iterator = std::make_unique( - *configuration.client, configuration.url, predicate, virtual_columns, local_context, + *configuration.client, configuration.url, predicate, getVirtualsList(), local_context, /* read_keys */nullptr, configuration.request_settings); + return std::make_shared(files_metadata, std::move(glob_iterator), s3queue_settings->s3queue_current_shard_num, shutdown_called); } diff --git a/src/Storages/S3Queue/StorageS3Queue.h b/src/Storages/S3Queue/StorageS3Queue.h index fd3b4bb4914..bdd3ab7b687 100644 --- a/src/Storages/S3Queue/StorageS3Queue.h +++ b/src/Storages/S3Queue/StorageS3Queue.h @@ -51,8 +51,6 @@ public: size_t max_block_size, size_t num_streams) override; - NamesAndTypesList getVirtuals() const override { return virtual_columns; } - const auto & getFormatName() const { return configuration.format; } const fs::path & getZooKeeperPath() const { return zk_path; } @@ -71,7 +69,6 @@ private: Configuration configuration; const std::optional format_settings; - NamesAndTypesList virtual_columns; BackgroundSchedulePool::TaskHolder task; std::atomic stream_cancelled{false}; diff --git a/src/Storages/StorageAzureBlob.cpp b/src/Storages/StorageAzureBlob.cpp index 1f0fba99f84..bac9aa1cbdf 100644 --- a/src/Storages/StorageAzureBlob.cpp +++ b/src/Storages/StorageAzureBlob.cpp @@ -491,12 +491,11 @@ StorageAzureBlob::StorageAzureBlob( storage_metadata.setConstraints(constraints_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); StoredObjects objects; for (const auto & key : configuration.blobs_paths) objects.emplace_back(key); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); } void StorageAzureBlob::truncate(const ASTPtr &, const StorageMetadataPtr &, ContextPtr, TableExclusiveLockHolder &) @@ -736,7 +735,7 @@ void StorageAzureBlob::read( auto this_ptr = std::static_pointer_cast(shared_from_this()); - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context), getVirtuals()); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context)); bool need_only_count = (query_info.optimize_trivial_count || read_from_format_info.requested_columns.empty()) && local_context->getSettingsRef().optimize_count_from_files; @@ -772,13 +771,13 @@ void ReadFromAzureBlob::createIterator(const ActionsDAG::Node * predicate) /// Iterate through disclosed globs and make a source for each file iterator_wrapper = std::make_shared( storage->object_storage.get(), configuration.container, configuration.blob_path, - predicate, storage->virtual_columns, context, nullptr, context->getFileProgressCallback()); + predicate, storage->getVirtualsList(), context, nullptr, context->getFileProgressCallback()); } else { iterator_wrapper = std::make_shared( storage->object_storage.get(), configuration.container, configuration.blobs_paths, - predicate, storage->virtual_columns, context, nullptr, context->getFileProgressCallback()); + predicate, storage->getVirtualsList(), context, nullptr, context->getFileProgressCallback()); } } @@ -886,16 +885,6 @@ SinkToStoragePtr StorageAzureBlob::write(const ASTPtr & query, const StorageMeta } } -NamesAndTypesList StorageAzureBlob::getVirtuals() const -{ - return virtual_columns; -} - -Names StorageAzureBlob::getVirtualColumnNames() -{ - return VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage({}).getNames(); -} - bool StorageAzureBlob::supportsPartitionBy() const { return true; diff --git a/src/Storages/StorageAzureBlob.h b/src/Storages/StorageAzureBlob.h index e1d1c3abd33..63fd489dcaf 100644 --- a/src/Storages/StorageAzureBlob.h +++ b/src/Storages/StorageAzureBlob.h @@ -94,9 +94,6 @@ public: void truncate(const ASTPtr & query, const StorageMetadataPtr & metadata_snapshot, ContextPtr local_context, TableExclusiveLockHolder &) override; - NamesAndTypesList getVirtuals() const override; - static Names getVirtualColumnNames(); - bool supportsPartitionBy() const override; bool supportsSubcolumns() const override { return true; } @@ -136,7 +133,6 @@ private: std::string name; Configuration configuration; std::unique_ptr object_storage; - NamesAndTypesList virtual_columns; const bool distributed_processing; std::optional format_settings; diff --git a/src/Storages/StorageAzureBlobCluster.cpp b/src/Storages/StorageAzureBlobCluster.cpp index 32445556611..a80d121567a 100644 --- a/src/Storages/StorageAzureBlobCluster.cpp +++ b/src/Storages/StorageAzureBlobCluster.cpp @@ -63,8 +63,7 @@ StorageAzureBlobCluster::StorageAzureBlobCluster( storage_metadata.setConstraints(constraints_); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } void StorageAzureBlobCluster::updateQueryToSendIfNeeded(DB::ASTPtr & query, const DB::StorageSnapshotPtr & storage_snapshot, const DB::ContextPtr & context) @@ -81,17 +80,12 @@ RemoteQueryExecutor::Extension StorageAzureBlobCluster::getTaskIteratorExtension { auto iterator = std::make_shared( object_storage.get(), configuration.container, configuration.blob_path, - predicate, virtual_columns, context, nullptr); + predicate, getVirtualsList(), context, nullptr); + auto callback = std::make_shared>([iterator]() mutable -> String{ return iterator->next().relative_path; }); return RemoteQueryExecutor::Extension{ .task_iterator = std::move(callback) }; } -NamesAndTypesList StorageAzureBlobCluster::getVirtuals() const -{ - return virtual_columns; -} - - } #endif diff --git a/src/Storages/StorageAzureBlobCluster.h b/src/Storages/StorageAzureBlobCluster.h index 476f21c6742..545e568a772 100644 --- a/src/Storages/StorageAzureBlobCluster.h +++ b/src/Storages/StorageAzureBlobCluster.h @@ -31,8 +31,6 @@ public: std::string getName() const override { return "AzureBlobStorageCluster"; } - NamesAndTypesList getVirtuals() const override; - RemoteQueryExecutor::Extension getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const override; bool supportsSubcolumns() const override { return true; } @@ -45,7 +43,6 @@ private: void updateQueryToSendIfNeeded(ASTPtr & query, const StorageSnapshotPtr & storage_snapshot, const ContextPtr & context) override; StorageAzureBlob::Configuration configuration; - NamesAndTypesList virtual_columns; std::unique_ptr object_storage; }; diff --git a/src/Storages/StorageDistributed.cpp b/src/Storages/StorageDistributed.cpp index 92e7dcdf4f2..3b766ac8d26 100644 --- a/src/Storages/StorageDistributed.cpp +++ b/src/Storages/StorageDistributed.cpp @@ -19,6 +19,7 @@ #include #include #include +#include #include @@ -104,11 +105,8 @@ #include #include -#include - #include #include -#include #include @@ -290,22 +288,17 @@ size_t getClusterQueriedNodes(const Settings & settings, const ClusterPtr & clus StorageDistributed::~StorageDistributed() = default; -NamesAndTypesList StorageDistributed::getVirtuals() const +VirtualColumnsDescription StorageDistributed::createVirtuals() { - /// NOTE This is weird. Most of these virtual columns are part of MergeTree + /// NOTE: This is weird. + /// Most of these virtual columns are part of MergeTree /// tables info. But Distributed is general-purpose engine. - return NamesAndTypesList{ - NameAndTypePair("_table", std::make_shared(std::make_shared())), - NameAndTypePair("_part", std::make_shared(std::make_shared())), - NameAndTypePair("_part_index", std::make_shared()), - NameAndTypePair("_part_uuid", std::make_shared()), - NameAndTypePair("_partition_id", std::make_shared(std::make_shared())), - NameAndTypePair("_sample_factor", std::make_shared()), - NameAndTypePair("_part_offset", std::make_shared()), - NameAndTypePair("_row_exists", std::make_shared()), - NameAndTypePair(BlockNumberColumn::name, BlockNumberColumn::type), - NameAndTypePair("_shard_num", std::make_shared()), /// deprecated - }; + StorageInMemoryMetadata metadata; + auto desc = MergeTreeData::createVirtuals(metadata); + + desc.addEphemeral("_shard_num", std::make_shared(), "Deprecated. Use function shardNum instead"); + + return desc; } StorageDistributed::StorageDistributed( @@ -354,6 +347,7 @@ StorageDistributed::StorageDistributed( storage_metadata.setConstraints(constraints_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals()); if (sharding_key_) { @@ -744,6 +738,32 @@ StorageSnapshotPtr StorageDistributed::getStorageSnapshotForQuery( namespace { +class ReplaseAliasColumnsVisitor : public InDepthQueryTreeVisitor +{ + static QueryTreeNodePtr getColumnNodeAliasExpression(const QueryTreeNodePtr & node) + { + const auto * column_node = node->as(); + if (!column_node || !column_node->hasExpression()) + return nullptr; + + const auto & column_source = column_node->getColumnSourceOrNull(); + if (!column_source || column_source->getNodeType() == QueryTreeNodeType::JOIN + || column_source->getNodeType() == QueryTreeNodeType::ARRAY_JOIN) + return nullptr; + + auto column_expression = column_node->getExpression(); + column_expression->setAlias(column_node->getColumnName()); + return column_expression; + } + +public: + void visitImpl(QueryTreeNodePtr & node) + { + if (auto column_expression = getColumnNodeAliasExpression(node)) + node = column_expression; + } +}; + QueryTreeNodePtr buildQueryTreeDistributed(SelectQueryInfo & query_info, const StorageSnapshotPtr & distributed_storage_snapshot, const StorageID & remote_storage_id, @@ -796,6 +816,8 @@ QueryTreeNodePtr buildQueryTreeDistributed(SelectQueryInfo & query_info, replacement_table_expression->setAlias(query_info.table_expression->getAlias()); auto query_tree_to_modify = query_info.query_tree->cloneAndReplace(query_info.table_expression, std::move(replacement_table_expression)); + ReplaseAliasColumnsVisitor replase_alias_columns_visitor; + replase_alias_columns_visitor.visit(query_tree_to_modify); return buildQueryTreeForShard(query_info.planner_context, query_tree_to_modify); } @@ -941,7 +963,7 @@ SinkToStoragePtr StorageDistributed::write(const ASTPtr &, const StorageMetadata else columns_to_send = metadata_snapshot->getSampleBlockNonMaterialized().getNames(); - /// DistributedSink will not own cluster, but will own ConnectionPools of the cluster + /// DistributedSink will not own cluster return std::make_shared( local_context, *this, metadata_snapshot, cluster, insert_sync, timeout, StorageID{remote_database, remote_table}, columns_to_send); diff --git a/src/Storages/StorageDistributed.h b/src/Storages/StorageDistributed.h index c00dd8cea04..cf5b78305a5 100644 --- a/src/Storages/StorageDistributed.h +++ b/src/Storages/StorageDistributed.h @@ -146,8 +146,6 @@ public: ActionLock getActionLock(StorageActionBlockType type) override; - NamesAndTypesList getVirtuals() const override; - /// Used by InterpreterInsertQuery std::string getRemoteDatabaseName() const { return remote_database; } std::string getRemoteTableName() const { return remote_table; } @@ -234,6 +232,8 @@ private: std::optional distributedWriteFromClusterStorage(const IStorageCluster & src_storage_cluster, const ASTInsertQuery & query, ContextPtr context) const; std::optional distributedWriteBetweenDistributedTables(const StorageDistributed & src_distributed, const ASTInsertQuery & query, ContextPtr context) const; + static VirtualColumnsDescription createVirtuals(); + String remote_database; String remote_table; ASTPtr remote_table_function_ptr; diff --git a/src/Storages/StorageFile.cpp b/src/Storages/StorageFile.cpp index 7ed99446122..0d220f2fd5d 100644 --- a/src/Storages/StorageFile.cpp +++ b/src/Storages/StorageFile.cpp @@ -1097,8 +1097,7 @@ void StorageFile::setStorageMetadata(CommonArguments args) storage_metadata.setConstraints(args.constraints); storage_metadata.setComment(args.comment); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } @@ -1581,7 +1580,7 @@ void StorageFile::read( auto this_ptr = std::static_pointer_cast(shared_from_this()); - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(context), getVirtuals()); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(context)); bool need_only_count = (query_info.optimize_trivial_count || read_from_format_info.requested_columns.empty()) && context->getSettingsRef().optimize_count_from_files; @@ -1609,7 +1608,7 @@ void ReadFromFile::createIterator(const ActionsDAG::Node * predicate) storage->paths, storage->archive_info, predicate, - storage->virtual_columns, + storage->getVirtualsList(), context, storage->distributed_processing); } @@ -2256,9 +2255,4 @@ StorageFile::ArchiveInfo StorageFile::getArchiveInfo( return archive_info; } -Names StorageFile::getVirtualColumnNames() -{ - return VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage({}).getNames(); -} - } diff --git a/src/Storages/StorageFile.h b/src/Storages/StorageFile.h index 1db33b72d77..93c263008a6 100644 --- a/src/Storages/StorageFile.h +++ b/src/Storages/StorageFile.h @@ -80,10 +80,6 @@ public: bool storesDataOnDisk() const override; Strings getDataPaths() const override; - NamesAndTypesList getVirtuals() const override { return virtual_columns; } - - static Names getVirtualColumnNames(); - static Strings getPathsList(const String & table_path, const String & user_files_path, const ContextPtr & context, size_t & total_bytes_to_read); /// Check if the format supports reading only some subset of columns. @@ -197,8 +193,6 @@ private: std::atomic readers_counter = 0; FileRenamer file_renamer; bool was_renamed = false; - - NamesAndTypesList virtual_columns; bool distributed_processing = false; }; diff --git a/src/Storages/StorageFileCluster.cpp b/src/Storages/StorageFileCluster.cpp index 0cc18abef5f..d43e242f70c 100644 --- a/src/Storages/StorageFileCluster.cpp +++ b/src/Storages/StorageFileCluster.cpp @@ -61,8 +61,7 @@ StorageFileCluster::StorageFileCluster( storage_metadata.setConstraints(constraints_); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } void StorageFileCluster::updateQueryToSendIfNeeded(DB::ASTPtr & query, const StorageSnapshotPtr & storage_snapshot, const DB::ContextPtr & context) @@ -77,7 +76,7 @@ void StorageFileCluster::updateQueryToSendIfNeeded(DB::ASTPtr & query, const Sto RemoteQueryExecutor::Extension StorageFileCluster::getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const { - auto iterator = std::make_shared(paths, std::nullopt, predicate, virtual_columns, context); + auto iterator = std::make_shared(paths, std::nullopt, predicate, getVirtualsList(), context); auto callback = std::make_shared([iter = std::move(iterator)]() mutable -> String { return iter->next(); }); return RemoteQueryExecutor::Extension{.task_iterator = std::move(callback)}; } diff --git a/src/Storages/StorageFileCluster.h b/src/Storages/StorageFileCluster.h index 2803c8b6e5b..3acbc71ba7e 100644 --- a/src/Storages/StorageFileCluster.h +++ b/src/Storages/StorageFileCluster.h @@ -28,8 +28,6 @@ public: std::string getName() const override { return "FileCluster"; } - NamesAndTypesList getVirtuals() const override { return virtual_columns; } - RemoteQueryExecutor::Extension getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const override; bool supportsSubcolumns() const override { return true; } @@ -42,7 +40,6 @@ private: Strings paths; String filename; String format_name; - NamesAndTypesList virtual_columns; }; } diff --git a/src/Storages/StorageKeeperMap.cpp b/src/Storages/StorageKeeperMap.cpp index 40a4190a413..b0b7afdfe8d 100644 --- a/src/Storages/StorageKeeperMap.cpp +++ b/src/Storages/StorageKeeperMap.cpp @@ -331,6 +331,10 @@ StorageKeeperMap::StorageKeeperMap( setInMemoryMetadata(metadata); + VirtualColumnsDescription virtuals; + virtuals.addEphemeral(String(version_column_name), std::make_shared(), ""); + setVirtuals(std::move(virtuals)); + WriteBufferFromOwnString out; out << "KeeperMap metadata format version: 1\n" << "columns: " << metadata.columns.toString() @@ -634,12 +638,6 @@ void StorageKeeperMap::drop() dropTable(client, metadata_drop_lock); } -NamesAndTypesList StorageKeeperMap::getVirtuals() const -{ - return NamesAndTypesList{ - {std::string{version_column_name}, std::make_shared()}}; -} - namespace { diff --git a/src/Storages/StorageKeeperMap.h b/src/Storages/StorageKeeperMap.h index d65548ed428..d4556792c48 100644 --- a/src/Storages/StorageKeeperMap.h +++ b/src/Storages/StorageKeeperMap.h @@ -50,8 +50,6 @@ public: void truncate(const ASTPtr &, const StorageMetadataPtr &, ContextPtr, TableExclusiveLockHolder &) override; void drop() override; - NamesAndTypesList getVirtuals() const override; - std::string getName() const override { return "KeeperMap"; } Names getPrimaryKey() const override { return {primary_key}; } diff --git a/src/Storages/StorageLog.cpp b/src/Storages/StorageLog.cpp index 7459760b0f5..549cfca1b6c 100644 --- a/src/Storages/StorageLog.cpp +++ b/src/Storages/StorageLog.cpp @@ -35,7 +35,6 @@ #include #include #include -#include #include #include @@ -48,8 +47,6 @@ namespace DB { - CompressionCodecPtr getCompressionCodecDelta(UInt8 delta_bytes_size); - namespace ErrorCodes { extern const int TIMEOUT_EXCEEDED; @@ -299,6 +296,7 @@ public: : SinkToStorage(metadata_snapshot_->getSampleBlock()) , storage(storage_) , metadata_snapshot(metadata_snapshot_) + , storage_snapshot(std::make_shared(storage, metadata_snapshot)) , lock(std::move(lock_)) { if (!lock) @@ -343,6 +341,7 @@ public: private: StorageLog & storage; StorageMetadataPtr metadata_snapshot; + StorageSnapshotPtr storage_snapshot; WriteLock lock; bool done = false; @@ -476,13 +475,7 @@ void LogSink::writeData(const NameAndTypePair & name_and_type, const IColumn & c throw Exception(ErrorCodes::LOGICAL_ERROR, "No information about file {} in StorageLog", data_file_name); const auto & data_file = *data_file_it->second; - const auto & columns = metadata_snapshot->getColumns(); - - CompressionCodecPtr compression; - if (name_and_type.name == BlockNumberColumn::name) - compression = BlockNumberColumn::compression_codec; - else - compression = columns.getCodecOrDefault(name_and_type.name); + auto compression = storage_snapshot->getCodecOrDefault(name_and_type.name); it = streams.try_emplace(data_file.name, storage.disk, data_file.path, storage.file_checker.getFileSize(data_file.path), diff --git a/src/Storages/StorageMaterializedMySQL.cpp b/src/Storages/StorageMaterializedMySQL.cpp index 0dc0b1bff0b..887c58ff816 100644 --- a/src/Storages/StorageMaterializedMySQL.cpp +++ b/src/Storages/StorageMaterializedMySQL.cpp @@ -22,9 +22,8 @@ namespace DB StorageMaterializedMySQL::StorageMaterializedMySQL(const StoragePtr & nested_storage_, const IDatabase * database_) : StorageProxy(nested_storage_->getStorageID()), nested_storage(nested_storage_), database(database_) { - StorageInMemoryMetadata in_memory_metadata; - in_memory_metadata = nested_storage->getInMemoryMetadata(); - setInMemoryMetadata(in_memory_metadata); + setInMemoryMetadata(nested_storage->getInMemoryMetadata()); + setVirtuals(*nested_storage->getVirtualsPtr()); } bool StorageMaterializedMySQL::needRewriteQueryWithFinal(const Names & column_names) const @@ -49,14 +48,6 @@ void StorageMaterializedMySQL::read( query_info, context, processed_stage, max_block_size, num_streams); } -NamesAndTypesList StorageMaterializedMySQL::getVirtuals() const -{ - if (const auto * db = typeid_cast(database)) - db->rethrowExceptionIfNeeded(); - - return nested_storage->getVirtuals(); -} - IStorage::ColumnSizeByName StorageMaterializedMySQL::getColumnSizes() const { auto sizes = nested_storage->getColumnSizes(); diff --git a/src/Storages/StorageMaterializedMySQL.h b/src/Storages/StorageMaterializedMySQL.h index 2cd589bfd75..9f5d157ce3b 100644 --- a/src/Storages/StorageMaterializedMySQL.h +++ b/src/Storages/StorageMaterializedMySQL.h @@ -34,7 +34,6 @@ public: SinkToStoragePtr write(const ASTPtr &, const StorageMetadataPtr &, ContextPtr, bool) override { throwNotAllowed(); } - NamesAndTypesList getVirtuals() const override; ColumnSizeByName getColumnSizes() const override; StoragePtr getNested() const override { return nested_storage; } diff --git a/src/Storages/StorageMaterializedView.cpp b/src/Storages/StorageMaterializedView.cpp index 1d0898a2f11..02cba1cf753 100644 --- a/src/Storages/StorageMaterializedView.cpp +++ b/src/Storages/StorageMaterializedView.cpp @@ -171,6 +171,12 @@ QueryProcessingStage::Enum StorageMaterializedView::getQueryProcessingStage( return getTargetTable()->getQueryProcessingStage(local_context, to_stage, getTargetTable()->getStorageSnapshot(target_metadata, local_context), query_info); } +StorageSnapshotPtr StorageMaterializedView::getStorageSnapshot(const StorageMetadataPtr & metadata_snapshot, ContextPtr) const +{ + /// We cannot set virtuals at table creation because target table may not exist at that time. + return std::make_shared(*this, metadata_snapshot, getTargetTable()->getVirtualsPtr()); +} + void StorageMaterializedView::read( QueryPlan & query_plan, const Names & column_names, @@ -536,11 +542,6 @@ StoragePtr StorageMaterializedView::tryGetTargetTable() const return DatabaseCatalog::instance().tryGetTable(getTargetTableId(), getContext()); } -NamesAndTypesList StorageMaterializedView::getVirtuals() const -{ - return getTargetTable()->getVirtuals(); -} - Strings StorageMaterializedView::getDataPaths() const { if (auto table = tryGetTargetTable()) diff --git a/src/Storages/StorageMaterializedView.h b/src/Storages/StorageMaterializedView.h index 4d574a821ec..198b7a642ee 100644 --- a/src/Storages/StorageMaterializedView.h +++ b/src/Storages/StorageMaterializedView.h @@ -75,12 +75,11 @@ public: StoragePtr tryGetTargetTable() const; StorageID getTargetTableId() const; - /// Get the virtual column of the target table; - NamesAndTypesList getVirtuals() const override; - ActionLock getActionLock(StorageActionBlockType type) override; void onActionLockRemove(StorageActionBlockType action_type) override; + StorageSnapshotPtr getStorageSnapshot(const StorageMetadataPtr & metadata_snapshot, ContextPtr) const override; + void read( QueryPlan & query_plan, const Names & column_names, diff --git a/src/Storages/StorageMerge.cpp b/src/Storages/StorageMerge.cpp index b827670bd4c..8410f0a8df8 100644 --- a/src/Storages/StorageMerge.cpp +++ b/src/Storages/StorageMerge.cpp @@ -63,10 +63,12 @@ #include #include +namespace DB +{ + namespace { -using namespace DB; bool columnIsPhysical(ColumnDefaultKind kind) { return kind == ColumnDefaultKind::Default || kind == ColumnDefaultKind::Materialized; @@ -82,10 +84,23 @@ bool columnDefaultKindHasSameType(ColumnDefaultKind lhs, ColumnDefaultKind rhs) return false; } +/// Adds to the select query section `WITH value AS column_name` +/// +/// For example: +/// - `WITH 9000 as _port`. +void rewriteEntityInAst(ASTPtr ast, const String & column_name, const Field & value) +{ + auto & select = ast->as(); + if (!select.with()) + select.setExpression(ASTSelectQuery::Expression::WITH, std::make_shared()); + + auto literal = std::make_shared(value); + literal->alias = column_name; + literal->prefer_alias_to_column_name = true; + select.with()->children.push_back(literal); } -namespace DB -{ +} namespace ErrorCodes { @@ -132,6 +147,7 @@ StorageMerge::StorageMerge( storage_metadata.setColumns(columns_.empty() ? getColumnsDescriptionFromSourceTables() : columns_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals()); } StorageMerge::StorageMerge( @@ -154,6 +170,7 @@ StorageMerge::StorageMerge( storage_metadata.setColumns(columns_.empty() ? getColumnsDescriptionFromSourceTables() : columns_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals()); } StorageMerge::DatabaseTablesIterators StorageMerge::getDatabaseIterators(ContextPtr context_) const @@ -306,6 +323,37 @@ QueryProcessingStage::Enum StorageMerge::getQueryProcessingStage( return selected_table_size == 1 ? stage_in_source_tables : std::min(stage_in_source_tables, QueryProcessingStage::WithMergeableState); } +VirtualColumnsDescription StorageMerge::createVirtuals() +{ + VirtualColumnsDescription desc; + + desc.addEphemeral("_database", std::make_shared(std::make_shared()), ""); + desc.addEphemeral("_table", std::make_shared(std::make_shared()), ""); + + return desc; +} + +StorageSnapshotPtr StorageMerge::getStorageSnapshot(const StorageMetadataPtr & metadata_snapshot, ContextPtr) const +{ + static const auto common_virtuals = createVirtuals(); + + auto virtuals = common_virtuals; + if (auto first_table = getFirstTable([](auto && table) { return table; })) + { + auto table_virtuals = first_table->getVirtualsPtr(); + for (const auto & column : *table_virtuals) + { + if (virtuals.has(column.name)) + continue; + + virtuals.add(column); + } + } + + auto virtuals_ptr = std::make_shared(std::move(virtuals)); + return std::make_shared(*this, metadata_snapshot, std::move(virtuals_ptr)); +} + void StorageMerge::read( QueryPlan & query_plan, const Names & column_names, @@ -897,7 +945,6 @@ SelectQueryInfo ReadFromMerge::getModifiedQueryInfo(const ContextPtr & modified_ column_node = std::make_shared(NameAndTypePair{column, storage_columns.getColumn(get_column_options, column).type }, modified_query_info.table_expression); } - PlannerActionsVisitor actions_visitor(modified_query_info.planner_context, false /*use_column_identifier_as_action_node_name*/); actions_visitor.visit(filter_actions_dag, column_node); } @@ -928,8 +975,8 @@ SelectQueryInfo ReadFromMerge::getModifiedQueryInfo(const ContextPtr & modified_ if (!is_storage_merge_engine) { - VirtualColumnUtils::rewriteEntityInAst(modified_query_info.query, "_table", current_storage_id.table_name); - VirtualColumnUtils::rewriteEntityInAst(modified_query_info.query, "_database", current_storage_id.database_name); + rewriteEntityInAst(modified_query_info.query, "_table", current_storage_id.table_name); + rewriteEntityInAst(modified_query_info.query, "_database", current_storage_id.database_name); } } @@ -1000,7 +1047,7 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( Block pipe_header = builder->getHeader(); - if (has_database_virtual_column && !pipe_header.has("_database")) + if (has_database_virtual_column && common_header.has("_database") && !pipe_header.has("_database")) { ColumnWithTypeAndName column; column.name = "_database"; @@ -1015,7 +1062,7 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( { return std::make_shared(stream_header, adding_column_actions); }); } - if (has_table_virtual_column && !pipe_header.has("_table")) + if (has_table_virtual_column && common_header.has("_table") && !pipe_header.has("_table")) { ColumnWithTypeAndName column; column.name = "_table"; @@ -1375,6 +1422,7 @@ void StorageMerge::alter( params.apply(storage_metadata, local_context); DatabaseCatalog::instance().getDatabase(table_id.database_name)->alterTable(local_context, table_id, storage_metadata); setInMemoryMetadata(storage_metadata); + setVirtuals(createVirtuals()); } void ReadFromMerge::convertAndFilterSourceStream( @@ -1634,20 +1682,4 @@ void registerStorageMerge(StorageFactory & factory) }); } -NamesAndTypesList StorageMerge::getVirtuals() const -{ - NamesAndTypesList virtuals{ - {"_database", std::make_shared(std::make_shared())}, - {"_table", std::make_shared(std::make_shared())}}; - - auto first_table = getFirstTable([](auto && table) { return table; }); - if (first_table) - { - auto table_virtuals = first_table->getVirtuals(); - virtuals.insert(virtuals.end(), table_virtuals.begin(), table_virtuals.end()); - } - - return virtuals; -} - } diff --git a/src/Storages/StorageMerge.h b/src/Storages/StorageMerge.h index 8b062a392d4..556649f622d 100644 --- a/src/Storages/StorageMerge.h +++ b/src/Storages/StorageMerge.h @@ -55,6 +55,8 @@ public: QueryProcessingStage::Enum getQueryProcessingStage(ContextPtr, QueryProcessingStage::Enum, const StorageSnapshotPtr &, SelectQueryInfo &) const override; + StorageSnapshotPtr getStorageSnapshot(const StorageMetadataPtr & metadata_snapshot, ContextPtr) const override; + void read( QueryPlan & query_plan, const Names & column_names, @@ -116,11 +118,12 @@ private: template void forEachTable(F && func) const; - NamesAndTypesList getVirtuals() const override; ColumnSizeByName getColumnSizes() const override; ColumnsDescription getColumnsDescriptionFromSourceTables() const; + static VirtualColumnsDescription createVirtuals(); + bool tableSupportsPrewhere() const; template @@ -281,6 +284,8 @@ private: ContextPtr query_context, bool filter_by_database_virtual_column, bool filter_by_table_virtual_column) const; + + // static VirtualColumnsDescription createVirtuals(StoragePtr first_table); }; } diff --git a/src/Storages/StorageMergeTreeIndex.cpp b/src/Storages/StorageMergeTreeIndex.cpp index d875611bb50..5f89849e920 100644 --- a/src/Storages/StorageMergeTreeIndex.cpp +++ b/src/Storages/StorageMergeTreeIndex.cpp @@ -15,6 +15,9 @@ #include #include #include +#include +#include +#include namespace DB { @@ -216,7 +219,6 @@ StorageMergeTreeIndex::StorageMergeTreeIndex( : IStorage(table_id_) , source_table(source_table_) , with_marks(with_marks_) - , log(&Poco::Logger::get("StorageMergeTreeIndex")) { const auto * merge_tree = dynamic_cast(source_table.get()); if (!merge_tree) @@ -230,7 +232,47 @@ StorageMergeTreeIndex::StorageMergeTreeIndex( setInMemoryMetadata(storage_metadata); } -Pipe StorageMergeTreeIndex::read( +class ReadFromMergeTreeIndex : public SourceStepWithFilter +{ +public: + std::string getName() const override { return "ReadFromMergeTreeIndex"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; + + ReadFromMergeTreeIndex( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , storage(std::move(storage_)) + , log(&Poco::Logger::get("StorageMergeTreeIndex")) + { + } + + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +private: + std::shared_ptr storage; + Poco::Logger * log; + const ActionsDAG::Node * predicate = nullptr; +}; + +void ReadFromMergeTreeIndex::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +void StorageMergeTreeIndex::read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -265,21 +307,32 @@ Pipe StorageMergeTreeIndex::read( context->checkAccess(AccessType::SELECT, source_table->getStorageID(), columns_from_storage); - auto header = storage_snapshot->getSampleBlockForColumns(column_names); - auto filtered_parts = getFilteredDataParts(query_info, context); + auto sample_block = storage_snapshot->getSampleBlockForColumns(column_names); - LOG_DEBUG(log, "Reading index{}from {} parts of table {}", - with_marks ? " with marks " : " ", - filtered_parts.size(), - source_table->getStorageID().getNameForLogs()); + auto this_ptr = std::static_pointer_cast(shared_from_this()); - return Pipe(std::make_shared(std::move(header), key_sample_block, std::move(filtered_parts), context, with_marks)); + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(sample_block), std::move(this_ptr)); + + query_plan.addStep(std::move(reading)); } -MergeTreeData::DataPartsVector StorageMergeTreeIndex::getFilteredDataParts(SelectQueryInfo & query_info, const ContextPtr & context) const +void ReadFromMergeTreeIndex::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) { - const auto * select_query = query_info.query->as(); - if (!select_query || !select_query->where()) + auto filtered_parts = storage->getFilteredDataParts(predicate, context); + + LOG_DEBUG(log, "Reading index{}from {} parts of table {}", + storage->with_marks ? " with marks " : " ", + filtered_parts.size(), + storage->source_table->getStorageID().getNameForLogs()); + + pipeline.init(Pipe(std::make_shared(getOutputStream().header, storage->key_sample_block, std::move(filtered_parts), context, storage->with_marks))); +} + +MergeTreeData::DataPartsVector StorageMergeTreeIndex::getFilteredDataParts(const ActionsDAG::Node * predicate, const ContextPtr & context) const +{ + if (!predicate) return data_parts; auto all_part_names = ColumnString::create(); @@ -287,7 +340,7 @@ MergeTreeData::DataPartsVector StorageMergeTreeIndex::getFilteredDataParts(Selec all_part_names->insert(part->name); Block filtered_block{{std::move(all_part_names), std::make_shared(), part_name_column.name}}; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) return {}; diff --git a/src/Storages/StorageMergeTreeIndex.h b/src/Storages/StorageMergeTreeIndex.h index b610d391655..a1fb61d5a56 100644 --- a/src/Storages/StorageMergeTreeIndex.h +++ b/src/Storages/StorageMergeTreeIndex.h @@ -21,7 +21,8 @@ public: const ColumnsDescription & columns, bool with_marks_); - Pipe read( + void read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -33,14 +34,15 @@ public: String getName() const override { return "MergeTreeIndex"; } private: - MergeTreeData::DataPartsVector getFilteredDataParts(SelectQueryInfo & query_info, const ContextPtr & context) const; + friend class ReadFromMergeTreeIndex; + + MergeTreeData::DataPartsVector getFilteredDataParts(const ActionsDAG::Node * predicate, const ContextPtr & context) const; StoragePtr source_table; bool with_marks; MergeTreeData::DataPartsVector data_parts; Block key_sample_block; - Poco::Logger * log; }; } diff --git a/src/Storages/StorageProxy.h b/src/Storages/StorageProxy.h index 18a1f9086ae..a8e7fd528dd 100644 --- a/src/Storages/StorageProxy.h +++ b/src/Storages/StorageProxy.h @@ -30,7 +30,6 @@ public: bool hasEvenlyDistributedRead() const override { return getNested()->hasEvenlyDistributedRead(); } ColumnSizeByName getColumnSizes() const override { return getNested()->getColumnSizes(); } - NamesAndTypesList getVirtuals() const override { return getNested()->getVirtuals(); } QueryProcessingStage::Enum getQueryProcessingStage( ContextPtr context, diff --git a/src/Storages/StorageReplicatedMergeTree.cpp b/src/Storages/StorageReplicatedMergeTree.cpp index 1c66a5f6ecd..73c7e484bcb 100644 --- a/src/Storages/StorageReplicatedMergeTree.cpp +++ b/src/Storages/StorageReplicatedMergeTree.cpp @@ -665,6 +665,34 @@ void StorageReplicatedMergeTree::waitMutationToFinishOnReplicas( } } +namespace +{ + +std::vector getAncestors(const String & path) +{ + auto full_path = fs::path(path); + if (full_path.empty()) + return {}; + + std::vector result; + fs::path ancestor = *full_path.begin(); + for (auto it = ++full_path.begin(); it != full_path.end(); it++) + { + /// If there is a directory separator after the last file-name in the path, + /// the last element before the end iterator is an empty element. + /// We do not what to create path with the / at the end + if (!it->empty()) + { + ancestor = ancestor / *it; + result.push_back(ancestor); + } + } + + return result; +} + +} + void StorageReplicatedMergeTree::createNewZooKeeperNodes() { auto zookeeper = getZooKeeper(); @@ -687,44 +715,14 @@ void StorageReplicatedMergeTree::createNewZooKeeperNodes() const auto settings = getSettings(); if (settings->allow_remote_fs_zero_copy_replication) { - auto disks = getStoragePolicy()->getDisks(); - std::set disk_types_for_zero_copy; - for (auto & disk : disks) + for (const auto & zero_copy_locks_root : getZookeeperZeroCopyLockPaths()) { - if (!disk->supportZeroCopyReplication()) - continue; - - disk_types_for_zero_copy.insert(disk->getDataSourceDescription().toString()); - } - - const auto table_shared_id_ = getTableSharedID(); - - if (!disk_types_for_zero_copy.empty()) - { - /// create all ancestors of remote_fs_zero_copy_zookeeper_path - auto remote_fs_zero_copy_zookeeper_path = fs::path(settings->remote_fs_zero_copy_zookeeper_path.toString()); - fs::path ancestor = *remote_fs_zero_copy_zookeeper_path.begin(); - for (auto it = ++remote_fs_zero_copy_zookeeper_path.begin(); it != remote_fs_zero_copy_zookeeper_path.end(); it++) + for (const auto & ancestor : getAncestors(zero_copy_locks_root)) { - /// If there is a directory separator after the last file-name in the path, - /// the last element before the end iterator is an empty element. - /// We do not what to create path with the / at the end - if (!it->empty()) - { - ancestor = ancestor / *it; - futures.push_back(zookeeper->asyncTryCreateNoThrow(ancestor, String(), zkutil::CreateMode::Persistent)); - } + futures.push_back(zookeeper->asyncTryCreateNoThrow(ancestor, String(), zkutil::CreateMode::Persistent)); } } - for (const auto & disk_type : disk_types_for_zero_copy) - { - auto zero_copy = fmt::format("zero_copy_{}", disk_type); - auto zero_copy_path = fs::path(settings->remote_fs_zero_copy_zookeeper_path.toString()) / zero_copy; - futures.push_back(zookeeper->asyncTryCreateNoThrow(zero_copy_path, String(), zkutil::CreateMode::Persistent)); - futures.push_back(zookeeper->asyncTryCreateNoThrow(zero_copy_path / table_shared_id_, String(), zkutil::CreateMode::Persistent)); - } - futures.push_back(zookeeper->asyncTryCreateNoThrow(zookeeper_path + "/zero_copy_s3", String(), zkutil::CreateMode::Persistent)); futures.push_back(zookeeper->asyncTryCreateNoThrow(zookeeper_path + "/zero_copy_s3/shared", String(), zkutil::CreateMode::Persistent)); futures.push_back(zookeeper->asyncTryCreateNoThrow(zookeeper_path + "/zero_copy_hdfs", String(), zkutil::CreateMode::Persistent)); @@ -1067,6 +1065,63 @@ zkutil::ZooKeeperPtr StorageReplicatedMergeTree::getZooKeeperIfTableShutDown() c return maybe_new_zookeeper; } +std::vector StorageReplicatedMergeTree::getZookeeperZeroCopyLockPaths() const +{ + const auto settings = getSettings(); + if (!settings->allow_remote_fs_zero_copy_replication) + { + return {}; + } + + const auto & disks = getStoragePolicy()->getDisks(); + std::set disk_types_with_zero_copy; + for (const auto & disk : disks) + { + if (!disk->supportZeroCopyReplication()) + continue; + + disk_types_with_zero_copy.insert(disk->getDataSourceDescription().toString()); + } + + const auto actual_table_shared_id = getTableSharedID(); + + std::vector result; + result.reserve(disk_types_with_zero_copy.size()); + + for (const auto & disk_type: disk_types_with_zero_copy) + { + auto zero_copy = fmt::format("zero_copy_{}", disk_type); + auto zero_copy_path = fs::path(settings->remote_fs_zero_copy_zookeeper_path.toString()) / zero_copy; + + result.push_back(zero_copy_path / actual_table_shared_id); + } + + return result; +} + +void StorageReplicatedMergeTree::dropZookeeperZeroCopyLockPaths(zkutil::ZooKeeperPtr zookeeper, std::vector zero_copy_locks_paths, + LoggerPtr logger) +{ + for (const auto & zero_copy_locks_root : zero_copy_locks_paths) + { + auto code = zookeeper->tryRemove(zero_copy_locks_root); + if (code == Coordination::Error::ZNOTEMPTY) + { + LOG_WARNING(logger, "Zero copy locks are not empty for {}. There are some lost locks inside." + "Removing them all.", zero_copy_locks_root); + zookeeper->tryRemoveRecursive(zero_copy_locks_root); + } + else if (code == Coordination::Error::ZNONODE) + { + LOG_INFO(logger, "Zero copy locks directory {} is absent on ZooKeeper.", zero_copy_locks_root); + } + else + { + chassert(code == Coordination::Error::ZOK); + } + } +} + void StorageReplicatedMergeTree::drop() { /// There is also the case when user has configured ClickHouse to wrong ZooKeeper cluster @@ -1098,6 +1153,11 @@ void StorageReplicatedMergeTree::drop() loadOutdatedDataParts(/*is_async=*/ false); } + + /// getZookeeperZeroCopyLockPaths has to be called before dropAllData + /// otherwise table_shared_id is unknown + auto zero_copy_locks_paths = getZookeeperZeroCopyLockPaths(); + dropAllData(); if (maybe_has_metadata_in_zookeeper) @@ -1114,12 +1174,18 @@ void StorageReplicatedMergeTree::drop() if (lost_part_count > 0) LOG_INFO(log, "Dropping table with non-zero lost_part_count equal to {}", lost_part_count); } - dropReplica(zookeeper, zookeeper_path, replica_name, log.load(), getSettings(), &has_metadata_in_zookeeper); + + bool last_replica_dropped = dropReplica(zookeeper, zookeeper_path, replica_name, log.load(), getSettings(), &has_metadata_in_zookeeper); + if (last_replica_dropped) + { + dropZookeeperZeroCopyLockPaths(zookeeper, zero_copy_locks_paths, log.load()); + } } } -void StorageReplicatedMergeTree::dropReplica(zkutil::ZooKeeperPtr zookeeper, const String & zookeeper_path, const String & replica, - LoggerPtr logger, MergeTreeSettingsPtr table_settings, std::optional * has_metadata_out) + +bool StorageReplicatedMergeTree::dropReplica(zkutil::ZooKeeperPtr zookeeper, const String & zookeeper_path, const String & replica, + LoggerPtr logger, MergeTreeSettingsPtr table_settings, std::optional * has_metadata_out) { if (zookeeper->expired()) throw Exception(ErrorCodes::TABLE_WAS_NOT_DROPPED, "Table was not dropped because ZooKeeper session has expired."); @@ -1136,7 +1202,7 @@ void StorageReplicatedMergeTree::dropReplica(zkutil::ZooKeeperPtr zookeeper, con if (!zookeeper->exists(remote_replica_path)) { LOG_INFO(logger, "Removing replica {} does not exist", remote_replica_path); - return; + return false; } { @@ -1191,7 +1257,7 @@ void StorageReplicatedMergeTree::dropReplica(zkutil::ZooKeeperPtr zookeeper, con /// Check that `zookeeper_path` exists: it could have been deleted by another replica after execution of previous line. Strings replicas; if (Coordination::Error::ZOK != zookeeper->tryGetChildren(zookeeper_path + "/replicas", replicas) || !replicas.empty()) - return; + return false; LOG_INFO(logger, "{} is the last replica, will remove table", remote_replica_path); @@ -1220,10 +1286,12 @@ void StorageReplicatedMergeTree::dropReplica(zkutil::ZooKeeperPtr zookeeper, con if (code == Coordination::Error::ZNONODE || code == Coordination::Error::ZNODEEXISTS) { LOG_WARNING(logger, "Table {} is already started to be removing by another replica right now", remote_replica_path); + return false; } else if (code == Coordination::Error::ZNOTEMPTY) { LOG_WARNING(logger, "Another replica was suddenly created, will keep the table {}", remote_replica_path); + return false; } else if (code != Coordination::Error::ZOK) { @@ -1235,9 +1303,12 @@ void StorageReplicatedMergeTree::dropReplica(zkutil::ZooKeeperPtr zookeeper, con LOG_INFO(logger, "Removing table {} (this might take several minutes)", zookeeper_path); removeTableNodesFromZooKeeper(zookeeper, zookeeper_path, metadata_drop_lock, logger); } + + return true; } -void StorageReplicatedMergeTree::dropReplica(const String & drop_zookeeper_path, const String & drop_replica, LoggerPtr logger) + +bool StorageReplicatedMergeTree::dropReplica(const String & drop_zookeeper_path, const String & drop_replica, LoggerPtr logger) { zkutil::ZooKeeperPtr zookeeper = getZooKeeperIfTableShutDown(); @@ -1247,7 +1318,7 @@ void StorageReplicatedMergeTree::dropReplica(const String & drop_zookeeper_path, if (zookeeper->exists(drop_zookeeper_path + "/replicas/" + drop_replica + "/is_active")) throw Exception(ErrorCodes::TABLE_WAS_NOT_DROPPED, "Can't drop replica: {}, because it's active", drop_replica); - dropReplica(zookeeper, drop_zookeeper_path, drop_replica, logger); + return dropReplica(zookeeper, drop_zookeeper_path, drop_replica, logger); } diff --git a/src/Storages/StorageReplicatedMergeTree.h b/src/Storages/StorageReplicatedMergeTree.h index f2679f8f7dd..d8aefdf5b4c 100644 --- a/src/Storages/StorageReplicatedMergeTree.h +++ b/src/Storages/StorageReplicatedMergeTree.h @@ -241,11 +241,12 @@ public: void restoreDataFromBackup(RestorerFromBackup & restorer, const String & data_path_in_backup, const std::optional & partitions) override; /** Remove a specific replica from zookeeper. + * returns true if there are no replicas left */ - static void dropReplica(zkutil::ZooKeeperPtr zookeeper, const String & zookeeper_path, const String & replica, + static bool dropReplica(zkutil::ZooKeeperPtr zookeeper, const String & zookeeper_path, const String & replica, LoggerPtr logger, MergeTreeSettingsPtr table_settings = nullptr, std::optional * has_metadata_out = nullptr); - void dropReplica(const String & drop_zookeeper_path, const String & drop_replica, LoggerPtr logger); + bool dropReplica(const String & drop_zookeeper_path, const String & drop_replica, LoggerPtr logger); /// Removes table from ZooKeeper after the last replica was dropped static bool removeTableNodesFromZooKeeper(zkutil::ZooKeeperPtr zookeeper, const String & zookeeper_path, @@ -983,6 +984,10 @@ private: void startupImpl(bool from_attach_thread); + std::vector getZookeeperZeroCopyLockPaths() const; + static void dropZookeeperZeroCopyLockPaths(zkutil::ZooKeeperPtr zookeeper, + std::vector zero_copy_locks_paths, LoggerPtr logger); + struct DataValidationTasks : public IStorage::DataValidationTasksBase { explicit DataValidationTasks(DataPartsVector && parts_, std::unique_lock && parts_check_lock_) diff --git a/src/Storages/StorageS3.cpp b/src/Storages/StorageS3.cpp index 6892e75f506..11da394feec 100644 --- a/src/Storages/StorageS3.cpp +++ b/src/Storages/StorageS3.cpp @@ -161,7 +161,7 @@ public: , num_streams(num_streams_) { query_configuration = storage.updateConfigurationAndGetCopy(context); - virtual_columns = storage.getVirtuals(); + virtual_columns = storage.getVirtualsList(); } private: @@ -258,45 +258,46 @@ private: KeyWithInfoPtr nextAssumeLocked() { - if (buffer_iter != buffer.end()) + do { - auto answer = *buffer_iter; - ++buffer_iter; - - /// If url doesn't contain globs, we didn't list s3 bucket and didn't get object info for the key. - /// So we get object info lazily here on 'next()' request. - if (!answer->info) + if (buffer_iter != buffer.end()) { - answer->info = S3::getObjectInfo(*client, globbed_uri.bucket, answer->key, globbed_uri.version_id, request_settings); - if (file_progress_callback) - file_progress_callback(FileProgress(0, answer->info->size)); + auto answer = *buffer_iter; + ++buffer_iter; + + /// If url doesn't contain globs, we didn't list s3 bucket and didn't get object info for the key. + /// So we get object info lazily here on 'next()' request. + if (!answer->info) + { + answer->info = S3::getObjectInfo(*client, globbed_uri.bucket, answer->key, globbed_uri.version_id, request_settings); + if (file_progress_callback) + file_progress_callback(FileProgress(0, answer->info->size)); + } + + return answer; } - return answer; - } + if (is_finished) + return {}; - if (is_finished) - return {}; - - try - { - fillInternalBufferAssumeLocked(); - } - catch (...) - { - /// In case of exception thrown while listing new batch of files - /// iterator may be partially initialized and its further using may lead to UB. - /// Iterator is used by several processors from several threads and - /// it may take some time for threads to stop processors and they - /// may still use this iterator after exception is thrown. - /// To avoid this UB, reset the buffer and return defaults for further calls. - is_finished = true; - buffer.clear(); - buffer_iter = buffer.begin(); - throw; - } - - return nextAssumeLocked(); + try + { + fillInternalBufferAssumeLocked(); + } + catch (...) + { + /// In case of exception thrown while listing new batch of files + /// iterator may be partially initialized and its further using may lead to UB. + /// Iterator is used by several processors from several threads and + /// it may take some time for threads to stop processors and they + /// may still use this iterator after exception is thrown. + /// To avoid this UB, reset the buffer and return defaults for further calls. + is_finished = true; + buffer.clear(); + buffer_iter = buffer.begin(); + throw; + } + } while (true); } void fillInternalBufferAssumeLocked() @@ -1083,8 +1084,7 @@ StorageS3::StorageS3( storage_metadata.setConstraints(constraints_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } static std::shared_ptr createFileIterator( @@ -1151,7 +1151,7 @@ void StorageS3::read( size_t max_block_size, size_t num_streams) { - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context), virtual_columns); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context)); bool need_only_count = (query_info.optimize_trivial_count || read_from_format_info.requested_columns.empty()) && local_context->getSettingsRef().optimize_count_from_files; @@ -1420,8 +1420,10 @@ void StorageS3::Configuration::connect(const ContextPtr & context) url.uri.getScheme()); client_configuration.endpointOverride = url.endpoint; + /// seems as we don't use it client_configuration.maxConnections = static_cast(request_settings.max_connections); - client_configuration.http_connection_pool_size = global_settings.s3_http_connection_pool_size; + client_configuration.connectTimeoutMs = local_settings.s3_connect_timeout_ms; + auto headers = auth_settings.headers; if (!headers_from_ast.empty()) headers.insert(headers.end(), headers_from_ast.begin(), headers_from_ast.end()); @@ -1984,16 +1986,6 @@ void registerStorageOSS(StorageFactory & factory) return registerStorageS3Impl("OSS", factory); } -NamesAndTypesList StorageS3::getVirtuals() const -{ - return virtual_columns; -} - -Names StorageS3::getVirtualColumnNames() -{ - return VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage({}).getNames(); -} - bool StorageS3::supportsPartitionBy() const { return true; diff --git a/src/Storages/StorageS3.h b/src/Storages/StorageS3.h index 65fb3b51be2..7c5db76aadf 100644 --- a/src/Storages/StorageS3.h +++ b/src/Storages/StorageS3.h @@ -336,9 +336,6 @@ public: void truncate(const ASTPtr & query, const StorageMetadataPtr & metadata_snapshot, ContextPtr local_context, TableExclusiveLockHolder &) override; - NamesAndTypesList getVirtuals() const override; - static Names getVirtualColumnNames(); - bool supportsPartitionBy() const override; static void processNamedCollectionResult(StorageS3::Configuration & configuration, const NamedCollection & collection); @@ -378,7 +375,6 @@ private: Configuration configuration; std::mutex configuration_update_mutex; - NamesAndTypesList virtual_columns; String name; const bool distributed_processing; diff --git a/src/Storages/StorageS3Cluster.cpp b/src/Storages/StorageS3Cluster.cpp index 31c241a5b13..6b22771b38f 100644 --- a/src/Storages/StorageS3Cluster.cpp +++ b/src/Storages/StorageS3Cluster.cpp @@ -67,8 +67,7 @@ StorageS3Cluster::StorageS3Cluster( storage_metadata.setConstraints(constraints_); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } void StorageS3Cluster::updateQueryToSendIfNeeded(DB::ASTPtr & query, const DB::StorageSnapshotPtr & storage_snapshot, const DB::ContextPtr & context) @@ -92,7 +91,7 @@ void StorageS3Cluster::updateConfigurationIfChanged(ContextPtr local_context) RemoteQueryExecutor::Extension StorageS3Cluster::getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const { auto iterator = std::make_shared( - *s3_configuration.client, s3_configuration.url, predicate, virtual_columns, context, nullptr, s3_configuration.request_settings, context->getFileProgressCallback()); + *s3_configuration.client, s3_configuration.url, predicate, getVirtualsList(), context, nullptr, s3_configuration.request_settings, context->getFileProgressCallback()); auto callback = std::make_shared>([iterator]() mutable -> String { @@ -103,12 +102,6 @@ RemoteQueryExecutor::Extension StorageS3Cluster::getTaskIteratorExtension(const return RemoteQueryExecutor::Extension{ .task_iterator = std::move(callback) }; } -NamesAndTypesList StorageS3Cluster::getVirtuals() const -{ - return virtual_columns; -} - - } #endif diff --git a/src/Storages/StorageS3Cluster.h b/src/Storages/StorageS3Cluster.h index 03155b6e707..6a5b03e682f 100644 --- a/src/Storages/StorageS3Cluster.h +++ b/src/Storages/StorageS3Cluster.h @@ -28,8 +28,6 @@ public: std::string getName() const override { return "S3Cluster"; } - NamesAndTypesList getVirtuals() const override; - RemoteQueryExecutor::Extension getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const override; bool supportsSubcolumns() const override { return true; } @@ -45,7 +43,6 @@ private: void updateQueryToSendIfNeeded(ASTPtr & query, const StorageSnapshotPtr & storage_snapshot, const ContextPtr & context) override; StorageS3::Configuration s3_configuration; - NamesAndTypesList virtual_columns; }; diff --git a/src/Storages/StorageSnapshot.cpp b/src/Storages/StorageSnapshot.cpp index 34c092c7208..8b087a4a2bc 100644 --- a/src/Storages/StorageSnapshot.cpp +++ b/src/Storages/StorageSnapshot.cpp @@ -1,6 +1,4 @@ #include -#include -#include #include #include #include @@ -18,6 +16,49 @@ namespace ErrorCodes extern const int COLUMN_QUERIED_MORE_THAN_ONCE; } +StorageSnapshot::StorageSnapshot( + const IStorage & storage_, + StorageMetadataPtr metadata_) + : storage(storage_) + , metadata(std::move(metadata_)) + , virtual_columns(storage_.getVirtualsPtr()) +{ +} + +StorageSnapshot::StorageSnapshot( + const IStorage & storage_, + StorageMetadataPtr metadata_, + VirtualsDescriptionPtr virtual_columns_) + : storage(storage_) + , metadata(std::move(metadata_)) + , virtual_columns(std::move(virtual_columns_)) +{ +} + +StorageSnapshot::StorageSnapshot( + const IStorage & storage_, + StorageMetadataPtr metadata_, + ColumnsDescription object_columns_) + : storage(storage_) + , metadata(std::move(metadata_)) + , virtual_columns(storage_.getVirtualsPtr()) + , object_columns(std::move(object_columns_)) +{ +} + +StorageSnapshot::StorageSnapshot( + const IStorage & storage_, + StorageMetadataPtr metadata_, + ColumnsDescription object_columns_, + DataPtr data_) + : storage(storage_) + , metadata(std::move(metadata_)) + , virtual_columns(storage_.getVirtualsPtr()) + , object_columns(std::move(object_columns_)) + , data(std::move(data_)) +{ +} + std::shared_ptr StorageSnapshot::clone(DataPtr data_) const { auto res = std::make_shared(storage, metadata, object_columns); @@ -28,15 +69,12 @@ std::shared_ptr StorageSnapshot::clone(DataPtr data_) const return res; } -void StorageSnapshot::init() +ColumnsDescription StorageSnapshot::getAllColumnsDescription() const { - for (const auto & [name, type] : storage.getVirtuals()) - virtual_columns[name] = type; + auto get_column_options = GetColumnsOptions(GetColumnsOptions::All).withExtendedObjects().withVirtuals(); + auto column_names_and_types = getColumns(get_column_options); - if (storage.hasLightweightDeletedMask()) - system_columns[LightweightDeleteDescription::FILTER_COLUMN.name] = LightweightDeleteDescription::FILTER_COLUMN.type; - - system_columns[BlockNumberColumn::name] = BlockNumberColumn::type; + return ColumnsDescription{column_names_and_types}; } NamesAndTypesList StorageSnapshot::getColumns(const GetColumnsOptions & options) const @@ -46,35 +84,22 @@ NamesAndTypesList StorageSnapshot::getColumns(const GetColumnsOptions & options) if (options.with_extended_objects) extendObjectColumns(all_columns, object_columns, options.with_subcolumns); - NameSet column_names; - if (options.with_virtuals) + if (options.virtuals_kind != VirtualsKind::None && !virtual_columns->empty()) { - /// Virtual columns must be appended after ordinary, - /// because user can override them. - if (!virtual_columns.empty()) + NameSet column_names; + for (const auto & column : all_columns) + column_names.insert(column.name); + + auto virtuals_list = virtual_columns->getNamesAndTypesList(options.virtuals_kind); + for (const auto & column : virtuals_list) { - for (const auto & column : all_columns) - column_names.insert(column.name); + if (column_names.contains(column.name)) + continue; - for (const auto & [name, type] : virtual_columns) - if (!column_names.contains(name)) - all_columns.emplace_back(name, type); + all_columns.emplace_back(column.name, column.type); } } - if (options.with_system_columns) - { - if (!system_columns.empty() && column_names.empty()) - { - for (const auto & column : all_columns) - column_names.insert(column.name); - } - - for (const auto & [name, type] : system_columns) - if (!column_names.contains(name)) - all_columns.emplace_back(name, type); - } - return all_columns; } @@ -100,18 +125,11 @@ std::optional StorageSnapshot::tryGetColumn(const GetColumnsOpt return object_column; } - if (options.with_virtuals) + if (options.virtuals_kind != VirtualsKind::None) { - auto it = virtual_columns.find(column_name); - if (it != virtual_columns.end()) - return NameAndTypePair(column_name, it->second); - } - - if (options.with_system_columns) - { - auto it = system_columns.find(column_name); - if (it != system_columns.end()) - return NameAndTypePair(column_name, it->second); + auto virtual_column = virtual_columns->tryGet(column_name, options.virtuals_kind); + if (virtual_column) + return NameAndTypePair{virtual_column->name, virtual_column->type}; } return {}; @@ -126,6 +144,47 @@ NameAndTypePair StorageSnapshot::getColumn(const GetColumnsOptions & options, co return *column; } +CompressionCodecPtr StorageSnapshot::getCodecOrDefault(const String & column_name, CompressionCodecPtr default_codec) const +{ + auto get_codec_or_default = [&](const auto & column_desc) + { + return column_desc.codec + ? CompressionCodecFactory::instance().get(column_desc.codec, column_desc.type, default_codec) + : default_codec; + }; + + const auto & columns = metadata->getColumns(); + if (const auto * column_desc = columns.tryGet(column_name)) + return get_codec_or_default(*column_desc); + + if (const auto * virtual_desc = virtual_columns->tryGetDescription(column_name)) + return get_codec_or_default(*virtual_desc); + + return default_codec; +} + +CompressionCodecPtr StorageSnapshot::getCodecOrDefault(const String & column_name) const +{ + return getCodecOrDefault(column_name, CompressionCodecFactory::instance().getDefaultCodec()); +} + +ASTPtr StorageSnapshot::getCodecDescOrDefault(const String & column_name, CompressionCodecPtr default_codec) const +{ + auto get_codec_or_default = [&](const auto & column_desc) + { + return column_desc.codec ? column_desc.codec : default_codec->getFullCodecDesc(); + }; + + const auto & columns = metadata->getColumns(); + if (const auto * column_desc = columns.tryGet(column_name)) + return get_codec_or_default(*column_desc); + + if (const auto * virtual_desc = virtual_columns->tryGetDescription(column_name)) + return get_codec_or_default(*virtual_desc); + + return default_codec->getFullCodecDesc(); +} + Block StorageSnapshot::getSampleBlockForColumns(const Names & column_names) const { Block res; @@ -143,11 +202,11 @@ Block StorageSnapshot::getSampleBlockForColumns(const Names & column_names) cons { res.insert({object_column->type->createColumn(), object_column->type, column_name}); } - else if (auto it = virtual_columns.find(column_name); it != virtual_columns.end()) + else if (auto virtual_column = virtual_columns->tryGet(column_name)) { /// Virtual columns must be appended after ordinary, because user can /// override them. - const auto & type = it->second; + const auto & type = virtual_column->type; res.insert({type->createColumn(), type, column_name}); } else @@ -175,12 +234,11 @@ ColumnsDescription StorageSnapshot::getDescriptionForColumns(const Names & colum { res.add(*object_column, "", false, false); } - else if (auto it = virtual_columns.find(name); it != virtual_columns.end()) + else if (auto virtual_column = virtual_columns->tryGet(name)) { /// Virtual columns must be appended after ordinary, because user can /// override them. - const auto & type = it->second; - res.add({name, type}); + res.add({name, virtual_column->type}); } else { @@ -216,7 +274,7 @@ void StorageSnapshot::check(const Names & column_names) const { bool has_column = columns.hasColumnOrSubcolumn(GetColumnsOptions::AllPhysical, name) || object_columns.hasColumnOrSubcolumn(GetColumnsOptions::AllPhysical, name) - || virtual_columns.contains(name); + || virtual_columns->has(name); if (!has_column) { diff --git a/src/Storages/StorageSnapshot.h b/src/Storages/StorageSnapshot.h index d62e118e1f2..89e97f2abb8 100644 --- a/src/Storages/StorageSnapshot.h +++ b/src/Storages/StorageSnapshot.h @@ -1,10 +1,14 @@ #pragma once #include +#include namespace DB { class IStorage; +class ICompressionCodec; + +using CompressionCodecPtr = std::shared_ptr; /// Snapshot of storage that fixes set columns that can be read in query. /// There are 3 sources of columns: regular columns from metadata, @@ -13,6 +17,7 @@ struct StorageSnapshot { const IStorage & storage; const StorageMetadataPtr metadata; + const VirtualsDescriptionPtr virtual_columns; const ColumnsDescription object_columns; /// Additional data, on which set of columns may depend. @@ -30,38 +35,29 @@ struct StorageSnapshot StorageSnapshot( const IStorage & storage_, - StorageMetadataPtr metadata_) - : storage(storage_), metadata(std::move(metadata_)) - { - init(); - } + StorageMetadataPtr metadata_); StorageSnapshot( const IStorage & storage_, StorageMetadataPtr metadata_, - ColumnsDescription object_columns_) - : storage(storage_) - , metadata(std::move(metadata_)) - , object_columns(std::move(object_columns_)) - { - init(); - } + VirtualsDescriptionPtr virtual_columns_); + + StorageSnapshot( + const IStorage & storage_, + StorageMetadataPtr metadata_, + ColumnsDescription object_columns_); StorageSnapshot( const IStorage & storage_, StorageMetadataPtr metadata_, ColumnsDescription object_columns_, - DataPtr data_) - : storage(storage_) - , metadata(std::move(metadata_)) - , object_columns(std::move(object_columns_)) - , data(std::move(data_)) - { - init(); - } + DataPtr data_); std::shared_ptr clone(DataPtr data_) const; + /// Get columns description + ColumnsDescription getAllColumnsDescription() const; + /// Get all available columns with types according to options. NamesAndTypesList getColumns(const GetColumnsOptions & options) const; @@ -72,6 +68,10 @@ struct StorageSnapshot std::optional tryGetColumn(const GetColumnsOptions & options, const String & column_name) const; NameAndTypePair getColumn(const GetColumnsOptions & options, const String & column_name) const; + CompressionCodecPtr getCodecOrDefault(const String & column_name, CompressionCodecPtr default_codec) const; + CompressionCodecPtr getCodecOrDefault(const String & column_name) const; + ASTPtr getCodecDescOrDefault(const String & column_name, CompressionCodecPtr default_codec) const; + /// Block with ordinary + materialized + aliases + virtuals + subcolumns. Block getSampleBlockForColumns(const Names & column_names) const; @@ -87,15 +87,6 @@ struct StorageSnapshot /// If we have a projection then we should use its metadata. StorageMetadataPtr getMetadataForQuery() const { return projection ? projection->metadata : metadata; } - -private: - void init(); - - std::unordered_map virtual_columns; - - /// System columns are not visible in the schema but might be persisted in the data. - /// One example of such column is lightweight delete mask '_row_exists'. - std::unordered_map system_columns; }; using StorageSnapshotPtr = std::shared_ptr; diff --git a/src/Storages/StorageURL.cpp b/src/Storages/StorageURL.cpp index 40acee57e7c..adcb83218fd 100644 --- a/src/Storages/StorageURL.cpp +++ b/src/Storages/StorageURL.cpp @@ -152,8 +152,7 @@ IStorageURLBase::IStorageURLBase( storage_metadata.setConstraints(constraints_); storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } @@ -461,22 +460,20 @@ std::pair> StorageURLSource: try { - auto res = std::make_unique( - request_uri, - http_method, - callback, - timeouts, - credentials, - settings.max_http_get_redirects, - settings.max_read_buffer_size, - read_settings, - headers, - &context_->getRemoteHostFilter(), - delay_initialization, - /* use_external_buffer */ false, - /* skip_url_not_found_error */ skip_url_not_found_error, - /* file_info */ std::nullopt, - proxy_config); + auto res = BuilderRWBufferFromHTTP(request_uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(http_method) + .withProxy(proxy_config) + .withSettings(read_settings) + .withTimeouts(timeouts) + .withHostFilter(&context_->getRemoteHostFilter()) + .withBufSize(settings.max_read_buffer_size) + .withRedirects(settings.max_http_get_redirects) + .withOutCallback(callback) + .withSkipNotFound(skip_url_not_found_error) + .withHeaders(headers) + .withDelayInit(delay_initialization) + .create(credentials); if (context_->getSettingsRef().engine_url_skip_empty_files && res->eof() && option != std::prev(end)) { @@ -548,7 +545,7 @@ StorageURLSink::StorageURLSink( auto proxy_config = getProxyConfiguration(http_method); auto write_buffer = std::make_unique( - Poco::URI(uri), http_method, content_type, content_encoding, headers, timeouts, DBMS_DEFAULT_BUFFER_SIZE, proxy_config + HTTPConnectionGroupType::STORAGE, Poco::URI(uri), http_method, content_type, content_encoding, headers, timeouts, DBMS_DEFAULT_BUFFER_SIZE, proxy_config ); const auto & settings = context->getSettingsRef(); @@ -1059,7 +1056,7 @@ void IStorageURLBase::read( size_t num_streams) { auto params = getReadURIParams(column_names, storage_snapshot, query_info, local_context, processed_stage, max_block_size); - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context), getVirtuals()); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context)); bool need_only_count = (query_info.optimize_trivial_count || read_from_format_info.requested_columns.empty()) && local_context->getSettingsRef().optimize_count_from_files; @@ -1127,7 +1124,7 @@ void ReadFromURL::createIterator(const ActionsDAG::Node * predicate) else if (is_url_with_globs) { /// Iterate through disclosed globs and make a source for each file - auto glob_iterator = std::make_shared(storage->uri, max_addresses, predicate, storage->virtual_columns, context); + auto glob_iterator = std::make_shared(storage->uri, max_addresses, predicate, storage->getVirtualsList(), context); /// check if we filtered out all the paths if (glob_iterator->size() == 0) @@ -1230,7 +1227,7 @@ void StorageURLWithFailover::read( size_t num_streams) { auto params = getReadURIParams(column_names, storage_snapshot, query_info, local_context, processed_stage, max_block_size); - auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context), getVirtuals()); + auto read_from_format_info = prepareReadingFromFormat(column_names, storage_snapshot, supportsSubsetOfColumns(local_context)); bool need_only_count = (query_info.optimize_trivial_count || read_from_format_info.requested_columns.empty()) && local_context->getSettingsRef().optimize_count_from_files; @@ -1303,16 +1300,6 @@ SinkToStoragePtr IStorageURLBase::write(const ASTPtr & query, const StorageMetad } } -NamesAndTypesList IStorageURLBase::getVirtuals() const -{ - return virtual_columns; -} - -Names IStorageURLBase::getVirtualColumnNames() -{ - return VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage({}).getNames(); -} - SchemaCache & IStorageURLBase::getSchemaCache(const ContextPtr & context) { static SchemaCache schema_cache(context->getConfigRef().getUInt("schema_inference_cache_max_elements_for_url", DEFAULT_SCHEMA_CACHE_ELEMENTS)); @@ -1331,24 +1318,17 @@ std::optional IStorageURLBase::tryGetLastModificationTime( auto proxy_config = getProxyConfiguration(uri.getScheme()); - ReadWriteBufferFromHTTP buf( - uri, - Poco::Net::HTTPRequest::HTTP_GET, - {}, - getHTTPTimeouts(context), - credentials, - settings.max_http_get_redirects, - settings.max_read_buffer_size, - context->getReadSettings(), - headers, - &context->getRemoteHostFilter(), - true, - false, - false, - std::nullopt, - proxy_config); + auto buf = BuilderRWBufferFromHTTP(uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withSettings(context->getReadSettings()) + .withTimeouts(getHTTPTimeouts(context)) + .withHostFilter(&context->getRemoteHostFilter()) + .withBufSize(settings.max_read_buffer_size) + .withRedirects(settings.max_http_get_redirects) + .withHeaders(headers) + .create(credentials); - return buf.tryGetLastModificationTime(); + return buf->tryGetLastModificationTime(); } StorageURL::StorageURL( diff --git a/src/Storages/StorageURL.h b/src/Storages/StorageURL.h index 480670475dd..294b1f828bb 100644 --- a/src/Storages/StorageURL.h +++ b/src/Storages/StorageURL.h @@ -48,9 +48,6 @@ public: bool supportsPartitionBy() const override { return true; } - NamesAndTypesList getVirtuals() const override; - static Names getVirtualColumnNames(); - static ColumnsDescription getTableStructureFromData( const String & format, const String & uri, @@ -106,8 +103,6 @@ protected: ASTPtr partition_by; bool distributed_processing; - NamesAndTypesList virtual_columns; - virtual std::string getReadMethod() const; virtual std::vector> getReadURIParams( diff --git a/src/Storages/StorageURLCluster.cpp b/src/Storages/StorageURLCluster.cpp index d0df74d7521..2e7c63d0097 100644 --- a/src/Storages/StorageURLCluster.cpp +++ b/src/Storages/StorageURLCluster.cpp @@ -75,8 +75,7 @@ StorageURLCluster::StorageURLCluster( storage_metadata.setConstraints(constraints_); setInMemoryMetadata(storage_metadata); - - virtual_columns = VirtualColumnUtils::getPathFileAndSizeVirtualsForStorage(storage_metadata.getSampleBlock().getNamesAndTypesList()); + setVirtuals(VirtualColumnUtils::getVirtualsForFileLikeStorage(storage_metadata.getColumns())); } void StorageURLCluster::updateQueryToSendIfNeeded(ASTPtr & query, const StorageSnapshotPtr & storage_snapshot, const ContextPtr & context) @@ -91,7 +90,7 @@ void StorageURLCluster::updateQueryToSendIfNeeded(ASTPtr & query, const StorageS RemoteQueryExecutor::Extension StorageURLCluster::getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const { - auto iterator = std::make_shared(uri, context->getSettingsRef().glob_expansion_max_elements, predicate, virtual_columns, context); + auto iterator = std::make_shared(uri, context->getSettingsRef().glob_expansion_max_elements, predicate, getVirtualsList(), context); auto callback = std::make_shared([iter = std::move(iterator)]() mutable -> String { return iter->next(); }); return RemoteQueryExecutor::Extension{.task_iterator = std::move(callback)}; } diff --git a/src/Storages/StorageURLCluster.h b/src/Storages/StorageURLCluster.h index f57d262f434..dce2e0106ea 100644 --- a/src/Storages/StorageURLCluster.h +++ b/src/Storages/StorageURLCluster.h @@ -31,8 +31,6 @@ public: std::string getName() const override { return "URLCluster"; } - NamesAndTypesList getVirtuals() const override { return virtual_columns; } - RemoteQueryExecutor::Extension getTaskIteratorExtension(const ActionsDAG::Node * predicate, const ContextPtr & context) const override; bool supportsSubcolumns() const override { return true; } @@ -44,7 +42,6 @@ private: String uri; String format_name; - NamesAndTypesList virtual_columns; }; diff --git a/src/Storages/StorageValues.cpp b/src/Storages/StorageValues.cpp index 300b11b7346..a930ffd1307 100644 --- a/src/Storages/StorageValues.cpp +++ b/src/Storages/StorageValues.cpp @@ -12,12 +12,13 @@ StorageValues::StorageValues( const StorageID & table_id_, const ColumnsDescription & columns_, const Block & res_block_, - const NamesAndTypesList & virtuals_) - : IStorage(table_id_), res_block(res_block_), virtuals(virtuals_) + VirtualColumnsDescription virtuals_) + : IStorage(table_id_), res_block(res_block_) { StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(columns_); setInMemoryMetadata(storage_metadata); + setVirtuals(std::move(virtuals_)); } Pipe StorageValues::read( diff --git a/src/Storages/StorageValues.h b/src/Storages/StorageValues.h index af1c134077b..ebb182ab667 100644 --- a/src/Storages/StorageValues.h +++ b/src/Storages/StorageValues.h @@ -11,8 +11,14 @@ namespace DB class StorageValues final : public IStorage { public: + /// Why we may have virtual columns in the storage from a single block? + /// Because it used as tmp storage for pushing blocks into views, and some + /// views may contain virtual columns from original storage. StorageValues( - const StorageID & table_id_, const ColumnsDescription & columns_, const Block & res_block_, const NamesAndTypesList & virtuals_ = {}); + const StorageID & table_id_, + const ColumnsDescription & columns_, + const Block & res_block_, + VirtualColumnsDescription virtuals_ = {}); std::string getName() const override { return "Values"; } @@ -25,13 +31,6 @@ public: size_t max_block_size, size_t num_streams) override; - /// Why we may have virtual columns in the storage from a single block? - /// Because it used as tmp storage for pushing blocks into views, and some - /// views may contain virtual columns from original storage. - NamesAndTypesList getVirtuals() const override - { - return virtuals; - } /// FIXME probably it should return false, but StorageValues is used in ExecutingInnerQueryFromViewTransform (whatever it is) bool supportsTransactions() const override { return true; } @@ -40,7 +39,6 @@ public: private: Block res_block; - NamesAndTypesList virtuals; }; } diff --git a/src/Storages/System/IStorageSystemOneBlock.cpp b/src/Storages/System/IStorageSystemOneBlock.cpp new file mode 100644 index 00000000000..53399654c8d --- /dev/null +++ b/src/Storages/System/IStorageSystemOneBlock.cpp @@ -0,0 +1,99 @@ +#include +// #include +// #include +// #include +// #include +#include +#include +#include +#include +#include +#include +#include + +namespace DB +{ + +class ReadFromSystemOneBlock : public SourceStepWithFilter +{ +public: + std::string getName() const override { return "ReadFromSystemOneBlock"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; + + ReadFromSystemOneBlock( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_, + std::vector columns_mask_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , storage(std::move(storage_)) + , columns_mask(std::move(columns_mask_)) + { + } + + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +private: + std::shared_ptr storage; + std::vector columns_mask; + const ActionsDAG::Node * predicate = nullptr; +}; + +void IStorageSystemOneBlock::read( + QueryPlan & query_plan, + const Names & column_names, + const StorageSnapshotPtr & storage_snapshot, + SelectQueryInfo & query_info, + ContextPtr context, + QueryProcessingStage::Enum /*processed_stage*/, + size_t /*max_block_size*/, + size_t /*num_streams*/) +{ + storage_snapshot->check(column_names); + Block sample_block = storage_snapshot->metadata->getSampleBlockWithVirtuals(getVirtualsList()); + std::vector columns_mask; + + if (supportsColumnsMask()) + { + auto [columns_mask_, header] = getQueriedColumnsMaskAndHeader(sample_block, column_names); + columns_mask = std::move(columns_mask_); + sample_block = std::move(header); + } + + auto this_ptr = std::static_pointer_cast(shared_from_this()); + + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(sample_block), std::move(this_ptr), std::move(columns_mask)); + + query_plan.addStep(std::move(reading)); +} + +void ReadFromSystemOneBlock::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) +{ + const auto & sample_block = getOutputStream().header; + MutableColumns res_columns = sample_block.cloneEmptyColumns(); + storage->fillData(res_columns, context, predicate, std::move(columns_mask)); + + UInt64 num_rows = res_columns.at(0)->size(); + Chunk chunk(std::move(res_columns), num_rows); + + pipeline.init(Pipe(std::make_shared(sample_block, std::move(chunk)))); +} + +void ReadFromSystemOneBlock::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +} diff --git a/src/Storages/System/IStorageSystemOneBlock.h b/src/Storages/System/IStorageSystemOneBlock.h index 3b2807965a4..a20434fd97e 100644 --- a/src/Storages/System/IStorageSystemOneBlock.h +++ b/src/Storages/System/IStorageSystemOneBlock.h @@ -1,13 +1,6 @@ #pragma once -#include -#include -#include #include -#include -#include -#include -#include namespace DB { @@ -17,8 +10,8 @@ class Context; /** IStorageSystemOneBlock is base class for system tables whose all columns can be synchronously fetched. * - * Client class need to provide static method static NamesAndTypesList getNamesAndTypes() that will return list of column names and - * their types. IStorageSystemOneBlock during read will create result columns in same order as result of getNamesAndTypes + * Client class need to provide columns_description. + * IStorageSystemOneBlock during read will create result columns in same order as in columns_description * and pass it with fillData method. * * Client also must override fillData and fill result columns. @@ -26,49 +19,32 @@ class Context; * If subclass want to support virtual columns, it should override getVirtuals method of IStorage interface. * IStorageSystemOneBlock will add virtuals columns at the end of result columns of fillData method. */ -template class IStorageSystemOneBlock : public IStorage { protected: - virtual void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const = 0; + virtual void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector columns_mask) const = 0; virtual bool supportsColumnsMask() const { return false; } + friend class ReadFromSystemOneBlock; + public: - explicit IStorageSystemOneBlock(const StorageID & table_id_) : IStorage(table_id_) + explicit IStorageSystemOneBlock(const StorageID & table_id_, ColumnsDescription columns_description) : IStorage(table_id_) { StorageInMemoryMetadata storage_metadata; - storage_metadata.setColumns(Self::getColumnsDescription()); + storage_metadata.setColumns(std::move(columns_description)); setInMemoryMetadata(storage_metadata); } - Pipe read( + void read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, ContextPtr context, QueryProcessingStage::Enum /*processed_stage*/, size_t /*max_block_size*/, - size_t /*num_streams*/) override - { - storage_snapshot->check(column_names); - Block sample_block = storage_snapshot->metadata->getSampleBlockWithVirtuals(getVirtuals()); - - if (supportsColumnsMask()) - { - auto [columns_mask, header] = getQueriedColumnsMaskAndHeader(sample_block, column_names); - query_info.columns_mask = std::move(columns_mask); - sample_block = std::move(header); - } - - MutableColumns res_columns = sample_block.cloneEmptyColumns(); - fillData(res_columns, context, query_info); - - UInt64 num_rows = res_columns.at(0)->size(); - Chunk chunk(std::move(res_columns), num_rows); - - return Pipe(std::make_shared(sample_block, std::move(chunk))); - } + size_t /*num_streams*/) override; bool isSystemStorage() const override { return true; } diff --git a/src/Storages/System/StorageSystemAggregateFunctionCombinators.cpp b/src/Storages/System/StorageSystemAggregateFunctionCombinators.cpp index 8e32a137fcb..7f3994528a9 100644 --- a/src/Storages/System/StorageSystemAggregateFunctionCombinators.cpp +++ b/src/Storages/System/StorageSystemAggregateFunctionCombinators.cpp @@ -13,7 +13,7 @@ ColumnsDescription StorageSystemAggregateFunctionCombinators::getColumnsDescript }; } -void StorageSystemAggregateFunctionCombinators::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemAggregateFunctionCombinators::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { const auto & combinators = AggregateFunctionCombinatorFactory::instance().getAllAggregateFunctionCombinators(); for (const auto & pair : combinators) diff --git a/src/Storages/System/StorageSystemAggregateFunctionCombinators.h b/src/Storages/System/StorageSystemAggregateFunctionCombinators.h index 6f3f90b58af..45036043636 100644 --- a/src/Storages/System/StorageSystemAggregateFunctionCombinators.h +++ b/src/Storages/System/StorageSystemAggregateFunctionCombinators.h @@ -6,10 +6,10 @@ namespace DB { -class StorageSystemAggregateFunctionCombinators final : public IStorageSystemOneBlock +class StorageSystemAggregateFunctionCombinators final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; public: diff --git a/src/Storages/System/StorageSystemAsyncLoader.cpp b/src/Storages/System/StorageSystemAsyncLoader.cpp index c56a3c3ce78..a7ffa282429 100644 --- a/src/Storages/System/StorageSystemAsyncLoader.cpp +++ b/src/Storages/System/StorageSystemAsyncLoader.cpp @@ -74,7 +74,7 @@ ColumnsDescription StorageSystemAsyncLoader::getColumnsDescription() }; } -void StorageSystemAsyncLoader::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemAsyncLoader::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { TimePoint now = std::chrono::system_clock::now(); diff --git a/src/Storages/System/StorageSystemAsyncLoader.h b/src/Storages/System/StorageSystemAsyncLoader.h index fa0ce11efe3..685db264a10 100644 --- a/src/Storages/System/StorageSystemAsyncLoader.h +++ b/src/Storages/System/StorageSystemAsyncLoader.h @@ -10,7 +10,7 @@ namespace DB class Context; /// system.asynchronous_loader table. Takes data from context.getAsyncLoader() -class StorageSystemAsyncLoader final : public IStorageSystemOneBlock +class StorageSystemAsyncLoader final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemAsyncLoader"; } @@ -20,7 +20,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemAsynchronousInserts.cpp b/src/Storages/System/StorageSystemAsynchronousInserts.cpp index 20ba4d1cdfb..b5f413f2e20 100644 --- a/src/Storages/System/StorageSystemAsynchronousInserts.cpp +++ b/src/Storages/System/StorageSystemAsynchronousInserts.cpp @@ -30,7 +30,7 @@ ColumnsDescription StorageSystemAsynchronousInserts::getColumnsDescription() }; } -void StorageSystemAsynchronousInserts::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemAsynchronousInserts::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { using namespace std::chrono; diff --git a/src/Storages/System/StorageSystemAsynchronousInserts.h b/src/Storages/System/StorageSystemAsynchronousInserts.h index 891494ffbeb..748937cffb8 100644 --- a/src/Storages/System/StorageSystemAsynchronousInserts.h +++ b/src/Storages/System/StorageSystemAsynchronousInserts.h @@ -8,7 +8,7 @@ namespace DB /** Implements the system table `asynhronous_inserts`, * which contains information about pending asynchronous inserts in queue. */ -class StorageSystemAsynchronousInserts final : public IStorageSystemOneBlock +class StorageSystemAsynchronousInserts final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemAsynchronousInserts"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemAsynchronousMetrics.cpp b/src/Storages/System/StorageSystemAsynchronousMetrics.cpp index 58940a7e52e..a54b577d72b 100644 --- a/src/Storages/System/StorageSystemAsynchronousMetrics.cpp +++ b/src/Storages/System/StorageSystemAsynchronousMetrics.cpp @@ -19,11 +19,11 @@ ColumnsDescription StorageSystemAsynchronousMetrics::getColumnsDescription() StorageSystemAsynchronousMetrics::StorageSystemAsynchronousMetrics(const StorageID & table_id_, const AsynchronousMetrics & async_metrics_) - : IStorageSystemOneBlock(table_id_), async_metrics(async_metrics_) + : IStorageSystemOneBlock(table_id_, getColumnsDescription()), async_metrics(async_metrics_) { } -void StorageSystemAsynchronousMetrics::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemAsynchronousMetrics::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { auto async_metrics_values = async_metrics.getValues(); for (const auto & name_value : async_metrics_values) diff --git a/src/Storages/System/StorageSystemAsynchronousMetrics.h b/src/Storages/System/StorageSystemAsynchronousMetrics.h index 026377c77a0..3543b7684d7 100644 --- a/src/Storages/System/StorageSystemAsynchronousMetrics.h +++ b/src/Storages/System/StorageSystemAsynchronousMetrics.h @@ -11,7 +11,7 @@ class Context; /** Implements system table asynchronous_metrics, which allows to get values of periodically (asynchronously) updated metrics. */ -class StorageSystemAsynchronousMetrics final : public IStorageSystemOneBlock +class StorageSystemAsynchronousMetrics final : public IStorageSystemOneBlock { public: StorageSystemAsynchronousMetrics(const StorageID & table_id_, const AsynchronousMetrics & async_metrics_); @@ -24,7 +24,7 @@ private: const AsynchronousMetrics & async_metrics; protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemBackups.cpp b/src/Storages/System/StorageSystemBackups.cpp index 0063d9e308f..fec92229556 100644 --- a/src/Storages/System/StorageSystemBackups.cpp +++ b/src/Storages/System/StorageSystemBackups.cpp @@ -39,7 +39,7 @@ ColumnsDescription StorageSystemBackups::getColumnsDescription() } -void StorageSystemBackups::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemBackups::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { size_t column_index = 0; auto & column_id = assert_cast(*res_columns[column_index++]); diff --git a/src/Storages/System/StorageSystemBackups.h b/src/Storages/System/StorageSystemBackups.h index a081bd52b4c..48e03d12e18 100644 --- a/src/Storages/System/StorageSystemBackups.h +++ b/src/Storages/System/StorageSystemBackups.h @@ -7,7 +7,7 @@ namespace DB { /// Implements `grants` system table, which allows you to get information about grants. -class StorageSystemBackups final : public IStorageSystemOneBlock +class StorageSystemBackups final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemBackups"; } @@ -15,7 +15,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemBuildOptions.cpp b/src/Storages/System/StorageSystemBuildOptions.cpp index c12935078af..a7144299eca 100644 --- a/src/Storages/System/StorageSystemBuildOptions.cpp +++ b/src/Storages/System/StorageSystemBuildOptions.cpp @@ -17,7 +17,7 @@ ColumnsDescription StorageSystemBuildOptions::getColumnsDescription() }; } -void StorageSystemBuildOptions::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemBuildOptions::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (auto * it = auto_config_build; *it; it += 2) { diff --git a/src/Storages/System/StorageSystemBuildOptions.h b/src/Storages/System/StorageSystemBuildOptions.h index 7c0bbf6b5fd..dbe651d7513 100644 --- a/src/Storages/System/StorageSystemBuildOptions.h +++ b/src/Storages/System/StorageSystemBuildOptions.h @@ -11,10 +11,10 @@ class Context; /** System table "build_options" with many params used for clickhouse building */ -class StorageSystemBuildOptions final : public IStorageSystemOneBlock +class StorageSystemBuildOptions final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemCertificates.cpp b/src/Storages/System/StorageSystemCertificates.cpp index c372e4b32b0..5ec683e1784 100644 --- a/src/Storages/System/StorageSystemCertificates.cpp +++ b/src/Storages/System/StorageSystemCertificates.cpp @@ -19,19 +19,18 @@ namespace DB ColumnsDescription StorageSystemCertificates::getColumnsDescription() { - /// TODO: Fill in all the comments. return ColumnsDescription { - {"version", std::make_shared>()}, - {"serial_number", std::make_shared(std::make_shared())}, - {"signature_algo", std::make_shared(std::make_shared())}, - {"issuer", std::make_shared(std::make_shared())}, - {"not_before", std::make_shared(std::make_shared())}, - {"not_after", std::make_shared(std::make_shared())}, - {"subject", std::make_shared(std::make_shared())}, - {"pkey_algo", std::make_shared(std::make_shared())}, - {"path", std::make_shared()}, - {"default", std::make_shared>()} + {"version", std::make_shared>(), "Version of the certificate. Values are 0 for v1, 1 for v2, 2 for v3."}, + {"serial_number", std::make_shared(std::make_shared()), "Serial Number of the certificate assigned by the issuer."}, + {"signature_algo", std::make_shared(std::make_shared()), "Signature Algorithm - an algorithm used by the issuer to sign this certificate."}, + {"issuer", std::make_shared(std::make_shared()), "Issuer - an unique identifier for the Certificate Authority issuing this certificate."}, + {"not_before", std::make_shared(std::make_shared()), "The beginning of the time window when this certificate is valid."}, + {"not_after", std::make_shared(std::make_shared()), "The end of the time window when this certificate is valid."}, + {"subject", std::make_shared(std::make_shared()), "Subject - identifies the owner of the public key."}, + {"pkey_algo", std::make_shared(std::make_shared()), "Public Key Algorithm defines the algorithm the public key can be used with."}, + {"path", std::make_shared(), "Path to the file or directory containing this certificate."}, + {"default", std::make_shared>(), "Certificate is in the default certificate location."} }; } @@ -169,7 +168,7 @@ static void enumCertificates(const std::string & dir, bool def, MutableColumns & #endif -void StorageSystemCertificates::fillData([[maybe_unused]] MutableColumns & res_columns, ContextPtr/* context*/, const SelectQueryInfo &) const +void StorageSystemCertificates::fillData([[maybe_unused]] MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { #if USE_SSL const auto & ca_paths = Poco::Net::SSLManager::instance().defaultServerContext()->getCAPaths(); diff --git a/src/Storages/System/StorageSystemCertificates.h b/src/Storages/System/StorageSystemCertificates.h index f8c8477c998..bc3fef83853 100644 --- a/src/Storages/System/StorageSystemCertificates.h +++ b/src/Storages/System/StorageSystemCertificates.h @@ -13,7 +13,7 @@ class Cluster; * that allows to obtain information about available certificates * and their sources. */ -class StorageSystemCertificates final : public IStorageSystemOneBlock +class StorageSystemCertificates final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemCertificates"; } @@ -23,7 +23,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemClusters.cpp b/src/Storages/System/StorageSystemClusters.cpp index 6172d4329cd..3c01b4717cc 100644 --- a/src/Storages/System/StorageSystemClusters.cpp +++ b/src/Storages/System/StorageSystemClusters.cpp @@ -39,7 +39,7 @@ ColumnsDescription StorageSystemClusters::getColumnsDescription() return description; } -void StorageSystemClusters::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemClusters::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { for (const auto & name_and_cluster : context->getClusters()) writeCluster(res_columns, name_and_cluster, {}); diff --git a/src/Storages/System/StorageSystemClusters.h b/src/Storages/System/StorageSystemClusters.h index 7b568641cb2..0f7c792261d 100644 --- a/src/Storages/System/StorageSystemClusters.h +++ b/src/Storages/System/StorageSystemClusters.h @@ -15,7 +15,7 @@ class Cluster; * that allows to obtain information about available clusters * (which may be specified in Distributed tables). */ -class StorageSystemClusters final : public IStorageSystemOneBlock +class StorageSystemClusters final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemClusters"; } @@ -26,7 +26,7 @@ protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; using NameAndCluster = std::pair>; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; static void writeCluster(MutableColumns & res_columns, const NameAndCluster & name_and_cluster, const std::vector & is_active); }; diff --git a/src/Storages/System/StorageSystemCollations.cpp b/src/Storages/System/StorageSystemCollations.cpp index 2e5d11943af..5ef075e53be 100644 --- a/src/Storages/System/StorageSystemCollations.cpp +++ b/src/Storages/System/StorageSystemCollations.cpp @@ -1,6 +1,8 @@ #include #include +#include #include +#include namespace DB { @@ -14,7 +16,7 @@ ColumnsDescription StorageSystemCollations::getColumnsDescription() }; } -void StorageSystemCollations::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemCollations::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (const auto & [locale, lang]: AvailableCollationLocales::instance().getAvailableCollations()) { diff --git a/src/Storages/System/StorageSystemCollations.h b/src/Storages/System/StorageSystemCollations.h index 1fc0ff0e024..449e0d8ac06 100644 --- a/src/Storages/System/StorageSystemCollations.h +++ b/src/Storages/System/StorageSystemCollations.h @@ -5,10 +5,10 @@ namespace DB { -class StorageSystemCollations final : public IStorageSystemOneBlock +class StorageSystemCollations final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemColumns.cpp b/src/Storages/System/StorageSystemColumns.cpp index ad65f6f5476..5c96c6502af 100644 --- a/src/Storages/System/StorageSystemColumns.cpp +++ b/src/Storages/System/StorageSystemColumns.cpp @@ -16,6 +16,9 @@ #include #include #include +#include +#include +#include namespace DB @@ -28,27 +31,32 @@ StorageSystemColumns::StorageSystemColumns(const StorageID & table_id_) StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription( { - { "database", std::make_shared() }, - { "table", std::make_shared() }, - { "name", std::make_shared() }, - { "type", std::make_shared() }, - { "position", std::make_shared() }, - { "default_kind", std::make_shared() }, - { "default_expression", std::make_shared() }, - { "data_compressed_bytes", std::make_shared() }, - { "data_uncompressed_bytes", std::make_shared() }, - { "marks_bytes", std::make_shared() }, - { "comment", std::make_shared() }, - { "is_in_partition_key", std::make_shared() }, - { "is_in_sorting_key", std::make_shared() }, - { "is_in_primary_key", std::make_shared() }, - { "is_in_sampling_key", std::make_shared() }, - { "compression_codec", std::make_shared() }, - { "character_octet_length", std::make_shared(std::make_shared()) }, - { "numeric_precision", std::make_shared(std::make_shared()) }, - { "numeric_precision_radix", std::make_shared(std::make_shared()) }, - { "numeric_scale", std::make_shared(std::make_shared()) }, - { "datetime_precision", std::make_shared(std::make_shared()) }, + { "database", std::make_shared(), "Database name."}, + { "table", std::make_shared(), "Table name."}, + { "name", std::make_shared(), "Column name."}, + { "type", std::make_shared(), "Column type."}, + { "position", std::make_shared(), "Ordinal position of a column in a table starting with 1."}, + { "default_kind", std::make_shared(), "Expression type (DEFAULT, MATERIALIZED, ALIAS) for the default value, or an empty string if it is not defined."}, + { "default_expression", std::make_shared(), "Expression for the default value, or an empty string if it is not defined."}, + { "data_compressed_bytes", std::make_shared(), "The size of compressed data, in bytes."}, + { "data_uncompressed_bytes", std::make_shared(), "The size of decompressed data, in bytes."}, + { "marks_bytes", std::make_shared(), "The size of marks, in bytes."}, + { "comment", std::make_shared(), "Comment on the column, or an empty string if it is not defined."}, + { "is_in_partition_key", std::make_shared(), "Flag that indicates whether the column is in the partition expression."}, + { "is_in_sorting_key", std::make_shared(), "Flag that indicates whether the column is in the sorting key expression."}, + { "is_in_primary_key", std::make_shared(), "Flag that indicates whether the column is in the primary key expression."}, + { "is_in_sampling_key", std::make_shared(), "Flag that indicates whether the column is in the sampling key expression."}, + { "compression_codec", std::make_shared(), "Compression codec name."}, + { "character_octet_length", std::make_shared(std::make_shared()), + "Maximum length in bytes for binary data, character data, or text data and images. In ClickHouse makes sense only for FixedString data type. Otherwise, the NULL value is returned."}, + { "numeric_precision", std::make_shared(std::make_shared()), + "Accuracy of approximate numeric data, exact numeric data, integer data, or monetary data. In ClickHouse it is bit width for integer types and decimal precision for Decimal types. Otherwise, the NULL value is returned."}, + { "numeric_precision_radix", std::make_shared(std::make_shared()), + "The base of the number system is the accuracy of approximate numeric data, exact numeric data, integer data or monetary data. In ClickHouse it's 2 for integer types and 10 for Decimal types. Otherwise, the NULL value is returned."}, + { "numeric_scale", std::make_shared(std::make_shared()), + "The scale of approximate numeric data, exact numeric data, integer data, or monetary data. In ClickHouse makes sense only for Decimal types. Otherwise, the NULL value is returned."}, + { "datetime_precision", std::make_shared(std::make_shared()), + "Decimal precision of DateTime64 data type. For other data types, the NULL value is returned."}, })); setInMemoryMetadata(storage_metadata); @@ -291,8 +299,51 @@ private: std::chrono::milliseconds lock_acquire_timeout; }; +class ReadFromSystemColumns : public SourceStepWithFilter +{ +public: + std::string getName() const override { return "ReadFromSystemColumns"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; -Pipe StorageSystemColumns::read( + ReadFromSystemColumns( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_, + std::vector columns_mask_, + size_t max_block_size_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , storage(std::move(storage_)) + , columns_mask(std::move(columns_mask_)) + , max_block_size(max_block_size_) + { + } + + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +private: + std::shared_ptr storage; + std::vector columns_mask; + const size_t max_block_size; + const ActionsDAG::Node * predicate = nullptr; +}; + +void ReadFromSystemColumns::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +void StorageSystemColumns::read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -306,9 +357,22 @@ Pipe StorageSystemColumns::read( auto [columns_mask, header] = getQueriedColumnsMaskAndHeader(sample_block, column_names); + + auto this_ptr = std::static_pointer_cast(shared_from_this()); + + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(header), std::move(this_ptr), std::move(columns_mask), max_block_size); + + query_plan.addStep(std::move(reading)); +} + +void ReadFromSystemColumns::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) +{ Block block_to_filter; Storages storages; Pipes pipes; + auto header = getOutputStream().header; { /// Add `database` column. @@ -338,12 +402,13 @@ Pipe StorageSystemColumns::read( block_to_filter.insert(ColumnWithTypeAndName(std::move(database_column_mut), std::make_shared(), "database")); /// Filter block with `database` column. - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block_to_filter, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block_to_filter, context); if (!block_to_filter.rows()) { - pipes.emplace_back(std::make_shared(header)); - return Pipe::unitePipes(std::move(pipes)); + pipes.emplace_back(std::make_shared(std::move(header))); + pipeline.init(Pipe::unitePipes(std::move(pipes))); + return; } ColumnPtr & database_column = block_to_filter.getByName("database").column; @@ -384,12 +449,13 @@ Pipe StorageSystemColumns::read( } /// Filter block with `database` and `table` columns. - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block_to_filter, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block_to_filter, context); if (!block_to_filter.rows()) { - pipes.emplace_back(std::make_shared(header)); - return Pipe::unitePipes(std::move(pipes)); + pipes.emplace_back(std::make_shared(std::move(header))); + pipeline.init(Pipe::unitePipes(std::move(pipes))); + return; } ColumnPtr filtered_database_column = block_to_filter.getByName("database").column; @@ -400,7 +466,7 @@ Pipe StorageSystemColumns::read( std::move(filtered_database_column), std::move(filtered_table_column), std::move(storages), context)); - return Pipe::unitePipes(std::move(pipes)); + pipeline.init(Pipe::unitePipes(std::move(pipes))); } } diff --git a/src/Storages/System/StorageSystemColumns.h b/src/Storages/System/StorageSystemColumns.h index 7b4b5dd8fb3..22b2541d93f 100644 --- a/src/Storages/System/StorageSystemColumns.h +++ b/src/Storages/System/StorageSystemColumns.h @@ -17,7 +17,8 @@ public: std::string getName() const override { return "SystemColumns"; } - Pipe read( + void read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, diff --git a/src/Storages/System/StorageSystemContributors.cpp b/src/Storages/System/StorageSystemContributors.cpp index 860a96c4388..f928337e1bd 100644 --- a/src/Storages/System/StorageSystemContributors.cpp +++ b/src/Storages/System/StorageSystemContributors.cpp @@ -17,7 +17,7 @@ ColumnsDescription StorageSystemContributors::getColumnsDescription() }; } -void StorageSystemContributors::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemContributors::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { std::vector contributors; for (auto * it = auto_contributors; *it; ++it) diff --git a/src/Storages/System/StorageSystemContributors.h b/src/Storages/System/StorageSystemContributors.h index ed983c5e61f..f8d082542d2 100644 --- a/src/Storages/System/StorageSystemContributors.h +++ b/src/Storages/System/StorageSystemContributors.h @@ -9,10 +9,10 @@ class Context; /** System table "contributors" with list of clickhouse contributors */ -class StorageSystemContributors final : public IStorageSystemOneBlock +class StorageSystemContributors final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemCurrentRoles.cpp b/src/Storages/System/StorageSystemCurrentRoles.cpp index 88bdf088175..bfa3a7c2e18 100644 --- a/src/Storages/System/StorageSystemCurrentRoles.cpp +++ b/src/Storages/System/StorageSystemCurrentRoles.cpp @@ -22,7 +22,7 @@ ColumnsDescription StorageSystemCurrentRoles::getColumnsDescription() } -void StorageSystemCurrentRoles::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemCurrentRoles::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto roles_info = context->getRolesInfo(); auto user = context->getUser(); diff --git a/src/Storages/System/StorageSystemCurrentRoles.h b/src/Storages/System/StorageSystemCurrentRoles.h index 4cc9b11d3f4..db1245e0ea7 100644 --- a/src/Storages/System/StorageSystemCurrentRoles.h +++ b/src/Storages/System/StorageSystemCurrentRoles.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `current_roles` system table, which allows you to get information about current roles. -class StorageSystemCurrentRoles final : public IStorageSystemOneBlock +class StorageSystemCurrentRoles final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemCurrentRoles"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemDDLWorkerQueue.cpp b/src/Storages/System/StorageSystemDDLWorkerQueue.cpp index c133a1b597c..ac5dd6c05d0 100644 --- a/src/Storages/System/StorageSystemDDLWorkerQueue.cpp +++ b/src/Storages/System/StorageSystemDDLWorkerQueue.cpp @@ -204,7 +204,7 @@ static void fillStatusColumns(MutableColumns & res_columns, size_t & col, } -void StorageSystemDDLWorkerQueue::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemDDLWorkerQueue::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto& ddl_worker = context->getDDLWorker(); fs::path ddl_zookeeper_path = ddl_worker.getQueueDir(); diff --git a/src/Storages/System/StorageSystemDDLWorkerQueue.h b/src/Storages/System/StorageSystemDDLWorkerQueue.h index 871bb706f94..eaffb488322 100644 --- a/src/Storages/System/StorageSystemDDLWorkerQueue.h +++ b/src/Storages/System/StorageSystemDDLWorkerQueue.h @@ -11,10 +11,10 @@ class Context; /** System table "distributed_ddl_queue" with list of queries that are currently in the DDL worker queue. */ -class StorageSystemDDLWorkerQueue final : public IStorageSystemOneBlock +class StorageSystemDDLWorkerQueue final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemDNSCache.cpp b/src/Storages/System/StorageSystemDNSCache.cpp index 867d1fd55b9..6fb83597908 100644 --- a/src/Storages/System/StorageSystemDNSCache.cpp +++ b/src/Storages/System/StorageSystemDNSCache.cpp @@ -32,7 +32,7 @@ ColumnsDescription StorageSystemDNSCache::getColumnsDescription() }; } -void StorageSystemDNSCache::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemDNSCache::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { using HostIPPair = std::pair; std::unordered_set>> reported_elements; diff --git a/src/Storages/System/StorageSystemDNSCache.h b/src/Storages/System/StorageSystemDNSCache.h index 340060335b8..dd24d2f35f6 100644 --- a/src/Storages/System/StorageSystemDNSCache.h +++ b/src/Storages/System/StorageSystemDNSCache.h @@ -9,7 +9,7 @@ namespace DB class Context; /// system.dns_cache table. -class StorageSystemDNSCache final : public IStorageSystemOneBlock +class StorageSystemDNSCache final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemDNSCache"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemDashboards.cpp b/src/Storages/System/StorageSystemDashboards.cpp index 7c9e8b73519..23d8fcfc481 100644 --- a/src/Storages/System/StorageSystemDashboards.cpp +++ b/src/Storages/System/StorageSystemDashboards.cpp @@ -22,7 +22,7 @@ String trim(const char * text) return String(view); } -void StorageSystemDashboards::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemDashboards::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { static const std::vector> dashboards { diff --git a/src/Storages/System/StorageSystemDashboards.h b/src/Storages/System/StorageSystemDashboards.h index 83a8664ad27..f3e957e06c5 100644 --- a/src/Storages/System/StorageSystemDashboards.h +++ b/src/Storages/System/StorageSystemDashboards.h @@ -12,7 +12,7 @@ namespace DB class Context; -class StorageSystemDashboards final : public IStorageSystemOneBlock +class StorageSystemDashboards final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemDashboards"; } @@ -22,7 +22,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemDataSkippingIndices.cpp b/src/Storages/System/StorageSystemDataSkippingIndices.cpp index 0c4eb197efd..ff782647c79 100644 --- a/src/Storages/System/StorageSystemDataSkippingIndices.cpp +++ b/src/Storages/System/StorageSystemDataSkippingIndices.cpp @@ -12,7 +12,10 @@ #include #include #include +#include +#include #include +#include namespace DB @@ -23,16 +26,16 @@ StorageSystemDataSkippingIndices::StorageSystemDataSkippingIndices(const Storage StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription( { - { "database", std::make_shared() }, - { "table", std::make_shared() }, - { "name", std::make_shared() }, - { "type", std::make_shared() }, - { "type_full", std::make_shared() }, - { "expr", std::make_shared() }, - { "granularity", std::make_shared() }, - { "data_compressed_bytes", std::make_shared() }, - { "data_uncompressed_bytes", std::make_shared() }, - { "marks", std::make_shared()} + { "database", std::make_shared(), "Database name."}, + { "table", std::make_shared(), "Table name."}, + { "name", std::make_shared(), "Index name."}, + { "type", std::make_shared(), "Index type."}, + { "type_full", std::make_shared(), "Index type expression from create statement."}, + { "expr", std::make_shared(), "Expression for the index calculation."}, + { "granularity", std::make_shared(), "The number of granules in the block."}, + { "data_compressed_bytes", std::make_shared(), "The size of compressed data, in bytes."}, + { "data_uncompressed_bytes", std::make_shared(), "The size of decompressed data, in bytes."}, + { "marks", std::make_shared(), "The size of marks, in bytes."} })); setInMemoryMetadata(storage_metadata); } @@ -176,7 +179,51 @@ private: DatabaseTablesIteratorPtr tables_it; }; -Pipe StorageSystemDataSkippingIndices::read( +class ReadFromSystemDataSkippingIndices : public SourceStepWithFilter +{ +public: + std::string getName() const override { return "ReadFromSystemDataSkippingIndices"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; + + ReadFromSystemDataSkippingIndices( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_, + std::vector columns_mask_, + size_t max_block_size_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , storage(std::move(storage_)) + , columns_mask(std::move(columns_mask_)) + , max_block_size(max_block_size_) + { + } + + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +private: + std::shared_ptr storage; + std::vector columns_mask; + const size_t max_block_size; + const ActionsDAG::Node * predicate = nullptr; +}; + +void ReadFromSystemDataSkippingIndices::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +void StorageSystemDataSkippingIndices::read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -190,6 +237,17 @@ Pipe StorageSystemDataSkippingIndices::read( auto [columns_mask, header] = getQueriedColumnsMaskAndHeader(sample_block, column_names); + auto this_ptr = std::static_pointer_cast(shared_from_this()); + + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(header), std::move(this_ptr), std::move(columns_mask), max_block_size); + + query_plan.addStep(std::move(reading)); +} + +void ReadFromSystemDataSkippingIndices::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) +{ MutableColumnPtr column = ColumnString::create(); const auto databases = DatabaseCatalog::instance().getDatabases(); @@ -207,11 +265,11 @@ Pipe StorageSystemDataSkippingIndices::read( /// Condition on "database" in a query acts like an index. Block block { ColumnWithTypeAndName(std::move(column), std::make_shared(), "database") }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block, context); ColumnPtr & filtered_databases = block.getByPosition(0).column; - return Pipe(std::make_shared( - std::move(columns_mask), std::move(header), max_block_size, std::move(filtered_databases), context)); + pipeline.init(Pipe(std::make_shared( + std::move(columns_mask), getOutputStream().header, max_block_size, std::move(filtered_databases), context))); } } diff --git a/src/Storages/System/StorageSystemDataSkippingIndices.h b/src/Storages/System/StorageSystemDataSkippingIndices.h index 8a1e8c159b4..8bf1da98368 100644 --- a/src/Storages/System/StorageSystemDataSkippingIndices.h +++ b/src/Storages/System/StorageSystemDataSkippingIndices.h @@ -14,7 +14,8 @@ public: std::string getName() const override { return "SystemDataSkippingIndices"; } - Pipe read( + void read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, diff --git a/src/Storages/System/StorageSystemDataTypeFamilies.cpp b/src/Storages/System/StorageSystemDataTypeFamilies.cpp index 821a43ed530..da607ade410 100644 --- a/src/Storages/System/StorageSystemDataTypeFamilies.cpp +++ b/src/Storages/System/StorageSystemDataTypeFamilies.cpp @@ -17,7 +17,7 @@ ColumnsDescription StorageSystemDataTypeFamilies::getColumnsDescription() }; } -void StorageSystemDataTypeFamilies::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemDataTypeFamilies::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { const auto & factory = DataTypeFactory::instance(); auto names = factory.getAllRegisteredNames(); diff --git a/src/Storages/System/StorageSystemDataTypeFamilies.h b/src/Storages/System/StorageSystemDataTypeFamilies.h index 2cb834f6931..6322c3bcfdd 100644 --- a/src/Storages/System/StorageSystemDataTypeFamilies.h +++ b/src/Storages/System/StorageSystemDataTypeFamilies.h @@ -5,10 +5,10 @@ namespace DB { -class StorageSystemDataTypeFamilies final : public IStorageSystemOneBlock +class StorageSystemDataTypeFamilies final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemDatabaseEngines.cpp b/src/Storages/System/StorageSystemDatabaseEngines.cpp index 514ca6d0ab3..59ba60fa871 100644 --- a/src/Storages/System/StorageSystemDatabaseEngines.cpp +++ b/src/Storages/System/StorageSystemDatabaseEngines.cpp @@ -14,7 +14,7 @@ ColumnsDescription StorageSystemDatabaseEngines::getColumnsDescription() }; } -void StorageSystemDatabaseEngines::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemDatabaseEngines::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (const auto & [engine, _] : DatabaseFactory::instance().getDatabaseEngines()) { diff --git a/src/Storages/System/StorageSystemDatabaseEngines.h b/src/Storages/System/StorageSystemDatabaseEngines.h index 16b517c91e6..3af13598c17 100644 --- a/src/Storages/System/StorageSystemDatabaseEngines.h +++ b/src/Storages/System/StorageSystemDatabaseEngines.h @@ -6,10 +6,10 @@ namespace DB { -class StorageSystemDatabaseEngines final : public IStorageSystemOneBlock +class StorageSystemDatabaseEngines final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemDatabases.cpp b/src/Storages/System/StorageSystemDatabases.cpp index 51ecb8f17ca..f5537b508ba 100644 --- a/src/Storages/System/StorageSystemDatabases.cpp +++ b/src/Storages/System/StorageSystemDatabases.cpp @@ -72,7 +72,7 @@ static String getEngineFull(const ContextPtr & ctx, const DatabasePtr & database return engine_full; } -static ColumnPtr getFilteredDatabases(const Databases & databases, const SelectQueryInfo & query_info, ContextPtr context) +static ColumnPtr getFilteredDatabases(const Databases & databases, const ActionsDAG::Node * predicate, ContextPtr context) { MutableColumnPtr name_column = ColumnString::create(); MutableColumnPtr engine_column = ColumnString::create(); @@ -94,17 +94,17 @@ static ColumnPtr getFilteredDatabases(const Databases & databases, const SelectQ ColumnWithTypeAndName(std::move(engine_column), std::make_shared(), "engine"), ColumnWithTypeAndName(std::move(uuid_column), std::make_shared(), "uuid") }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block, context); return block.getByPosition(0).column; } -void StorageSystemDatabases::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const +void StorageSystemDatabases::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector columns_mask) const { const auto access = context->getAccess(); const bool check_access_for_databases = !access->isGranted(AccessType::SHOW_DATABASES); const auto databases = DatabaseCatalog::instance().getDatabases(); - ColumnPtr filtered_databases_column = getFilteredDatabases(databases, query_info, context); + ColumnPtr filtered_databases_column = getFilteredDatabases(databases, predicate, context); for (size_t i = 0; i < filtered_databases_column->size(); ++i) { @@ -120,7 +120,6 @@ void StorageSystemDatabases::fillData(MutableColumns & res_columns, ContextPtr c size_t src_index = 0; size_t res_index = 0; - const auto & columns_mask = query_info.columns_mask; if (columns_mask[src_index++]) res_columns[res_index++]->insert(database_name); if (columns_mask[src_index++]) diff --git a/src/Storages/System/StorageSystemDatabases.h b/src/Storages/System/StorageSystemDatabases.h index 2fd9ccdc970..fa55f0aea32 100644 --- a/src/Storages/System/StorageSystemDatabases.h +++ b/src/Storages/System/StorageSystemDatabases.h @@ -11,7 +11,7 @@ class Context; /** Implements `databases` system table, which allows you to get information about all databases. */ -class StorageSystemDatabases final : public IStorageSystemOneBlock +class StorageSystemDatabases final : public IStorageSystemOneBlock { public: std::string getName() const override @@ -26,7 +26,7 @@ protected: bool supportsColumnsMask() const override { return true; } - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector columns_mask) const override; }; } diff --git a/src/Storages/System/StorageSystemDetachedParts.cpp b/src/Storages/System/StorageSystemDetachedParts.cpp index 3dae43976f7..1eb79744022 100644 --- a/src/Storages/System/StorageSystemDetachedParts.cpp +++ b/src/Storages/System/StorageSystemDetachedParts.cpp @@ -12,6 +12,8 @@ #include #include #include +#include +#include #include @@ -269,23 +271,69 @@ StorageSystemDetachedParts::StorageSystemDetachedParts(const StorageID & table_i { StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription{{ - {"database", std::make_shared()}, - {"table", std::make_shared()}, - {"partition_id", std::make_shared(std::make_shared())}, - {"name", std::make_shared()}, - {"bytes_on_disk", std::make_shared()}, - {"modification_time",std::make_shared()}, - {"disk", std::make_shared()}, - {"path", std::make_shared()}, - {"reason", std::make_shared(std::make_shared())}, - {"min_block_number", std::make_shared(std::make_shared())}, - {"max_block_number", std::make_shared(std::make_shared())}, - {"level", std::make_shared(std::make_shared())}, + {"database", std::make_shared(), "The name of the database this part belongs to."}, + {"table", std::make_shared(), "The name of the table this part belongs to."}, + {"partition_id", std::make_shared(std::make_shared()), "The identifier of the partition this part belongs to."}, + {"name", std::make_shared(), "The name of the part."}, + {"bytes_on_disk", std::make_shared(), "Total size of all the data part files in bytes."}, + {"modification_time",std::make_shared(), "The time the directory with the data part was modified. This usually corresponds to the time when detach happened."}, + {"disk", std::make_shared(), "The name of the disk that stores this data part."}, + {"path", std::make_shared(), "The path of the disk to the file of this data part."}, + {"reason", std::make_shared(std::make_shared()), "The explanation why this part was detached."}, + {"min_block_number", std::make_shared(std::make_shared()), "The minimum number of data parts that make up the current part after merging."}, + {"max_block_number", std::make_shared(std::make_shared()), "The maximum number of data parts that make up the current part after merging."}, + {"level", std::make_shared(std::make_shared()), "Depth of the merge tree. Zero means that the current part was created by insert rather than by merging other parts."}, }}); setInMemoryMetadata(storage_metadata); } -Pipe StorageSystemDetachedParts::read( +class ReadFromSystemDetachedParts : public SourceStepWithFilter +{ +public: + ReadFromSystemDetachedParts( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_, + std::vector columns_mask_, + size_t max_block_size_, + size_t num_streams_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , storage(std::move(storage_)) + , columns_mask(std::move(columns_mask_)) + , max_block_size(max_block_size_) + , num_streams(num_streams_) + {} + + std::string getName() const override { return "ReadFromSystemDetachedParts"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +protected: + std::shared_ptr storage; + std::vector columns_mask; + + const ActionsDAG::Node * predicate = nullptr; + const size_t max_block_size; + const size_t num_streams; +}; + +void ReadFromSystemDetachedParts::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +void StorageSystemDetachedParts::read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -299,17 +347,28 @@ Pipe StorageSystemDetachedParts::read( auto [columns_mask, header] = getQueriedColumnsMaskAndHeader(sample_block, column_names); - auto state = std::make_shared(StoragesInfoStream(query_info, context)); + auto this_ptr = std::static_pointer_cast(shared_from_this()); + + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(header), std::move(this_ptr), std::move(columns_mask), max_block_size, num_streams); + + query_plan.addStep(std::move(reading)); +} + +void ReadFromSystemDetachedParts::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) +{ + auto state = std::make_shared(StoragesInfoStream(predicate, context)); Pipe pipe; for (size_t i = 0; i < num_streams; ++i) { - auto source = std::make_shared(header.cloneEmpty(), state, columns_mask, max_block_size); + auto source = std::make_shared(getOutputStream().header, state, columns_mask, max_block_size); pipe.addSource(std::move(source)); } - return pipe; + pipeline.init(std::move(pipe)); } } diff --git a/src/Storages/System/StorageSystemDetachedParts.h b/src/Storages/System/StorageSystemDetachedParts.h index 20ac69f0eea..52b964e3b3c 100644 --- a/src/Storages/System/StorageSystemDetachedParts.h +++ b/src/Storages/System/StorageSystemDetachedParts.h @@ -20,14 +20,15 @@ public: bool isSystemStorage() const override { return true; } protected: - Pipe read( - const Names & /* column_names */, - const StorageSnapshotPtr & storage_snapshot, - SelectQueryInfo & query_info, - ContextPtr context, - QueryProcessingStage::Enum /*processed_stage*/, - size_t /*max_block_size*/, - size_t /*num_streams*/) override; + void read( + QueryPlan & query_plan, + const Names & /* column_names */, + const StorageSnapshotPtr & storage_snapshot, + SelectQueryInfo & query_info, + ContextPtr context, + QueryProcessingStage::Enum /*processed_stage*/, + size_t /*max_block_size*/, + size_t /*num_streams*/) override; }; } diff --git a/src/Storages/System/StorageSystemDictionaries.cpp b/src/Storages/System/StorageSystemDictionaries.cpp index c2ed35c5510..8b528b4a298 100644 --- a/src/Storages/System/StorageSystemDictionaries.cpp +++ b/src/Storages/System/StorageSystemDictionaries.cpp @@ -16,7 +16,6 @@ #include #include -#include namespace DB { @@ -52,6 +51,14 @@ catch (const DB::Exception &) } +StorageSystemDictionaries::StorageSystemDictionaries(const StorageID & storage_id_, ColumnsDescription columns_description_) + : IStorageSystemOneBlock(storage_id_, std::move(columns_description_)) +{ + VirtualColumnsDescription virtuals; + virtuals.addEphemeral("key", std::make_shared(), ""); + setVirtuals(std::move(virtuals)); +} + ColumnsDescription StorageSystemDictionaries::getColumnsDescription() { return ColumnsDescription @@ -75,7 +82,7 @@ ColumnsDescription StorageSystemDictionaries::getColumnsDescription() {"attribute.names", std::make_shared(std::make_shared()), "Array of attribute names provided by the dictionary."}, {"attribute.types", std::make_shared(std::make_shared()), "Corresponding array of attribute types provided by the dictionary."}, {"bytes_allocated", std::make_shared(), "Amount of RAM allocated for the dictionary."}, - {"hierarchical_index_bytes_allocated", std::make_shared(), ""}, + {"hierarchical_index_bytes_allocated", std::make_shared(), "Amount of RAM allocated for hierarchical index."}, {"query_count", std::make_shared(), "Number of queries since the dictionary was loaded or since the last successful reboot."}, {"hit_rate", std::make_shared(), "For cache dictionaries, the percentage of uses for which the value was in the cache."}, {"found_rate", std::make_shared(), "The percentage of uses for which the value was found."}, @@ -92,14 +99,7 @@ ColumnsDescription StorageSystemDictionaries::getColumnsDescription() }; } -NamesAndTypesList StorageSystemDictionaries::getVirtuals() const -{ - return { - {"key", std::make_shared()} - }; -} - -void StorageSystemDictionaries::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & /*query_info*/) const +void StorageSystemDictionaries::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_dictionaries = access->isGranted(AccessType::SHOW_DICTIONARIES); diff --git a/src/Storages/System/StorageSystemDictionaries.h b/src/Storages/System/StorageSystemDictionaries.h index 792b3c0dd30..058b8b163d9 100644 --- a/src/Storages/System/StorageSystemDictionaries.h +++ b/src/Storages/System/StorageSystemDictionaries.h @@ -9,19 +9,17 @@ namespace DB class Context; -class StorageSystemDictionaries final : public IStorageSystemOneBlock +class StorageSystemDictionaries final : public IStorageSystemOneBlock { public: + StorageSystemDictionaries(const StorageID & storage_id_, ColumnsDescription columns_description_); + std::string getName() const override { return "SystemDictionaries"; } static ColumnsDescription getColumnsDescription(); - NamesAndTypesList getVirtuals() const override; - protected: - using IStorageSystemOneBlock::IStorageSystemOneBlock; - - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemDisks.cpp b/src/Storages/System/StorageSystemDisks.cpp index 30d64156b22..b010eff2fe3 100644 --- a/src/Storages/System/StorageSystemDisks.cpp +++ b/src/Storages/System/StorageSystemDisks.cpp @@ -18,19 +18,19 @@ StorageSystemDisks::StorageSystemDisks(const StorageID & table_id_) StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription( { - {"name", std::make_shared()}, - {"path", std::make_shared()}, - {"free_space", std::make_shared()}, - {"total_space", std::make_shared()}, - {"unreserved_space", std::make_shared()}, - {"keep_free_space", std::make_shared()}, - {"type", std::make_shared()}, - {"is_encrypted", std::make_shared()}, - {"is_read_only", std::make_shared()}, - {"is_write_once", std::make_shared()}, - {"is_remote", std::make_shared()}, - {"is_broken", std::make_shared()}, - {"cache_path", std::make_shared()}, + {"name", std::make_shared(), "Name of a disk in the server configuration."}, + {"path", std::make_shared(), "Path to the mount point in the file system."}, + {"free_space", std::make_shared(), "Free space on disk in bytes."}, + {"total_space", std::make_shared(), "Disk volume in bytes."}, + {"unreserved_space", std::make_shared(), "Free space which is not taken by reservations (free_space minus the size of reservations taken by merges, inserts, and other disk write operations currently running)."}, + {"keep_free_space", std::make_shared(), "Amount of disk space that should stay free on disk in bytes. Defined in the keep_free_space_bytes parameter of disk configuration."}, + {"type", std::make_shared(), "The disk type which tells where this disk stores the data - RAM, local drive or remote storage."}, + {"is_encrypted", std::make_shared(), "Flag which shows whether this disk ecrypts the underlying data. "}, + {"is_read_only", std::make_shared(), "Flag which indicates that you can only perform read operations with this disk."}, + {"is_write_once", std::make_shared(), "Flag which indicates if disk is write-once. Which means that it does support BACKUP to this disk, but does not support INSERT into MergeTree table on this disk."}, + {"is_remote", std::make_shared(), "Flag which indicated what operations with this disk involve network interaction."}, + {"is_broken", std::make_shared(), "Flag which indicates if disk is broken. Broken disks will have 0 space and cannot be used."}, + {"cache_path", std::make_shared(), "The path to the cache directory on local drive in case when the disk supports caching."}, })); setInMemoryMetadata(storage_metadata); } diff --git a/src/Storages/System/StorageSystemDistributionQueue.cpp b/src/Storages/System/StorageSystemDistributionQueue.cpp index f0247275a8b..50c6436f316 100644 --- a/src/Storages/System/StorageSystemDistributionQueue.cpp +++ b/src/Storages/System/StorageSystemDistributionQueue.cpp @@ -107,7 +107,7 @@ ColumnsDescription StorageSystemDistributionQueue::getColumnsDescription() } -void StorageSystemDistributionQueue::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const +void StorageSystemDistributionQueue::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_databases = !access->isGranted(AccessType::SHOW_TABLES); @@ -159,7 +159,7 @@ void StorageSystemDistributionQueue::fillData(MutableColumns & res_columns, Cont { col_table_to_filter, std::make_shared(), "table" }, }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) return; diff --git a/src/Storages/System/StorageSystemDistributionQueue.h b/src/Storages/System/StorageSystemDistributionQueue.h index 477a9d6e245..159a86bf082 100644 --- a/src/Storages/System/StorageSystemDistributionQueue.h +++ b/src/Storages/System/StorageSystemDistributionQueue.h @@ -11,7 +11,7 @@ class Context; /** Implements the `distribution_queue` system table, which allows you to view the INSERT queues for the Distributed tables. */ -class StorageSystemDistributionQueue final : public IStorageSystemOneBlock +class StorageSystemDistributionQueue final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemDistributionQueue"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemDroppedTables.cpp b/src/Storages/System/StorageSystemDroppedTables.cpp index a95127847f8..528f4d8995d 100644 --- a/src/Storages/System/StorageSystemDroppedTables.cpp +++ b/src/Storages/System/StorageSystemDroppedTables.cpp @@ -29,7 +29,7 @@ ColumnsDescription StorageSystemDroppedTables::getColumnsDescription() } -void StorageSystemDroppedTables::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemDroppedTables::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { auto tables_mark_dropped = DatabaseCatalog::instance().getTablesMarkedDropped(); diff --git a/src/Storages/System/StorageSystemDroppedTables.h b/src/Storages/System/StorageSystemDroppedTables.h index d7c3569eb62..4fc620ab962 100644 --- a/src/Storages/System/StorageSystemDroppedTables.h +++ b/src/Storages/System/StorageSystemDroppedTables.h @@ -6,7 +6,7 @@ namespace DB { -class StorageSystemDroppedTables final : public IStorageSystemOneBlock +class StorageSystemDroppedTables final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemMarkedDroppedTables"; } @@ -14,7 +14,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemDroppedTablesParts.cpp b/src/Storages/System/StorageSystemDroppedTablesParts.cpp index bbe99c688c9..20baeee1d3b 100644 --- a/src/Storages/System/StorageSystemDroppedTablesParts.cpp +++ b/src/Storages/System/StorageSystemDroppedTablesParts.cpp @@ -10,7 +10,7 @@ namespace DB { -StoragesDroppedInfoStream::StoragesDroppedInfoStream(const SelectQueryInfo & query_info, ContextPtr context) +StoragesDroppedInfoStream::StoragesDroppedInfoStream(const ActionsDAG::Node * predicate, ContextPtr context) : StoragesInfoStreamBase(context) { /// Will apply WHERE to subset of columns and then add more columns. @@ -73,7 +73,7 @@ StoragesDroppedInfoStream::StoragesDroppedInfoStream(const SelectQueryInfo & que if (block_to_filter.rows()) { /// Filter block_to_filter with columns 'database', 'table', 'engine', 'active'. - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block_to_filter, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block_to_filter, context); rows = block_to_filter.rows(); } diff --git a/src/Storages/System/StorageSystemDroppedTablesParts.h b/src/Storages/System/StorageSystemDroppedTablesParts.h index f548697a6a9..a44abea7285 100644 --- a/src/Storages/System/StorageSystemDroppedTablesParts.h +++ b/src/Storages/System/StorageSystemDroppedTablesParts.h @@ -9,7 +9,7 @@ namespace DB class StoragesDroppedInfoStream : public StoragesInfoStreamBase { public: - StoragesDroppedInfoStream(const SelectQueryInfo & query_info, ContextPtr context); + StoragesDroppedInfoStream(const ActionsDAG::Node * predicate, ContextPtr context); protected: bool tryLockTable(StoragesInfo &) override { @@ -30,9 +30,9 @@ public: std::string getName() const override { return "SystemDroppedTablesParts"; } protected: - std::unique_ptr getStoragesInfoStream(const SelectQueryInfo & query_info, ContextPtr context) override + std::unique_ptr getStoragesInfoStream(const ActionsDAG::Node * predicate, ContextPtr context) override { - return std::make_unique(query_info, context); + return std::make_unique(predicate, context); } }; diff --git a/src/Storages/System/StorageSystemEnabledRoles.cpp b/src/Storages/System/StorageSystemEnabledRoles.cpp index 6dbb6f18488..42806a832cf 100644 --- a/src/Storages/System/StorageSystemEnabledRoles.cpp +++ b/src/Storages/System/StorageSystemEnabledRoles.cpp @@ -23,7 +23,7 @@ ColumnsDescription StorageSystemEnabledRoles::getColumnsDescription() } -void StorageSystemEnabledRoles::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemEnabledRoles::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto roles_info = context->getRolesInfo(); auto user = context->getUser(); diff --git a/src/Storages/System/StorageSystemEnabledRoles.h b/src/Storages/System/StorageSystemEnabledRoles.h index 5367b2ccbea..1b11d025367 100644 --- a/src/Storages/System/StorageSystemEnabledRoles.h +++ b/src/Storages/System/StorageSystemEnabledRoles.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `enabled_roles` system table, which allows you to get information about enabled roles. -class StorageSystemEnabledRoles final : public IStorageSystemOneBlock +class StorageSystemEnabledRoles final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemEnabledRoles"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemErrors.cpp b/src/Storages/System/StorageSystemErrors.cpp index 730e4cf05da..2da268305f8 100644 --- a/src/Storages/System/StorageSystemErrors.cpp +++ b/src/Storages/System/StorageSystemErrors.cpp @@ -25,7 +25,7 @@ ColumnsDescription StorageSystemErrors::getColumnsDescription() } -void StorageSystemErrors::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemErrors::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto add_row = [&](std::string_view name, size_t code, const auto & error, bool remote) { diff --git a/src/Storages/System/StorageSystemErrors.h b/src/Storages/System/StorageSystemErrors.h index 9e8ec628bac..bc86c085ff1 100644 --- a/src/Storages/System/StorageSystemErrors.h +++ b/src/Storages/System/StorageSystemErrors.h @@ -13,7 +13,7 @@ class Context; * Implements the `errors` system table, which shows the error code and the number of times it happens * (i.e. Exception with this code had been thrown). */ -class StorageSystemErrors final : public IStorageSystemOneBlock +class StorageSystemErrors final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemErrors"; } @@ -23,7 +23,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemEvents.cpp b/src/Storages/System/StorageSystemEvents.cpp index 5d9bc3a773a..822d5c77788 100644 --- a/src/Storages/System/StorageSystemEvents.cpp +++ b/src/Storages/System/StorageSystemEvents.cpp @@ -23,7 +23,7 @@ ColumnsDescription StorageSystemEvents::getColumnsDescription() return description; } -void StorageSystemEvents::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemEvents::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { for (ProfileEvents::Event i = ProfileEvents::Event(0), end = ProfileEvents::end(); i < end; ++i) { diff --git a/src/Storages/System/StorageSystemEvents.h b/src/Storages/System/StorageSystemEvents.h index cbd92f90d7a..9217fdfb53e 100644 --- a/src/Storages/System/StorageSystemEvents.h +++ b/src/Storages/System/StorageSystemEvents.h @@ -10,7 +10,7 @@ class Context; /** Implements `events` system table, which allows you to obtain information for profiling. */ -class StorageSystemEvents final : public IStorageSystemOneBlock +class StorageSystemEvents final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemEvents"; } @@ -20,7 +20,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemFilesystemCache.cpp b/src/Storages/System/StorageSystemFilesystemCache.cpp index d2bd085f934..cfb388bc232 100644 --- a/src/Storages/System/StorageSystemFilesystemCache.cpp +++ b/src/Storages/System/StorageSystemFilesystemCache.cpp @@ -17,30 +17,30 @@ ColumnsDescription StorageSystemFilesystemCache::getColumnsDescription() /// TODO: Fill in all the comments. return ColumnsDescription { - {"cache_name", std::make_shared()}, - {"cache_base_path", std::make_shared()}, - {"cache_path", std::make_shared()}, - {"key", std::make_shared()}, - {"file_segment_range_begin", std::make_shared()}, - {"file_segment_range_end", std::make_shared()}, - {"size", std::make_shared()}, - {"state", std::make_shared()}, - {"cache_hits", std::make_shared()}, - {"references", std::make_shared()}, - {"downloaded_size", std::make_shared()}, - {"kind", std::make_shared()}, - {"unbound", std::make_shared>()}, - {"user_id", std::make_shared()}, - {"file_size", std::make_shared(std::make_shared())}, + {"cache_name", std::make_shared(), "Name of the cache object. Can be used in `SYSTEM DESCRIBE FILESYSTEM CACHE `, `SYSTEM DROP FILESYSTEM CACHE ` commands"}, + {"cache_base_path", std::make_shared(), "Path to the base directory where all caches files (of a cache identidied by `cache_name`) are stored."}, + {"cache_path", std::make_shared(), "Path to a particular cache file, corresponding to a file segment in a source file"}, + {"key", std::make_shared(), "Cache key of the file segment"}, + {"file_segment_range_begin", std::make_shared(), "Offset corresponding to the beginning of the file segment range"}, + {"file_segment_range_end", std::make_shared(), "Offset corresponding to the (including) end of the file segment range"}, + {"size", std::make_shared(), "Size of the file segment"}, + {"state", std::make_shared(), "File segment state (DOWNLOADED, DOWNLOADING, PARTIALLY_DOWNLOADED, ...)"}, + {"cache_hits", std::make_shared(), "Number of cache hits of corresponding file segment"}, + {"references", std::make_shared(), "Number of references to corresponding file segment. Value 1 means that nobody uses it at the moment (the only existing reference is in cache storage itself)"}, + {"downloaded_size", std::make_shared(), "Downloaded size of the file segment"}, + {"kind", std::make_shared(), "File segment kind (used to distringuish between file segments added as a part of 'Temporary data in cache')"}, + {"unbound", std::make_shared>(), "Internal implementation flag"}, + {"user_id", std::make_shared(), "User id of the user which created the file segment"}, + {"file_size", std::make_shared(std::make_shared()), "File size of the file to which current file segment belongs"}, }; } StorageSystemFilesystemCache::StorageSystemFilesystemCache(const StorageID & table_id_) - : IStorageSystemOneBlock(table_id_) + : IStorageSystemOneBlock(table_id_, getColumnsDescription()) { } -void StorageSystemFilesystemCache::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemFilesystemCache::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { auto caches = FileCacheFactory::instance().getAll(); diff --git a/src/Storages/System/StorageSystemFilesystemCache.h b/src/Storages/System/StorageSystemFilesystemCache.h index 4b13b375f95..ea49fd16ba2 100644 --- a/src/Storages/System/StorageSystemFilesystemCache.h +++ b/src/Storages/System/StorageSystemFilesystemCache.h @@ -29,7 +29,7 @@ namespace DB * FORMAT Vertical */ -class StorageSystemFilesystemCache final : public IStorageSystemOneBlock +class StorageSystemFilesystemCache final : public IStorageSystemOneBlock { public: explicit StorageSystemFilesystemCache(const StorageID & table_id_); @@ -39,7 +39,7 @@ public: static ColumnsDescription getColumnsDescription(); protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemFormats.cpp b/src/Storages/System/StorageSystemFormats.cpp index 849e4eadf78..3c6a19c4f95 100644 --- a/src/Storages/System/StorageSystemFormats.cpp +++ b/src/Storages/System/StorageSystemFormats.cpp @@ -18,7 +18,7 @@ ColumnsDescription StorageSystemFormats::getColumnsDescription() }; } -void StorageSystemFormats::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemFormats::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { const auto & formats = FormatFactory::instance().getAllFormats(); for (const auto & pair : formats) diff --git a/src/Storages/System/StorageSystemFormats.h b/src/Storages/System/StorageSystemFormats.h index 9f9d1df1bde..f93641ee8a8 100644 --- a/src/Storages/System/StorageSystemFormats.h +++ b/src/Storages/System/StorageSystemFormats.h @@ -4,10 +4,10 @@ namespace DB { -class StorageSystemFormats final : public IStorageSystemOneBlock +class StorageSystemFormats final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemFunctions.cpp b/src/Storages/System/StorageSystemFunctions.cpp index ac1e03515c9..967132e4d4a 100644 --- a/src/Storages/System/StorageSystemFunctions.cpp +++ b/src/Storages/System/StorageSystemFunctions.cpp @@ -133,7 +133,7 @@ ColumnsDescription StorageSystemFunctions::getColumnsDescription() }; } -void StorageSystemFunctions::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemFunctions::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto & functions_factory = FunctionFactory::instance(); const auto & function_names = functions_factory.getAllRegisteredNames(); diff --git a/src/Storages/System/StorageSystemFunctions.h b/src/Storages/System/StorageSystemFunctions.h index ac1129e8127..c6f85d436fc 100644 --- a/src/Storages/System/StorageSystemFunctions.h +++ b/src/Storages/System/StorageSystemFunctions.h @@ -12,7 +12,7 @@ class Context; /** Implements `functions`system table, which allows you to get a list * all normal and aggregate functions. */ -class StorageSystemFunctions final : public IStorageSystemOneBlock +class StorageSystemFunctions final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemFunctions"; } @@ -25,7 +25,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemGrants.cpp b/src/Storages/System/StorageSystemGrants.cpp index f5f3fa07e53..afa49536983 100644 --- a/src/Storages/System/StorageSystemGrants.cpp +++ b/src/Storages/System/StorageSystemGrants.cpp @@ -38,7 +38,7 @@ ColumnsDescription StorageSystemGrants::getColumnsDescription() } -void StorageSystemGrants::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemGrants::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemGrants.h b/src/Storages/System/StorageSystemGrants.h index 2202b52ad5f..6bf3793c3dc 100644 --- a/src/Storages/System/StorageSystemGrants.h +++ b/src/Storages/System/StorageSystemGrants.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `grants` system table, which allows you to get information about grants. -class StorageSystemGrants final : public IStorageSystemOneBlock +class StorageSystemGrants final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemGrants"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemGraphite.cpp b/src/Storages/System/StorageSystemGraphite.cpp index ffeb1b6c890..eaa386763c2 100644 --- a/src/Storages/System/StorageSystemGraphite.cpp +++ b/src/Storages/System/StorageSystemGraphite.cpp @@ -11,7 +11,11 @@ ColumnsDescription StorageSystemGraphite::getColumnsDescription() return ColumnsDescription { {"config_name", std::make_shared(), "graphite_rollup parameter name."}, - {"rule_type", std::make_shared(), ""}, + {"rule_type", std::make_shared(), + "The rule type. Possible values: RuleTypeAll = 0 - default, with regex, compatible with old scheme; " + "RuleTypePlain = 1 - plain metrics, with regex, compatible with old scheme; " + "RuleTypeTagged = 2 - tagged metrics, with regex, compatible with old scheme; " + "RuleTypeTagList = 3 - tagged metrics, with regex (converted to RuleTypeTagged from string like 'retention=10min ; env=(staging|prod)')"}, {"regexp", std::make_shared(), "A pattern for the metric name."}, {"function", std::make_shared(), "The name of the aggregating function."}, {"age", std::make_shared(), "The minimum age of the data in seconds."}, @@ -75,7 +79,7 @@ static StorageSystemGraphite::Configs getConfigs(ContextPtr context) return graphite_configs; } -void StorageSystemGraphite::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemGraphite::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { Configs graphite_configs = getConfigs(context); diff --git a/src/Storages/System/StorageSystemGraphite.h b/src/Storages/System/StorageSystemGraphite.h index be101181cf7..78379afac9d 100644 --- a/src/Storages/System/StorageSystemGraphite.h +++ b/src/Storages/System/StorageSystemGraphite.h @@ -10,7 +10,7 @@ namespace DB { /// Provides information about Graphite configuration. -class StorageSystemGraphite final : public IStorageSystemOneBlock +class StorageSystemGraphite final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemGraphite"; } @@ -30,7 +30,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemJemalloc.cpp b/src/Storages/System/StorageSystemJemalloc.cpp index 15543208dd9..b55e32c479c 100644 --- a/src/Storages/System/StorageSystemJemalloc.cpp +++ b/src/Storages/System/StorageSystemJemalloc.cpp @@ -115,7 +115,7 @@ Pipe StorageSystemJemallocBins::read( { storage_snapshot->check(column_names); - auto header = storage_snapshot->metadata->getSampleBlockWithVirtuals(getVirtuals()); + auto header = storage_snapshot->metadata->getSampleBlockWithVirtuals(getVirtualsList()); MutableColumns res_columns = header.cloneEmptyColumns(); fillJemallocBins(res_columns); diff --git a/src/Storages/System/StorageSystemKafkaConsumers.cpp b/src/Storages/System/StorageSystemKafkaConsumers.cpp index 95962d8de8b..86713632339 100644 --- a/src/Storages/System/StorageSystemKafkaConsumers.cpp +++ b/src/Storages/System/StorageSystemKafkaConsumers.cpp @@ -41,13 +41,13 @@ ColumnsDescription StorageSystemKafkaConsumers::getColumnsDescription() {"last_rebalance_time", std::make_shared(), "Timestamp of the most recent Kafka rebalance."}, {"num_rebalance_revocations", std::make_shared(), "Number of times the consumer was revoked its partitions."}, {"num_rebalance_assignments", std::make_shared(), "Number of times the consumer was assigned to Kafka cluster."}, - {"is_currently_used", std::make_shared(), "Consumer is in use."}, - {"last_used", std::make_shared(6)}, + {"is_currently_used", std::make_shared(), "The flag which shows whether the consumer is in use."}, + {"last_used", std::make_shared(6), "The last time this consumer was in use."}, {"rdkafka_stat", std::make_shared(), "Library internal statistic. Set statistics_interval_ms to 0 disable, default is 3000 (once in three seconds)."}, }; } -void StorageSystemKafkaConsumers::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemKafkaConsumers::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto tables_mark_dropped = DatabaseCatalog::instance().getTablesMarkedDropped(); diff --git a/src/Storages/System/StorageSystemKafkaConsumers.h b/src/Storages/System/StorageSystemKafkaConsumers.h index ae2c726849d..8d1fd504810 100644 --- a/src/Storages/System/StorageSystemKafkaConsumers.h +++ b/src/Storages/System/StorageSystemKafkaConsumers.h @@ -11,7 +11,7 @@ namespace DB { -class StorageSystemKafkaConsumers final : public IStorageSystemOneBlock +class StorageSystemKafkaConsumers final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemKafkaConsumers"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemLicenses.cpp b/src/Storages/System/StorageSystemLicenses.cpp index c5c04b4eb94..2faf772aca4 100644 --- a/src/Storages/System/StorageSystemLicenses.cpp +++ b/src/Storages/System/StorageSystemLicenses.cpp @@ -19,7 +19,7 @@ ColumnsDescription StorageSystemLicenses::getColumnsDescription() }; } -void StorageSystemLicenses::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemLicenses::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (const auto * it = library_licenses; *it; it += 4) { diff --git a/src/Storages/System/StorageSystemLicenses.h b/src/Storages/System/StorageSystemLicenses.h index 57a3ff201a2..ab74a590dea 100644 --- a/src/Storages/System/StorageSystemLicenses.h +++ b/src/Storages/System/StorageSystemLicenses.h @@ -10,10 +10,10 @@ class Context; /** System table "licenses" with list of licenses of 3rd party libraries */ -class StorageSystemLicenses final : public IStorageSystemOneBlock +class StorageSystemLicenses final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemMacros.cpp b/src/Storages/System/StorageSystemMacros.cpp index 6c1a24d152a..82408fd5a7e 100644 --- a/src/Storages/System/StorageSystemMacros.cpp +++ b/src/Storages/System/StorageSystemMacros.cpp @@ -15,7 +15,7 @@ ColumnsDescription StorageSystemMacros::getColumnsDescription() }; } -void StorageSystemMacros::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemMacros::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto macros = context->getMacros(); diff --git a/src/Storages/System/StorageSystemMacros.h b/src/Storages/System/StorageSystemMacros.h index ffbeb70796e..c272985e978 100644 --- a/src/Storages/System/StorageSystemMacros.h +++ b/src/Storages/System/StorageSystemMacros.h @@ -12,7 +12,7 @@ class Context; /** Information about macros for introspection. */ -class StorageSystemMacros final : public IStorageSystemOneBlock +class StorageSystemMacros final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemMacros"; } @@ -22,7 +22,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemMergeTreeSettings.cpp b/src/Storages/System/StorageSystemMergeTreeSettings.cpp index 85caa572edd..7781e3789a4 100644 --- a/src/Storages/System/StorageSystemMergeTreeSettings.cpp +++ b/src/Storages/System/StorageSystemMergeTreeSettings.cpp @@ -31,7 +31,7 @@ ColumnsDescription SystemMergeTreeSettings::getColumnsDescription() } template -void SystemMergeTreeSettings::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void SystemMergeTreeSettings::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto & settings = replicated ? context->getReplicatedMergeTreeSettings() : context->getMergeTreeSettings(); auto constraints_and_current_profiles = context->getSettingsConstraintsAndCurrentProfiles(); diff --git a/src/Storages/System/StorageSystemMergeTreeSettings.h b/src/Storages/System/StorageSystemMergeTreeSettings.h index 48e83f0a880..e2913a7e55b 100644 --- a/src/Storages/System/StorageSystemMergeTreeSettings.h +++ b/src/Storages/System/StorageSystemMergeTreeSettings.h @@ -14,7 +14,7 @@ class Context; * which allows to get information about the current MergeTree settings. */ template -class SystemMergeTreeSettings final : public IStorageSystemOneBlock> +class SystemMergeTreeSettings final : public IStorageSystemOneBlock { public: std::string getName() const override { return replicated ? "SystemReplicatedMergeTreeSettings" : "SystemMergeTreeSettings"; } @@ -22,9 +22,9 @@ public: static ColumnsDescription getColumnsDescription(); protected: - using IStorageSystemOneBlock>::IStorageSystemOneBlock; + using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemMerges.cpp b/src/Storages/System/StorageSystemMerges.cpp index 4129e4c235b..0fca5dc84a2 100644 --- a/src/Storages/System/StorageSystemMerges.cpp +++ b/src/Storages/System/StorageSystemMerges.cpp @@ -16,12 +16,12 @@ ColumnsDescription StorageSystemMerges::getColumnsDescription() {"elapsed", std::make_shared(), "The time elapsed (in seconds) since the merge started."}, {"progress", std::make_shared(), "The percentage of completed work from 0 to 1."}, {"num_parts", std::make_shared(), "The number of parts to be merged."}, - {"source_part_names", std::make_shared(std::make_shared()), ""}, + {"source_part_names", std::make_shared(std::make_shared()), "The list of source parts names."}, {"result_part_name", std::make_shared(), "The name of the part that will be formed as the result of merging."}, - {"source_part_paths", std::make_shared(std::make_shared()), ""}, - {"result_part_path", std::make_shared(), ""}, - {"partition_id", std::make_shared()}, - {"partition", std::make_shared()}, + {"source_part_paths", std::make_shared(std::make_shared()), "The list of paths for each source part."}, + {"result_part_path", std::make_shared(), "The path of the part that will be formed as the result of merging."}, + {"partition_id", std::make_shared(), "The identifier of the partition where the merge is happening."}, + {"partition", std::make_shared(), "The name of the partition"}, {"is_mutation", std::make_shared(), "1 if this process is a part mutation."}, {"total_size_bytes_compressed", std::make_shared(), "The total size of the compressed data in the merged chunks."}, {"total_size_bytes_uncompressed", std::make_shared(), "The total size of compressed data in the merged chunks."}, @@ -39,7 +39,7 @@ ColumnsDescription StorageSystemMerges::getColumnsDescription() } -void StorageSystemMerges::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemMerges::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_tables = !access->isGranted(AccessType::SHOW_TABLES); diff --git a/src/Storages/System/StorageSystemMerges.h b/src/Storages/System/StorageSystemMerges.h index 961d28daf9a..fd9077c56d5 100644 --- a/src/Storages/System/StorageSystemMerges.h +++ b/src/Storages/System/StorageSystemMerges.h @@ -12,7 +12,7 @@ namespace DB class Context; -class StorageSystemMerges final : public IStorageSystemOneBlock +class StorageSystemMerges final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemMerges"; } @@ -22,7 +22,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemMetrics.cpp b/src/Storages/System/StorageSystemMetrics.cpp index 30544d66070..ae34a04cd87 100644 --- a/src/Storages/System/StorageSystemMetrics.cpp +++ b/src/Storages/System/StorageSystemMetrics.cpp @@ -25,7 +25,7 @@ ColumnsDescription StorageSystemMetrics::getColumnsDescription() return description; } -void StorageSystemMetrics::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemMetrics::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (size_t i = 0, end = CurrentMetrics::end(); i < end; ++i) { diff --git a/src/Storages/System/StorageSystemMetrics.h b/src/Storages/System/StorageSystemMetrics.h index ec0c67cf6b7..829fc231a79 100644 --- a/src/Storages/System/StorageSystemMetrics.h +++ b/src/Storages/System/StorageSystemMetrics.h @@ -11,7 +11,7 @@ class Context; /** Implements `metrics` system table, which provides information about the operation of the server. */ -class StorageSystemMetrics final : public IStorageSystemOneBlock +class StorageSystemMetrics final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemMetrics"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemModels.cpp b/src/Storages/System/StorageSystemModels.cpp index 03b659d79bd..8960d0625ba 100644 --- a/src/Storages/System/StorageSystemModels.cpp +++ b/src/Storages/System/StorageSystemModels.cpp @@ -13,17 +13,16 @@ namespace DB ColumnsDescription StorageSystemModels::getColumnsDescription() { - /// TODO: Fill in all the comments. return ColumnsDescription { - { "model_path", std::make_shared() }, - { "type", std::make_shared() }, - { "loading_start_time", std::make_shared() }, - { "loading_duration", std::make_shared() }, + { "model_path", std::make_shared(), "Path to trained model."}, + { "type", std::make_shared(), "Model type. Now catboost only."}, + { "loading_start_time", std::make_shared(), "The time when the loading of the model started."}, + { "loading_duration", std::make_shared(), "How much time did it take to load the model."}, }; } -void StorageSystemModels::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemModels::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto bridge_helper = std::make_unique(context); ExternalModelInfos infos = bridge_helper->listModels(); diff --git a/src/Storages/System/StorageSystemModels.h b/src/Storages/System/StorageSystemModels.h index 91fa3761743..419b623dac0 100644 --- a/src/Storages/System/StorageSystemModels.h +++ b/src/Storages/System/StorageSystemModels.h @@ -9,7 +9,7 @@ namespace DB class Context; -class StorageSystemModels final : public IStorageSystemOneBlock +class StorageSystemModels final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemModels"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemMoves.cpp b/src/Storages/System/StorageSystemMoves.cpp index 9b749218283..2c43043d820 100644 --- a/src/Storages/System/StorageSystemMoves.cpp +++ b/src/Storages/System/StorageSystemMoves.cpp @@ -23,7 +23,7 @@ ColumnsDescription StorageSystemMoves::getColumnsDescription() } -void StorageSystemMoves::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemMoves::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_tables = !access->isGranted(AccessType::SHOW_TABLES); diff --git a/src/Storages/System/StorageSystemMoves.h b/src/Storages/System/StorageSystemMoves.h index acdd9642f8f..45b1e9c6121 100644 --- a/src/Storages/System/StorageSystemMoves.h +++ b/src/Storages/System/StorageSystemMoves.h @@ -12,7 +12,7 @@ namespace DB class Context; -class StorageSystemMoves final : public IStorageSystemOneBlock +class StorageSystemMoves final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemMoves"; } @@ -22,7 +22,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemMutations.cpp b/src/Storages/System/StorageSystemMutations.cpp index 50545a55c7f..60b80e0b0ad 100644 --- a/src/Storages/System/StorageSystemMutations.cpp +++ b/src/Storages/System/StorageSystemMutations.cpp @@ -38,7 +38,7 @@ ColumnsDescription StorageSystemMutations::getColumnsDescription() "1 if the mutation is completed, " "0 if the mutation is still in process. " }, - { "is_killed", std::make_shared() }, + { "is_killed", std::make_shared(), "Only available in ClickHouse Cloud."}, { "latest_failed_part", std::make_shared(), "The name of the most recent part that could not be mutated."}, { "latest_fail_time", std::make_shared(), "The date and time of the most recent part mutation failure."}, { "latest_fail_reason", std::make_shared(), "The exception message that caused the most recent part mutation failure."}, @@ -46,7 +46,7 @@ ColumnsDescription StorageSystemMutations::getColumnsDescription() } -void StorageSystemMutations::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const +void StorageSystemMutations::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_databases = !access->isGranted(AccessType::SHOW_TABLES); @@ -100,7 +100,7 @@ void StorageSystemMutations::fillData(MutableColumns & res_columns, ContextPtr c { col_table, std::make_shared(), "table" }, }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) return; diff --git a/src/Storages/System/StorageSystemMutations.h b/src/Storages/System/StorageSystemMutations.h index 2db6e0c17f1..c60157cd853 100644 --- a/src/Storages/System/StorageSystemMutations.h +++ b/src/Storages/System/StorageSystemMutations.h @@ -11,7 +11,7 @@ class Context; /// Implements the `mutations` system table, which provides information about the status of mutations /// in the MergeTree tables. -class StorageSystemMutations final : public IStorageSystemOneBlock +class StorageSystemMutations final : public IStorageSystemOneBlock { public: String getName() const override { return "SystemMutations"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemNamedCollections.cpp b/src/Storages/System/StorageSystemNamedCollections.cpp index 25401bb751b..156fa5e5a9b 100644 --- a/src/Storages/System/StorageSystemNamedCollections.cpp +++ b/src/Storages/System/StorageSystemNamedCollections.cpp @@ -25,11 +25,11 @@ ColumnsDescription StorageSystemNamedCollections::getColumnsDescription() } StorageSystemNamedCollections::StorageSystemNamedCollections(const StorageID & table_id_) - : IStorageSystemOneBlock(table_id_) + : IStorageSystemOneBlock(table_id_, getColumnsDescription()) { } -void StorageSystemNamedCollections::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemNamedCollections::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto & access = context->getAccess(); diff --git a/src/Storages/System/StorageSystemNamedCollections.h b/src/Storages/System/StorageSystemNamedCollections.h index 596df99be83..ab302b400fc 100644 --- a/src/Storages/System/StorageSystemNamedCollections.h +++ b/src/Storages/System/StorageSystemNamedCollections.h @@ -5,7 +5,7 @@ namespace DB { -class StorageSystemNamedCollections final : public IStorageSystemOneBlock +class StorageSystemNamedCollections final : public IStorageSystemOneBlock { public: explicit StorageSystemNamedCollections(const StorageID & table_id_); @@ -15,7 +15,7 @@ public: static ColumnsDescription getColumnsDescription(); protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemNumbers.cpp b/src/Storages/System/StorageSystemNumbers.cpp index c5d8b307368..10898f79d10 100644 --- a/src/Storages/System/StorageSystemNumbers.cpp +++ b/src/Storages/System/StorageSystemNumbers.cpp @@ -19,6 +19,8 @@ StorageSystemNumbers::StorageSystemNumbers(const StorageID & table_id, bool mult : IStorage(table_id), multithreaded(multithreaded_), limit(limit_), offset(offset_) { StorageInMemoryMetadata storage_metadata; + /// This column doesn't have a comment, because otherwise it will be added to all the tables which were created via + /// CREATE TABLE test as numbers(5) storage_metadata.setColumns(ColumnsDescription({{"number", std::make_shared()}})); setInMemoryMetadata(storage_metadata); } diff --git a/src/Storages/System/StorageSystemOne.cpp b/src/Storages/System/StorageSystemOne.cpp index 3091ffdb51a..936d55e61a0 100644 --- a/src/Storages/System/StorageSystemOne.cpp +++ b/src/Storages/System/StorageSystemOne.cpp @@ -15,6 +15,8 @@ StorageSystemOne::StorageSystemOne(const StorageID & table_id_) : IStorage(table_id_) { StorageInMemoryMetadata storage_metadata; + /// This column doesn't have a comment, because otherwise it will be added to all tables created via: + /// CREATE TABLE test (dummy UInt8) ENGINE = Distributed(`default`, `system.one`) storage_metadata.setColumns(ColumnsDescription({{"dummy", std::make_shared()}})); setInMemoryMetadata(storage_metadata); } diff --git a/src/Storages/System/StorageSystemPartMovesBetweenShards.cpp b/src/Storages/System/StorageSystemPartMovesBetweenShards.cpp index 1a2646d3295..9cba92bca12 100644 --- a/src/Storages/System/StorageSystemPartMovesBetweenShards.cpp +++ b/src/Storages/System/StorageSystemPartMovesBetweenShards.cpp @@ -18,33 +18,32 @@ namespace DB ColumnsDescription StorageSystemPartMovesBetweenShards::getColumnsDescription() { - /// TODO: Fill in all the comments return ColumnsDescription { /// Table properties. - { "database", std::make_shared() }, - { "table", std::make_shared() }, + { "database", std::make_shared(), "The name of the database where move is performed."}, + { "table", std::make_shared(), "The name of the table where move is performed."}, /// Constant element properties. - { "task_name", std::make_shared() }, - { "task_uuid", std::make_shared() }, - { "create_time", std::make_shared() }, - { "part_name", std::make_shared() }, - { "part_uuid", std::make_shared() }, - { "to_shard", std::make_shared() }, - { "dst_part_name", std::make_shared() }, + { "task_name", std::make_shared(), "The name of the moving task."}, + { "task_uuid", std::make_shared(), "The identifier of the moving task."}, + { "create_time", std::make_shared(), "The time when the task was created."}, + { "part_name", std::make_shared(), "The name of the part which is in a process of moving."}, + { "part_uuid", std::make_shared(), "The UUID of the part which is in a process of moving."}, + { "to_shard", std::make_shared(), "The name of the destination shard."}, + { "dst_part_name", std::make_shared(), "The result part name."}, /// Processing status of item. - { "update_time", std::make_shared() }, - { "state", std::make_shared() }, - { "rollback", std::make_shared() }, - { "num_tries", std::make_shared() }, - { "last_exception", std::make_shared() }, + { "update_time", std::make_shared(), "The last time update was performed."}, + { "state", std::make_shared(), "The current state of the move."}, + { "rollback", std::make_shared(), "The flag which indicated whether the operation was rolled back."}, + { "num_tries", std::make_shared(), "The number of tries to complete the operation."}, + { "last_exception", std::make_shared(), "The last exception name if any."}, }; } -void StorageSystemPartMovesBetweenShards::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const +void StorageSystemPartMovesBetweenShards::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_databases = !access->isGranted(AccessType::SHOW_TABLES); @@ -95,7 +94,7 @@ void StorageSystemPartMovesBetweenShards::fillData(MutableColumns & res_columns, { col_table_to_filter, std::make_shared(), "table" }, }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) return; diff --git a/src/Storages/System/StorageSystemPartMovesBetweenShards.h b/src/Storages/System/StorageSystemPartMovesBetweenShards.h index 93a26bcd1b7..6a859d4de80 100644 --- a/src/Storages/System/StorageSystemPartMovesBetweenShards.h +++ b/src/Storages/System/StorageSystemPartMovesBetweenShards.h @@ -9,7 +9,7 @@ namespace DB class Context; -class StorageSystemPartMovesBetweenShards final : public IStorageSystemOneBlock +class StorageSystemPartMovesBetweenShards final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemShardMoves"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemParts.cpp b/src/Storages/System/StorageSystemParts.cpp index b6e4ee4161e..1b800fd64a9 100644 --- a/src/Storages/System/StorageSystemParts.cpp +++ b/src/Storages/System/StorageSystemParts.cpp @@ -94,33 +94,33 @@ StorageSystemParts::StorageSystemParts(const StorageID & table_id_) {"move_ttl_info.min", std::make_shared(std::make_shared()), "Array of date and time values. Each element describes the minimum key value for a TTL MOVE rule."}, {"move_ttl_info.max", std::make_shared(std::make_shared()), "Array of date and time values. Each element describes the maximum key value for a TTL MOVE rule."}, - {"default_compression_codec", std::make_shared()}, + {"default_compression_codec", std::make_shared(), "The name of the codec used to compress this data part (in case when there is no explicit codec for columns)."}, - {"recompression_ttl_info.expression", std::make_shared(std::make_shared())}, - {"recompression_ttl_info.min", std::make_shared(std::make_shared())}, - {"recompression_ttl_info.max", std::make_shared(std::make_shared())}, + {"recompression_ttl_info.expression", std::make_shared(std::make_shared()), "The TTL expression."}, + {"recompression_ttl_info.min", std::make_shared(std::make_shared()), "The minimum value of the calculated TTL expression within this part. Used to understand whether we have at least one row with expired TTL."}, + {"recompression_ttl_info.max", std::make_shared(std::make_shared()), "The maximum value of the calculated TTL expression within this part. Used to understand whether we have all rows with expired TTL."}, - {"group_by_ttl_info.expression", std::make_shared(std::make_shared())}, - {"group_by_ttl_info.min", std::make_shared(std::make_shared())}, - {"group_by_ttl_info.max", std::make_shared(std::make_shared())}, + {"group_by_ttl_info.expression", std::make_shared(std::make_shared()), "The TTL expression."}, + {"group_by_ttl_info.min", std::make_shared(std::make_shared()), "The minimum value of the calculated TTL expression within this part. Used to understand whether we have at least one row with expired TTL."}, + {"group_by_ttl_info.max", std::make_shared(std::make_shared()), "The maximum value of the calculated TTL expression within this part. Used to understand whether we have all rows with expired TTL."}, - {"rows_where_ttl_info.expression", std::make_shared(std::make_shared())}, - {"rows_where_ttl_info.min", std::make_shared(std::make_shared())}, - {"rows_where_ttl_info.max", std::make_shared(std::make_shared())}, + {"rows_where_ttl_info.expression", std::make_shared(std::make_shared()), "The TTL expression."}, + {"rows_where_ttl_info.min", std::make_shared(std::make_shared()), "The minimum value of the calculated TTL expression within this part. Used to understand whether we have at least one row with expired TTL."}, + {"rows_where_ttl_info.max", std::make_shared(std::make_shared()), "The maximum value of the calculated TTL expression within this part. Used to understand whether we have all rows with expired TTL."}, - {"projections", std::make_shared(std::make_shared())}, + {"projections", std::make_shared(std::make_shared()), "The list of projection names calculated for this part."}, - {"visible", std::make_shared()}, - {"creation_tid", getTransactionIDDataType()}, - {"removal_tid_lock", std::make_shared()}, - {"removal_tid", getTransactionIDDataType()}, - {"creation_csn", std::make_shared()}, - {"removal_csn", std::make_shared()}, + {"visible", std::make_shared(), "Flag which indicated whether this part is visible for SELECT queries."}, + {"creation_tid", getTransactionIDDataType(), "ID of transaction that has created/is trying to create this object."}, + {"removal_tid_lock", std::make_shared(), "Hash of removal_tid, used to lock an object for removal."}, + {"removal_tid", getTransactionIDDataType(), "ID of transaction that has removed/is trying to remove this object"}, + {"creation_csn", std::make_shared(), "CSN of transaction that has created this object"}, + {"removal_csn", std::make_shared(), "CSN of transaction that has removed this object"}, - {"has_lightweight_delete", std::make_shared()}, + {"has_lightweight_delete", std::make_shared(), "The flag which indicated whether the part has lightweight delete mask."}, - {"last_removal_attempt_time", std::make_shared()}, - {"removal_state", std::make_shared()}, + {"last_removal_attempt_time", std::make_shared(), "The last time the server tried to delete this part."}, + {"removal_state", std::make_shared(), "The current state of part removal process."}, } ) { diff --git a/src/Storages/System/StorageSystemPartsBase.cpp b/src/Storages/System/StorageSystemPartsBase.cpp index 3babdd5c7c1..6bdfdd357e8 100644 --- a/src/Storages/System/StorageSystemPartsBase.cpp +++ b/src/Storages/System/StorageSystemPartsBase.cpp @@ -2,6 +2,9 @@ #include #include #include +#include +#include +#include #include #include #include @@ -79,7 +82,7 @@ StoragesInfo::getProjectionParts(MergeTreeData::DataPartStateVector & state, boo return data->getProjectionPartsVectorForInternalUsage({State::Active}, &state); } -StoragesInfoStream::StoragesInfoStream(const SelectQueryInfo & query_info, ContextPtr context) +StoragesInfoStream::StoragesInfoStream(const ActionsDAG::Node * predicate, ContextPtr context) : StoragesInfoStreamBase(context) { /// Will apply WHERE to subset of columns and then add more columns. @@ -111,7 +114,7 @@ StoragesInfoStream::StoragesInfoStream(const SelectQueryInfo & query_info, Conte std::move(database_column_mut), std::make_shared(), "database")); /// Filter block_to_filter with column 'database'. - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block_to_filter, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block_to_filter, context); rows = block_to_filter.rows(); /// Block contains new columns, update database_column. @@ -190,7 +193,7 @@ StoragesInfoStream::StoragesInfoStream(const SelectQueryInfo & query_info, Conte if (rows) { /// Filter block_to_filter with columns 'database', 'table', 'engine', 'active'. - VirtualColumnUtils::filterBlockWithQuery(query_info.query, block_to_filter, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, block_to_filter, context); rows = block_to_filter.rows(); } @@ -200,8 +203,61 @@ StoragesInfoStream::StoragesInfoStream(const SelectQueryInfo & query_info, Conte storage_uuid_column = block_to_filter.getByName("uuid").column; } +class ReadFromSystemPartsBase : public SourceStepWithFilter +{ +public: + std::string getName() const override { return "ReadFromSystemPartsBase"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; -Pipe StorageSystemPartsBase::read( + ReadFromSystemPartsBase( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_, + std::vector columns_mask_, + bool has_state_column_); + + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +protected: + std::shared_ptr storage; + std::vector columns_mask; + const bool has_state_column; + const ActionsDAG::Node * predicate = nullptr; +}; + +ReadFromSystemPartsBase::ReadFromSystemPartsBase( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::shared_ptr storage_, + std::vector columns_mask_, + bool has_state_column_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , storage(std::move(storage_)) + , columns_mask(std::move(columns_mask_)) + , has_state_column(has_state_column_) +{ +} + +void ReadFromSystemPartsBase::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +void StorageSystemPartsBase::read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -212,29 +268,39 @@ Pipe StorageSystemPartsBase::read( { bool has_state_column = hasStateColumn(column_names, storage_snapshot); - auto stream = getStoragesInfoStream(query_info, context); - /// Create the result. Block sample = storage_snapshot->metadata->getSampleBlock(); auto [columns_mask, header] = getQueriedColumnsMaskAndHeader(sample, column_names); - MutableColumns res_columns = header.cloneEmptyColumns(); if (has_state_column) - res_columns.push_back(ColumnString::create()); + header.insert(ColumnWithTypeAndName(std::make_shared(), "_state")); + + auto this_ptr = std::static_pointer_cast(shared_from_this()); + + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(header), std::move(this_ptr), std::move(columns_mask), has_state_column); + + query_plan.addStep(std::move(reading)); +} + +void ReadFromSystemPartsBase::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) +{ + auto stream = storage->getStoragesInfoStream(predicate, context); + auto header = getOutputStream().header; + + MutableColumns res_columns = header.cloneEmptyColumns(); while (StoragesInfo info = stream->next()) { - processNextStorage(context, res_columns, columns_mask, info, has_state_column); + storage->processNextStorage(context, res_columns, columns_mask, info, has_state_column); } - if (has_state_column) - header.insert(ColumnWithTypeAndName(std::make_shared(), "_state")); - UInt64 num_rows = res_columns.at(0)->size(); Chunk chunk(std::move(res_columns), num_rows); - return Pipe(std::make_shared(std::move(header), std::move(chunk))); + pipeline.init(Pipe(std::make_shared(std::move(header), std::move(chunk)))); } @@ -259,12 +325,10 @@ StorageSystemPartsBase::StorageSystemPartsBase(const StorageID & table_id_, Colu StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(columns); setInMemoryMetadata(storage_metadata); + + VirtualColumnsDescription virtuals; + virtuals.addEphemeral("_state", std::make_shared(), ""); + setVirtuals(std::move(virtuals)); } -NamesAndTypesList StorageSystemPartsBase::getVirtuals() const -{ - return NamesAndTypesList{ - NameAndTypePair("_state", std::make_shared()) - }; -} } diff --git a/src/Storages/System/StorageSystemPartsBase.h b/src/Storages/System/StorageSystemPartsBase.h index 1127a8906e5..0527a99b841 100644 --- a/src/Storages/System/StorageSystemPartsBase.h +++ b/src/Storages/System/StorageSystemPartsBase.h @@ -3,6 +3,7 @@ #include #include #include +#include namespace DB @@ -114,7 +115,7 @@ protected: class StoragesInfoStream : public StoragesInfoStreamBase { public: - StoragesInfoStream(const SelectQueryInfo & query_info, ContextPtr context); + StoragesInfoStream(const ActionsDAG::Node * predicate, ContextPtr context); }; /** Implements system table 'parts' which allows to get information about data parts for tables of MergeTree family. @@ -122,7 +123,8 @@ public: class StorageSystemPartsBase : public IStorage { public: - Pipe read( + void read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -131,21 +133,21 @@ public: size_t max_block_size, size_t num_streams) override; - NamesAndTypesList getVirtuals() const override; - bool isSystemStorage() const override { return true; } private: static bool hasStateColumn(const Names & column_names, const StorageSnapshotPtr & storage_snapshot); protected: + friend class ReadFromSystemPartsBase; + const FormatSettings format_settings = {}; StorageSystemPartsBase(const StorageID & table_id_, ColumnsDescription && columns); - virtual std::unique_ptr getStoragesInfoStream(const SelectQueryInfo & query_info, ContextPtr context) + virtual std::unique_ptr getStoragesInfoStream(const ActionsDAG::Node * predicate, ContextPtr context) { - return std::make_unique(query_info, context); + return std::make_unique(predicate, context); } virtual void diff --git a/src/Storages/System/StorageSystemPartsColumns.cpp b/src/Storages/System/StorageSystemPartsColumns.cpp index 833a5e1ec16..f34b0e0cfda 100644 --- a/src/Storages/System/StorageSystemPartsColumns.cpp +++ b/src/Storages/System/StorageSystemPartsColumns.cpp @@ -22,59 +22,60 @@ namespace DB StorageSystemPartsColumns::StorageSystemPartsColumns(const StorageID & table_id_) : StorageSystemPartsBase(table_id_, ColumnsDescription{ - {"partition", std::make_shared()}, - {"name", std::make_shared()}, - {"uuid", std::make_shared()}, - {"part_type", std::make_shared()}, - {"active", std::make_shared()}, - {"marks", std::make_shared()}, - {"rows", std::make_shared()}, - {"bytes_on_disk", std::make_shared()}, - {"data_compressed_bytes", std::make_shared()}, - {"data_uncompressed_bytes", std::make_shared()}, - {"marks_bytes", std::make_shared()}, - {"modification_time", std::make_shared()}, - {"remove_time", std::make_shared()}, - {"refcount", std::make_shared()}, - {"min_date", std::make_shared()}, - {"max_date", std::make_shared()}, - {"min_time", std::make_shared()}, - {"max_time", std::make_shared()}, - {"partition_id", std::make_shared()}, - {"min_block_number", std::make_shared()}, - {"max_block_number", std::make_shared()}, - {"level", std::make_shared()}, - {"data_version", std::make_shared()}, - {"primary_key_bytes_in_memory", std::make_shared()}, - {"primary_key_bytes_in_memory_allocated", std::make_shared()}, + {"partition", std::make_shared(), "The partition name."}, + {"name", std::make_shared(), "Name of the data part."}, + {"uuid", std::make_shared(), "The parts UUID."}, + {"part_type", std::make_shared(), "The data part storing format. " + "Possible values: Wide — Each column is stored in a separate file in a filesystem, Compact — All columns are stored in one file in a filesystem."}, + {"active", std::make_shared(), "Flag that indicates whether the data part is active. If a data part is active, it’s used in a table. Otherwise, it’s deleted. Inactive data parts remain after merging."}, + {"marks", std::make_shared(), "The number of marks. To get the approximate number of rows in a data part, multiply marks by the index granularity (usually 8192) (this hint does not work for adaptive granularity)."}, + {"rows", std::make_shared(), "The number of rows."}, + {"bytes_on_disk", std::make_shared(), "Total size of all the data part files in bytes."}, + {"data_compressed_bytes", std::make_shared(), "Total size of compressed data in the data part. All the auxiliary files (for example, files with marks) are not included."}, + {"data_uncompressed_bytes", std::make_shared(), "Total size of uncompressed data in the data part. All the auxiliary files (for example, files with marks) are not included."}, + {"marks_bytes", std::make_shared(),"The size of the file with marks."}, + {"modification_time", std::make_shared(), "The time the directory with the data part was modified. This usually corresponds to the time of data part creation."}, + {"remove_time", std::make_shared(), "The time when the data part became inactive."}, + {"refcount", std::make_shared(), "The number of places where the data part is used. A value greater than 2 indicates that the data part is used in queries or merges."}, + {"min_date", std::make_shared(), "The minimum value for the Date column if that is included in the partition key."}, + {"max_date", std::make_shared(), "The maximum value for the Date column if that is included in the partition key."}, + {"min_time", std::make_shared(), "The minimum value for the DateTime column if that is included in the partition key."}, + {"max_time", std::make_shared(), "The maximum value for the DateTime column if that is included in the partition key."}, + {"partition_id", std::make_shared(), "ID of the partition."}, + {"min_block_number", std::make_shared(), "The minimum number of data parts that make up the current part after merging."}, + {"max_block_number", std::make_shared(), "The maximum number of data parts that make up the current part after merging."}, + {"level", std::make_shared(), "Depth of the merge tree. Zero means that the current part was created by insert rather than by merging other parts."}, + {"data_version", std::make_shared(), "Number that is used to determine which mutations should be applied to the data part (mutations with a version higher than data_version)."}, + {"primary_key_bytes_in_memory", std::make_shared(), "The amount of memory (in bytes) used by primary key values."}, + {"primary_key_bytes_in_memory_allocated", std::make_shared(), "The amount of memory (in bytes) reserved for primary key values."}, - {"database", std::make_shared()}, - {"table", std::make_shared()}, - {"engine", std::make_shared()}, - {"disk_name", std::make_shared()}, - {"path", std::make_shared()}, + {"database", std::make_shared(), "Name of the database."}, + {"table", std::make_shared(), "Name of the table."}, + {"engine", std::make_shared(), "Name of the table engine without parameters."}, + {"disk_name", std::make_shared(), "Name of a disk that stores the data part."}, + {"path", std::make_shared(), "Absolute path to the folder with data part files."}, - {"column", std::make_shared()}, - {"type", std::make_shared()}, - {"column_position", std::make_shared()}, - {"default_kind", std::make_shared()}, - {"default_expression", std::make_shared()}, - {"column_bytes_on_disk", std::make_shared()}, - {"column_data_compressed_bytes", std::make_shared()}, - {"column_data_uncompressed_bytes", std::make_shared()}, - {"column_marks_bytes", std::make_shared()}, - {"column_modification_time", std::make_shared(std::make_shared())}, + {"column", std::make_shared(), "Name of the column."}, + {"type", std::make_shared(), "Column type."}, + {"column_position", std::make_shared(), "Ordinal position of a column in a table starting with 1."}, + {"default_kind", std::make_shared(), "Expression type (DEFAULT, MATERIALIZED, ALIAS) for the default value, or an empty string if it is not defined."}, + {"default_expression", std::make_shared(), "Expression for the default value, or an empty string if it is not defined."}, + {"column_bytes_on_disk", std::make_shared(), "Total size of the column in bytes."}, + {"column_data_compressed_bytes", std::make_shared(), "Total size of the compressed data in the column, in bytes."}, + {"column_data_uncompressed_bytes", std::make_shared(), "Total size of the decompressed data in the column, in bytes."}, + {"column_marks_bytes", std::make_shared(), "The size of the marks for column, in bytes."}, + {"column_modification_time", std::make_shared(std::make_shared()), "The last time the column was modified."}, - {"serialization_kind", std::make_shared()}, - {"substreams", std::make_shared(std::make_shared())}, - {"filenames", std::make_shared(std::make_shared())}, - {"subcolumns.names", std::make_shared(std::make_shared())}, - {"subcolumns.types", std::make_shared(std::make_shared())}, - {"subcolumns.serializations", std::make_shared(std::make_shared())}, - {"subcolumns.bytes_on_disk", std::make_shared(std::make_shared())}, - {"subcolumns.data_compressed_bytes", std::make_shared(std::make_shared())}, - {"subcolumns.data_uncompressed_bytes", std::make_shared(std::make_shared())}, - {"subcolumns.marks_bytes", std::make_shared(std::make_shared())}, + {"serialization_kind", std::make_shared(), "Kind of serialization of a column"}, + {"substreams", std::make_shared(std::make_shared()), "Names of substreams to which column is serialized"}, + {"filenames", std::make_shared(std::make_shared()), "Names of files for each substream of a column respectively"}, + {"subcolumns.names", std::make_shared(std::make_shared()), "Names of subcolumns of a column"}, + {"subcolumns.types", std::make_shared(std::make_shared()), "Types of subcolumns of a column"}, + {"subcolumns.serializations", std::make_shared(std::make_shared()), "Kinds of serialization of subcolumns of a column"}, + {"subcolumns.bytes_on_disk", std::make_shared(std::make_shared()), "Sizes in bytes for each subcolumn"}, + {"subcolumns.data_compressed_bytes", std::make_shared(std::make_shared()), "Sizes of the compressed data for each subcolumn, in bytes"}, + {"subcolumns.data_uncompressed_bytes", std::make_shared(std::make_shared()), "Sizes of the decompressed data for each subcolumn, in bytes"}, + {"subcolumns.marks_bytes", std::make_shared(std::make_shared()), "Sizes of the marks for each subcolumn of a column, in bytes"}, } ) { diff --git a/src/Storages/System/StorageSystemPrivileges.cpp b/src/Storages/System/StorageSystemPrivileges.cpp index a2d3e699c17..58dcf62115e 100644 --- a/src/Storages/System/StorageSystemPrivileges.cpp +++ b/src/Storages/System/StorageSystemPrivileges.cpp @@ -68,18 +68,26 @@ const std::vector> & StorageSystemPrivileges::getAccess ColumnsDescription StorageSystemPrivileges::getColumnsDescription() { - /// TODO: Fill in all the comments. - return ColumnsDescription - { - {"privilege", std::make_shared(getAccessTypeEnumValues())}, - {"aliases", std::make_shared(std::make_shared())}, - {"level", std::make_shared(std::make_shared(getLevelEnumValues()))}, - {"parent_group", std::make_shared(std::make_shared(getAccessTypeEnumValues()))}, + return ColumnsDescription{ + {"privilege", + std::make_shared(getAccessTypeEnumValues()), + "Name of a privilege which can be used in the GRANT command."}, + {"aliases", + std::make_shared(std::make_shared()), + "List of aliases which can be used instead of the name of the privilege."}, + {"level", + std::make_shared(std::make_shared(getLevelEnumValues())), + "Level of the privilege. GLOBAL privileges can be granted only globally (ON *.*), DATABASE privileges can be granted " + "on a specific database (ON .*) or globally (ON *.*), TABLE privileges can be granted either on a specific table or " + "on a specific database or globally, and COLUMN privileges can be granted like TABLE privileges but also allow to specify columns."}, + {"parent_group", std::make_shared(std::make_shared(getAccessTypeEnumValues())), + "Parent privilege - if the parent privilege is granted then all its children privileges are considered as granted too." + }, }; } -void StorageSystemPrivileges::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemPrivileges::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { size_t column_index = 0; auto & column_access_type = assert_cast(*res_columns[column_index++]).getData(); diff --git a/src/Storages/System/StorageSystemPrivileges.h b/src/Storages/System/StorageSystemPrivileges.h index 4441cf78d5c..eaef7f0db6d 100644 --- a/src/Storages/System/StorageSystemPrivileges.h +++ b/src/Storages/System/StorageSystemPrivileges.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `privileges` system table, which allows you to get information about access types. -class StorageSystemPrivileges final : public IStorageSystemOneBlock +class StorageSystemPrivileges final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemPrivileges"; } @@ -17,7 +17,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemProcesses.cpp b/src/Storages/System/StorageSystemProcesses.cpp index 6702e68b81e..bef98e59687 100644 --- a/src/Storages/System/StorageSystemProcesses.cpp +++ b/src/Storages/System/StorageSystemProcesses.cpp @@ -21,35 +21,35 @@ ColumnsDescription StorageSystemProcesses::getColumnsDescription() { auto description = ColumnsDescription { - {"is_initial_query", std::make_shared()}, + {"is_initial_query", std::make_shared(), "Whether this query comes directly from user or was issues by ClickHouse server in a scope of distributed query execution."}, {"user", std::make_shared(), "The user who made the query. Keep in mind that for distributed processing, queries are sent to remote servers under the default user. The field contains the username for a specific query, not for a query that this query initiated."}, {"query_id", std::make_shared(), "Query ID, if defined."}, - {"address", DataTypeFactory::instance().get("IPv6"), "The IP address the request was made from. The same for distributed processing. To track where a distributed query was originally made from, look at system.processes on the query requestor server."}, - {"port", std::make_shared()}, + {"address", DataTypeFactory::instance().get("IPv6"), "The IP address the query was made from. The same for distributed processing. To track where a distributed query was originally made from, look at system.processes on the query requestor server."}, + {"port", std::make_shared(), "The client port the query was made from."}, - {"initial_user", std::make_shared()}, - {"initial_query_id", std::make_shared()}, - {"initial_address", DataTypeFactory::instance().get("IPv6")}, - {"initial_port", std::make_shared()}, + {"initial_user", std::make_shared(), "Name of the user who ran the initial query (for distributed query execution)."}, + {"initial_query_id", std::make_shared(), "ID of the initial query (for distributed query execution)."}, + {"initial_address", DataTypeFactory::instance().get("IPv6"), "IP address that the parent query was launched from."}, + {"initial_port", std::make_shared(), "The client port that was used to make the parent query."}, - {"interface", std::make_shared()}, + {"interface", std::make_shared(), "The interface which was used to send the query. TCP = 1, HTTP = 2, GRPC = 3, MYSQL = 4, POSTGRESQL = 5, LOCAL = 6, TCP_INTERSERVER = 7."}, - {"os_user", std::make_shared()}, - {"client_hostname", std::make_shared()}, - {"client_name", std::make_shared()}, - {"client_revision", std::make_shared()}, - {"client_version_major", std::make_shared()}, - {"client_version_minor", std::make_shared()}, - {"client_version_patch", std::make_shared()}, + {"os_user", std::make_shared(), "Operating system username who runs clickhouse-client."}, + {"client_hostname", std::make_shared(), "Hostname of the client machine where the clickhouse-client or another TCP client is run."}, + {"client_name", std::make_shared(), "The clickhouse-client or another TCP client name."}, + {"client_revision", std::make_shared(), "Revision of the clickhouse-client or another TCP client."}, + {"client_version_major", std::make_shared(), "Major version of the clickhouse-client or another TCP client."}, + {"client_version_minor", std::make_shared(), "Minor version of the clickhouse-client or another TCP client."}, + {"client_version_patch", std::make_shared(), "Patch component of the clickhouse-client or another TCP client version."}, - {"http_method", std::make_shared()}, - {"http_user_agent", std::make_shared()}, - {"http_referer", std::make_shared()}, - {"forwarded_for", std::make_shared()}, + {"http_method", std::make_shared(), "HTTP method that initiated the query. Possible values: 0 — The query was launched from the TCP interface. 1 — GET method was used. 2 — POST method was used."}, + {"http_user_agent", std::make_shared(), "HTTP header UserAgent passed in the HTTP query."}, + {"http_referer", std::make_shared(), "HTTP header Referer passed in the HTTP query (contains an absolute or partial address of the page making the query)."}, + {"forwarded_for", std::make_shared(), "HTTP header X-Forwarded-For passed in the HTTP query."}, - {"quota_key", std::make_shared()}, - {"distributed_depth", std::make_shared()}, + {"quota_key", std::make_shared(), "The quota key specified in the quotas setting (see keyed)."}, + {"distributed_depth", std::make_shared(), "The number of times query was retransmitted between server nodes internally."}, {"elapsed", std::make_shared(), "The time in seconds since request execution started."}, {"is_cancelled", std::make_shared(), "Was query cancelled."}, @@ -57,18 +57,18 @@ ColumnsDescription StorageSystemProcesses::getColumnsDescription() {"read_rows", std::make_shared(), "The number of rows read from the table. For distributed processing, on the requestor server, this is the total for all remote servers."}, {"read_bytes", std::make_shared(), "The number of uncompressed bytes read from the table. For distributed processing, on the requestor server, this is the total for all remote servers."}, {"total_rows_approx", std::make_shared(), "The approximation of the total number of rows that should be read. For distributed processing, on the requestor server, this is the total for all remote servers. It can be updated during request processing, when new sources to process become known."}, - {"written_rows", std::make_shared()}, - {"written_bytes", std::make_shared()}, + {"written_rows", std::make_shared(), "The amount of rows written to the storage."}, + {"written_bytes", std::make_shared(), "The amount of bytes written to the storage."}, {"memory_usage", std::make_shared(), "Amount of RAM the query uses. It might not include some types of dedicated memory"}, - {"peak_memory_usage", std::make_shared()}, + {"peak_memory_usage", std::make_shared(), "The current peak of memory usage."}, {"query", std::make_shared(), "The query text. For INSERT, it does not include the data to insert."}, - {"query_kind", std::make_shared()}, + {"query_kind", std::make_shared(), "The type of the query - SELECT, INSERT, etc."}, - {"thread_ids", std::make_shared(std::make_shared())}, - {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared())}, - {"Settings", std::make_shared(std::make_shared(), std::make_shared())}, + {"thread_ids", std::make_shared(std::make_shared()), "The list of identificators of all threads which executed this query."}, + {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared()), "ProfileEvents calculated for this query."}, + {"Settings", std::make_shared(std::make_shared(), std::make_shared()), "The list of modified user-level settings."}, - {"current_database", std::make_shared()}, + {"current_database", std::make_shared(), "The name of the current database."}, }; description.setAliases({ @@ -81,7 +81,7 @@ ColumnsDescription StorageSystemProcesses::getColumnsDescription() return description; } -void StorageSystemProcesses::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemProcesses::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { ProcessList::Info info = context->getProcessList().getInfo(true, true, true); diff --git a/src/Storages/System/StorageSystemProcesses.h b/src/Storages/System/StorageSystemProcesses.h index 3017f9fd367..eb241d4b1ae 100644 --- a/src/Storages/System/StorageSystemProcesses.h +++ b/src/Storages/System/StorageSystemProcesses.h @@ -11,7 +11,7 @@ class Context; /** Implements `processes` system table, which allows you to get information about the queries that are currently executing. */ -class StorageSystemProcesses final : public IStorageSystemOneBlock +class StorageSystemProcesses final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemProcesses"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemProjectionParts.cpp b/src/Storages/System/StorageSystemProjectionParts.cpp index 016705f4e66..4bdcea67313 100644 --- a/src/Storages/System/StorageSystemProjectionParts.cpp +++ b/src/Storages/System/StorageSystemProjectionParts.cpp @@ -19,71 +19,71 @@ namespace DB StorageSystemProjectionParts::StorageSystemProjectionParts(const StorageID & table_id_) : StorageSystemPartsBase(table_id_, ColumnsDescription{ - {"partition", std::make_shared()}, - {"name", std::make_shared()}, - {"part_type", std::make_shared()}, - {"parent_name", std::make_shared()}, - {"parent_uuid", std::make_shared()}, - {"parent_part_type", std::make_shared()}, - {"active", std::make_shared()}, - {"marks", std::make_shared()}, - {"rows", std::make_shared()}, - {"bytes_on_disk", std::make_shared()}, - {"data_compressed_bytes", std::make_shared()}, - {"data_uncompressed_bytes", std::make_shared()}, - {"marks_bytes", std::make_shared()}, - {"parent_marks", std::make_shared()}, - {"parent_rows", std::make_shared()}, - {"parent_bytes_on_disk", std::make_shared()}, - {"parent_data_compressed_bytes", std::make_shared()}, - {"parent_data_uncompressed_bytes", std::make_shared()}, - {"parent_marks_bytes", std::make_shared()}, - {"modification_time", std::make_shared()}, - {"remove_time", std::make_shared()}, - {"refcount", std::make_shared()}, - {"min_date", std::make_shared()}, - {"max_date", std::make_shared()}, - {"min_time", std::make_shared()}, - {"max_time", std::make_shared()}, - {"partition_id", std::make_shared()}, - {"min_block_number", std::make_shared()}, - {"max_block_number", std::make_shared()}, - {"level", std::make_shared()}, - {"data_version", std::make_shared()}, - {"primary_key_bytes_in_memory", std::make_shared()}, - {"primary_key_bytes_in_memory_allocated", std::make_shared()}, - {"is_frozen", std::make_shared()}, + {"partition", std::make_shared(), "The partition name."}, + {"name", std::make_shared(), "Name of the data part."}, + {"part_type", std::make_shared(), "The data part storing format. Possible Values: Wide (a file per column) and Compact (a single file for all columns)."}, + {"parent_name", std::make_shared(), "The name of the source (parent) data part."}, + {"parent_uuid", std::make_shared(), "The UUID of the source (parent) data part."}, + {"parent_part_type", std::make_shared(), "The source (parent) data part storing format."}, + {"active", std::make_shared(), "Flag that indicates whether the data part is active. If a data part is active, it's used in a table. Otherwise, it's about to be deleted. Inactive data parts appear after merging and mutating operations."}, + {"marks", std::make_shared(), "The number of marks. To get the approximate number of rows in a data part, multiply marks by the index granularity (usually 8192) (this hint does not work for adaptive granularity)."}, + {"rows", std::make_shared(), "The number of rows."}, + {"bytes_on_disk", std::make_shared(), "Total size of all the data part files in bytes."}, + {"data_compressed_bytes", std::make_shared(), "Total size of compressed data in the data part. All the auxiliary files (for example, files with marks) are not included."}, + {"data_uncompressed_bytes", std::make_shared(), "Total size of uncompressed data in the data part. All the auxiliary files (for example, files with marks) are not included."}, + {"marks_bytes", std::make_shared(), "The size of the file with marks."}, + {"parent_marks", std::make_shared(), "The number of marks in the source (parent) part."}, + {"parent_rows", std::make_shared(), "The number of rows in the source (parent) part."}, + {"parent_bytes_on_disk", std::make_shared(), "Total size of all the source (parent) data part files in bytes."}, + {"parent_data_compressed_bytes", std::make_shared(), "Total size of compressed data in the source (parent) data part."}, + {"parent_data_uncompressed_bytes", std::make_shared(), "Total size of uncompressed data in the source (parent) data part."}, + {"parent_marks_bytes", std::make_shared(), "The size of the file with marks in the source (parent) data part."}, + {"modification_time", std::make_shared(), "The time the directory with the data part was modified. This usually corresponds to the time of data part creation."}, + {"remove_time", std::make_shared(), "The time when the data part became inactive."}, + {"refcount", std::make_shared(), "The number of places where the data part is used. A value greater than 2 indicates that the data part is used in queries or merges."}, + {"min_date", std::make_shared(), "The minimum value of the date key in the data part."}, + {"max_date", std::make_shared(), "The maximum value of the date key in the data part."}, + {"min_time", std::make_shared(), "The minimum value of the date and time key in the data part."}, + {"max_time", std::make_shared(), "The maximum value of the date and time key in the data part."}, + {"partition_id", std::make_shared(), "ID of the partition."}, + {"min_block_number", std::make_shared(), "The minimum number of data parts that make up the current part after merging."}, + {"max_block_number", std::make_shared(), "The maximum number of data parts that make up the current part after merging."}, + {"level", std::make_shared(), "Depth of the merge tree. Zero means that the current part was created by insert rather than by merging other parts."}, + {"data_version", std::make_shared(), "Number that is used to determine which mutations should be applied to the data part (mutations with a version higher than data_version)."}, + {"primary_key_bytes_in_memory", std::make_shared(), "The amount of memory (in bytes) used by primary key values."}, + {"primary_key_bytes_in_memory_allocated", std::make_shared(), "The amount of memory (in bytes) reserved for primary key values."}, + {"is_frozen", std::make_shared(), "Flag that shows that a partition data backup exists. 1, the backup exists. 0, the backup does not exist. "}, - {"database", std::make_shared()}, - {"table", std::make_shared()}, - {"engine", std::make_shared()}, - {"disk_name", std::make_shared()}, - {"path", std::make_shared()}, + {"database", std::make_shared(), "Name of the database."}, + {"table", std::make_shared(), "Name of the table."}, + {"engine", std::make_shared(), "Name of the table engine without parameters."}, + {"disk_name", std::make_shared(), "Name of a disk that stores the data part."}, + {"path", std::make_shared(), "Absolute path to the folder with data part files."}, - {"hash_of_all_files", std::make_shared()}, - {"hash_of_uncompressed_files", std::make_shared()}, - {"uncompressed_hash_of_compressed_files", std::make_shared()}, + {"hash_of_all_files", std::make_shared(), "sipHash128 of compressed files."}, + {"hash_of_uncompressed_files", std::make_shared(), "sipHash128 of uncompressed files (files with marks, index file etc.)."}, + {"uncompressed_hash_of_compressed_files", std::make_shared(), "sipHash128 of data in the compressed files as if they were uncompressed."}, - {"delete_ttl_info_min", std::make_shared()}, - {"delete_ttl_info_max", std::make_shared()}, + {"delete_ttl_info_min", std::make_shared(), "The minimum value of the date and time key for TTL DELETE rule."}, + {"delete_ttl_info_max", std::make_shared(), "The maximum value of the date and time key for TTL DELETE rule."}, - {"move_ttl_info.expression", std::make_shared(std::make_shared())}, - {"move_ttl_info.min", std::make_shared(std::make_shared())}, - {"move_ttl_info.max", std::make_shared(std::make_shared())}, + {"move_ttl_info.expression", std::make_shared(std::make_shared()), "Array of expressions. Each expression defines a TTL MOVE rule."}, + {"move_ttl_info.min", std::make_shared(std::make_shared()), "Array of date and time values. Each element describes the minimum key value for a TTL MOVE rule."}, + {"move_ttl_info.max", std::make_shared(std::make_shared()), "Array of date and time values. Each element describes the maximum key value for a TTL MOVE rule."}, - {"default_compression_codec", std::make_shared()}, + {"default_compression_codec", std::make_shared(), "The name of the codec used to compress this data part (in case when there is no explicit codec for columns)."}, - {"recompression_ttl_info.expression", std::make_shared(std::make_shared())}, - {"recompression_ttl_info.min", std::make_shared(std::make_shared())}, - {"recompression_ttl_info.max", std::make_shared(std::make_shared())}, + {"recompression_ttl_info.expression", std::make_shared(std::make_shared()), "The TTL expression."}, + {"recompression_ttl_info.min", std::make_shared(std::make_shared()), "The minimum value of the calculated TTL expression within this part. Used to understand whether we have at least one row with expired TTL."}, + {"recompression_ttl_info.max", std::make_shared(std::make_shared()), "The maximum value of the calculated TTL expression within this part. Used to understand whether we have all rows with expired TTL."}, - {"group_by_ttl_info.expression", std::make_shared(std::make_shared())}, - {"group_by_ttl_info.min", std::make_shared(std::make_shared())}, - {"group_by_ttl_info.max", std::make_shared(std::make_shared())}, + {"group_by_ttl_info.expression", std::make_shared(std::make_shared()), "The TTL expression."}, + {"group_by_ttl_info.min", std::make_shared(std::make_shared()), "The minimum value of the calculated TTL expression within this part. Used to understand whether we have at least one row with expired TTL."}, + {"group_by_ttl_info.max", std::make_shared(std::make_shared()), "The maximum value of the calculated TTL expression within this part. Used to understand whether we have all rows with expired TTL."}, - {"rows_where_ttl_info.expression", std::make_shared(std::make_shared())}, - {"rows_where_ttl_info.min", std::make_shared(std::make_shared())}, - {"rows_where_ttl_info.max", std::make_shared(std::make_shared())} + {"rows_where_ttl_info.expression", std::make_shared(std::make_shared()), "The TTL expression."}, + {"rows_where_ttl_info.min", std::make_shared(std::make_shared()), "The minimum value of the calculated TTL expression within this part. Used to understand whether we have at least one row with expired TTL."}, + {"rows_where_ttl_info.max", std::make_shared(std::make_shared()), "The maximum value of the calculated TTL expression within this part. Used to understand whether we have all rows with expired TTL."}, } ) { diff --git a/src/Storages/System/StorageSystemProjectionPartsColumns.cpp b/src/Storages/System/StorageSystemProjectionPartsColumns.cpp index 2ff25f86366..9521605688d 100644 --- a/src/Storages/System/StorageSystemProjectionPartsColumns.cpp +++ b/src/Storages/System/StorageSystemProjectionPartsColumns.cpp @@ -19,56 +19,56 @@ namespace DB StorageSystemProjectionPartsColumns::StorageSystemProjectionPartsColumns(const StorageID & table_id_) : StorageSystemPartsBase(table_id_, ColumnsDescription{ - {"partition", std::make_shared()}, - {"name", std::make_shared()}, - {"part_type", std::make_shared()}, - {"parent_name", std::make_shared()}, - {"parent_uuid", std::make_shared()}, - {"parent_part_type", std::make_shared()}, - {"active", std::make_shared()}, - {"marks", std::make_shared()}, - {"rows", std::make_shared()}, - {"bytes_on_disk", std::make_shared()}, - {"data_compressed_bytes", std::make_shared()}, - {"data_uncompressed_bytes", std::make_shared()}, - {"marks_bytes", std::make_shared()}, - {"parent_marks", std::make_shared()}, - {"parent_rows", std::make_shared()}, - {"parent_bytes_on_disk", std::make_shared()}, - {"parent_data_compressed_bytes", std::make_shared()}, - {"parent_data_uncompressed_bytes", std::make_shared()}, - {"parent_marks_bytes", std::make_shared()}, - {"modification_time", std::make_shared()}, - {"remove_time", std::make_shared()}, - {"refcount", std::make_shared()}, - {"min_date", std::make_shared()}, - {"max_date", std::make_shared()}, - {"min_time", std::make_shared()}, - {"max_time", std::make_shared()}, - {"partition_id", std::make_shared()}, - {"min_block_number", std::make_shared()}, - {"max_block_number", std::make_shared()}, - {"level", std::make_shared()}, - {"data_version", std::make_shared()}, - {"primary_key_bytes_in_memory", std::make_shared()}, - {"primary_key_bytes_in_memory_allocated", std::make_shared()}, + {"partition", std::make_shared(), "The partition name. "}, + {"name", std::make_shared(), "Name of the data part."}, + {"part_type", std::make_shared(), "The data part storing format."}, + {"parent_name", std::make_shared(), "The name of the source (parent) data part."}, + {"parent_uuid", std::make_shared(), "The UUID of the source (parent) data part."}, + {"parent_part_type", std::make_shared(), "The source (parent) data part storing format."}, + {"active", std::make_shared(), "Flag that indicates whether the data part is active"}, + {"marks", std::make_shared(), "The number of marks."}, + {"rows", std::make_shared(), "The number of rows."}, + {"bytes_on_disk", std::make_shared(), "Total size of all the data part files in bytes."}, + {"data_compressed_bytes", std::make_shared(), "Total size of compressed data in the data part. All the auxiliary files (for example, files with marks) are not included."}, + {"data_uncompressed_bytes", std::make_shared(), "Total size of uncompressed data in the data part. All the auxiliary files (for example, files with marks) are not included."}, + {"marks_bytes", std::make_shared(), "The size of the file with marks."}, + {"parent_marks", std::make_shared(), "The number of marks in the source (parent) part."}, + {"parent_rows", std::make_shared(), "The number of rows in the source (parent) part."}, + {"parent_bytes_on_disk", std::make_shared(), "Total size of all the source (parent) data part files in bytes."}, + {"parent_data_compressed_bytes", std::make_shared(), "Total size of compressed data in the source (parent) data part."}, + {"parent_data_uncompressed_bytes", std::make_shared(), "Total size of uncompressed data in the source (parent) data part."}, + {"parent_marks_bytes", std::make_shared(), "The size of the file with marks in the source (parent) data part."}, + {"modification_time", std::make_shared(), "The time the directory with the data part was modified. This usually corresponds to the time of data part creation."}, + {"remove_time", std::make_shared(), "The time when the data part became inactive."}, + {"refcount", std::make_shared(), "The number of places where the data part is used. A value greater than 2 indicates that the data part is used in queries or merges."}, + {"min_date", std::make_shared(), "The minimum value for the Date column if that is included in the partition key."}, + {"max_date", std::make_shared(), "The maximum value for the Date column if that is included in the partition key."}, + {"min_time", std::make_shared(), "The minimum value for the DateTime column if that is included in the partition key."}, + {"max_time", std::make_shared(), "The maximum value for the DateTime column if that is included in the partition key."}, + {"partition_id", std::make_shared(), "ID of the partition."}, + {"min_block_number", std::make_shared(), "The minimum number of data parts that make up the current part after merging."}, + {"max_block_number", std::make_shared(), "The maximum number of data parts that make up the current part after merging."}, + {"level", std::make_shared(), "Depth of the merge tree. Zero means that the current part was created by insert rather than by merging other parts."}, + {"data_version", std::make_shared(), "Number that is used to determine which mutations should be applied to the data part (mutations with a version higher than data_version)."}, + {"primary_key_bytes_in_memory", std::make_shared(), "The amount of memory (in bytes) used by primary key values."}, + {"primary_key_bytes_in_memory_allocated", std::make_shared(), "The amount of memory (in bytes) reserved for primary key values."}, - {"database", std::make_shared()}, - {"table", std::make_shared()}, - {"engine", std::make_shared()}, - {"disk_name", std::make_shared()}, - {"path", std::make_shared()}, + {"database", std::make_shared(), "Name of the database."}, + {"table", std::make_shared(), "Name of the table."}, + {"engine", std::make_shared(), "Name of the table engine without parameters."}, + {"disk_name", std::make_shared(), "Name of a disk that stores the data part."}, + {"path", std::make_shared(), "Absolute path to the folder with data part files."}, - {"column", std::make_shared()}, - {"type", std::make_shared()}, - {"column_position", std::make_shared()}, - {"default_kind", std::make_shared()}, - {"default_expression", std::make_shared()}, - {"column_bytes_on_disk", std::make_shared()}, - {"column_data_compressed_bytes", std::make_shared()}, - {"column_data_uncompressed_bytes", std::make_shared()}, - {"column_marks_bytes", std::make_shared()}, - {"column_modification_time", std::make_shared(std::make_shared())}, + {"column", std::make_shared(), "Name of the column."}, + {"type", std::make_shared(), "Column type."}, + {"column_position", std::make_shared(), "Ordinal position of a column in a table starting with 1."}, + {"default_kind", std::make_shared(), "Expression type (DEFAULT, MATERIALIZED, ALIAS) for the default value, or an empty string if it is not defined."}, + {"default_expression", std::make_shared(), "Expression for the default value, or an empty string if it is not defined."}, + {"column_bytes_on_disk", std::make_shared(), "Total size of the column in bytes."}, + {"column_data_compressed_bytes", std::make_shared(), "Total size of compressed data in the column, in bytes."}, + {"column_data_uncompressed_bytes", std::make_shared(), "Total size of the decompressed data in the column, in bytes."}, + {"column_marks_bytes", std::make_shared(), "The size of the column with marks, in bytes."}, + {"column_modification_time", std::make_shared(std::make_shared()), "The last time the column was modified."}, } ) { diff --git a/src/Storages/System/StorageSystemQueryCache.cpp b/src/Storages/System/StorageSystemQueryCache.cpp index 03111755904..e454012da3a 100644 --- a/src/Storages/System/StorageSystemQueryCache.cpp +++ b/src/Storages/System/StorageSystemQueryCache.cpp @@ -24,11 +24,11 @@ ColumnsDescription StorageSystemQueryCache::getColumnsDescription() } StorageSystemQueryCache::StorageSystemQueryCache(const StorageID & table_id_) - : IStorageSystemOneBlock(table_id_) + : IStorageSystemOneBlock(table_id_, getColumnsDescription()) { } -void StorageSystemQueryCache::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemQueryCache::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { QueryCachePtr query_cache = context->getQueryCache(); diff --git a/src/Storages/System/StorageSystemQueryCache.h b/src/Storages/System/StorageSystemQueryCache.h index 08ad30afb81..22856c2b1bb 100644 --- a/src/Storages/System/StorageSystemQueryCache.h +++ b/src/Storages/System/StorageSystemQueryCache.h @@ -5,7 +5,7 @@ namespace DB { -class StorageSystemQueryCache final : public IStorageSystemOneBlock +class StorageSystemQueryCache final : public IStorageSystemOneBlock { public: explicit StorageSystemQueryCache(const StorageID & table_id_); @@ -15,7 +15,7 @@ public: static ColumnsDescription getColumnsDescription(); protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemQuotaLimits.cpp b/src/Storages/System/StorageSystemQuotaLimits.cpp index f125a990a88..65acfba0c1b 100644 --- a/src/Storages/System/StorageSystemQuotaLimits.cpp +++ b/src/Storages/System/StorageSystemQuotaLimits.cpp @@ -68,14 +68,15 @@ ColumnsDescription StorageSystemQuotaLimits::getColumnsDescription() data_type = std::make_shared(); else data_type = std::make_shared(); - result.add({column_name, std::make_shared(data_type)}); + + result.add({column_name, std::make_shared(data_type), type_info.max_allowed_usage_description}); } return result; } -void StorageSystemQuotaLimits::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemQuotaLimits::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemQuotaLimits.h b/src/Storages/System/StorageSystemQuotaLimits.h index acc977d0df7..a8385e878ca 100644 --- a/src/Storages/System/StorageSystemQuotaLimits.h +++ b/src/Storages/System/StorageSystemQuotaLimits.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `quota_limits` system table, which allows you to get information about the limits set for quotas. -class StorageSystemQuotaLimits final : public IStorageSystemOneBlock +class StorageSystemQuotaLimits final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemQuotaLimits"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemQuotaUsage.cpp b/src/Storages/System/StorageSystemQuotaUsage.cpp index a91e8b7b2c1..da9c174b0d3 100644 --- a/src/Storages/System/StorageSystemQuotaUsage.cpp +++ b/src/Storages/System/StorageSystemQuotaUsage.cpp @@ -81,15 +81,15 @@ ColumnsDescription StorageSystemQuotaUsage::getColumnsDescriptionImpl(bool add_c data_type = std::make_shared(); else data_type = std::make_shared(); - description.add({column_name, std::make_shared(data_type)}); - description.add({String("max_") + column_name, std::make_shared(data_type)}); + description.add({column_name, std::make_shared(data_type), type_info.current_usage_description}); + description.add({String("max_") + column_name, std::make_shared(data_type), type_info.max_allowed_usage_description}); } return description; } -void StorageSystemQuotaUsage::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemQuotaUsage::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemQuotaUsage.h b/src/Storages/System/StorageSystemQuotaUsage.h index a3109e9ca31..3100098fe87 100644 --- a/src/Storages/System/StorageSystemQuotaUsage.h +++ b/src/Storages/System/StorageSystemQuotaUsage.h @@ -12,7 +12,7 @@ struct QuotaUsage; /** Implements the `quota_usage` system table, which allows you to get information about * how the current user uses the quota. */ -class StorageSystemQuotaUsage final : public IStorageSystemOneBlock +class StorageSystemQuotaUsage final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemQuotaUsage"; } @@ -23,7 +23,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemQuotas.cpp b/src/Storages/System/StorageSystemQuotas.cpp index ee302f2f163..641bbb319d5 100644 --- a/src/Storages/System/StorageSystemQuotas.cpp +++ b/src/Storages/System/StorageSystemQuotas.cpp @@ -66,7 +66,7 @@ ColumnsDescription StorageSystemQuotas::getColumnsDescription() } -void StorageSystemQuotas::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemQuotas::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemQuotas.h b/src/Storages/System/StorageSystemQuotas.h index cafd8b921fa..76e6f1df536 100644 --- a/src/Storages/System/StorageSystemQuotas.h +++ b/src/Storages/System/StorageSystemQuotas.h @@ -9,7 +9,7 @@ class Context; /** Implements the `quotas` system tables, which allows you to get information about quotas. */ -class StorageSystemQuotas final : public IStorageSystemOneBlock +class StorageSystemQuotas final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemQuotas"; } @@ -20,7 +20,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemQuotasUsage.cpp b/src/Storages/System/StorageSystemQuotasUsage.cpp index ed22f73dd50..1587048e7e7 100644 --- a/src/Storages/System/StorageSystemQuotasUsage.cpp +++ b/src/Storages/System/StorageSystemQuotasUsage.cpp @@ -13,7 +13,7 @@ ColumnsDescription StorageSystemQuotasUsage::getColumnsDescription() return StorageSystemQuotaUsage::getColumnsDescriptionImpl(/* add_column_is_current = */ true); } -void StorageSystemQuotasUsage::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemQuotasUsage::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemQuotasUsage.h b/src/Storages/System/StorageSystemQuotasUsage.h index ecdc62865d1..516e722f7df 100644 --- a/src/Storages/System/StorageSystemQuotasUsage.h +++ b/src/Storages/System/StorageSystemQuotasUsage.h @@ -10,7 +10,7 @@ class Context; /** Implements the `quotas_usage` system table, which allows you to get information about * how all users use the quotas. */ -class StorageSystemQuotasUsage final : public IStorageSystemOneBlock +class StorageSystemQuotasUsage final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemQuotasUsage"; } @@ -18,7 +18,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemRemoteDataPaths.cpp b/src/Storages/System/StorageSystemRemoteDataPaths.cpp index 87b7a84e8ba..7cc1951fc05 100644 --- a/src/Storages/System/StorageSystemRemoteDataPaths.cpp +++ b/src/Storages/System/StorageSystemRemoteDataPaths.cpp @@ -6,6 +6,7 @@ #include #include #include +#include #include #include @@ -19,14 +20,14 @@ StorageSystemRemoteDataPaths::StorageSystemRemoteDataPaths(const StorageID & tab StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription( { - {"disk_name", std::make_shared()}, - {"path", std::make_shared()}, - {"cache_base_path", std::make_shared()}, - {"local_path", std::make_shared()}, - {"remote_path", std::make_shared()}, - {"size", std::make_shared()}, - {"common_prefix_for_blobs", std::make_shared()}, - {"cache_paths", std::make_shared(std::make_shared())}, + {"disk_name", std::make_shared(), "Disk name."}, + {"path", std::make_shared(), "Disk path."}, + {"cache_base_path", std::make_shared(), "Base directory of cache files."}, + {"local_path", std::make_shared(), "Path of ClickHouse file, also used as metadata path."}, + {"remote_path", std::make_shared(), "Blob path in object storage, with which ClickHouse file is associated with."}, + {"size", std::make_shared(), "Size of the file (compressed)."}, + {"common_prefix_for_blobs", std::make_shared(), "Common prefix for blobs in object storage."}, + {"cache_paths", std::make_shared(std::make_shared()), "Cache files for corresponding blob."}, })); setInMemoryMetadata(storage_metadata); } diff --git a/src/Storages/System/StorageSystemReplicas.cpp b/src/Storages/System/StorageSystemReplicas.cpp index eeb3db342b4..af8d67cbc21 100644 --- a/src/Storages/System/StorageSystemReplicas.cpp +++ b/src/Storages/System/StorageSystemReplicas.cpp @@ -14,6 +14,9 @@ #include #include #include +#include +#include +#include #include #include #include @@ -182,13 +185,6 @@ public: , requests_with_zk_fields(max_threads) {} - Pipe read( - const Names & column_names, - const StorageSnapshotPtr & storage_snapshot, - SelectQueryInfo & query_info, - ContextPtr context); - -private: StatusRequestsPool requests_without_zk_fields; StatusRequestsPool requests_with_zk_fields; }; @@ -200,49 +196,98 @@ StorageSystemReplicas::StorageSystemReplicas(const StorageID & table_id_) { StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription({ - { "database", std::make_shared() }, - { "table", std::make_shared() }, - { "engine", std::make_shared() }, - { "is_leader", std::make_shared() }, - { "can_become_leader", std::make_shared() }, - { "is_readonly", std::make_shared() }, - { "is_session_expired", std::make_shared() }, - { "future_parts", std::make_shared() }, - { "parts_to_check", std::make_shared() }, - { "zookeeper_name", std::make_shared() }, - { "zookeeper_path", std::make_shared() }, - { "replica_name", std::make_shared() }, - { "replica_path", std::make_shared() }, - { "columns_version", std::make_shared() }, - { "queue_size", std::make_shared() }, - { "inserts_in_queue", std::make_shared() }, - { "merges_in_queue", std::make_shared() }, - { "part_mutations_in_queue", std::make_shared() }, - { "queue_oldest_time", std::make_shared() }, - { "inserts_oldest_time", std::make_shared() }, - { "merges_oldest_time", std::make_shared() }, - { "part_mutations_oldest_time", std::make_shared() }, - { "oldest_part_to_get", std::make_shared() }, - { "oldest_part_to_merge_to", std::make_shared() }, - { "oldest_part_to_mutate_to", std::make_shared() }, - { "log_max_index", std::make_shared() }, - { "log_pointer", std::make_shared() }, - { "last_queue_update", std::make_shared() }, - { "absolute_delay", std::make_shared() }, - { "total_replicas", std::make_shared() }, - { "active_replicas", std::make_shared() }, - { "lost_part_count", std::make_shared() }, - { "last_queue_update_exception", std::make_shared() }, - { "zookeeper_exception", std::make_shared() }, - { "replica_is_active", std::make_shared(std::make_shared(), std::make_shared()) } + { "database", std::make_shared(), "Database name."}, + { "table", std::make_shared(), "Table name."}, + { "engine", std::make_shared(), "Table engine name."}, + { "is_leader", std::make_shared(), "Whether the replica is the leader. Multiple replicas can be leaders at the same time. " + "A replica can be prevented from becoming a leader using the merge_tree setting replicated_can_become_leader. " + "The leaders are responsible for scheduling background merges. " + "Note that writes can be performed to any replica that is available and has a session in ZK, regardless of whether it is a leader."}, + { "can_become_leader", std::make_shared(), "Whether the replica can be a leader."}, + { "is_readonly", std::make_shared(), "Whether the replica is in read-only mode. This mode is turned on if the config does not have sections with ClickHouse Keeper, " + "if an unknown error occurred when reinitializing sessions in ClickHouse Keeper, and during session reinitialization in ClickHouse Keeper."}, + { "is_session_expired", std::make_shared(), "Whether the session with ClickHouse Keeper has expired. Basically the same as `is_readonly`."}, + { "future_parts", std::make_shared(), "The number of data parts that will appear as the result of INSERTs or merges that haven't been done yet."}, + { "parts_to_check", std::make_shared(), "The number of data parts in the queue for verification. A part is put in the verification queue if there is suspicion that it might be damaged."}, + { "zookeeper_name", std::make_shared(), "The name of the the [Zoo]Keeper cluster (possibly auxiliary one) where the table's metadata is stored"}, + { "zookeeper_path", std::make_shared(), "Path to table data in ClickHouse Keeper."}, + { "replica_name", std::make_shared(), "Replica name in ClickHouse Keeper. Different replicas of the same table have different names."}, + { "replica_path", std::make_shared(), "Path to replica data in ClickHouse Keeper. The same as concatenating 'zookeeper_path/replicas/replica_path'."}, + { "columns_version", std::make_shared(), "Version number of the table structure. Indicates how many times ALTER was performed. " + "If replicas have different versions, it means some replicas haven't made all of the ALTERs yet."}, + { "queue_size", std::make_shared(), "Size of the queue for operations waiting to be performed. Operations include inserting blocks of data, merges, and certain other actions. It usually coincides with future_parts."}, + { "inserts_in_queue", std::make_shared(), "Number of inserts of blocks of data that need to be made. Insertions are usually replicated fairly quickly. If this number is large, it means something is wrong."}, + { "merges_in_queue", std::make_shared(), "The number of merges waiting to be made. Sometimes merges are lengthy, so this value may be greater than zero for a long time."}, + { "part_mutations_in_queue", std::make_shared(), "The number of mutations waiting to be made."}, + { "queue_oldest_time", std::make_shared(), "If `queue_size` greater than 0, shows when the oldest operation was added to the queue."}, + { "inserts_oldest_time", std::make_shared(), "See `queue_oldest_time`."}, + { "merges_oldest_time", std::make_shared(), "See `queue_oldest_time`."}, + { "part_mutations_oldest_time", std::make_shared(), "See `queue_oldest_time`."}, + { "oldest_part_to_get", std::make_shared(), "The name of the part to fetch from other replicas obtained from the oldest GET_PARTS entry in the replication queue."}, + { "oldest_part_to_merge_to", std::make_shared(), "The result part name to merge to obtained from the oldest MERGE_PARTS entry in the replication queue."}, + { "oldest_part_to_mutate_to", std::make_shared(), "The result part name to mutate to obtained from the oldest MUTATE_PARTS entry in the replication queue."}, + { "log_max_index", std::make_shared(), "Maximum entry number in the log of general activity."}, + { "log_pointer", std::make_shared(), "Maximum entry number in the log of general activity that the replica copied to its execution queue, plus one. " + "If log_pointer is much smaller than log_max_index, something is wrong."}, + { "last_queue_update", std::make_shared(), "When the queue was updated last time."}, + { "absolute_delay", std::make_shared(), "How big lag in seconds the current replica has."}, + { "total_replicas", std::make_shared(), "The total number of known replicas of this table."}, + { "active_replicas", std::make_shared(), "The number of replicas of this table that have a session in ClickHouse Keeper (i.e., the number of functioning replicas)."}, + { "lost_part_count", std::make_shared(), "The number of data parts lost in the table by all replicas in total since table creation. Value is persisted in ClickHouse Keeper and can only increase."}, + { "last_queue_update_exception", std::make_shared(), "When the queue contains broken entries. Especially important when ClickHouse breaks backward compatibility between versions and log entries written by newer versions aren't parseable by old versions."}, + { "zookeeper_exception", std::make_shared(), "The last exception message, got if the error happened when fetching the info from ClickHouse Keeper."}, + { "replica_is_active", std::make_shared(std::make_shared(), std::make_shared()), "Map between replica name and is replica active."} })); setInMemoryMetadata(storage_metadata); } StorageSystemReplicas::~StorageSystemReplicas() = default; +class ReadFromSystemReplicas : public SourceStepWithFilter +{ +public: + std::string getName() const override { return "ReadFromSystemReplicas"; } + void initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override; -Pipe StorageSystemReplicas::read( + ReadFromSystemReplicas( + const Names & column_names_, + const SelectQueryInfo & query_info_, + const StorageSnapshotPtr & storage_snapshot_, + const ContextPtr & context_, + Block sample_block, + std::map> replicated_tables_, + bool with_zk_fields_, + std::shared_ptr impl_) + : SourceStepWithFilter( + DataStream{.header = std::move(sample_block)}, + column_names_, + query_info_, + storage_snapshot_, + context_) + , replicated_tables(std::move(replicated_tables_)) + , with_zk_fields(with_zk_fields_) + , impl(std::move(impl_)) + { + } + + void applyFilters(ActionDAGNodes added_filter_nodes) override; + +private: + std::map> replicated_tables; + const bool with_zk_fields; + std::shared_ptr impl; + const ActionsDAG::Node * predicate = nullptr; +}; + +void ReadFromSystemReplicas::applyFilters(ActionDAGNodes added_filter_nodes) +{ + filter_actions_dag = ActionsDAG::buildFilterActionsDAG(added_filter_nodes.nodes); + if (filter_actions_dag) + predicate = filter_actions_dag->getOutputs().at(0); +} + +void StorageSystemReplicas::read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -250,15 +295,6 @@ Pipe StorageSystemReplicas::read( QueryProcessingStage::Enum /*processed_stage*/, const size_t /*max_block_size*/, const size_t /*num_streams*/) -{ - return impl->read(column_names, storage_snapshot, query_info, context); -} - -Pipe StorageSystemReplicasImpl::read( - const Names & column_names, - const StorageSnapshotPtr & storage_snapshot, - SelectQueryInfo & query_info, - ContextPtr context) { storage_snapshot->check(column_names); @@ -305,6 +341,18 @@ Pipe StorageSystemReplicasImpl::read( } } + auto header = storage_snapshot->metadata->getSampleBlock(); + auto reading = std::make_unique( + column_names, query_info, storage_snapshot, + std::move(context), std::move(header), std::move(replicated_tables), with_zk_fields, impl); // /*std::move(this_ptr),*/ std::move(columns_mask), max_block_size); + + query_plan.addStep(std::move(reading)); +} + +void ReadFromSystemReplicas::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) +{ + auto header = getOutputStream().header; + MutableColumnPtr col_database_mut = ColumnString::create(); MutableColumnPtr col_table_mut = ColumnString::create(); MutableColumnPtr col_engine_mut = ColumnString::create(); @@ -332,10 +380,14 @@ Pipe StorageSystemReplicasImpl::read( { col_engine, std::make_shared(), "engine" }, }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) - return {}; + { + auto source = std::make_shared(std::move(header)); + pipeline.init(Pipe(std::move(source))); + return; + } col_database = filtered_block.getByName("database").column; col_table = filtered_block.getByName("table").column; @@ -347,7 +399,7 @@ Pipe StorageSystemReplicasImpl::read( size_t tables_size = col_database->size(); /// Use separate queues for requests with and without ZooKeeper fields. - StatusRequestsPool & get_status_requests = with_zk_fields ? requests_with_zk_fields : requests_without_zk_fields; + StatusRequestsPool & get_status_requests = with_zk_fields ? impl->requests_with_zk_fields : impl->requests_without_zk_fields; QueryStatusPtr query_status = context ? context->getProcessListElement() : nullptr; @@ -435,7 +487,7 @@ Pipe StorageSystemReplicasImpl::read( UInt64 num_rows = fin_columns.at(0)->size(); Chunk chunk(std::move(fin_columns), num_rows); - return Pipe(std::make_shared(storage_snapshot->metadata->getSampleBlock(), std::move(chunk))); + pipeline.init(Pipe(std::make_shared(header, std::move(chunk)))); } diff --git a/src/Storages/System/StorageSystemReplicas.h b/src/Storages/System/StorageSystemReplicas.h index f6bdfc937ea..11c5371310f 100644 --- a/src/Storages/System/StorageSystemReplicas.h +++ b/src/Storages/System/StorageSystemReplicas.h @@ -1,7 +1,7 @@ #pragma once #include -#include +#include namespace DB @@ -20,7 +20,8 @@ public: std::string getName() const override { return "SystemReplicas"; } - Pipe read( + void read( + QueryPlan & query_plan, const Names & column_names, const StorageSnapshotPtr & storage_snapshot, SelectQueryInfo & query_info, @@ -32,7 +33,7 @@ public: bool isSystemStorage() const override { return true; } private: - std::unique_ptr impl; + std::shared_ptr impl; }; } diff --git a/src/Storages/System/StorageSystemReplicatedFetches.cpp b/src/Storages/System/StorageSystemReplicatedFetches.cpp index e643cb9f86a..6913665a8d2 100644 --- a/src/Storages/System/StorageSystemReplicatedFetches.cpp +++ b/src/Storages/System/StorageSystemReplicatedFetches.cpp @@ -34,7 +34,7 @@ ColumnsDescription StorageSystemReplicatedFetches::getColumnsDescription() }; } -void StorageSystemReplicatedFetches::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemReplicatedFetches::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_tables = !access->isGranted(AccessType::SHOW_TABLES); diff --git a/src/Storages/System/StorageSystemReplicatedFetches.h b/src/Storages/System/StorageSystemReplicatedFetches.h index a176912cac0..dba9124b39d 100644 --- a/src/Storages/System/StorageSystemReplicatedFetches.h +++ b/src/Storages/System/StorageSystemReplicatedFetches.h @@ -10,7 +10,7 @@ namespace DB class Context; /// system.replicated_fetches table. Takes data from context.getReplicatedFetchList() -class StorageSystemReplicatedFetches final : public IStorageSystemOneBlock +class StorageSystemReplicatedFetches final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemReplicatedFetches"; } @@ -20,7 +20,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemReplicationQueue.cpp b/src/Storages/System/StorageSystemReplicationQueue.cpp index 194a2ae6fb8..14b641f46c7 100644 --- a/src/Storages/System/StorageSystemReplicationQueue.cpp +++ b/src/Storages/System/StorageSystemReplicationQueue.cpp @@ -62,7 +62,7 @@ ColumnsDescription StorageSystemReplicationQueue::getColumnsDescription() } -void StorageSystemReplicationQueue::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const +void StorageSystemReplicationQueue::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const { const auto access = context->getAccess(); const bool check_access_for_databases = !access->isGranted(AccessType::SHOW_TABLES); @@ -113,7 +113,7 @@ void StorageSystemReplicationQueue::fillData(MutableColumns & res_columns, Conte { col_table_to_filter, std::make_shared(), "table" }, }; - VirtualColumnUtils::filterBlockWithQuery(query_info.query, filtered_block, context); + VirtualColumnUtils::filterBlockWithPredicate(predicate, filtered_block, context); if (!filtered_block.rows()) return; diff --git a/src/Storages/System/StorageSystemReplicationQueue.h b/src/Storages/System/StorageSystemReplicationQueue.h index 003e4eeb927..a9e57851be1 100644 --- a/src/Storages/System/StorageSystemReplicationQueue.h +++ b/src/Storages/System/StorageSystemReplicationQueue.h @@ -11,7 +11,7 @@ class Context; /** Implements the `replication_queue` system table, which allows you to view the replication queues for the replicated tables. */ -class StorageSystemReplicationQueue final : public IStorageSystemOneBlock +class StorageSystemReplicationQueue final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemReplicationQueue"; } @@ -20,7 +20,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node * predicate, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemRoleGrants.cpp b/src/Storages/System/StorageSystemRoleGrants.cpp index 241481275c4..e5baeed4873 100644 --- a/src/Storages/System/StorageSystemRoleGrants.cpp +++ b/src/Storages/System/StorageSystemRoleGrants.cpp @@ -40,7 +40,7 @@ ColumnsDescription StorageSystemRoleGrants::getColumnsDescription() } -void StorageSystemRoleGrants::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemRoleGrants::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemRoleGrants.h b/src/Storages/System/StorageSystemRoleGrants.h index 969f82f85d5..f82aece3f24 100644 --- a/src/Storages/System/StorageSystemRoleGrants.h +++ b/src/Storages/System/StorageSystemRoleGrants.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `role_grants` system table, which allows you to get information about granted roles. -class StorageSystemRoleGrants final : public IStorageSystemOneBlock +class StorageSystemRoleGrants final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemRoleGrants"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemRoles.cpp b/src/Storages/System/StorageSystemRoles.cpp index ec26a50db77..9bfddc25ebf 100644 --- a/src/Storages/System/StorageSystemRoles.cpp +++ b/src/Storages/System/StorageSystemRoles.cpp @@ -25,7 +25,7 @@ ColumnsDescription StorageSystemRoles::getColumnsDescription() } -void StorageSystemRoles::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemRoles::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemRoles.h b/src/Storages/System/StorageSystemRoles.h index 2b4ae93a932..e96bb70d9cd 100644 --- a/src/Storages/System/StorageSystemRoles.h +++ b/src/Storages/System/StorageSystemRoles.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `roles` system table, which allows you to get information about roles. -class StorageSystemRoles final : public IStorageSystemOneBlock +class StorageSystemRoles final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemRoles"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemRowPolicies.cpp b/src/Storages/System/StorageSystemRowPolicies.cpp index 5a959cdf9af..93c5ba60a7f 100644 --- a/src/Storages/System/StorageSystemRowPolicies.cpp +++ b/src/Storages/System/StorageSystemRowPolicies.cpp @@ -38,8 +38,8 @@ ColumnsDescription StorageSystemRowPolicies::getColumnsDescription() for (auto filter_type : collections::range(RowPolicyFilterType::MAX)) { - const String & column_name = RowPolicyFilterTypeInfo::get(filter_type).name; - description.add({column_name, std::make_shared(std::make_shared())}); + const auto & filter_type_info = RowPolicyFilterTypeInfo::get(filter_type); + description.add({filter_type_info.name, std::make_shared(std::make_shared()), filter_type_info.description}); } description.add({"is_restrictive", std::make_shared(), @@ -61,7 +61,7 @@ ColumnsDescription StorageSystemRowPolicies::getColumnsDescription() } -void StorageSystemRowPolicies::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemRowPolicies::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemRowPolicies.h b/src/Storages/System/StorageSystemRowPolicies.h index f8aa5618126..cdbc4731000 100644 --- a/src/Storages/System/StorageSystemRowPolicies.h +++ b/src/Storages/System/StorageSystemRowPolicies.h @@ -10,7 +10,7 @@ class Context; /// Implements `row_policies` system table, which allows you to get information about row policies. -class StorageSystemRowPolicies final : public IStorageSystemOneBlock +class StorageSystemRowPolicies final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemRowPolicies"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemS3Queue.cpp b/src/Storages/System/StorageSystemS3Queue.cpp index 3637734b225..a6bb7da2b6e 100644 --- a/src/Storages/System/StorageSystemS3Queue.cpp +++ b/src/Storages/System/StorageSystemS3Queue.cpp @@ -25,23 +25,23 @@ ColumnsDescription StorageSystemS3Queue::getColumnsDescription() /// TODO: Fill in all the comments return ColumnsDescription { - {"zookeeper_path", std::make_shared()}, - {"file_name", std::make_shared()}, - {"rows_processed", std::make_shared()}, - {"status", std::make_shared()}, - {"processing_start_time", std::make_shared(std::make_shared())}, - {"processing_end_time", std::make_shared(std::make_shared())}, - {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared())}, - {"exception", std::make_shared()}, + {"zookeeper_path", std::make_shared(), "Path in zookeeper to S3Queue metadata"}, + {"file_name", std::make_shared(), "File name of a file which is being processed by S3Queue"}, + {"rows_processed", std::make_shared(), "Currently processed number of rows"}, + {"status", std::make_shared(), "Status of processing: Processed, Processing, Failed"}, + {"processing_start_time", std::make_shared(std::make_shared()), "Time at which processing of the file started"}, + {"processing_end_time", std::make_shared(std::make_shared()), "Time at which processing of the file ended"}, + {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared()), "Profile events collected during processing of the file"}, + {"exception", std::make_shared(), "Exception which happened during processing"}, }; } StorageSystemS3Queue::StorageSystemS3Queue(const StorageID & table_id_) - : IStorageSystemOneBlock(table_id_) + : IStorageSystemOneBlock(table_id_, getColumnsDescription()) { } -void StorageSystemS3Queue::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemS3Queue::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (const auto & [zookeeper_path, metadata] : S3QueueMetadataFactory::instance().getAll()) { diff --git a/src/Storages/System/StorageSystemS3Queue.h b/src/Storages/System/StorageSystemS3Queue.h index 1dc5c521941..c89e18f458f 100644 --- a/src/Storages/System/StorageSystemS3Queue.h +++ b/src/Storages/System/StorageSystemS3Queue.h @@ -7,7 +7,7 @@ namespace DB { -class StorageSystemS3Queue final : public IStorageSystemOneBlock +class StorageSystemS3Queue final : public IStorageSystemOneBlock { public: explicit StorageSystemS3Queue(const StorageID & table_id_); @@ -17,7 +17,7 @@ public: static ColumnsDescription getColumnsDescription(); protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemScheduler.cpp b/src/Storages/System/StorageSystemScheduler.cpp index ba07d44dbf9..cae42011fc5 100644 --- a/src/Storages/System/StorageSystemScheduler.cpp +++ b/src/Storages/System/StorageSystemScheduler.cpp @@ -80,7 +80,7 @@ ColumnsDescription StorageSystemScheduler::getColumnsDescription() } -void StorageSystemScheduler::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemScheduler::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { context->getResourceManager()->forEachNode([&] (const String & resource, const String & path, const String & type, const SchedulerNodePtr & node) { diff --git a/src/Storages/System/StorageSystemScheduler.h b/src/Storages/System/StorageSystemScheduler.h index 1de72a85e9b..c6a259e5b51 100644 --- a/src/Storages/System/StorageSystemScheduler.h +++ b/src/Storages/System/StorageSystemScheduler.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `system.scheduler` table, which allows you to get information about scheduling nodes. -class StorageSystemScheduler final : public IStorageSystemOneBlock +class StorageSystemScheduler final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemScheduler"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemSchemaInferenceCache.cpp b/src/Storages/System/StorageSystemSchemaInferenceCache.cpp index 1426ea83800..634089bd1cd 100644 --- a/src/Storages/System/StorageSystemSchemaInferenceCache.cpp +++ b/src/Storages/System/StorageSystemSchemaInferenceCache.cpp @@ -72,7 +72,7 @@ static void fillDataImpl(MutableColumns & res_columns, SchemaCache & schema_cach } } -void StorageSystemSchemaInferenceCache::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemSchemaInferenceCache::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { fillDataImpl(res_columns, StorageFile::getSchemaCache(context), "File"); #if USE_AWS_S3 diff --git a/src/Storages/System/StorageSystemSchemaInferenceCache.h b/src/Storages/System/StorageSystemSchemaInferenceCache.h index e6d306f8252..3e12f4b850b 100644 --- a/src/Storages/System/StorageSystemSchemaInferenceCache.h +++ b/src/Storages/System/StorageSystemSchemaInferenceCache.h @@ -6,7 +6,7 @@ namespace DB { -class StorageSystemSchemaInferenceCache final : public IStorageSystemOneBlock +class StorageSystemSchemaInferenceCache final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemSettingsChanges"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemServerSettings.cpp b/src/Storages/System/StorageSystemServerSettings.cpp index f390985546b..b75f4280877 100644 --- a/src/Storages/System/StorageSystemServerSettings.cpp +++ b/src/Storages/System/StorageSystemServerSettings.cpp @@ -55,7 +55,7 @@ ColumnsDescription StorageSystemServerSettings::getColumnsDescription() }; } -void StorageSystemServerSettings::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemServerSettings::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// When the server configuration file is periodically re-loaded from disk, the server components (e.g. memory tracking) are updated /// with new the setting values but the settings themselves are not stored between re-loads. As a result, if one wants to know the diff --git a/src/Storages/System/StorageSystemServerSettings.h b/src/Storages/System/StorageSystemServerSettings.h index 276f21d674b..03c363c9920 100644 --- a/src/Storages/System/StorageSystemServerSettings.h +++ b/src/Storages/System/StorageSystemServerSettings.h @@ -11,7 +11,7 @@ class Context; /** implements system table "settings", which allows to get information about the current settings. */ -class StorageSystemServerSettings final : public IStorageSystemOneBlock +class StorageSystemServerSettings final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemServerSettings"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemSettings.cpp b/src/Storages/System/StorageSystemSettings.cpp index 43877582af6..b437108b00e 100644 --- a/src/Storages/System/StorageSystemSettings.cpp +++ b/src/Storages/System/StorageSystemSettings.cpp @@ -34,7 +34,7 @@ ColumnsDescription StorageSystemSettings::getColumnsDescription() }; } -void StorageSystemSettings::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemSettings::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const Settings & settings = context->getSettingsRef(); auto constraints_and_current_profiles = context->getSettingsConstraintsAndCurrentProfiles(); diff --git a/src/Storages/System/StorageSystemSettings.h b/src/Storages/System/StorageSystemSettings.h index 6749f9b20a4..fae0d69ada8 100644 --- a/src/Storages/System/StorageSystemSettings.h +++ b/src/Storages/System/StorageSystemSettings.h @@ -11,7 +11,7 @@ class Context; /** implements system table "settings", which allows to get information about the current settings. */ -class StorageSystemSettings final : public IStorageSystemOneBlock +class StorageSystemSettings final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemSettings"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemSettingsChanges.cpp b/src/Storages/System/StorageSystemSettingsChanges.cpp index b1942ea9ac6..de47ec52031 100644 --- a/src/Storages/System/StorageSystemSettingsChanges.cpp +++ b/src/Storages/System/StorageSystemSettingsChanges.cpp @@ -12,7 +12,7 @@ ColumnsDescription StorageSystemSettingsChanges::getColumnsDescription() /// TODO: Fill in all the comments return ColumnsDescription { - {"version", std::make_shared()}, + {"version", std::make_shared(), "The ClickHouse server version."}, {"changes", std::make_shared(std::make_shared( DataTypes{ @@ -20,11 +20,11 @@ ColumnsDescription StorageSystemSettingsChanges::getColumnsDescription() std::make_shared(), std::make_shared(), std::make_shared()}, - Names{"name", "previous_value", "new_value", "reason"}))}, + Names{"name", "previous_value", "new_value", "reason"})), "The list of changes in settings which changed the behaviour of ClickHouse."}, }; } -void StorageSystemSettingsChanges::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemSettingsChanges::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (auto it = settings_changes_history.rbegin(); it != settings_changes_history.rend(); ++it) { diff --git a/src/Storages/System/StorageSystemSettingsChanges.h b/src/Storages/System/StorageSystemSettingsChanges.h index 3a1a8ce23d1..9d8899797fe 100644 --- a/src/Storages/System/StorageSystemSettingsChanges.h +++ b/src/Storages/System/StorageSystemSettingsChanges.h @@ -12,7 +12,7 @@ class Context; /** Implements system table "settings_changes", which allows to get information * about the settings changes through different ClickHouse versions. */ -class StorageSystemSettingsChanges final : public IStorageSystemOneBlock +class StorageSystemSettingsChanges final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemSettingsChanges"; } @@ -22,7 +22,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemSettingsProfileElements.cpp b/src/Storages/System/StorageSystemSettingsProfileElements.cpp index a530bd2c1b3..2af3e6dfd05 100644 --- a/src/Storages/System/StorageSystemSettingsProfileElements.cpp +++ b/src/Storages/System/StorageSystemSettingsProfileElements.cpp @@ -42,7 +42,7 @@ ColumnsDescription StorageSystemSettingsProfileElements::getColumnsDescription() {"value", std::make_shared(std::make_shared()), "Setting value."}, {"min", std::make_shared(std::make_shared()), "The minimum value of the setting. NULL if not set."}, {"max", std::make_shared(std::make_shared()), "The maximum value of the setting. NULL if not set."}, - {"writability", std::make_shared(std::make_shared(getSettingConstraintWritabilityEnumValues()))}, + {"writability", std::make_shared(std::make_shared(getSettingConstraintWritabilityEnumValues())), "The property which shows whether a setting can be changed or not."}, {"inherit_profile", std::make_shared(std::make_shared()), "A parent profile for this setting profile. NULL if not set. " "Setting profile will inherit all the settings' values and constraints (min, max, readonly) from its parent profiles." @@ -51,7 +51,7 @@ ColumnsDescription StorageSystemSettingsProfileElements::getColumnsDescription() } -void StorageSystemSettingsProfileElements::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemSettingsProfileElements::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemSettingsProfileElements.h b/src/Storages/System/StorageSystemSettingsProfileElements.h index 1dedd616c82..8b08c463071 100644 --- a/src/Storages/System/StorageSystemSettingsProfileElements.h +++ b/src/Storages/System/StorageSystemSettingsProfileElements.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `settings_profile_elements` system table, which allows you to get information about elements of settings profiles. -class StorageSystemSettingsProfileElements final : public IStorageSystemOneBlock +class StorageSystemSettingsProfileElements final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemSettingsProfileElements"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemSettingsProfiles.cpp b/src/Storages/System/StorageSystemSettingsProfiles.cpp index 01041bee445..795152e31f3 100644 --- a/src/Storages/System/StorageSystemSettingsProfiles.cpp +++ b/src/Storages/System/StorageSystemSettingsProfiles.cpp @@ -36,7 +36,7 @@ ColumnsDescription StorageSystemSettingsProfiles::getColumnsDescription() } -void StorageSystemSettingsProfiles::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemSettingsProfiles::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemSettingsProfiles.h b/src/Storages/System/StorageSystemSettingsProfiles.h index b0c8fc8658c..056666ae4c7 100644 --- a/src/Storages/System/StorageSystemSettingsProfiles.h +++ b/src/Storages/System/StorageSystemSettingsProfiles.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `settings_profiles` system table, which allows you to get information about profiles. -class StorageSystemSettingsProfiles final : public IStorageSystemOneBlock +class StorageSystemSettingsProfiles final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemSettingsProfiles"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemStackTrace.cpp b/src/Storages/System/StorageSystemStackTrace.cpp index 74864bb50e1..ba7433fb9ae 100644 --- a/src/Storages/System/StorageSystemStackTrace.cpp +++ b/src/Storages/System/StorageSystemStackTrace.cpp @@ -507,11 +507,11 @@ StorageSystemStackTrace::StorageSystemStackTrace(const StorageID & table_id_) { StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription({ - { "thread_name", std::make_shared() }, - { "thread_id", std::make_shared() }, - { "query_id", std::make_shared() }, - { "trace", std::make_shared(std::make_shared()) }, - }, { /* aliases */ })); + {"thread_name", std::make_shared(), "The name of the thread."}, + {"thread_id", std::make_shared(), "The thread identifier"}, + {"query_id", std::make_shared(), "The ID of the query this thread belongs to."}, + {"trace", std::make_shared(std::make_shared()), "The stacktrace of this thread. Basically just an array of addresses."}, + })); setInMemoryMetadata(storage_metadata); notification_pipe.open(); diff --git a/src/Storages/System/StorageSystemStoragePolicies.cpp b/src/Storages/System/StorageSystemStoragePolicies.cpp index 39e99884e1d..21251136f7d 100644 --- a/src/Storages/System/StorageSystemStoragePolicies.cpp +++ b/src/Storages/System/StorageSystemStoragePolicies.cpp @@ -37,16 +37,16 @@ StorageSystemStoragePolicies::StorageSystemStoragePolicies(const StorageID & tab StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns( ColumnsDescription({ - {"policy_name", std::make_shared()}, - {"volume_name", std::make_shared()}, - {"volume_priority", std::make_shared()}, - {"disks", std::make_shared(std::make_shared())}, - {"volume_type", std::make_shared(getTypeEnumValues())}, - {"max_data_part_size", std::make_shared()}, - {"move_factor", std::make_shared()}, - {"prefer_not_to_merge", std::make_shared()}, - {"perform_ttl_move_on_insert", std::make_shared()}, - {"load_balancing", std::make_shared(getTypeEnumValues())} + {"policy_name", std::make_shared(), "The name of the storage policy."}, + {"volume_name", std::make_shared(), "The name of the volume."}, + {"volume_priority", std::make_shared(), "The priority of the volume."}, + {"disks", std::make_shared(std::make_shared()), "The list of all disks names which are a part of this storage policy."}, + {"volume_type", std::make_shared(getTypeEnumValues()), "The type of the volume - JBOD or a single disk."}, + {"max_data_part_size", std::make_shared(), "the maximum size of a part that can be stored on any of the volumes disks."}, + {"move_factor", std::make_shared(), "When the amount of available space gets lower than this factor, data automatically starts to move on the next volume if any (by default, 0.1)."}, + {"prefer_not_to_merge", std::make_shared(), "You should not use this setting. Disables merging of data parts on this volume (this is harmful and leads to performance degradation)."}, + {"perform_ttl_move_on_insert", std::make_shared(), "Disables TTL move on data part INSERT. By default (if enabled) if we insert a data part that already expired by the TTL move rule it immediately goes to a volume/disk declared in move rule."}, + {"load_balancing", std::make_shared(getTypeEnumValues()), "Policy for disk balancing, `round_robin` or `least_used`."} })); // TODO: Add string column with custom volume-type-specific options setInMemoryMetadata(storage_metadata); diff --git a/src/Storages/System/StorageSystemSymbols.cpp b/src/Storages/System/StorageSystemSymbols.cpp index 56195544448..6b313407e61 100644 --- a/src/Storages/System/StorageSystemSymbols.cpp +++ b/src/Storages/System/StorageSystemSymbols.cpp @@ -22,9 +22,9 @@ StorageSystemSymbols::StorageSystemSymbols(const StorageID & table_id_) StorageInMemoryMetadata storage_metadata; storage_metadata.setColumns(ColumnsDescription( { - {"symbol", std::make_shared()}, - {"address_begin", std::make_shared()}, - {"address_end", std::make_shared()}, + {"symbol", std::make_shared(), "Symbol name in the binary. It is mangled. You can apply demangle(symbol) to obtain a readable name."}, + {"address_begin", std::make_shared(), "Start address of the symbol in the binary."}, + {"address_end", std::make_shared(), "End address of the symbol in the binary."}, })); setInMemoryMetadata(storage_metadata); } diff --git a/src/Storages/System/StorageSystemTableEngines.cpp b/src/Storages/System/StorageSystemTableEngines.cpp index c0cf95423d9..e27546aa2a4 100644 --- a/src/Storages/System/StorageSystemTableEngines.cpp +++ b/src/Storages/System/StorageSystemTableEngines.cpp @@ -26,7 +26,7 @@ ColumnsDescription StorageSystemTableEngines::getColumnsDescription() }; } -void StorageSystemTableEngines::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemTableEngines::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (const auto & pair : StorageFactory::instance().getAllStorages()) { diff --git a/src/Storages/System/StorageSystemTableEngines.h b/src/Storages/System/StorageSystemTableEngines.h index 258b9d210b1..d7af471bb2d 100644 --- a/src/Storages/System/StorageSystemTableEngines.h +++ b/src/Storages/System/StorageSystemTableEngines.h @@ -6,10 +6,10 @@ namespace DB { -class StorageSystemTableEngines final : public IStorageSystemOneBlock +class StorageSystemTableEngines final : public IStorageSystemOneBlock { protected: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemTableFunctions.cpp b/src/Storages/System/StorageSystemTableFunctions.cpp index 9fb8e11e4d1..94b7d73a67e 100644 --- a/src/Storages/System/StorageSystemTableFunctions.cpp +++ b/src/Storages/System/StorageSystemTableFunctions.cpp @@ -20,7 +20,7 @@ ColumnsDescription StorageSystemTableFunctions::getColumnsDescription() }; } -void StorageSystemTableFunctions::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemTableFunctions::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { const auto & factory = TableFunctionFactory::instance(); const auto & functions_names = factory.getAllRegisteredNames(); diff --git a/src/Storages/System/StorageSystemTableFunctions.h b/src/Storages/System/StorageSystemTableFunctions.h index 804c3b51940..07ef8857135 100644 --- a/src/Storages/System/StorageSystemTableFunctions.h +++ b/src/Storages/System/StorageSystemTableFunctions.h @@ -6,12 +6,12 @@ namespace DB { -class StorageSystemTableFunctions final : public IStorageSystemOneBlock +class StorageSystemTableFunctions final : public IStorageSystemOneBlock { protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; public: std::string getName() const override diff --git a/src/Storages/System/StorageSystemTables.cpp b/src/Storages/System/StorageSystemTables.cpp index 6904eba8ab5..639c1455b83 100644 --- a/src/Storages/System/StorageSystemTables.cpp +++ b/src/Storages/System/StorageSystemTables.cpp @@ -67,9 +67,9 @@ StorageSystemTables::StorageSystemTables(const StorageID & table_id_) "Total number of uncompressed bytes, if it's possible to quickly determine the exact number " "of bytes from the part checksums for the table on storage, otherwise NULL (does not take underlying storage (if any) into account)." }, - {"parts", std::make_shared(std::make_shared())}, - {"active_parts", std::make_shared(std::make_shared())}, - {"total_marks", std::make_shared(std::make_shared())}, + {"parts", std::make_shared(std::make_shared()), "The total number of parts in this table."}, + {"active_parts", std::make_shared(std::make_shared()), "The number of active parts in this table."}, + {"total_marks", std::make_shared(std::make_shared()), "The total number of marks in all parts in this table."}, {"lifetime_rows", std::make_shared(std::make_shared()), "Total number of rows INSERTed since server start (only for Buffer tables)." }, diff --git a/src/Storages/System/StorageSystemTimeZones.cpp b/src/Storages/System/StorageSystemTimeZones.cpp index 14f4ce0f5de..af997c6423f 100644 --- a/src/Storages/System/StorageSystemTimeZones.cpp +++ b/src/Storages/System/StorageSystemTimeZones.cpp @@ -16,7 +16,7 @@ ColumnsDescription StorageSystemTimeZones::getColumnsDescription() }; } -void StorageSystemTimeZones::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemTimeZones::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { for (auto * it = auto_time_zones; *it; ++it) res_columns[0]->insert(String(*it)); diff --git a/src/Storages/System/StorageSystemTimeZones.h b/src/Storages/System/StorageSystemTimeZones.h index f3743a1ef09..160e8854e3e 100644 --- a/src/Storages/System/StorageSystemTimeZones.h +++ b/src/Storages/System/StorageSystemTimeZones.h @@ -10,10 +10,10 @@ class Context; /** System table "time_zones" with list of timezones pulled from /contrib/cctz/testdata/zoneinfo */ -class StorageSystemTimeZones final : public IStorageSystemOneBlock +class StorageSystemTimeZones final : public IStorageSystemOneBlock { public: - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; using IStorageSystemOneBlock::IStorageSystemOneBlock; diff --git a/src/Storages/System/StorageSystemTransactions.cpp b/src/Storages/System/StorageSystemTransactions.cpp index 47e44688c14..295d93edf7d 100644 --- a/src/Storages/System/StorageSystemTransactions.cpp +++ b/src/Storages/System/StorageSystemTransactions.cpp @@ -23,18 +23,17 @@ static DataTypePtr getStateEnumType() ColumnsDescription StorageSystemTransactions::getColumnsDescription() { - /// TODO: Fill in all the comments. return ColumnsDescription { - {"tid", getTransactionIDDataType()}, - {"tid_hash", std::make_shared()}, - {"elapsed", std::make_shared()}, - {"is_readonly", std::make_shared()}, - {"state", getStateEnumType()}, + {"tid", getTransactionIDDataType(), "The identifier of the transaction."}, + {"tid_hash", std::make_shared(), "The hash of the identifier."}, + {"elapsed", std::make_shared(), "The amount of time the transaction being processed."}, + {"is_readonly", std::make_shared(), "The flag which shows whether the transaction has executed any write operation."}, + {"state", getStateEnumType(), "The state of the transaction. Possible values: RUNNING, COMMITTING, COMMITTED, ROLLED_BACK."}, }; } -void StorageSystemTransactions::fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const +void StorageSystemTransactions::fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const { auto list = TransactionLog::instance().getTransactionsList(); for (const auto & elem : list) diff --git a/src/Storages/System/StorageSystemTransactions.h b/src/Storages/System/StorageSystemTransactions.h index b5a538b7b55..1c400619c1a 100644 --- a/src/Storages/System/StorageSystemTransactions.h +++ b/src/Storages/System/StorageSystemTransactions.h @@ -8,7 +8,7 @@ namespace DB class Context; -class StorageSystemTransactions final : public IStorageSystemOneBlock +class StorageSystemTransactions final : public IStorageSystemOneBlock { public: String getName() const override { return "SystemTransactions"; } @@ -18,7 +18,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemUserDirectories.cpp b/src/Storages/System/StorageSystemUserDirectories.cpp index 620c9746a4a..1b3469e7597 100644 --- a/src/Storages/System/StorageSystemUserDirectories.cpp +++ b/src/Storages/System/StorageSystemUserDirectories.cpp @@ -11,18 +11,17 @@ namespace DB { ColumnsDescription StorageSystemUserDirectories::getColumnsDescription() { - /// TODO: Fill in all the comments. return ColumnsDescription { - {"name", std::make_shared()}, - {"type", std::make_shared()}, - {"params", std::make_shared()}, - {"precedence", std::make_shared()}, + {"name", std::make_shared(), "The name of the directory."}, + {"type", std::make_shared(), "The type of the access storage e.g. users.xml or replicated or memory etc."}, + {"params", std::make_shared(), "JSON with the parameters of the access storage."}, + {"precedence", std::make_shared(), "The order in which this directory is declared in the config. The same order is used when ClickHouse tries to find a user or role."}, }; } -void StorageSystemUserDirectories::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemUserDirectories::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto & access_control = context->getAccessControl(); auto storages = access_control.getStorages(); diff --git a/src/Storages/System/StorageSystemUserDirectories.h b/src/Storages/System/StorageSystemUserDirectories.h index bca6a9b5aa6..3cdaa877b81 100644 --- a/src/Storages/System/StorageSystemUserDirectories.h +++ b/src/Storages/System/StorageSystemUserDirectories.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `users_directories` system table, which allows you to get information about user directories. -class StorageSystemUserDirectories final : public IStorageSystemOneBlock +class StorageSystemUserDirectories final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemUserDirectories"; } @@ -16,7 +16,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemUserProcesses.cpp b/src/Storages/System/StorageSystemUserProcesses.cpp index 4fbbd7ab54d..65fbeedf406 100644 --- a/src/Storages/System/StorageSystemUserProcesses.cpp +++ b/src/Storages/System/StorageSystemUserProcesses.cpp @@ -18,10 +18,10 @@ ColumnsDescription StorageSystemUserProcesses::getColumnsDescription() { auto description = ColumnsDescription { - {"user", std::make_shared()}, - {"memory_usage", std::make_shared()}, - {"peak_memory_usage", std::make_shared()}, - {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared())}, + {"user", std::make_shared(), "User name."}, + {"memory_usage", std::make_shared(), "Sum of RAM used by all processes of the user. It might not include some types of dedicated memory. See the max_memory_usage setting."}, + {"peak_memory_usage", std::make_shared(), "The peak of memory usage of the user. It can be reset when no queries are run for the user."}, + {"ProfileEvents", std::make_shared(std::make_shared(), std::make_shared()), "Summary of ProfileEvents that measure different metrics for the user. The description of them could be found in the table system.events"}, }; description.setAliases({ @@ -32,7 +32,7 @@ ColumnsDescription StorageSystemUserProcesses::getColumnsDescription() return description; } -void StorageSystemUserProcesses::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemUserProcesses::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { const auto user_info = context->getProcessList().getUserInfo(true); diff --git a/src/Storages/System/StorageSystemUserProcesses.h b/src/Storages/System/StorageSystemUserProcesses.h index 6eb12e30559..3141eae9662 100644 --- a/src/Storages/System/StorageSystemUserProcesses.h +++ b/src/Storages/System/StorageSystemUserProcesses.h @@ -11,7 +11,7 @@ class Context; /** Implements `processes` system table, which allows you to get information about the queries that are currently executing. */ -class StorageSystemUserProcesses final : public IStorageSystemOneBlock +class StorageSystemUserProcesses final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemUserProcesses"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemUsers.cpp b/src/Storages/System/StorageSystemUsers.cpp index cf114a85645..0c34f04844d 100644 --- a/src/Storages/System/StorageSystemUsers.cpp +++ b/src/Storages/System/StorageSystemUsers.cpp @@ -69,15 +69,15 @@ ColumnsDescription StorageSystemUsers::getColumnsDescription() {"default_roles_except", std::make_shared(std::make_shared()), "All the granted roles set as default excepting of the listed ones." }, - {"grantees_any", std::make_shared()}, - {"grantees_list", std::make_shared(std::make_shared())}, - {"grantees_except", std::make_shared(std::make_shared())}, - {"default_database", std::make_shared()}, + {"grantees_any", std::make_shared(), "The flag that indicates whether a user with any grant option can grant it to anyone."}, + {"grantees_list", std::make_shared(std::make_shared()), "The list of users or roles to which this user is allowed to grant options to."}, + {"grantees_except", std::make_shared(std::make_shared()), "The list of users or roles to which this user is forbidden from grant options to."}, + {"default_database", std::make_shared(), "The name of the default database for this user."}, }; } -void StorageSystemUsers::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemUsers::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { /// If "select_from_system_db_requires_grant" is enabled the access rights were already checked in InterpreterSelectQuery. const auto & access_control = context->getAccessControl(); diff --git a/src/Storages/System/StorageSystemUsers.h b/src/Storages/System/StorageSystemUsers.h index cfa5947d370..a1a3d717b35 100644 --- a/src/Storages/System/StorageSystemUsers.h +++ b/src/Storages/System/StorageSystemUsers.h @@ -8,7 +8,7 @@ namespace DB class Context; /// Implements `users` system table, which allows you to get information about users. -class StorageSystemUsers final : public IStorageSystemOneBlock +class StorageSystemUsers final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemUsers"; } @@ -19,7 +19,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemViewRefreshes.cpp b/src/Storages/System/StorageSystemViewRefreshes.cpp index d2b933e65a8..30539ed6b6a 100644 --- a/src/Storages/System/StorageSystemViewRefreshes.cpp +++ b/src/Storages/System/StorageSystemViewRefreshes.cpp @@ -36,20 +36,20 @@ ColumnsDescription StorageSystemViewRefreshes::getColumnsDescription() }, {"refresh_count", std::make_shared(), "Number of successful refreshes since last server restart or table creation."}, {"progress", std::make_shared(), "Progress of the current refresh, between 0 and 1."}, - {"elapsed", std::make_shared()}, - {"read_rows", std::make_shared(), "Number of rows read by the current refresh so far."}, - {"read_bytes", std::make_shared()}, + {"elapsed", std::make_shared(), "The amount of nanoseconds the current refresh took."}, + {"read_rows", std::make_shared(), "Number of rows read during the current refresh."}, + {"read_bytes", std::make_shared(), "Number of bytes read during the current refresh."}, {"total_rows", std::make_shared(), "Estimated total number of rows that need to be read by the current refresh."}, - {"total_bytes", std::make_shared()}, - {"written_rows", std::make_shared()}, - {"written_bytes", std::make_shared()}, - {"result_rows", std::make_shared()}, - {"result_bytes", std::make_shared()}, + {"total_bytes", std::make_shared(), "Estimated total number of bytes that need to be read by the current refresh."}, + {"written_rows", std::make_shared(), "Number of rows written during the current refresh."}, + {"written_bytes", std::make_shared(), "Number rof bytes written during the current refresh."}, + {"result_rows", std::make_shared(), "Estimated total number of rows in the result set of the SELECT query."}, + {"result_bytes", std::make_shared(), "Estimated total number of bytes in the result set of the SELECT query."}, }; } void StorageSystemViewRefreshes::fillData( - MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const + MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { auto access = context->getAccess(); auto valid_access = AccessType::SHOW_TABLES; diff --git a/src/Storages/System/StorageSystemViewRefreshes.h b/src/Storages/System/StorageSystemViewRefreshes.h index 02d3a39dfff..5a29f3a3bc8 100644 --- a/src/Storages/System/StorageSystemViewRefreshes.h +++ b/src/Storages/System/StorageSystemViewRefreshes.h @@ -10,7 +10,7 @@ namespace DB { -class StorageSystemViewRefreshes final : public IStorageSystemOneBlock +class StorageSystemViewRefreshes final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemViewRefreshes"; } @@ -20,7 +20,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemWarnings.cpp b/src/Storages/System/StorageSystemWarnings.cpp index e2579372b4d..01e96e980a8 100644 --- a/src/Storages/System/StorageSystemWarnings.cpp +++ b/src/Storages/System/StorageSystemWarnings.cpp @@ -1,5 +1,7 @@ #include +#include #include +#include namespace DB @@ -13,7 +15,7 @@ ColumnsDescription StorageSystemWarnings::getColumnsDescription() }; } -void StorageSystemWarnings::fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo &) const +void StorageSystemWarnings::fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const { for (const auto & warning : context->getWarnings()) res_columns[0]->insert(warning); diff --git a/src/Storages/System/StorageSystemWarnings.h b/src/Storages/System/StorageSystemWarnings.h index 42948a765ea..685fb60b430 100644 --- a/src/Storages/System/StorageSystemWarnings.h +++ b/src/Storages/System/StorageSystemWarnings.h @@ -11,7 +11,7 @@ class Context; /** Implements system.warnings table that contains warnings about server configuration * to be displayed in clickhouse-client. */ -class StorageSystemWarnings final : public IStorageSystemOneBlock +class StorageSystemWarnings final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemWarnings"; } @@ -21,6 +21,6 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr, const SelectQueryInfo &) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/StorageSystemZooKeeper.cpp b/src/Storages/System/StorageSystemZooKeeper.cpp index 2845e194614..d1bf86ba8ef 100644 --- a/src/Storages/System/StorageSystemZooKeeper.cpp +++ b/src/Storages/System/StorageSystemZooKeeper.cpp @@ -233,9 +233,9 @@ private: StorageSystemZooKeeper::StorageSystemZooKeeper(const StorageID & table_id_) : IStorage(table_id_) { - StorageInMemoryMetadata storage_metadata; - storage_metadata.setColumns(getColumnsDescription()); - setInMemoryMetadata(storage_metadata); + StorageInMemoryMetadata storage_metadata; + storage_metadata.setColumns(getColumnsDescription()); + setInMemoryMetadata(storage_metadata); } void StorageSystemZooKeeper::read( @@ -248,7 +248,7 @@ void StorageSystemZooKeeper::read( size_t max_block_size, size_t /*num_streams*/) { - auto header = storage_snapshot->metadata->getSampleBlockWithVirtuals(getVirtuals()); + auto header = storage_snapshot->metadata->getSampleBlockWithVirtuals(getVirtualsList()); auto read_step = std::make_unique( column_names, query_info, diff --git a/src/Storages/System/StorageSystemZooKeeperConnection.cpp b/src/Storages/System/StorageSystemZooKeeperConnection.cpp index c165bfa217d..8041370ee92 100644 --- a/src/Storages/System/StorageSystemZooKeeperConnection.cpp +++ b/src/Storages/System/StorageSystemZooKeeperConnection.cpp @@ -39,7 +39,7 @@ ColumnsDescription StorageSystemZooKeeperConnection::getColumnsDescription() } void StorageSystemZooKeeperConnection::fillData(MutableColumns & res_columns, ContextPtr context, - const SelectQueryInfo &) const + const ActionsDAG::Node *, std::vector) const { const auto add_enabled_feature_flags = [&](const auto & zookeeper) { diff --git a/src/Storages/System/StorageSystemZooKeeperConnection.h b/src/Storages/System/StorageSystemZooKeeperConnection.h index 2b6d3b2e516..f8263e1f1bc 100644 --- a/src/Storages/System/StorageSystemZooKeeperConnection.h +++ b/src/Storages/System/StorageSystemZooKeeperConnection.h @@ -11,7 +11,7 @@ class Context; /** Implements `zookeeper_connection` system table, which allows you to get information about the connected zookeeper info. */ -class StorageSystemZooKeeperConnection final : public IStorageSystemOneBlock +class StorageSystemZooKeeperConnection final : public IStorageSystemOneBlock { public: std::string getName() const override { return "SystemZooKeeperConnection"; } @@ -21,7 +21,7 @@ public: protected: using IStorageSystemOneBlock::IStorageSystemOneBlock; - void fillData(MutableColumns & res_columns, ContextPtr context, const SelectQueryInfo & query_info) const override; + void fillData(MutableColumns & res_columns, ContextPtr context, const ActionsDAG::Node *, std::vector) const override; }; } diff --git a/src/Storages/System/attachSystemTables.cpp b/src/Storages/System/attachSystemTables.cpp index f08d6972b98..79475b9695d 100644 --- a/src/Storages/System/attachSystemTables.cpp +++ b/src/Storages/System/attachSystemTables.cpp @@ -118,14 +118,14 @@ namespace DB void attachSystemTablesServer(ContextPtr context, IDatabase & system_database, bool has_zookeeper) { - attach(context, system_database, "one", "This table contains a single row with a single dummy UInt8 column containing the value 0. Used when the table is not specified explicitly, for example in queries like `SELECT 1`."); - attach(context, system_database, "numbers", "Generates all natural numbers, starting from 0 (to 2^64 - 1, and then again) in sorted order.", false); - attach(context, system_database, "numbers_mt", "Multithreaded version of `system.numbers`. Numbers order is not guaranteed.", true); - attach(context, system_database, "zeros", "Produces unlimited number of non-materialized zeros.", false); - attach(context, system_database, "zeros_mt", "Multithreaded version of system.zeros.", true); + attachNoDescription(context, system_database, "one", "This table contains a single row with a single dummy UInt8 column containing the value 0. Used when the table is not specified explicitly, for example in queries like `SELECT 1`."); + attachNoDescription(context, system_database, "numbers", "Generates all natural numbers, starting from 0 (to 2^64 - 1, and then again) in sorted order.", false); + attachNoDescription(context, system_database, "numbers_mt", "Multithreaded version of `system.numbers`. Numbers order is not guaranteed.", true); + attachNoDescription(context, system_database, "zeros", "Produces unlimited number of non-materialized zeros.", false); + attachNoDescription(context, system_database, "zeros_mt", "Multithreaded version of system.zeros.", true); attach(context, system_database, "databases", "Lists all databases of the current server."); - attach(context, system_database, "tables", "Lists all tables of the current server."); - attach(context, system_database, "columns", "Lists all columns from all tables of the current server."); + attachNoDescription(context, system_database, "tables", "Lists all tables of the current server."); + attachNoDescription(context, system_database, "columns", "Lists all columns from all tables of the current server."); attach(context, system_database, "functions", "Contains a list of all available ordinary and aggregate functions with their descriptions."); attach(context, system_database, "events", "Contains profiling events and their current value."); attach(context, system_database, "settings", "Contains a list of all user-level settings (which can be modified in a scope of query or session), their current and default values along with descriptions."); @@ -159,44 +159,44 @@ void attachSystemTablesServer(ContextPtr context, IDatabase & system_database, b attach(context, system_database, "privileges", "Contains a list of all available privileges that could be granted to a user or role."); attach(context, system_database, "errors", "Contains a list of all errors which have ever happened including the error code, last time and message with unsymbolized stacktrace."); attach(context, system_database, "warnings", "Contains warnings about server configuration to be displayed by clickhouse-client right after it connects to the server."); - attach(context, system_database, "data_skipping_indices", "Contains all the information about all the data skipping indices in tables, similar to system.columns."); + attachNoDescription(context, system_database, "data_skipping_indices", "Contains all the information about all the data skipping indices in tables, similar to system.columns."); attach(context, system_database, "licenses", "Contains licenses of third-party libraries that are located in the contrib directory of ClickHouse sources."); attach(context, system_database, "time_zones", "Contains a list of time zones that are supported by the ClickHouse server. This list of timezones might vary depending on the version of ClickHouse."); attach(context, system_database, "backups", "Contains a list of all BACKUP or RESTORE operations with their current states and other propertis. Note, that table is not persistent and it shows only operations executed after the last server restart."); attach(context, system_database, "schema_inference_cache", "Contains information about all cached file schemas."); attach(context, system_database, "dropped_tables", "Contains a list of tables which were dropped from Atomic databases but not completely removed yet."); - attach(context, system_database, "dropped_tables_parts", "Contains parts of system.dropped_tables tables "); + attachNoDescription(context, system_database, "dropped_tables_parts", "Contains parts of system.dropped_tables tables "); attach(context, system_database, "scheduler", "Contains information and status for scheduling nodes residing on the local server."); attach(context, system_database, "dns_cache", "Contains information about cached DNS records."); #if defined(__ELF__) && !defined(OS_FREEBSD) - attach(context, system_database, "symbols", "Contains information for introspection of ClickHouse binary. This table is only useful for C++ experts and ClickHouse engineers."); + attachNoDescription(context, system_database, "symbols", "Contains information for introspection of ClickHouse binary. This table is only useful for C++ experts and ClickHouse engineers."); #endif #if USE_RDKAFKA attach(context, system_database, "kafka_consumers", "Contains information about Kafka consumers. Applicable for Kafka table engine (native ClickHouse integration)."); #endif #ifdef OS_LINUX - attach(context, system_database, "stack_trace", "Allows to obtain an unsymbolized stacktrace from all the threads of the server process."); + attachNoDescription(context, system_database, "stack_trace", "Allows to obtain an unsymbolized stacktrace from all the threads of the server process."); #endif #if USE_ROCKSDB attach(context, system_database, "rocksdb", "Contains a list of metrics exposed from embedded RocksDB."); #endif #if USE_MYSQL - attach(context, system_database, "mysql_binlogs", "Shows a list of active binlogs for MaterializedMySQL."); + attachNoDescription(context, system_database, "mysql_binlogs", "Shows a list of active binlogs for MaterializedMySQL."); #endif - attach(context, system_database, "parts", "Contains a list of currently existing (both active and inactive) parts of all *-MergeTree tables. Each part is represented by a single row."); - attach(context, system_database, "projection_parts", "Contains a list of currently existing projection parts (a copy of some part containing aggregated data or just sorted in different order) created for all the projections for all tables within a cluster."); - attach(context, system_database, "detached_parts", "Contains a list of all parts which are being found in /detached directory along with a reason why it was detached. ClickHouse server doesn't use such parts anyhow."); - attach(context, system_database, "parts_columns", "Contains a list of columns of all currently existing parts of all MergeTree tables. Each column is represented by a single row."); - attach(context, system_database, "projection_parts_columns", "Contains a list of columns of all currently existing projection parts of all MergeTree tables. Each column is represented by a single row."); - attach(context, system_database, "disks", "Contains information about disks defined in the server configuration."); - attach(context, system_database, "storage_policies", "Contains information about storage policies and volumes defined in the server configuration."); + attachNoDescription(context, system_database, "parts", "Contains a list of currently existing (both active and inactive) parts of all *-MergeTree tables. Each part is represented by a single row."); + attachNoDescription(context, system_database, "projection_parts", "Contains a list of currently existing projection parts (a copy of some part containing aggregated data or just sorted in different order) created for all the projections for all tables within a cluster."); + attachNoDescription(context, system_database, "detached_parts", "Contains a list of all parts which are being found in /detached directory along with a reason why it was detached. ClickHouse server doesn't use such parts anyhow."); + attachNoDescription(context, system_database, "parts_columns", "Contains a list of columns of all currently existing parts of all MergeTree tables. Each column is represented by a single row."); + attachNoDescription(context, system_database, "projection_parts_columns", "Contains a list of columns of all currently existing projection parts of all MergeTree tables. Each column is represented by a single row."); + attachNoDescription(context, system_database, "disks", "Contains information about disks defined in the server configuration."); + attachNoDescription(context, system_database, "storage_policies", "Contains information about storage policies and volumes defined in the server configuration."); attach(context, system_database, "processes", "Contains a list of currently executing processes (queries) with their progress."); attach(context, system_database, "metrics", "Contains metrics which can be calculated instantly, or have a current value. For example, the number of simultaneously processed queries or the current replica delay. This table is always up to date."); attach(context, system_database, "merges", "Contains a list of merges currently executing merges of MergeTree tables and their progress. Each merge operation is represented by a single row."); attach(context, system_database, "moves", "Contains information about in-progress data part moves of MergeTree tables. Each data part movement is represented by a single row."); attach(context, system_database, "mutations", "Contains a list of mutations and their progress. Each mutation command is represented by a single row."); - attach(context, system_database, "replicas", "Contains information and status of all table replicas on current server. Each replica is represented by a single row."); + attachNoDescription(context, system_database, "replicas", "Contains information and status of all table replicas on current server. Each replica is represented by a single row."); attach(context, system_database, "replication_queue", "Contains information about tasks from replication queues stored in ClickHouse Keeper, or ZooKeeper, for each table replica."); attach(context, system_database, "distributed_ddl_queue", "Contains information about distributed DDL queries (ON CLUSTER clause) that were executed on a cluster."); attach(context, system_database, "distribution_queue", "Contains information about local files that are in the queue to be sent to the shards. These local files contain new parts that are created by inserting new data into the Distributed table in asynchronous mode."); @@ -208,21 +208,21 @@ void attachSystemTablesServer(ContextPtr context, IDatabase & system_database, b attach(context, system_database, "replicated_fetches", "Contains information about currently running background fetches."); attach(context, system_database, "part_moves_between_shards", "Contains information about parts which are currently in a process of moving between shards and their progress."); attach(context, system_database, "asynchronous_inserts", "Contains information about pending asynchronous inserts in queue in server's memory."); - attach(context, system_database, "filesystem_cache", "Contains information about all entries inside filesystem cache for remote objects."); - attach(context, system_database, "query_cache", "Contains information about all entries inside query cache in server's memory."); - attach(context, system_database, "remote_data_paths", "Contains a mapping from a filename on local filesystem to a blob name inside object storage."); + attachNoDescription(context, system_database, "filesystem_cache", "Contains information about all entries inside filesystem cache for remote objects."); + attachNoDescription(context, system_database, "query_cache", "Contains information about all entries inside query cache in server's memory."); + attachNoDescription(context, system_database, "remote_data_paths", "Contains a mapping from a filename on local filesystem to a blob name inside object storage."); attach(context, system_database, "certificates", "Contains information about available certificates and their sources."); - attach(context, system_database, "named_collections", "Contains a list of all named collections which were created via SQL query or parsed from configuration file."); + attachNoDescription(context, system_database, "named_collections", "Contains a list of all named collections which were created via SQL query or parsed from configuration file."); attach(context, system_database, "asynchronous_loader", "Contains information and status for recent asynchronous jobs (e.g. for tables loading). The table contains a row for every job."); attach(context, system_database, "user_processes", "This system table can be used to get overview of memory usage and ProfileEvents of users."); - attach(context, system_database, "jemalloc_bins", "Contains information about memory allocations done via jemalloc allocator in different size classes (bins) aggregated from all arenas. These statistics might not be absolutely accurate because of thread local caching in jemalloc."); - attach(context, system_database, "s3queue", "Contains in-memory state of S3Queue metadata and currently processed rows per file."); + attachNoDescription(context, system_database, "jemalloc_bins", "Contains information about memory allocations done via jemalloc allocator in different size classes (bins) aggregated from all arenas. These statistics might not be absolutely accurate because of thread local caching in jemalloc."); + attachNoDescription(context, system_database, "s3queue", "Contains in-memory state of S3Queue metadata and currently processed rows per file."); attach(context, system_database, "dashboards", "Contains queries used by /dashboard page accessible though HTTP interface. This table can be useful for monitoring and troubleshooting. The table contains a row for every chart in a dashboard."); attach(context, system_database, "view_refreshes", "Lists all Refreshable Materialized Views of current server."); if (has_zookeeper) { - attach(context, system_database, "zookeeper", "Exposes data from the [Zoo]Keeper cluster defined in the config. Allow to get the list of children for a particular node or read the value written inside it."); + attachNoDescription(context, system_database, "zookeeper", "Exposes data from the [Zoo]Keeper cluster defined in the config. Allow to get the list of children for a particular node or read the value written inside it."); attach(context, system_database, "zookeeper_connection", "Shows the information about current connections to [Zoo]Keeper (including auxiliary [ZooKeepers)"); } @@ -232,7 +232,7 @@ void attachSystemTablesServer(ContextPtr context, IDatabase & system_database, b void attachSystemTablesAsync(ContextPtr context, IDatabase & system_database, AsynchronousMetrics & async_metrics) { - attach(context, system_database, "asynchronous_metrics", "Contains metrics that are calculated periodically in the background. For example, the amount of RAM in use.", async_metrics); + attachNoDescription(context, system_database, "asynchronous_metrics", "Contains metrics that are calculated periodically in the background. For example, the amount of RAM in use.", async_metrics); } } diff --git a/src/Storages/System/attachSystemTablesImpl.h b/src/Storages/System/attachSystemTablesImpl.h index 9f2c4e8016d..d9ab164f2b3 100644 --- a/src/Storages/System/attachSystemTablesImpl.h +++ b/src/Storages/System/attachSystemTablesImpl.h @@ -10,8 +10,8 @@ namespace DB template using StringLiteral = const char(&)[Length]; -template -void attach(ContextPtr context, IDatabase & system_database, const String & table_name, StringLiteral comment, StorageArgs && ... args) +template +void attachImpl(ContextPtr context, IDatabase & system_database, const String & table_name, StringLiteral comment, StorageArgs && ... args) { static_assert(CommentSize > 15, "The comment for a system table is too short or empty"); assert(system_database.getDatabaseName() == DatabaseCatalog::SYSTEM_DATABASE); @@ -21,7 +21,10 @@ void attach(ContextPtr context, IDatabase & system_database, const String & tabl { /// Attach to Ordinary database. table_id = StorageID(DatabaseCatalog::SYSTEM_DATABASE, table_name); - system_database.attachTable(context, table_name, std::make_shared(table_id, std::forward(args)...)); + if constexpr (with_description) + system_database.attachTable(context, table_name, std::make_shared(table_id, StorageT::getColumnsDescription(), std::forward(args)...)); + else + system_database.attachTable(context, table_name, std::make_shared(table_id, std::forward(args)...)); } else { @@ -31,7 +34,10 @@ void attach(ContextPtr context, IDatabase & system_database, const String & tabl table_id = StorageID(DatabaseCatalog::SYSTEM_DATABASE, table_name, UUIDHelpers::generateV4()); DatabaseCatalog::instance().addUUIDMapping(table_id.uuid); String path = "store/" + DatabaseCatalog::getPathForUUID(table_id.uuid); - system_database.attachTable(context, table_name, std::make_shared(table_id, std::forward(args)...), path); + if constexpr (with_description) + system_database.attachTable(context, table_name, std::make_shared(table_id, StorageT::getColumnsDescription(), std::forward(args)...), path); + else + system_database.attachTable(context, table_name, std::make_shared(table_id, std::forward(args)...), path); } /// Set the comment @@ -42,4 +48,17 @@ void attach(ContextPtr context, IDatabase & system_database, const String & tabl table->setInMemoryMetadata(metadata); } + +template +void attach(ContextPtr context, IDatabase & system_database, const String & table_name, StringLiteral comment, StorageArgs && ... args) +{ + attachImpl(context, system_database, table_name, comment, std::forward(args)...); +} + +template +void attachNoDescription(ContextPtr context, IDatabase & system_database, const String & table_name, StringLiteral comment, StorageArgs && ... args) +{ + attachImpl(context, system_database, table_name, comment, std::forward(args)...); +} + } diff --git a/src/Storages/VirtualColumnUtils.cpp b/src/Storages/VirtualColumnUtils.cpp index 07ac61c110d..22466b3d4c2 100644 --- a/src/Storages/VirtualColumnUtils.cpp +++ b/src/Storages/VirtualColumnUtils.cpp @@ -50,194 +50,9 @@ namespace DB { -namespace ErrorCodes -{ - extern const int LOGICAL_ERROR; -} - -namespace -{ - -/// Verifying that the function depends only on the specified columns -bool isValidFunction(const ASTPtr & expression, const std::function & is_constant) -{ - const auto * function = expression->as(); - if (function && functionIsInOrGlobalInOperator(function->name)) - { - // Second argument of IN can be a scalar subquery - return isValidFunction(function->arguments->children[0], is_constant); - } - else - return is_constant(expression); -} - -/// Extract all subfunctions of the main conjunction, but depending only on the specified columns -bool extractFunctions(const ASTPtr & expression, const std::function & is_constant, ASTs & result) -{ - const auto * function = expression->as(); - - if (function) - { - if (function->name == "and" || function->name == "indexHint") - { - bool ret = true; - for (const auto & child : function->arguments->children) - ret &= extractFunctions(child, is_constant, result); - return ret; - } - else if (function->name == "or") - { - bool ret = false; - ASTs or_args; - for (const auto & child : function->arguments->children) - ret |= extractFunctions(child, is_constant, or_args); - - if (!or_args.empty()) - { - /// In case of there are less number of arguments for which - /// is_constant() == true, we need to add always-true - /// implicitly to avoid breaking AND invariant. - /// - /// Consider the following: - /// - /// ((value = 10) OR (_table = 'v2')) AND ((_table = 'v1') OR (value = 20)) - /// - /// Without implicit always-true: - /// - /// (_table = 'v2') AND (_table = 'v1') - /// - /// With: - /// - /// (_table = 'v2' OR 1) AND (_table = 'v1' OR 1) -> (_table = 'v2') OR (_table = 'v1') - /// - if (or_args.size() != function->arguments->children.size()) - or_args.push_back(std::make_shared(Field(1))); - result.push_back(makeASTForLogicalOr(std::move(or_args))); - } - return ret; - } - } - - if (isValidFunction(expression, is_constant)) - { - result.push_back(expression->clone()); - return true; - } - else - return false; -} - -/// Construct a conjunction from given functions -ASTPtr buildWhereExpression(ASTs && functions) -{ - if (functions.empty()) - return nullptr; - if (functions.size() == 1) - return functions[0]; - return makeASTForLogicalAnd(std::move(functions)); -} - -} - namespace VirtualColumnUtils { -void rewriteEntityInAst(ASTPtr ast, const String & column_name, const Field & value, const String & func) -{ - auto & select = ast->as(); - if (!select.with()) - select.setExpression(ASTSelectQuery::Expression::WITH, std::make_shared()); - - if (func.empty()) - { - auto literal = std::make_shared(value); - literal->alias = column_name; - literal->prefer_alias_to_column_name = true; - select.with()->children.push_back(literal); - } - else - { - auto literal = std::make_shared(value); - literal->prefer_alias_to_column_name = true; - - auto function = makeASTFunction(func, literal); - function->alias = column_name; - function->prefer_alias_to_column_name = true; - select.with()->children.push_back(function); - } -} - -bool prepareFilterBlockWithQuery(const ASTPtr & query, ContextPtr context, Block block, ASTPtr & expression_ast) -{ - if (block.rows() == 0) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot prepare filter with empty block"); - - /// Take the first row of the input block to build a constant block - auto columns = block.getColumns(); - Columns const_columns(columns.size()); - for (size_t i = 0; i < columns.size(); ++i) - { - if (isColumnConst(*columns[i])) - const_columns[i] = columns[i]->cloneResized(1); - else - const_columns[i] = ColumnConst::create(columns[i]->cloneResized(1), 1); - } - - block.setColumns(const_columns); - - bool unmodified = true; - const auto & select = query->as(); - if (!select.where() && !select.prewhere()) - return unmodified; - - // Provide input columns as constant columns to check if an expression is - // constant and depends on the columns from provided block (the last is - // required to allow skipping some conditions for handling OR). - std::function is_constant = [&block, &context](const ASTPtr & expr) - { - auto actions = std::make_shared(block.getColumnsWithTypeAndName()); - PreparedSetsPtr prepared_sets = std::make_shared(); - const NamesAndTypesList source_columns; - const NamesAndTypesList aggregation_keys; - const ColumnNumbersList grouping_set_keys; - - ActionsVisitor::Data visitor_data( - context, SizeLimits{}, 1, source_columns, std::move(actions), prepared_sets, true, true, true, - { aggregation_keys, grouping_set_keys, GroupByKind::NONE }); - - ActionsVisitor(visitor_data).visit(expr); - actions = visitor_data.getActions(); - auto expr_column_name = expr->getColumnName(); - - const auto * expr_const_node = actions->tryFindInOutputs(expr_column_name); - if (!expr_const_node) - return false; - auto filter_actions = ActionsDAG::buildFilterActionsDAG({expr_const_node}); - const auto & nodes = filter_actions->getNodes(); - bool has_dependent_columns = std::any_of(nodes.begin(), nodes.end(), [&](const auto & node) - { - return block.has(node.result_name); - }); - if (!has_dependent_columns) - return false; - - auto expression_actions = std::make_shared(actions); - auto block_with_constants = block; - expression_actions->execute(block_with_constants, /*dry_run=*/ false, /*allow_duplicates_in_input=*/ true); - return block_with_constants.has(expr_column_name) && isColumnConst(*block_with_constants.getByName(expr_column_name).column); - }; - - /// Create an expression that evaluates the expressions in WHERE and PREWHERE, depending only on the existing columns. - ASTs functions; - if (select.where()) - unmodified &= extractFunctions(select.where(), is_constant, functions); - if (select.prewhere()) - unmodified &= extractFunctions(select.prewhere(), is_constant, functions); - - expression_ast = buildWhereExpression(std::move(functions)); - return unmodified; -} - static void makeSets(const ExpressionActionsPtr & actions, const ContextPtr & context) { for (const auto & node : actions->getNodes()) @@ -294,72 +109,28 @@ void filterBlockWithDAG(ActionsDAGPtr dag, Block & block, ContextPtr context) } } -void filterBlockWithQuery(const ASTPtr & query, Block & block, ContextPtr context, ASTPtr expression_ast) +NameSet getVirtualNamesForFileLikeStorage() { - if (block.rows() == 0) - return; - - if (!expression_ast) - prepareFilterBlockWithQuery(query, context, block, expression_ast); - - if (!expression_ast) - return; - - /// Let's analyze and calculate the prepared expression. - auto syntax_result = TreeRewriter(context).analyze(expression_ast, block.getNamesAndTypesList()); - ExpressionAnalyzer analyzer(expression_ast, syntax_result, context); - ExpressionActionsPtr actions = analyzer.getActions(false /* add alises */, true /* project result */, CompileExpressions::yes); - - makeSets(actions, context); - - Block block_with_filter = block; - actions->execute(block_with_filter, /*dry_run=*/ false, /*allow_duplicates_in_input=*/ true); - - /// Filter the block. - String filter_column_name = expression_ast->getColumnName(); - ColumnPtr filter_column = block_with_filter.getByName(filter_column_name).column->convertToFullIfNeeded(); - if (filter_column->getDataType() != TypeIndex::UInt8) - return; - - ConstantFilterDescription constant_filter(*filter_column); - - if (constant_filter.always_true) - { - return; - } - - if (constant_filter.always_false) - { - block = block.cloneEmpty(); - return; - } - - FilterDescription filter(*filter_column); - - for (size_t i = 0; i < block.columns(); ++i) - { - ColumnPtr & column = block.safeGetByPosition(i).column; - column = column->filter(*filter.data, -1); - } + return {"_path", "_file", "_size"}; } -NamesAndTypesList getPathFileAndSizeVirtualsForStorage(NamesAndTypesList storage_columns) +VirtualColumnsDescription getVirtualsForFileLikeStorage(const ColumnsDescription & storage_columns) { - auto default_virtuals = NamesAndTypesList{ - {"_path", std::make_shared(std::make_shared())}, - {"_file", std::make_shared(std::make_shared())}, - {"_size", makeNullable(std::make_shared())}}; + VirtualColumnsDescription desc; - default_virtuals.sort(); - storage_columns.sort(); + auto add_virtual = [&](const auto & name, const auto & type) + { + if (storage_columns.has(name)) + return; - NamesAndTypesList result_virtuals; - std::set_difference( - default_virtuals.begin(), default_virtuals.end(), storage_columns.begin(), storage_columns.end(), - std::back_inserter(result_virtuals), - [](const NameAndTypePair & lhs, const NameAndTypePair & rhs){ return lhs.name < rhs.name; }); + desc.addEphemeral(name, type, ""); + }; - return result_virtuals; + add_virtual("_path", std::make_shared(std::make_shared())); + add_virtual("_file", std::make_shared(std::make_shared())); + add_virtual("_size", makeNullable(std::make_shared())); + + return desc; } static void addPathAndFileToVirtualColumns(Block & block, const String & path, size_t idx) diff --git a/src/Storages/VirtualColumnUtils.h b/src/Storages/VirtualColumnUtils.h index 7a9b2605339..b5526fc5c7f 100644 --- a/src/Storages/VirtualColumnUtils.h +++ b/src/Storages/VirtualColumnUtils.h @@ -4,6 +4,7 @@ #include #include #include +#include #include @@ -17,23 +18,6 @@ class NamesAndTypesList; namespace VirtualColumnUtils { -/// Adds to the select query section `WITH value AS column_name`, and uses func -/// to wrap the value (if any) -/// -/// For example: -/// - `WITH 9000 as _port`. -/// - `WITH toUInt16(9000) as _port`. -void rewriteEntityInAst(ASTPtr ast, const String & column_name, const Field & value, const String & func = ""); - -/// Prepare `expression_ast` to filter block. Returns true if `expression_ast` is not trimmed, that is, -/// `block` provides all needed columns for `expression_ast`, else return false. -bool prepareFilterBlockWithQuery(const ASTPtr & query, ContextPtr context, Block block, ASTPtr & expression_ast); - -/// Leave in the block only the rows that fit under the WHERE clause and the PREWHERE clause of the query. -/// Only elements of the outer conjunction are considered, depending only on the columns present in the block. -/// If `expression_ast` is passed, use it to filter block. -void filterBlockWithQuery(const ASTPtr & query, Block & block, ContextPtr context, ASTPtr expression_ast = {}); - /// Similar to filterBlockWithQuery, but uses ActionsDAG as a predicate. /// Basically it is filterBlockWithDAG(splitFilterDagForAllowedInputs). void filterBlockWithPredicate(const ActionsDAG::Node * predicate, Block & block, ContextPtr context); @@ -56,7 +40,8 @@ auto extractSingleValueFromBlock(const Block & block, const String & name) return res; } -NamesAndTypesList getPathFileAndSizeVirtualsForStorage(NamesAndTypesList storage_columns); +NameSet getVirtualNamesForFileLikeStorage(); +VirtualColumnsDescription getVirtualsForFileLikeStorage(const ColumnsDescription & storage_columns); ActionsDAGPtr createPathAndFileFilterDAG(const ActionsDAG::Node * predicate, const NamesAndTypesList & virtual_columns); diff --git a/src/Storages/VirtualColumnsDescription.cpp b/src/Storages/VirtualColumnsDescription.cpp new file mode 100644 index 00000000000..64097224ed9 --- /dev/null +++ b/src/Storages/VirtualColumnsDescription.cpp @@ -0,0 +1,94 @@ +#include + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int DUPLICATE_COLUMN; + extern const int NO_SUCH_COLUMN_IN_TABLE; +} + +VirtualColumnDescription::VirtualColumnDescription( + String name_, DataTypePtr type_, ASTPtr codec_, String comment_, VirtualsKind kind_) + : ColumnDescription(std::move(name_), std::move(type_), std::move(codec_), std::move(comment_)) + , kind(kind_) +{ +} + +void VirtualColumnsDescription::add(VirtualColumnDescription desc) +{ + if (container.get<1>().contains(desc.name)) + throw Exception(ErrorCodes::DUPLICATE_COLUMN, "Virtual column {} already exists", desc.name); + + container.get<0>().push_back(std::move(desc)); +} + +void VirtualColumnsDescription::addEphemeral(String name, DataTypePtr type, String comment) +{ + add({std::move(name), std::move(type), nullptr, std::move(comment), VirtualsKind::Ephemeral}); +} + +void VirtualColumnsDescription::addPersistent(String name, DataTypePtr type, ASTPtr codec, String comment) +{ + add({std::move(name), std::move(type), std::move(codec), std::move(comment), VirtualsKind::Persistent}); +} + +std::optional VirtualColumnsDescription::tryGet(const String & name, VirtualsKind kind) const +{ + auto it = container.get<1>().find(name); + if (it != container.get<1>().end() && (static_cast(it->kind) & static_cast(kind))) + return NameAndTypePair{it->name, it->type}; + return {}; +} + +NameAndTypePair VirtualColumnsDescription::get(const String & name, VirtualsKind kind) const +{ + auto column = tryGet(name, kind); + if (!column) + throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "There is no virtual column {}", name); + return *column; +} + +const VirtualColumnDescription * VirtualColumnsDescription::tryGetDescription(const String & name, VirtualsKind kind) const +{ + auto it = container.get<1>().find(name); + if (it != container.get<1>().end() && (static_cast(it->kind) & static_cast(kind))) + return &(*it); + return nullptr; +} + +const VirtualColumnDescription & VirtualColumnsDescription::getDescription(const String & name, VirtualsKind kind) const +{ + const auto * column = tryGetDescription(name, kind); + if (!column) + throw Exception(ErrorCodes::NO_SUCH_COLUMN_IN_TABLE, "There is no virtual column {}", name); + return *column; +} + +Block VirtualColumnsDescription::getSampleBlock() const +{ + Block result; + for (const auto & desc : container) + result.insert({desc.type->createColumn(), desc.type, desc.name}); + return result; +} + +NamesAndTypesList VirtualColumnsDescription::getNamesAndTypesList() const +{ + NamesAndTypesList result; + for (const auto & desc : container) + result.emplace_back(desc.name, desc.type); + return result; +} + +NamesAndTypesList VirtualColumnsDescription::getNamesAndTypesList(VirtualsKind kind) const +{ + NamesAndTypesList result; + for (const auto & column : container) + if (static_cast(column.kind) & static_cast(kind)) + result.emplace_back(column.name, column.type); + return result; +} + +} diff --git a/src/Storages/VirtualColumnsDescription.h b/src/Storages/VirtualColumnsDescription.h new file mode 100644 index 00000000000..e11ba706718 --- /dev/null +++ b/src/Storages/VirtualColumnsDescription.h @@ -0,0 +1,71 @@ +#pragma once +#include +#include + +namespace DB +{ + +struct VirtualColumnDescription : public ColumnDescription +{ + using Self = VirtualColumnDescription; + VirtualsKind kind; + + VirtualColumnDescription() = default; + VirtualColumnDescription(String name_, DataTypePtr type_, ASTPtr codec_, String comment_, VirtualsKind kind_); + + bool isEphemeral() const { return kind == VirtualsKind::Ephemeral; } + bool isPersistent() const { return kind == VirtualsKind::Persistent; } + + /// This method is needed for boost::multi_index because field + /// of base class cannot be referenced in boost::multi_index::member. + const String & getName() const { return name; } +}; + +class VirtualColumnsDescription +{ +public: + using Container = boost::multi_index_container< + VirtualColumnDescription, + boost::multi_index::indexed_by< + boost::multi_index::sequenced<>, + boost::multi_index::ordered_unique>>>; + + using const_iterator = Container::const_iterator; + + const_iterator begin() const { return container.begin(); } + const_iterator end() const { return container.end(); } + + VirtualColumnsDescription() = default; + + void add(VirtualColumnDescription desc); + void addEphemeral(String name, DataTypePtr type, String comment); + void addPersistent(String name, DataTypePtr type, ASTPtr codec, String comment); + + size_t size() const { return container.size(); } + bool empty() const { return container.empty(); } + bool has(const String & name) const { return container.get<1>().contains(name); } + + NameAndTypePair get(const String & name, VirtualsKind kind) const; + std::optional tryGet(const String & name, VirtualsKind kind) const; + + NameAndTypePair get(const String & name) const { return get(name, VirtualsKind::All); } + std::optional tryGet(const String & name) const { return tryGet(name, VirtualsKind::All); } + + const VirtualColumnDescription * tryGetDescription(const String & name, VirtualsKind kind) const; + const VirtualColumnDescription & getDescription(const String & name, VirtualsKind kind) const; + + const VirtualColumnDescription * tryGetDescription(const String & name) const { return tryGetDescription(name, VirtualsKind::All); } + const VirtualColumnDescription & getDescription(const String & name) const { return getDescription(name, VirtualsKind::All); } + + Block getSampleBlock() const; + NamesAndTypesList getNamesAndTypesList() const; + NamesAndTypesList getNamesAndTypesList(VirtualsKind kind) const; + +private: + Container container; +}; + +using VirtualsDescriptionPtr = std::shared_ptr; +using MultiVersionVirtualsDescriptionPtr = MultiVersion; + +} diff --git a/src/Storages/prepareReadingFromFormat.cpp b/src/Storages/prepareReadingFromFormat.cpp index 6be4213ec6b..406b7f379f9 100644 --- a/src/Storages/prepareReadingFromFormat.cpp +++ b/src/Storages/prepareReadingFromFormat.cpp @@ -4,7 +4,7 @@ namespace DB { -ReadFromFormatInfo prepareReadingFromFormat(const Strings & requested_columns, const StorageSnapshotPtr & storage_snapshot, bool supports_subset_of_columns, const NamesAndTypesList & virtuals) +ReadFromFormatInfo prepareReadingFromFormat(const Strings & requested_columns, const StorageSnapshotPtr & storage_snapshot, bool supports_subset_of_columns) { ReadFromFormatInfo info; /// Collect requested virtual columns and remove them from requested columns. @@ -12,11 +12,11 @@ ReadFromFormatInfo prepareReadingFromFormat(const Strings & requested_columns, c for (const auto & column_name : requested_columns) { bool is_virtual = false; - for (const auto & virtual_column : virtuals) + for (const auto & virtual_column : *storage_snapshot->virtual_columns) { if (column_name == virtual_column.name) { - info.requested_virtual_columns.push_back(virtual_column); + info.requested_virtual_columns.emplace_back(virtual_column.name, virtual_column.type); is_virtual = true; break; } diff --git a/src/Storages/prepareReadingFromFormat.h b/src/Storages/prepareReadingFromFormat.h index c5f3959a550..e4d62c29ec6 100644 --- a/src/Storages/prepareReadingFromFormat.h +++ b/src/Storages/prepareReadingFromFormat.h @@ -22,5 +22,5 @@ namespace DB }; /// Get all needed information for reading from data in some input format. - ReadFromFormatInfo prepareReadingFromFormat(const Strings & requested_columns, const StorageSnapshotPtr & storage_snapshot, bool supports_subset_of_columns, const NamesAndTypesList & virtuals); + ReadFromFormatInfo prepareReadingFromFormat(const Strings & requested_columns, const StorageSnapshotPtr & storage_snapshot, bool supports_subset_of_columns); } diff --git a/src/TableFunctions/ITableFunctionFileLike.cpp b/src/TableFunctions/ITableFunctionFileLike.cpp index b697f3df925..1a58be4f75b 100644 --- a/src/TableFunctions/ITableFunctionFileLike.cpp +++ b/src/TableFunctions/ITableFunctionFileLike.cpp @@ -7,6 +7,7 @@ #include #include +#include #include @@ -37,6 +38,11 @@ bool ITableFunctionFileLike::supportsReadingSubsetOfColumns(const ContextPtr & c return format != "auto" && FormatFactory::instance().checkIfFormatSupportsSubsetOfColumns(format, context); } +NameSet ITableFunctionFileLike::getVirtualsToCheckBeforeUsingStructureHint() const +{ + return VirtualColumnUtils::getVirtualNamesForFileLikeStorage(); +} + void ITableFunctionFileLike::parseArguments(const ASTPtr & ast_function, ContextPtr context) { /// Parse args diff --git a/src/TableFunctions/ITableFunctionFileLike.h b/src/TableFunctions/ITableFunctionFileLike.h index c8412905e44..ba1b7d2bb3f 100644 --- a/src/TableFunctions/ITableFunctionFileLike.h +++ b/src/TableFunctions/ITableFunctionFileLike.h @@ -1,6 +1,7 @@ #pragma once #include +#include "Core/Names.h" #include "Parsers/IAST_fwd.h" namespace DB @@ -29,6 +30,8 @@ public: bool supportsReadingSubsetOfColumns(const ContextPtr & context) override; + NameSet getVirtualsToCheckBeforeUsingStructureHint() const override; + static size_t getMaxNumberOfArguments() { return 4; } static void updateStructureAndFormatArgumentsIfNeeded(ASTs & args, const String & structure, const String & format, const ContextPtr &); diff --git a/src/TableFunctions/ITableFunctionXDBC.cpp b/src/TableFunctions/ITableFunctionXDBC.cpp index ca6d40a05a3..a5c16b3a5aa 100644 --- a/src/TableFunctions/ITableFunctionXDBC.cpp +++ b/src/TableFunctions/ITableFunctionXDBC.cpp @@ -153,17 +153,16 @@ ColumnsDescription ITableFunctionXDBC::getActualTableStructure(ContextPtr contex columns_info_uri.addQueryParameter("external_table_functions_use_nulls", toString(use_nulls)); Poco::Net::HTTPBasicCredentials credentials{}; - ReadWriteBufferFromHTTP buf( - columns_info_uri, - Poco::Net::HTTPRequest::HTTP_POST, - {}, - ConnectionTimeouts::getHTTPTimeouts( - context->getSettingsRef(), - context->getServerSettings().keep_alive_timeout), - credentials); + auto buf = BuilderRWBufferFromHTTP(columns_info_uri) + .withConnectionGroup(HTTPConnectionGroupType::STORAGE) + .withMethod(Poco::Net::HTTPRequest::HTTP_POST) + .withTimeouts(ConnectionTimeouts::getHTTPTimeouts( + context->getSettingsRef(), + context->getServerSettings().keep_alive_timeout)) + .create(credentials); std::string columns_info; - readStringBinary(columns_info, buf); + readStringBinary(columns_info, *buf); NamesAndTypesList columns = NamesAndTypesList::parse(columns_info); return ColumnsDescription{columns}; diff --git a/src/TableFunctions/TableFunctionAzureBlobStorage.cpp b/src/TableFunctions/TableFunctionAzureBlobStorage.cpp index 8f558adb09b..275cd2a9cbb 100644 --- a/src/TableFunctions/TableFunctionAzureBlobStorage.cpp +++ b/src/TableFunctions/TableFunctionAzureBlobStorage.cpp @@ -16,6 +16,7 @@ #include #include #include +#include #include #include "registerTableFunctions.h" #include @@ -348,8 +349,7 @@ bool TableFunctionAzureBlobStorage::supportsReadingSubsetOfColumns(const Context std::unordered_set TableFunctionAzureBlobStorage::getVirtualsToCheckBeforeUsingStructureHint() const { - auto virtual_column_names = StorageAzureBlob::getVirtualColumnNames(); - return {virtual_column_names.begin(), virtual_column_names.end()}; + return VirtualColumnUtils::getVirtualNamesForFileLikeStorage(); } StoragePtr TableFunctionAzureBlobStorage::executeImpl(const ASTPtr & /*ast_function*/, ContextPtr context, const std::string & table_name, ColumnsDescription /*cached_columns*/, bool is_insert_query) const diff --git a/src/TableFunctions/TableFunctionExecutable.cpp b/src/TableFunctions/TableFunctionExecutable.cpp index 9fa5ddf15c2..2c3802e8667 100644 --- a/src/TableFunctions/TableFunctionExecutable.cpp +++ b/src/TableFunctions/TableFunctionExecutable.cpp @@ -71,6 +71,9 @@ std::vector TableFunctionExecutable::skipAnalysisForArguments(const Quer const auto & table_function_node_arguments = table_function_node.getArguments().getNodes(); size_t table_function_node_arguments_size = table_function_node_arguments.size(); + if (table_function_node_arguments_size <= 2) + return {}; + std::vector result_indexes; result_indexes.reserve(table_function_node_arguments_size - 2); for (size_t i = 2; i < table_function_node_arguments_size; ++i) diff --git a/src/TableFunctions/TableFunctionFile.cpp b/src/TableFunctions/TableFunctionFile.cpp index b481076e9b6..28bf72e07fb 100644 --- a/src/TableFunctions/TableFunctionFile.cpp +++ b/src/TableFunctions/TableFunctionFile.cpp @@ -112,12 +112,6 @@ ColumnsDescription TableFunctionFile::getActualTableStructure(ContextPtr context return parseColumnsListFromString(structure, context); } -std::unordered_set TableFunctionFile::getVirtualsToCheckBeforeUsingStructureHint() const -{ - auto virtual_column_names = StorageFile::getVirtualColumnNames(); - return {virtual_column_names.begin(), virtual_column_names.end()}; -} - void registerTableFunctionFile(TableFunctionFactory & factory) { factory.registerFunction(); diff --git a/src/TableFunctions/TableFunctionFile.h b/src/TableFunctions/TableFunctionFile.h index c1924028b49..aaf5ba8873a 100644 --- a/src/TableFunctions/TableFunctionFile.h +++ b/src/TableFunctions/TableFunctionFile.h @@ -22,8 +22,6 @@ public: ColumnsDescription getActualTableStructure(ContextPtr context, bool is_insert_query) const override; - std::unordered_set getVirtualsToCheckBeforeUsingStructureHint() const override; - protected: int fd = -1; String path_to_archive; diff --git a/src/TableFunctions/TableFunctionHDFS.cpp b/src/TableFunctions/TableFunctionHDFS.cpp index 2dac4398144..45829245551 100644 --- a/src/TableFunctions/TableFunctionHDFS.cpp +++ b/src/TableFunctions/TableFunctionHDFS.cpp @@ -41,12 +41,6 @@ ColumnsDescription TableFunctionHDFS::getActualTableStructure(ContextPtr context return parseColumnsListFromString(structure, context); } -std::unordered_set TableFunctionHDFS::getVirtualsToCheckBeforeUsingStructureHint() const -{ - auto virtual_column_names = StorageHDFS::getVirtualColumnNames(); - return {virtual_column_names.begin(), virtual_column_names.end()}; -} - void registerTableFunctionHDFS(TableFunctionFactory & factory) { factory.registerFunction(); diff --git a/src/TableFunctions/TableFunctionHDFS.h b/src/TableFunctions/TableFunctionHDFS.h index 3a719496b26..f1c0b8a7eae 100644 --- a/src/TableFunctions/TableFunctionHDFS.h +++ b/src/TableFunctions/TableFunctionHDFS.h @@ -36,8 +36,6 @@ public: ColumnsDescription getActualTableStructure(ContextPtr context, bool is_insert_query) const override; - std::unordered_set getVirtualsToCheckBeforeUsingStructureHint() const override; - private: StoragePtr getStorage( const String & source, const String & format_, const ColumnsDescription & columns, ContextPtr global_context, diff --git a/src/TableFunctions/TableFunctionS3.cpp b/src/TableFunctions/TableFunctionS3.cpp index c00b1e2e3e5..a8c100ebd44 100644 --- a/src/TableFunctions/TableFunctionS3.cpp +++ b/src/TableFunctions/TableFunctionS3.cpp @@ -16,6 +16,7 @@ #include #include #include +#include #include #include "registerTableFunctions.h" #include @@ -401,8 +402,7 @@ bool TableFunctionS3::supportsReadingSubsetOfColumns(const ContextPtr & context) std::unordered_set TableFunctionS3::getVirtualsToCheckBeforeUsingStructureHint() const { - auto virtual_column_names = StorageS3::getVirtualColumnNames(); - return {virtual_column_names.begin(), virtual_column_names.end()}; + return VirtualColumnUtils::getVirtualNamesForFileLikeStorage(); } StoragePtr TableFunctionS3::executeImpl(const ASTPtr & /*ast_function*/, ContextPtr context, const std::string & table_name, ColumnsDescription cached_columns, bool /*is_insert_query*/) const diff --git a/src/TableFunctions/TableFunctionURL.cpp b/src/TableFunctions/TableFunctionURL.cpp index a78b2affa9a..2bdc0b449e0 100644 --- a/src/TableFunctions/TableFunctionURL.cpp +++ b/src/TableFunctions/TableFunctionURL.cpp @@ -159,12 +159,6 @@ ColumnsDescription TableFunctionURL::getActualTableStructure(ContextPtr context, return parseColumnsListFromString(structure, context); } -std::unordered_set TableFunctionURL::getVirtualsToCheckBeforeUsingStructureHint() const -{ - auto virtual_column_names = StorageURL::getVirtualColumnNames(); - return {virtual_column_names.begin(), virtual_column_names.end()}; -} - std::optional TableFunctionURL::tryGetFormatFromFirstArgument() { return FormatFactory::instance().tryGetFormatFromFileName(Poco::URI(filename).getPath()); diff --git a/src/TableFunctions/TableFunctionURL.h b/src/TableFunctions/TableFunctionURL.h index 54e223283ba..a1efddb84c6 100644 --- a/src/TableFunctions/TableFunctionURL.h +++ b/src/TableFunctions/TableFunctionURL.h @@ -36,8 +36,6 @@ public: static void updateStructureAndFormatArgumentsIfNeeded(ASTs & args, const String & structure_, const String & format_, const ContextPtr & context); - std::unordered_set getVirtualsToCheckBeforeUsingStructureHint() const override; - protected: void parseArguments(const ASTPtr & ast, ContextPtr context) override; void parseArgumentsImpl(ASTs & args, const ContextPtr & context) override; diff --git a/tests/analyzer_integration_broken_tests.txt b/tests/analyzer_integration_broken_tests.txt index e819e134706..31527dc3476 100644 --- a/tests/analyzer_integration_broken_tests.txt +++ b/tests/analyzer_integration_broken_tests.txt @@ -2,5 +2,3 @@ test_build_sets_from_multiple_threads/test.py::test_set test_concurrent_backups_s3/test.py::test_concurrent_backups test_distributed_type_object/test.py::test_distributed_type_object test_merge_table_over_distributed/test.py::test_global_in -test_merge_table_over_distributed/test.py::test_select_table_name_from_merge_over_distributed -test_select_access_rights/test_main.py::test_alias_columns diff --git a/tests/ci/.mypy.ini b/tests/ci/.mypy.ini index bcaac49f190..9bc44025826 100644 --- a/tests/ci/.mypy.ini +++ b/tests/ci/.mypy.ini @@ -14,4 +14,4 @@ warn_unused_ignores = False warn_return_any = True no_implicit_reexport = True strict_equality = True -strict_concatenate = True +extra_checks = True diff --git a/tests/ci/ast_fuzzer_check.py b/tests/ci/ast_fuzzer_check.py index 6e3da7fa816..1ecf805cadc 100644 --- a/tests/ci/ast_fuzzer_check.py +++ b/tests/ci/ast_fuzzer_check.py @@ -67,14 +67,14 @@ def main(): build_name = get_build_name_for_check(check_name) urls = read_build_urls(build_name, reports_path) if not urls: - raise Exception("No build URLs found") + raise ValueError("No build URLs found") for url in urls: if url.endswith("/clickhouse"): build_url = url break else: - raise Exception("Cannot find the clickhouse binary among build results") + raise ValueError("Cannot find the clickhouse binary among build results") logging.info("Got build url %s", build_url) diff --git a/tests/ci/autoscale_runners_lambda/app.py b/tests/ci/autoscale_runners_lambda/app.py index 26a05ab0af4..1aa2e424320 100644 --- a/tests/ci/autoscale_runners_lambda/app.py +++ b/tests/ci/autoscale_runners_lambda/app.py @@ -51,7 +51,7 @@ class Queue: label: str -def get_scales(runner_type: str) -> Tuple[int, int]: +def get_scales() -> Tuple[int, int]: "returns the multipliers for scaling down and up ASG by types" # Scaling down is quicker on the lack of running jobs than scaling up on # queue @@ -95,7 +95,7 @@ def set_capacity( continue raise ValueError("Queue status is not in ['in_progress', 'queued']") - scale_down, scale_up = get_scales(runner_type) + scale_down, scale_up = get_scales() # With lyfecycle hooks some instances are actually free because some of # them are in 'Terminating:Wait' state effective_capacity = max( diff --git a/tests/ci/bugfix_validate_check.py b/tests/ci/bugfix_validate_check.py index ae7fce1f102..7aaf18e7765 100644 --- a/tests/ci/bugfix_validate_check.py +++ b/tests/ci/bugfix_validate_check.py @@ -1,37 +1,37 @@ #!/usr/bin/env python3 -from pathlib import Path -import subprocess -import sys -from typing import List, Sequence, Tuple import csv import logging +import subprocess +import sys +from pathlib import Path +from typing import List, Sequence, Tuple -from report import ( - ERROR, - FAILURE, - SKIPPED, - SUCCESS, - FAIL, - OK, - TestResult, - TestResults, - JobReport, -) -from env_helper import TEMP_PATH -from stopwatch import Stopwatch from ci_config import JobNames from ci_utils import normalize_string +from env_helper import TEMP_PATH from functional_test_check import NO_CHANGES_MSG +from report import ( + ERROR, + FAIL, + FAILURE, + OK, + SKIPPED, + SUCCESS, + JobReport, + TestResult, + TestResults, +) +from stopwatch import Stopwatch def post_commit_status_from_file(file_path: Path) -> List[str]: with open(file_path, "r", encoding="utf-8") as f: res = list(csv.reader(f, delimiter="\t")) if len(res) < 1: - raise Exception(f'Can\'t read from "{file_path}"') + raise IndexError(f'Can\'t read from "{file_path}"') if len(res[0]) != 3: - raise Exception(f'Can\'t read from "{file_path}"') + raise IndexError(f'Can\'t read from "{file_path}"') return res[0] diff --git a/tests/ci/build_download_helper.py b/tests/ci/build_download_helper.py index 0d24cb80021..66734f58b23 100644 --- a/tests/ci/build_download_helper.py +++ b/tests/ci/build_download_helper.py @@ -8,7 +8,10 @@ import time from pathlib import Path from typing import Any, Callable, List, Union -import requests # type: ignore +# isort: off +import requests + +# isort: on import get_robot_token as grt # we need an updated ROBOT_TOKEN from ci_config import CI_CONFIG @@ -30,9 +33,10 @@ def get_with_retries( "Getting URL with %i tries and sleep %i in between: %s", retries, sleep, url ) exc = Exception("A placeholder to satisfy typing and avoid nesting") + timeout = kwargs.pop("timeout", 30) for i in range(retries): try: - response = requests.get(url, **kwargs) + response = requests.get(url, timeout=timeout, **kwargs) response.raise_for_status() return response except Exception as e: @@ -74,10 +78,11 @@ def get_gh_api( token_is_set = "Authorization" in kwargs.get("headers", {}) exc = Exception("A placeholder to satisfy typing and avoid nesting") try_cnt = 0 + timeout = kwargs.pop("timeout", 30) while try_cnt < retries: try_cnt += 1 try: - response = requests.get(url, **kwargs) + response = requests.get(url, timeout=timeout, **kwargs) response.raise_for_status() return response except requests.HTTPError as e: @@ -85,7 +90,8 @@ def get_gh_api( ratelimit_exceeded = ( e.response.status_code == 403 and b"rate limit exceeded" - in e.response._content # pylint:disable=protected-access + # pylint:disable-next=protected-access + in (e.response._content or b"") ) try_auth = e.response.status_code == 404 if (ratelimit_exceeded or try_auth) and not token_is_set: diff --git a/tests/ci/build_report_check.py b/tests/ci/build_report_check.py index 94e429ad77b..48640f15ac0 100644 --- a/tests/ci/build_report_check.py +++ b/tests/ci/build_report_check.py @@ -51,7 +51,7 @@ def main(): builds_for_check = CI_CONFIG.get_builds_for_report( build_check_name, release=pr_info.is_release(), - backport=pr_info.head_ref.startswith("backport"), + backport=pr_info.head_ref.startswith("backport/"), ) required_builds = len(builds_for_check) missing_builds = 0 diff --git a/tests/ci/cancel_and_rerun_workflow_lambda/app.py b/tests/ci/cancel_and_rerun_workflow_lambda/app.py index 656198c6985..625936ec5c8 100644 --- a/tests/ci/cancel_and_rerun_workflow_lambda/app.py +++ b/tests/ci/cancel_and_rerun_workflow_lambda/app.py @@ -1,16 +1,15 @@ #!/usr/bin/env python3 import json -import re import time from base64 import b64decode from collections import namedtuple from queue import Queue from threading import Thread -from typing import Any, Dict, List, Optional, Tuple +from typing import Any, Dict, List, Optional -import requests # type: ignore -from lambda_shared.pr import CATEGORY_TO_LABEL, check_pr_description +import requests +from lambda_shared.pr import check_pr_description from lambda_shared.token import get_cached_access_token NEED_RERUN_OR_CANCELL_WORKFLOWS = { @@ -48,16 +47,18 @@ class Worker(Thread): def _exec_get_with_retry(url: str, token: str) -> dict: headers = {"Authorization": f"token {token}"} + e = Exception() for i in range(MAX_RETRY): try: - response = requests.get(url, headers=headers) + response = requests.get(url, headers=headers, timeout=30) response.raise_for_status() return response.json() # type: ignore except Exception as ex: print("Got exception executing request", ex) + e = ex time.sleep(i + 1) - raise Exception("Cannot execute GET request with retries") + raise requests.HTTPError("Cannot execute GET request with retries") from e WorkflowDescription = namedtuple( @@ -215,16 +216,18 @@ def get_workflow_description(workflow_url: str, token: str) -> WorkflowDescripti def _exec_post_with_retry(url: str, token: str, json: Optional[Any] = None) -> Any: headers = {"Authorization": f"token {token}"} + e = Exception() for i in range(MAX_RETRY): try: - response = requests.post(url, headers=headers, json=json) + response = requests.post(url, headers=headers, json=json, timeout=30) response.raise_for_status() return response.json() except Exception as ex: print("Got exception executing request", ex) + e = ex time.sleep(i + 1) - raise Exception("Cannot execute POST request with retry") + raise requests.HTTPError("Cannot execute POST request with retry") from e def exec_workflow_url(urls_to_post, token): diff --git a/tests/ci/cherry_pick.py b/tests/ci/cherry_pick.py index 2c40b2a4099..d92504e30bd 100644 --- a/tests/ci/cherry_pick.py +++ b/tests/ci/cherry_pick.py @@ -456,11 +456,13 @@ class Backport: tomorrow = date.today() + timedelta(days=1) logging.info("Receive PRs suppose to be backported") - query_args = dict( - query=f"type:pr repo:{self._fetch_from} -label:{self.backport_created_label}", - label=",".join(self.labels_to_backport + [self.must_create_backport_label]), - merged=[since_date, tomorrow], - ) + query_args = { + "query": f"type:pr repo:{self._fetch_from} -label:{self.backport_created_label}", + "label": ",".join( + self.labels_to_backport + [self.must_create_backport_label] + ), + "merged": [since_date, tomorrow], + } logging.info("Query to find the backport PRs:\n %s", query_args) self.prs_for_backport = self.gh.get_pulls_from_search(**query_args) logging.info( diff --git a/tests/ci/ci.py b/tests/ci/ci.py index 234eec48463..1eec9a6771b 100644 --- a/tests/ci/ci.py +++ b/tests/ci/ci.py @@ -17,7 +17,7 @@ from typing import Any, Dict, List, Optional, Sequence, Set, Union import docker_images_helper import upload_result_helper from build_check import get_release_or_pr -from ci_config import CI_CONFIG, Build, JobNames, Labels +from ci_config import CI_CONFIG, Build, CIStages, Labels, JobNames from ci_utils import GHActions, is_hex, normalize_string from clickhouse_helper import ( CiLogsCredentials, @@ -397,7 +397,7 @@ class CiCache: status.dump_to_file(record_file) elif record_type == self.RecordType.PENDING: assert isinstance(status, PendingState) - with open(record_file, "w") as json_file: + with open(record_file, "w", encoding="utf-8") as json_file: json.dump(asdict(status), json_file) else: assert False @@ -645,7 +645,8 @@ class CiCache: if not jobs_with_params: return {} poll_interval_sec = 300 - TIMEOUT = 3590 + # TIMEOUT * MAX_ROUNDS_TO_WAIT must be less than 6h (GH job timeout) with a room for rest RunConfig work + TIMEOUT = 3000 # 50 min MAX_ROUNDS_TO_WAIT = 6 MAX_JOB_NUM_TO_WAIT = 3 await_finished: Dict[str, List[int]] = {} @@ -1005,7 +1006,7 @@ def _mark_success_action( def _print_results(result: Any, outfile: Optional[str], pretty: bool = False) -> None: if outfile: - with open(outfile, "w") as f: + with open(outfile, "w", encoding="utf-8") as f: if isinstance(result, str): print(result, file=f) elif isinstance(result, dict): @@ -1110,7 +1111,7 @@ def _configure_jobs( digests: Dict[str, str] = {} print("::group::Job Digests") - for job in CI_CONFIG.job_generator(): + for job in CI_CONFIG.job_generator(pr_info.head_ref): digest = job_digester.get_job_digest(CI_CONFIG.get_digest_config(job)) digests[job] = digest print(f" job [{job.rjust(50)}] has digest [{digest}]") @@ -1118,15 +1119,14 @@ def _configure_jobs( ## b. check what we need to run ci_cache = None - if not ci_cache_disabled: + if not ci_cache_disabled and CI: ci_cache = CiCache(s3, digests).update() ci_cache.print_status() jobs_to_wait: Dict[str, Dict[str, Any]] = {} randomization_buckets = {} # type: Dict[str, Set[str]] - for job in digests: - digest = digests[job] + for job, digest in digests.items(): job_config = CI_CONFIG.get_job_config(job) num_batches: int = job_config.num_batches batches_to_do: List[int] = [] @@ -1282,6 +1282,29 @@ def _configure_jobs( } +def _generate_ci_stage_config(jobs_data: Dict[str, Any]) -> Dict[str, Dict[str, Any]]: + """ + populates GH Actions' workflow with real jobs + "Builds_1": [{"job_name": NAME, "runner_type": RUNER_TYPE}] + "Tests_1": [{"job_name": NAME, "runner_type": RUNER_TYPE}] + ... + """ + result = {} # type: Dict[str, Any] + stages_to_do = [] + for job in jobs_data["jobs_to_do"]: + stage_type = CI_CONFIG.get_job_ci_stage(job) + if stage_type == CIStages.NA: + continue + if stage_type not in result: + result[stage_type] = [] + stages_to_do.append(stage_type) + result[stage_type].append( + {"job_name": job, "runner_type": CI_CONFIG.get_runner_type(job)} + ) + result["stages_to_do"] = stages_to_do + return result + + def _create_gh_status( commit: Any, job: str, batch: int, num_batches: int, job_status: CommitStatusData ) -> None: @@ -1612,11 +1635,11 @@ def main() -> int: indata: Optional[Dict[str, Any]] = None if args.infile: - indata = ( - json.loads(args.infile) - if not os.path.isfile(args.infile) - else json.load(open(args.infile)) - ) + if os.path.isfile(args.infile): + with open(args.infile, encoding="utf-8") as jfd: + indata = json.load(jfd) + else: + indata = json.loads(args.infile) assert indata and isinstance(indata, dict), "Invalid --infile json" result: Dict[str, Any] = {} @@ -1733,6 +1756,7 @@ def main() -> int: result["build"] = build_digest result["docs"] = docs_digest result["ci_flags"] = ci_flags + result["stages_data"] = _generate_ci_stage_config(jobs_data) result["jobs_data"] = jobs_data result["docker_data"] = docker_data ### CONFIGURE action: end @@ -1771,24 +1795,29 @@ def main() -> int: print(build_result.as_json()) print("::endgroup::") else: - # this is a test job - check if GH commit status is present - - # rerun helper check - # FIXME: remove rerun_helper check and rely on ci cache only + # this is a test job - check if GH commit status or cache record is present commit = get_commit( Github(get_best_robot_token(), per_page=100), pr_info.sha ) - rerun_helper = RerunHelper(commit, check_name_with_group) - if rerun_helper.is_already_finished_by_status(): - status = rerun_helper.get_finished_status() - assert status - previous_status = status.state - print("::group::Commit Status") - print(status) - print("::endgroup::") + + # rerun helper check + # FIXME: remove rerun_helper check and rely on ci cache only + if check_name not in ( + # we might want to rerun reports' jobs - disable rerun check for them + JobNames.BUILD_CHECK, + JobNames.BUILD_CHECK_SPECIAL, + ): + rerun_helper = RerunHelper(commit, check_name_with_group) + if rerun_helper.is_already_finished_by_status(): + status = rerun_helper.get_finished_status() + assert status + previous_status = status.state + print("::group::Commit Status") + print(status) + print("::endgroup::") # ci cache check - elif not indata["ci_flags"][Labels.NO_CI_CACHE]: + if not previous_status and not indata["ci_flags"][Labels.NO_CI_CACHE]: ci_cache = CiCache(s3, indata["jobs_data"]["digests"]).update() job_config = CI_CONFIG.get_job_config(check_name) if ci_cache.is_successful( diff --git a/tests/ci/ci_config.py b/tests/ci/ci_config.py index b9c332e8e18..df8bfb1c2a8 100644 --- a/tests/ci/ci_config.py +++ b/tests/ci/ci_config.py @@ -1,5 +1,6 @@ #!/usr/bin/env python3 +from copy import deepcopy import logging from argparse import ArgumentDefaultsHelpFormatter, ArgumentParser from dataclasses import dataclass, field @@ -10,6 +11,32 @@ from ci_utils import WithIter from integration_test_images import IMAGES +class WorkFlows(metaclass=WithIter): + PULL_REQUEST = "PULL_REQUEST" + MASTER = "MASTER" + BACKPORT = "BACKPORT" + RELEASE = "RELEASE" + SYNC = "SYNC" + + +class CIStages(metaclass=WithIter): + NA = "UNKNOWN" + BUILDS_1 = "Builds_1" + BUILDS_2 = "Builds_2" + TESTS_1 = "Tests_1" + TESTS_2 = "Tests_2" + + +class Runners(metaclass=WithIter): + BUILDER = "builder" + STYLE_CHECKER = "style-checker" + STYLE_CHECKER_ARM = "style-checker-aarch64" + FUNC_TESTER = "func-tester" + FUNC_TESTER_ARM = "func-tester-aarch64" + STRESS_TESTER = "stress-tester" + FUZZER_UNIT_TESTER = "fuzzer-unit-tester" + + class Labels(metaclass=WithIter): """ Label names or commit tokens in normalized form @@ -207,6 +234,45 @@ class JobConfig: random_bucket: str = "" +builds_job_config = JobConfig( + required_on_release_branch=True, + digest=DigestConfig( + include_paths=[ + "./src", + "./contrib/*-cmake", + "./contrib/consistent-hashing", + "./contrib/murmurhash", + "./contrib/libfarmhash", + "./contrib/pdqsort", + "./contrib/cityhash102", + "./contrib/sparse-checkout", + "./contrib/libmetrohash", + "./contrib/update-submodules.sh", + "./contrib/CMakeLists.txt", + "./CMakeLists.txt", + "./PreLoad.cmake", + "./cmake", + "./base", + "./programs", + "./packages", + "./docker/packager/packager", + "./rust", + # FIXME: This is a WA to rebuild the CH and recreate the Performance.tar.zst artifact + # when there are changes in performance test scripts. + # Due to the current design of the perf test we need to rebuild CH when the performance test changes, + # otherwise the changes will not be visible in the PerformanceTest job in CI + "./tests/performance", + ], + exclude_files=[".md"], + docker=["clickhouse/binary-builder"], + git_submodules=True, + ), + run_command="build_check.py $BUILD_NAME", +) +fuzzer_build_job_config = deepcopy(builds_job_config) +fuzzer_build_job_config.run_by_label = Labels.libFuzzer + + @dataclass class BuildConfig: name: str @@ -222,43 +288,7 @@ class BuildConfig: sparse_checkout: bool = False comment: str = "" static_binary_name: str = "" - job_config: JobConfig = field( - default_factory=lambda: JobConfig( - required_on_release_branch=True, - digest=DigestConfig( - include_paths=[ - "./src", - "./contrib/*-cmake", - "./contrib/consistent-hashing", - "./contrib/murmurhash", - "./contrib/libfarmhash", - "./contrib/pdqsort", - "./contrib/cityhash102", - "./contrib/sparse-checkout", - "./contrib/libmetrohash", - "./contrib/update-submodules.sh", - "./contrib/CMakeLists.txt", - "./CMakeLists.txt", - "./PreLoad.cmake", - "./cmake", - "./base", - "./programs", - "./packages", - "./docker/packager/packager", - "./rust", - # FIXME: This is a WA to rebuild the CH and recreate the Performance.tar.zst artifact - # when there are changes in performance test scripts. - # Due to the current design of the perf test we need to rebuild CH when the performance test changes, - # otherwise the changes will not be visible in the PerformanceTest job in CI - "./tests/performance", - ], - exclude_files=[".md"], - docker=["clickhouse/binary-builder"], - git_submodules=True, - ), - run_command="build_check.py $BUILD_NAME", - ) - ) + job_config: JobConfig = field(default_factory=lambda: deepcopy(builds_job_config)) def export_env(self, export: bool = False) -> str: def process(field_name: str, field: Union[bool, str]) -> str: @@ -451,20 +481,36 @@ perf_test_common_params = { "digest": perf_check_digest, "run_command": "performance_comparison_check.py", } -sqllancer_test_common_params = { - "digest": sqllancer_check_digest, - "run_command": "sqlancer_check.py", - "run_always": True, +sqllancer_test_common_params = JobConfig( + digest=sqllancer_check_digest, + run_command="sqlancer_check.py", + release_only=True, + run_always=True, +) +sqllogic_test_params = JobConfig( + digest=sqllogic_check_digest, + run_command="sqllogic_test.py", + timeout=10800, + release_only=True, +) +sql_test_params = JobConfig( + digest=sqltest_check_digest, + run_command="sqltest.py", + timeout=10800, + release_only=True, +) +clickbench_test_params = { + "digest": DigestConfig( + include_paths=[ + "tests/ci/clickbench.py", + ], + docker=["clickhouse/clickbench"], + ), + "run_command": 'clickbench.py "$CHECK_NAME"', } -sqllogic_test_params = { - "digest": sqllogic_check_digest, - "run_command": "sqllogic_test.py", - "timeout": 10800, -} -sql_test_params = { - "digest": sqltest_check_digest, - "run_command": "sqltest.py", - "timeout": 10800, +install_test_params = { + "digest": install_check_digest, + "run_command": 'install_check.py "$CHECK_NAME"', } @@ -487,6 +533,45 @@ class CIConfig: return config return None + def get_job_ci_stage(self, job_name: str) -> str: + if job_name in [ + JobNames.STYLE_CHECK, + JobNames.FAST_TEST, + JobNames.JEPSEN_KEEPER, + JobNames.BUILD_CHECK, + JobNames.BUILD_CHECK_SPECIAL, + ]: + # FIXME: we can't currently handle Jepsen in the Stage as it's job has concurrency directive + # BUILD_CHECK and BUILD_CHECK_SPECIAL runs not in stage because we need them even if Builds stage failed + return CIStages.NA + stage_type = None + if self.is_build_job(job_name): + stage_type = CIStages.BUILDS_1 + if job_name in CI_CONFIG.get_builds_for_report( + JobNames.BUILD_CHECK_SPECIAL + ): + # special builds go to Build_2 stage to not delay Builds_1/Test_1 + stage_type = CIStages.BUILDS_2 + elif self.is_docs_job(job_name): + stage_type = CIStages.TESTS_1 + elif job_name == JobNames.BUILD_CHECK_SPECIAL: + stage_type = CIStages.TESTS_2 + elif self.is_test_job(job_name): + stage_type = CIStages.TESTS_1 + if job_name in CI_CONFIG.test_configs: + required_build = CI_CONFIG.test_configs[job_name].required_build + assert required_build + if required_build in CI_CONFIG.get_builds_for_report( + JobNames.BUILD_CHECK + ): + stage_type = CIStages.TESTS_1 + else: + stage_type = CIStages.TESTS_2 + else: + stage_type = CIStages.TESTS_1 + assert stage_type, f"BUG [{job_name}]" + return stage_type + def get_job_config(self, check_name: str) -> JobConfig: res = None for config in ( @@ -500,6 +585,63 @@ class CIConfig: break return res # type: ignore + def get_runner_type(self, check_name: str) -> str: + result = None + if self.is_build_job(check_name) or check_name == JobNames.FAST_TEST: + result = Runners.BUILDER + elif any( + words in check_name.lower() + for words in [ + "install packages", + "compatibility check", + "docker", + "build check", + "jepsen", + "style check", + ] + ): + result = Runners.STYLE_CHECKER + elif check_name == JobNames.DOCS_CHECK: + # docs job is demanding + result = Runners.FUNC_TESTER_ARM + elif any( + words in check_name.lower() + for words in [ + "stateless", + "stateful", + "clickbench", + "sqllogic test", + "libfuzzer", + "bugfix validation", + ] + ): + result = Runners.FUNC_TESTER + elif any( + words in check_name.lower() + for words in ["stress", "upgrade", "integration", "performance comparison"] + ): + result = Runners.STRESS_TESTER + elif any( + words in check_name.lower() + for words in ["ast fuzzer", "unit tests", "sqlancer", "sqltest"] + ): + result = Runners.FUZZER_UNIT_TESTER + + assert result, f"BUG, no runner for [{check_name}]" + + if ("aarch" in check_name or "arm" in check_name) and "aarch" not in result: + if result == Runners.STRESS_TESTER: + # FIXME: no arm stress tester group atm + result = Runners.FUNC_TESTER_ARM + elif result == Runners.BUILDER: + # crosscompile - no arm required + pass + else: + # switch to aarch64 runnner + result += "-aarch64" + + return result + @staticmethod def normalize_string(input_string: str) -> str: lowercase_string = input_string.lower() @@ -560,18 +702,18 @@ class CIConfig: ), f"Invalid check_name or CI_CONFIG outdated, config not found for [{check_name}]" return res # type: ignore - def job_generator(self) -> Iterable[str]: + def job_generator(self, branch: str) -> Iterable[str]: """ traverses all check names in CI pipeline """ + assert branch for config in ( self.other_jobs_configs, self.build_config, self.builds_report_config, self.test_configs, ): - for check_name in config: # type: ignore - yield check_name + yield from config # type: ignore def get_builds_for_report( self, report_name: str, release: bool = False, backport: bool = False @@ -586,7 +728,7 @@ class CIConfig: Build.PACKAGE_TSAN, Build.PACKAGE_DEBUG, ] - if release and report_name == JobNames.BUILD_CHECK_SPECIAL: + if (release or backport) and report_name == JobNames.BUILD_CHECK_SPECIAL: return [ Build.BINARY_DARWIN, Build.BINARY_DARWIN_AARCH64, @@ -600,11 +742,7 @@ class CIConfig: @classmethod def is_test_job(cls, job: str) -> bool: - return ( - not cls.is_build_job(job) - and not cls.is_build_job(job) - and job != JobNames.STYLE_CHECK - ) + return not cls.is_build_job(job) and job != JobNames.STYLE_CHECK @classmethod def is_docs_job(cls, job: str) -> bool: @@ -698,17 +836,15 @@ CI_CONFIG = CIConfig( job for job in JobNames if not any( - [ - nogo in job - for nogo in ( - "asan", - "tsan", - "msan", - "ubsan", - # skip build report jobs as not all builds will be done - "build check", - ) - ] + nogo in job + for nogo in ( + "asan", + "tsan", + "msan", + "ubsan", + # skip build report jobs as not all builds will be done + "build check", + ) ) ] ), @@ -845,7 +981,7 @@ CI_CONFIG = CIConfig( name=Build.FUZZERS, compiler="clang-17", package_type="fuzzers", - job_config=JobConfig(run_by_label=Labels.libFuzzer), + job_config=fuzzer_build_job_config, ), }, builds_report_config={ @@ -889,6 +1025,7 @@ CI_CONFIG = CIConfig( include_paths=["**/*.md", "./docs", "tests/ci/docs_check.py"], docker=["clickhouse/docs-builder"], ), + run_command="docs_check.py", ), ), JobNames.FAST_TEST: TestConfig( @@ -918,10 +1055,10 @@ CI_CONFIG = CIConfig( }, test_configs={ JobNames.INSTALL_TEST_AMD: TestConfig( - Build.PACKAGE_RELEASE, job_config=JobConfig(digest=install_check_digest) + Build.PACKAGE_RELEASE, job_config=JobConfig(**install_test_params) # type: ignore ), JobNames.INSTALL_TEST_ARM: TestConfig( - Build.PACKAGE_AARCH64, job_config=JobConfig(digest=install_check_digest) + Build.PACKAGE_AARCH64, job_config=JobConfig(**install_test_params) # type: ignore ), JobNames.STATEFUL_TEST_ASAN: TestConfig( Build.PACKAGE_ASAN, job_config=JobConfig(**stateful_test_common_params) # type: ignore @@ -1128,20 +1265,29 @@ CI_CONFIG = CIConfig( job_config=JobConfig(num_batches=4, run_by_label="pr-performance", **perf_test_common_params), # type: ignore ), JobNames.SQLANCER: TestConfig( - Build.PACKAGE_RELEASE, job_config=JobConfig(**sqllancer_test_common_params) # type: ignore + Build.PACKAGE_RELEASE, job_config=sqllancer_test_common_params ), JobNames.SQLANCER_DEBUG: TestConfig( - Build.PACKAGE_DEBUG, job_config=JobConfig(**sqllancer_test_common_params) # type: ignore + Build.PACKAGE_DEBUG, job_config=sqllancer_test_common_params ), JobNames.SQL_LOGIC_TEST: TestConfig( - Build.PACKAGE_RELEASE, job_config=JobConfig(**sqllogic_test_params) # type: ignore + Build.PACKAGE_RELEASE, job_config=sqllogic_test_params ), - JobNames.SQLTEST: TestConfig( - Build.PACKAGE_RELEASE, job_config=JobConfig(**sql_test_params) # type: ignore + JobNames.SQLTEST: TestConfig(Build.PACKAGE_RELEASE, job_config=sql_test_params), + JobNames.CLCIKBENCH_TEST: TestConfig( + Build.PACKAGE_RELEASE, job_config=JobConfig(**clickbench_test_params) # type: ignore ), - JobNames.CLCIKBENCH_TEST: TestConfig(Build.PACKAGE_RELEASE), - JobNames.CLCIKBENCH_TEST_ARM: TestConfig(Build.PACKAGE_AARCH64), - JobNames.LIBFUZZER_TEST: TestConfig(Build.FUZZERS, job_config=JobConfig(run_by_label=Labels.libFuzzer)), # type: ignore + JobNames.CLCIKBENCH_TEST_ARM: TestConfig( + Build.PACKAGE_AARCH64, job_config=JobConfig(**clickbench_test_params) # type: ignore + ), + JobNames.LIBFUZZER_TEST: TestConfig( + Build.FUZZERS, + job_config=JobConfig( + run_by_label=Labels.libFuzzer, + timeout=10800, + run_command='libfuzzer_test_check.py "$CHECK_NAME" 10800', + ), + ), # type: ignore }, ) CI_CONFIG.validate() diff --git a/tests/ci/ci_runners_metrics_lambda/app.py b/tests/ci/ci_runners_metrics_lambda/app.py index 5cb1e45dd14..47161215a97 100644 --- a/tests/ci/ci_runners_metrics_lambda/app.py +++ b/tests/ci/ci_runners_metrics_lambda/app.py @@ -8,23 +8,14 @@ Lambda function to: import argparse import sys -from datetime import datetime -from typing import Dict, List +from typing import Dict -import requests # type: ignore import boto3 # type: ignore -from botocore.exceptions import ClientError # type: ignore - -from lambda_shared import ( - RUNNER_TYPE_LABELS, - RunnerDescription, - RunnerDescriptions, - list_runners, -) +from lambda_shared import RUNNER_TYPE_LABELS, RunnerDescriptions, list_runners from lambda_shared.token import ( + get_access_token_by_key_app, get_cached_access_token, get_key_and_app_from_aws, - get_access_token_by_key_app, ) UNIVERSAL_LABEL = "universal" @@ -162,7 +153,7 @@ if __name__ == "__main__": if args.private_key: private_key = args.private_key elif args.private_key_path: - with open(args.private_key_path, "r") as key_file: + with open(args.private_key_path, "r", encoding="utf-8") as key_file: private_key = key_file.read() else: print("Attempt to get key and id from AWS secret manager") diff --git a/tests/ci/clean_lost_instances_lambda/app.py b/tests/ci/clean_lost_instances_lambda/app.py index 65f6ff78d4a..4accc14f7ae 100644 --- a/tests/ci/clean_lost_instances_lambda/app.py +++ b/tests/ci/clean_lost_instances_lambda/app.py @@ -8,14 +8,13 @@ Lambda function to: import argparse import sys -from datetime import datetime from dataclasses import dataclass +from datetime import datetime from typing import Dict, List -import requests # type: ignore import boto3 # type: ignore +import requests from botocore.exceptions import ClientError # type: ignore - from lambda_shared import ( RUNNER_TYPE_LABELS, RunnerDescription, @@ -23,9 +22,9 @@ from lambda_shared import ( list_runners, ) from lambda_shared.token import ( + get_access_token_by_key_app, get_cached_access_token, get_key_and_app_from_aws, - get_access_token_by_key_app, ) UNIVERSAL_LABEL = "universal" @@ -140,6 +139,7 @@ def delete_runner(access_token: str, runner: RunnerDescription) -> bool: response = requests.delete( f"https://api.github.com/orgs/ClickHouse/actions/runners/{runner.id}", headers=headers, + timeout=30, ) response.raise_for_status() print(f"Response code deleting {runner.name} is {response.status_code}") @@ -325,7 +325,7 @@ if __name__ == "__main__": if args.private_key: private_key = args.private_key elif args.private_key_path: - with open(args.private_key_path, "r") as key_file: + with open(args.private_key_path, "r", encoding="utf-8") as key_file: private_key = key_file.read() else: print("Attempt to get key and id from AWS secret manager") diff --git a/tests/ci/clickbench.py b/tests/ci/clickbench.py index 50c7bb85d28..a1988abb1f5 100644 --- a/tests/ci/clickbench.py +++ b/tests/ci/clickbench.py @@ -42,6 +42,7 @@ def get_run_command( f"{ci_logs_args}" f"--volume={result_path}:/test_output " f"--volume={server_log_path}:/var/log/clickhouse-server " + "--security-opt seccomp=unconfined " # required to issue io_uring sys-calls f"--cap-add=SYS_PTRACE {env_str} {image}" ) diff --git a/tests/ci/clickhouse_helper.py b/tests/ci/clickhouse_helper.py index f338a1e14c3..637c4519d3d 100644 --- a/tests/ci/clickhouse_helper.py +++ b/tests/ci/clickhouse_helper.py @@ -1,13 +1,12 @@ #!/usr/bin/env python3 -from pathlib import Path -from typing import Dict, List, Optional import fileinput import json import logging import time +from pathlib import Path +from typing import Dict, List, Optional -import requests # type: ignore - +import requests from get_robot_token import get_parameter_from_ssm from pr_info import PRInfo from report import TestResults @@ -72,11 +71,11 @@ class ClickHouseHelper: if args: url = args[0] url = kwargs.get("url", url) - kwargs["timeout"] = kwargs.get("timeout", 100) + timeout = kwargs.pop("timeout", 100) for i in range(5): try: - response = requests.post(*args, **kwargs) + response = requests.post(*args, timeout=timeout, **kwargs) except Exception as e: error = f"Received exception while sending data to {url} on {i} attempt: {e}" logging.warning(error) @@ -148,7 +147,9 @@ class ClickHouseHelper: for i in range(5): response = None try: - response = requests.get(self.url, params=params, headers=self.auth) + response = requests.get( + self.url, params=params, headers=self.auth, timeout=100 + ) response.raise_for_status() return response.text except Exception as ex: @@ -215,24 +216,24 @@ def prepare_tests_results_for_clickhouse( head_ref = pr_info.head_ref head_repo = pr_info.head_name - common_properties = dict( - pull_request_number=pr_info.number, - commit_sha=pr_info.sha, - commit_url=pr_info.commit_html_url, - check_name=check_name, - check_status=check_status, - check_duration_ms=int(float(check_duration) * 1000), - check_start_time=check_start_time, - report_url=report_url, - pull_request_url=pull_request_url, - base_ref=base_ref, - base_repo=base_repo, - head_ref=head_ref, - head_repo=head_repo, - task_url=pr_info.task_url, - instance_type=get_instance_type(), - instance_id=get_instance_id(), - ) + common_properties = { + "pull_request_number": pr_info.number, + "commit_sha": pr_info.sha, + "commit_url": pr_info.commit_html_url, + "check_name": check_name, + "check_status": check_status, + "check_duration_ms": int(float(check_duration) * 1000), + "check_start_time": check_start_time, + "report_url": report_url, + "pull_request_url": pull_request_url, + "base_ref": base_ref, + "base_repo": base_repo, + "head_ref": head_ref, + "head_repo": head_repo, + "task_url": pr_info.task_url, + "instance_type": get_instance_type(), + "instance_id": get_instance_id(), + } # Always publish a total record for all checks. For checks with individual # tests, also publish a record per test. diff --git a/tests/ci/commit_status_helper.py b/tests/ci/commit_status_helper.py index 9d2742ba5ba..1c2d8b2ade8 100644 --- a/tests/ci/commit_status_helper.py +++ b/tests/ci/commit_status_helper.py @@ -303,7 +303,7 @@ def post_commit_status_to_file( file_path: Path, description: str, state: str, report_url: str ) -> None: if file_path.exists(): - raise Exception(f'File "{file_path}" already exists!') + raise FileExistsError(f'File "{file_path}" already exists!') with open(file_path, "w", encoding="utf-8") as f: out = csv.writer(f, delimiter="\t") out.writerow([state, report_url, description]) @@ -329,7 +329,7 @@ class CommitStatusData: @classmethod def load_from_file(cls, file_path: Union[Path, str]): # type: ignore res = {} - with open(file_path, "r") as json_file: + with open(file_path, "r", encoding="utf-8") as json_file: res = json.load(json_file) return CommitStatusData(**cls._filter_dict(res)) @@ -347,7 +347,7 @@ class CommitStatusData: def dump_to_file(self, file_path: Union[Path, str]) -> None: file_path = Path(file_path) or STATUS_FILE_PATH - with open(file_path, "w") as json_file: + with open(file_path, "w", encoding="utf-8") as json_file: json.dump(asdict(self), json_file) def is_ok(self): diff --git a/tests/ci/compatibility_check.py b/tests/ci/compatibility_check.py index 8009ef24760..5e980660749 100644 --- a/tests/ci/compatibility_check.py +++ b/tests/ci/compatibility_check.py @@ -26,7 +26,7 @@ DOWNLOAD_RETRIES_COUNT = 5 def process_os_check(log_path: Path) -> TestResult: name = log_path.name - with open(log_path, "r") as log: + with open(log_path, "r", encoding="utf-8") as log: line = log.read().split("\n")[0].strip() if line != "OK": return TestResult(name, "FAIL") @@ -35,7 +35,7 @@ def process_os_check(log_path: Path) -> TestResult: def process_glibc_check(log_path: Path, max_glibc_version: str) -> TestResults: test_results = [] # type: TestResults - with open(log_path, "r") as log: + with open(log_path, "r", encoding="utf-8") as log: for line in log: if line.strip(): columns = line.strip().split(" ") @@ -204,7 +204,7 @@ def main(): elif "aarch64" in check_name: max_glibc_version = "2.18" # because of build with newer sysroot? else: - raise Exception("Can't determine max glibc version") + raise RuntimeError("Can't determine max glibc version") state, description, test_results, additional_logs = process_result( result_path, diff --git a/tests/ci/docker_images_check.py b/tests/ci/docker_images_check.py index ad497a00eba..b04a3975545 100644 --- a/tests/ci/docker_images_check.py +++ b/tests/ci/docker_images_check.py @@ -195,18 +195,21 @@ def main(): ok_cnt = 0 status = SUCCESS # type: StatusType - image_tags = ( - json.loads(args.image_tags) - if not os.path.isfile(args.image_tags) - else json.load(open(args.image_tags)) - ) - missing_images = ( - image_tags - if args.missing_images == "all" - else json.loads(args.missing_images) - if not os.path.isfile(args.missing_images) - else json.load(open(args.missing_images)) - ) + + if os.path.isfile(args.image_tags): + with open(args.image_tags, "r", encoding="utf-8") as jfd: + image_tags = json.load(jfd) + else: + image_tags = json.loads(args.image_tags) + + if args.missing_images == "all": + missing_images = image_tags + elif os.path.isfile(args.missing_images): + with open(args.missing_images, "r", encoding="utf-8") as jfd: + missing_images = json.load(jfd) + else: + missing_images = json.loads(args.missing_images) + images_build_list = get_images_oredered_list() for image in images_build_list: diff --git a/tests/ci/docker_manifests_merge.py b/tests/ci/docker_manifests_merge.py index fc00969d5d6..3c122545735 100644 --- a/tests/ci/docker_manifests_merge.py +++ b/tests/ci/docker_manifests_merge.py @@ -135,18 +135,20 @@ def main(): archs = args.suffixes assert len(archs) > 1, "arch suffix input param is invalid" - image_tags = ( - json.loads(args.image_tags) - if not os.path.isfile(args.image_tags) - else json.load(open(args.image_tags)) - ) - missing_images = ( - list(image_tags) - if args.missing_images == "all" - else json.loads(args.missing_images) - if not os.path.isfile(args.missing_images) - else json.load(open(args.missing_images)) - ) + if os.path.isfile(args.image_tags): + with open(args.image_tags, "r", encoding="utf-8") as jfd: + image_tags = json.load(jfd) + else: + image_tags = json.loads(args.image_tags) + + if args.missing_images == "all": + missing_images = image_tags + elif os.path.isfile(args.missing_images): + with open(args.missing_images, "r", encoding="utf-8") as jfd: + missing_images = json.load(jfd) + else: + missing_images = json.loads(args.missing_images) + test_results = [] status = SUCCESS # type: StatusType diff --git a/tests/ci/docker_server.py b/tests/ci/docker_server.py index 38d0ea6d86b..fda0dd16d0b 100644 --- a/tests/ci/docker_server.py +++ b/tests/ci/docker_server.py @@ -363,8 +363,8 @@ def main(): image = DockerImageData(image_path, image_repo, False) args.release_type = auto_release_type(args.version, args.release_type) tags = gen_tags(args.version, args.release_type) - repo_urls = dict() - direct_urls: Dict[str, List[str]] = dict() + repo_urls = {} + direct_urls: Dict[str, List[str]] = {} release_or_pr, _ = get_release_or_pr(pr_info, args.version) for arch, build_name in zip(ARCH, ("package_release", "package_aarch64")): diff --git a/tests/ci/download_release_packages.py b/tests/ci/download_release_packages.py index 26223de2f8a..550301e8fa2 100755 --- a/tests/ci/download_release_packages.py +++ b/tests/ci/download_release_packages.py @@ -3,7 +3,7 @@ import os import logging -import requests # type: ignore +import requests from requests.adapters import HTTPAdapter # type: ignore from urllib3.util.retry import Retry # type: ignore diff --git a/tests/ci/fast_test_check.py b/tests/ci/fast_test_check.py index 5d528bb4c48..a624ee06e85 100644 --- a/tests/ci/fast_test_check.py +++ b/tests/ci/fast_test_check.py @@ -37,9 +37,10 @@ def get_fasttest_cmd( ) -> str: return ( f"docker run --cap-add=SYS_PTRACE --user={os.geteuid()}:{os.getegid()} " + "--security-opt seccomp=unconfined " # required to issue io_uring sys-calls "--network=host " # required to get access to IAM credentials f"-e FASTTEST_WORKSPACE=/fasttest-workspace -e FASTTEST_OUTPUT=/test_output " - f"-e FASTTEST_SOURCE=/ClickHouse --cap-add=SYS_PTRACE " + f"-e FASTTEST_SOURCE=/ClickHouse " f"-e FASTTEST_CMAKE_FLAGS='-DCOMPILER_CACHE=sccache' " f"-e PULL_REQUEST_NUMBER={pr_number} -e COMMIT_SHA={commit_sha} " f"-e COPY_CLICKHOUSE_BINARY_TO_OUTPUT=1 " diff --git a/tests/ci/finish_check.py b/tests/ci/finish_check.py index e5268947304..eebc846f4b1 100644 --- a/tests/ci/finish_check.py +++ b/tests/ci/finish_check.py @@ -11,7 +11,7 @@ from commit_status_helper import ( get_commit, get_commit_filtered_statuses, post_commit_status, - update_mergeable_check, + trigger_mergeable_check, ) from get_robot_token import get_best_robot_token from pr_info import PRInfo @@ -24,14 +24,11 @@ def main(): pr_info = PRInfo(need_orgs=True) gh = Github(get_best_robot_token(), per_page=100) commit = get_commit(gh, pr_info.sha) - # Update the Mergeable Check at the final step - update_mergeable_check(commit, pr_info, CI_STATUS_NAME) + # Unconditionally update the Mergeable Check at the final step + statuses = get_commit_filtered_statuses(commit) + trigger_mergeable_check(commit, statuses) - statuses = [ - status - for status in get_commit_filtered_statuses(commit) - if status.context == CI_STATUS_NAME - ] + statuses = [s for s in statuses if s.context == CI_STATUS_NAME] if not statuses: return # Take the latest status diff --git a/tests/ci/functional_test_check.py b/tests/ci/functional_test_check.py index 888305aa166..4416b7863a6 100644 --- a/tests/ci/functional_test_check.py +++ b/tests/ci/functional_test_check.py @@ -12,10 +12,9 @@ from typing import List, Tuple from build_download_helper import download_all_deb_packages from clickhouse_helper import CiLogsCredentials - -from docker_images_helper import DockerImage, pull_image, get_docker_image +from docker_images_helper import DockerImage, get_docker_image, pull_image from download_release_packages import download_last_release -from env_helper import REPORT_PATH, TEMP_PATH, REPO_COPY +from env_helper import REPO_COPY, REPORT_PATH, TEMP_PATH from pr_info import PRInfo from report import ERROR, SUCCESS, JobReport, StatusType, TestResults, read_test_results from stopwatch import Stopwatch @@ -54,8 +53,7 @@ def get_image_name(check_name: str) -> str: return "clickhouse/stateless-test" if "stateful" in check_name.lower(): return "clickhouse/stateful-test" - else: - raise Exception(f"Cannot deduce image name based on check name {check_name}") + raise ValueError(f"Cannot deduce image name based on check name {check_name}") def get_run_command( @@ -107,6 +105,7 @@ def get_run_command( f"{volume_with_broken_test}" f"--volume={result_path}:/test_output " f"--volume={server_log_path}:/var/log/clickhouse-server " + "--security-opt seccomp=unconfined " # required to issue io_uring sys-calls f"--cap-add=SYS_PTRACE {env_str} {additional_options_str} {image}" ) diff --git a/tests/ci/get_previous_release_tag.py b/tests/ci/get_previous_release_tag.py index c2d279f7fec..bc0cb975ef5 100755 --- a/tests/ci/get_previous_release_tag.py +++ b/tests/ci/get_previous_release_tag.py @@ -1,10 +1,10 @@ #!/usr/bin/env python3 -import re import logging +import re from typing import List, Optional, Tuple -import requests # type: ignore +import requests CLICKHOUSE_TAGS_URL = "https://api.github.com/repos/ClickHouse/ClickHouse/tags" CLICKHOUSE_PACKAGE_URL = ( @@ -82,13 +82,14 @@ def get_previous_release(server_version: Optional[Version]) -> Optional[ReleaseI CLICKHOUSE_TAGS_URL, {"page": page, "per_page": 100}, timeout=10 ) if not response.ok: - raise Exception( - "Cannot load the list of tags from github: " + response.reason + logger.error( + "Cannot load the list of tags from github: %s", response.reason ) + response.raise_for_status() releases_str = set(re.findall(VERSION_PATTERN, response.text)) if len(releases_str) == 0: - raise Exception( + raise ValueError( "Cannot find previous release for " + str(server_version) + " server version" diff --git a/tests/ci/jepsen_check.py b/tests/ci/jepsen_check.py index fb7540abda3..011ecff635e 100644 --- a/tests/ci/jepsen_check.py +++ b/tests/ci/jepsen_check.py @@ -9,7 +9,7 @@ from pathlib import Path from typing import Any, List import boto3 # type: ignore -import requests # type: ignore +import requests from build_download_helper import ( download_build_with_progress, get_build_name_for_check, @@ -46,7 +46,7 @@ FAILED_TESTS_ANCHOR = "# Failed tests" def _parse_jepsen_output(path: Path) -> TestResults: test_results = [] # type: TestResults current_type = "" - with open(path, "r") as f: + with open(path, "r", encoding="utf-8") as f: for line in f: if SUCCESSFUL_TESTS_ANCHOR in line: current_type = "OK" @@ -101,7 +101,7 @@ def prepare_autoscaling_group_and_get_hostnames(count): instances = get_autoscaling_group_instances_ids(asg_client, JEPSEN_GROUP_NAME) counter += 1 if counter > 30: - raise Exception("Cannot wait autoscaling group") + raise RuntimeError("Cannot wait autoscaling group") ec2_client = boto3.client("ec2", region_name="us-east-1") return get_instances_addresses(ec2_client, instances) @@ -119,12 +119,12 @@ def clear_autoscaling_group(): instances = get_autoscaling_group_instances_ids(asg_client, JEPSEN_GROUP_NAME) counter += 1 if counter > 30: - raise Exception("Cannot wait autoscaling group") + raise RuntimeError("Cannot wait autoscaling group") def save_nodes_to_file(instances: List[Any], temp_path: Path) -> Path: nodes_path = temp_path / "nodes.txt" - with open(nodes_path, "w") as f: + with open(nodes_path, "w", encoding="utf-8") as f: f.write("\n".join(instances)) f.flush() return nodes_path @@ -159,7 +159,7 @@ def main(): ) args = parser.parse_args() - if args.program != "server" and args.program != "keeper": + if args.program not in ("server", "keeper"): logging.warning("Invalid argument '%s'", args.program) sys.exit(0) @@ -220,7 +220,7 @@ def main(): f"{S3_URL}/{S3_BUILDS_BUCKET}/{version}/{sha}/binary_release/clickhouse" ) print(f"Clickhouse version: [{version_full}], sha: [{sha}], url: [{build_url}]") - head = requests.head(build_url) + head = requests.head(build_url, timeout=60) assert head.status_code == 200, f"Clickhouse binary not found: {build_url}" else: build_name = get_build_name_for_check(check_name) diff --git a/tests/ci/lambda_shared_package/lambda_shared/__init__.py b/tests/ci/lambda_shared_package/lambda_shared/__init__.py index 043a0310d11..8b53f9dcb23 100644 --- a/tests/ci/lambda_shared_package/lambda_shared/__init__.py +++ b/tests/ci/lambda_shared_package/lambda_shared/__init__.py @@ -8,7 +8,7 @@ from collections import namedtuple from typing import Any, Dict, Iterable, List, Optional import boto3 # type: ignore -import requests # type: ignore +import requests RUNNER_TYPE_LABELS = [ "builder", diff --git a/tests/ci/lambda_shared_package/lambda_shared/token.py b/tests/ci/lambda_shared_package/lambda_shared/token.py index 6d5653f6a58..f9860f6ad2a 100644 --- a/tests/ci/lambda_shared_package/lambda_shared/token.py +++ b/tests/ci/lambda_shared_package/lambda_shared/token.py @@ -6,7 +6,7 @@ from typing import Tuple import boto3 # type: ignore import jwt -import requests # type: ignore +import requests from . import cached_value_is_valid diff --git a/tests/ci/libfuzzer_test_check.py b/tests/ci/libfuzzer_test_check.py index 5f41afe9fb6..62dadc6fea9 100644 --- a/tests/ci/libfuzzer_test_check.py +++ b/tests/ci/libfuzzer_test_check.py @@ -9,19 +9,13 @@ from pathlib import Path from typing import List from build_download_helper import download_fuzzers -from clickhouse_helper import ( - CiLogsCredentials, -) -from docker_images_helper import DockerImage, pull_image, get_docker_image - -from env_helper import REPORT_PATH, TEMP_PATH, REPO_COPY +from clickhouse_helper import CiLogsCredentials +from docker_images_helper import DockerImage, get_docker_image, pull_image +from env_helper import REPO_COPY, REPORT_PATH, TEMP_PATH from pr_info import PRInfo - from stopwatch import Stopwatch - from tee_popen import TeePopen - NO_CHANGES_MSG = "Nothing to run" @@ -81,6 +75,7 @@ def get_run_command( f"--volume={fuzzers_path}:/fuzzers " f"--volume={repo_path}/tests:/usr/share/clickhouse-test " f"--volume={result_path}:/test_output " + "--security-opt seccomp=unconfined " # required to issue io_uring sys-calls f"--cap-add=SYS_PTRACE {env_str} {additional_options_str} {image}" ) @@ -129,7 +124,8 @@ def main(): os.chmod(fuzzers_path / file, 0o777) elif file.endswith("_seed_corpus.zip"): corpus_path = fuzzers_path / (file.removesuffix("_seed_corpus.zip") + ".in") - zipfile.ZipFile(fuzzers_path / file, "r").extractall(corpus_path) + with zipfile.ZipFile(fuzzers_path / file, "r") as zfd: + zfd.extractall(corpus_path) result_path = temp_path / "result_path" result_path.mkdir(parents=True, exist_ok=True) diff --git a/tests/ci/release.py b/tests/ci/release.py index 2b3331938e7..679e65560f5 100755 --- a/tests/ci/release.py +++ b/tests/ci/release.py @@ -54,7 +54,7 @@ class Repo: elif protocol == "origin": self._url = protocol else: - raise Exception(f"protocol must be in {self.VALID}") + raise ValueError(f"protocol must be in {self.VALID}") def __str__(self): return self._repo @@ -144,7 +144,7 @@ class Release: for status in statuses: if status["context"] == RELEASE_READY_STATUS: if not status["state"] == SUCCESS: - raise Exception( + raise ValueError( f"the status {RELEASE_READY_STATUS} is {status['state']}" ", not success" ) @@ -153,7 +153,7 @@ class Release: page += 1 - raise Exception( + raise KeyError( f"the status {RELEASE_READY_STATUS} " f"is not found for commit {self.release_commit}" ) @@ -188,7 +188,7 @@ class Release: raise if check_run_from_master and self._git.branch != "master": - raise Exception("the script must be launched only from master") + raise RuntimeError("the script must be launched only from master") self.set_release_info() @@ -229,7 +229,7 @@ class Release: def check_no_tags_after(self): tags_after_commit = self.run(f"git tag --contains={self.release_commit}") if tags_after_commit: - raise Exception( + raise RuntimeError( f"Commit {self.release_commit} belongs to following tags:\n" f"{tags_after_commit}\nChoose another commit" ) @@ -253,7 +253,7 @@ class Release: ) output = self.run(f"git branch --contains={self.release_commit} {branch}") if branch not in output: - raise Exception( + raise RuntimeError( f"commit {self.release_commit} must belong to {branch} " f"for {self.release_type} release" ) @@ -464,9 +464,9 @@ class Release: logging.warning("Rolling back checked out %s for %s", ref, orig_ref) self.run(f"git reset --hard; git checkout -f {orig_ref}") raise - else: - if with_checkout_back and need_rollback: - self.run(rollback_cmd) + # Normal flow when we need to checkout back + if with_checkout_back and need_rollback: + self.run(rollback_cmd) @contextmanager def _create_branch(self, name: str, start_point: str = "") -> Iterator[None]: diff --git a/tests/ci/report.py b/tests/ci/report.py index 282c343eec3..669409d1729 100644 --- a/tests/ci/report.py +++ b/tests/ci/report.py @@ -22,8 +22,8 @@ from typing import ( from build_download_helper import get_gh_api from ci_config import CI_CONFIG, BuildConfig -from env_helper import REPORT_PATH, TEMP_PATH from ci_utils import normalize_string +from env_helper import REPORT_PATH, TEMP_PATH logger = logging.getLogger(__name__) @@ -296,7 +296,7 @@ class JobReport: def load(cls, from_file=None): # type: ignore res = {} from_file = from_file or JOB_REPORT_FILE - with open(from_file, "r") as json_file: + with open(from_file, "r", encoding="utf-8") as json_file: res = json.load(json_file) # Deserialize the nested lists of TestResult test_results_data = res.get("test_results", []) @@ -316,7 +316,7 @@ class JobReport: raise TypeError("Type not serializable") to_file = to_file or JOB_REPORT_FILE - with open(to_file, "w") as json_file: + with open(to_file, "w", encoding="utf-8") as json_file: json.dump(asdict(self), json_file, default=path_converter, indent=2) @@ -418,7 +418,7 @@ class BuildResult: def load_from_file(cls, file: Union[Path, str]): # type: ignore if not Path(file).exists(): return None - with open(file, "r") as json_file: + with open(file, "r", encoding="utf-8") as json_file: res = json.load(json_file) return BuildResult(**res) diff --git a/tests/ci/runner_token_rotation_lambda/app.py b/tests/ci/runner_token_rotation_lambda/app.py index 6544eee9581..445704648a9 100644 --- a/tests/ci/runner_token_rotation_lambda/app.py +++ b/tests/ci/runner_token_rotation_lambda/app.py @@ -4,9 +4,8 @@ import argparse import sys import boto3 # type: ignore -import requests # type: ignore - -from lambda_shared.token import get_cached_access_token, get_access_token_by_key_app +import requests +from lambda_shared.token import get_access_token_by_key_app, get_cached_access_token def get_runner_registration_token(access_token): @@ -17,6 +16,7 @@ def get_runner_registration_token(access_token): response = requests.post( "https://api.github.com/orgs/ClickHouse/actions/runners/registration-token", headers=headers, + timeout=30, ) response.raise_for_status() data = response.json() @@ -43,6 +43,7 @@ def main(access_token, push_to_ssm, ssm_parameter_name): def handler(event, context): + _, _ = event, context main(get_cached_access_token(), True, "github_runner_registration_token") @@ -85,7 +86,7 @@ if __name__ == "__main__": if args.private_key: private_key = args.private_key else: - with open(args.private_key_path, "r") as key_file: + with open(args.private_key_path, "r", encoding="utf-8") as key_file: private_key = key_file.read() token = get_access_token_by_key_app(private_key, args.app_id) diff --git a/tests/ci/slack_bot_ci_lambda/app.py b/tests/ci/slack_bot_ci_lambda/app.py index 45e14138335..94b71724b1c 100755 --- a/tests/ci/slack_bot_ci_lambda/app.py +++ b/tests/ci/slack_bot_ci_lambda/app.py @@ -17,12 +17,12 @@ It's deployed to slack-bot-ci-lambda in CI/CD account See also: https://aretestsgreenyet.com/ """ -import os -import json import base64 +import json +import os import random -import requests # type: ignore +import requests DRY_RUN_MARK = "" @@ -139,13 +139,11 @@ def get_play_url(query): def run_clickhouse_query(query): - url = "https://play.clickhouse.com/?user=play&query=" + requests.utils.quote(query) - res = requests.get(url) + url = "https://play.clickhouse.com/?user=play&query=" + requests.compat.quote(query) + res = requests.get(url, timeout=30) if res.status_code != 200: print("Failed to execute query: ", res.status_code, res.content) - raise Exception( - "Failed to execute query: {}: {}".format(res.status_code, res.content) - ) + res.raise_for_status() lines = res.text.strip().splitlines() return [x.split("\t") for x in lines] @@ -159,9 +157,9 @@ def split_broken_and_flaky_tests(failed_tests): flaky_tests = [] for name, report, count_prev_str, count_str in failed_tests: count_prev, count = int(count_prev_str), int(count_str) - if (2 <= count and count_prev < 2) or (count_prev == 1 and count == 1): + if (count_prev < 2 <= count) or (count_prev == count == 1): # It failed 2 times or more within extended time window, it's definitely broken. - # 2 <= count_prev means that it was not reported as broken on previous runs + # 2 <= count means that it was not reported as broken on previous runs broken_tests.append([name, report]) elif 0 < count and count_prev == 0: # It failed only once, can be a rare flaky test @@ -172,19 +170,18 @@ def split_broken_and_flaky_tests(failed_tests): def format_failed_tests_list(failed_tests, failure_type): if len(failed_tests) == 1: - res = "There is a new {} test:\n".format(failure_type) + res = f"There is a new {failure_type} test:\n" else: - res = "There are {} new {} tests:\n".format(len(failed_tests), failure_type) + res = f"There are {len(failed_tests)} new {failure_type} tests:\n" for name, report in failed_tests[:MAX_TESTS_TO_REPORT]: cidb_url = get_play_url(ALL_RECENT_FAILURES_QUERY.format(name)) - res += "- *{}* - <{}|Report> - <{}|CI DB> \n".format( - name, report, cidb_url - ) + res += f"- *{name}* - <{report}|Report> - <{cidb_url}|CI DB> \n" if MAX_TESTS_TO_REPORT < len(failed_tests): - res += "- and {} other tests... :this-is-fine-fire:".format( - len(failed_tests) - MAX_TESTS_TO_REPORT + res += ( + f"- and {len(failed_tests) - MAX_TESTS_TO_REPORT} other " + "tests... :this-is-fine-fire:" ) return res @@ -223,19 +220,16 @@ def get_too_many_failures_message_impl(failures_count): if random.random() < REPORT_NO_FAILURES_PROBABILITY: return None return "Wow, there are *no failures* at all... 0_o" - if curr_failures < MAX_FAILURES: + return_none = ( + curr_failures < MAX_FAILURES + or curr_failures < prev_failures + or (curr_failures - prev_failures) / prev_failures < 0.2 + ) + if return_none: return None if prev_failures < MAX_FAILURES: - return ":alert: *CI is broken: there are {} failures during the last 24 hours*".format( - curr_failures - ) - if curr_failures < prev_failures: - return None - if (curr_failures - prev_failures) / prev_failures < 0.2: - return None - return "CI is broken and it's getting worse: there are {} failures during the last 24 hours".format( - curr_failures - ) + return f":alert: *CI is broken: there are {curr_failures} failures during the last 24 hours*" + return "CI is broken and it's getting worse: there are {curr_failures} failures during the last 24 hours" def get_too_many_failures_message(failures_count): @@ -254,7 +248,7 @@ def get_failed_checks_percentage_message(percentage): return None msg = ":alert: " if p > 1 else "Only " if p < 0.5 else "" - msg += "*{0:.2f}%* of all checks in master have failed yesterday".format(p) + msg += f"*{p:.2f}%* of all checks in master have failed yesterday" return msg @@ -280,14 +274,10 @@ def send_to_slack_impl(message): payload = SLACK_MESSAGE_JSON.copy() payload["text"] = message - res = requests.post(SLACK_URL, json.dumps(payload)) + res = requests.post(SLACK_URL, json.dumps(payload), timeout=30) if res.status_code != 200: print("Failed to send a message to Slack: ", res.status_code, res.content) - raise Exception( - "Failed to send a message to Slack: {}: {}".format( - res.status_code, res.content - ) - ) + res.raise_for_status() def send_to_slack(message): @@ -303,7 +293,7 @@ def query_and_alert_if_needed(query, get_message_func): if msg is None: return - msg += "\nCI DB query: <{}|link>".format(get_play_url(query)) + msg += f"\nCI DB query: <{get_play_url(query)}|link>" print("Sending message to slack:", msg) send_to_slack(msg) @@ -317,6 +307,7 @@ def check_and_alert(): def handler(event, context): + _, _ = event, context try: check_and_alert() return {"statusCode": 200, "body": "OK"} diff --git a/tests/ci/sqlancer_check.py b/tests/ci/sqlancer_check.py index 59d2a3d6275..9d33c480598 100644 --- a/tests/ci/sqlancer_check.py +++ b/tests/ci/sqlancer_check.py @@ -46,14 +46,14 @@ def main(): build_name = get_build_name_for_check(check_name) urls = read_build_urls(build_name, reports_path) if not urls: - raise Exception("No build URLs found") + raise ValueError("No build URLs found") for url in urls: if url.endswith("/clickhouse"): build_url = url break else: - raise Exception("Cannot find binary clickhouse among build results") + raise ValueError("Cannot find the clickhouse binary among build results") logging.info("Got build url %s", build_url) diff --git a/tests/ci/sqllogic_test.py b/tests/ci/sqllogic_test.py index e9a109e425e..6ea6fa19d91 100755 --- a/tests/ci/sqllogic_test.py +++ b/tests/ci/sqllogic_test.py @@ -42,6 +42,7 @@ def get_run_command( f"--volume={repo_tests_path}:/clickhouse-tests " f"--volume={result_path}:/test_output " f"--volume={server_log_path}:/var/log/clickhouse-server " + "--security-opt seccomp=unconfined " # required to issue io_uring sys-calls f"--cap-add=SYS_PTRACE {image}" ) diff --git a/tests/ci/sqltest.py b/tests/ci/sqltest.py index 2fe6aabd69c..c8c2adbbd56 100644 --- a/tests/ci/sqltest.py +++ b/tests/ci/sqltest.py @@ -53,14 +53,14 @@ def main(): print(build_name) urls = read_build_urls(build_name, reports_path) if not urls: - raise Exception("No build URLs found") + raise ValueError("No build URLs found") for url in urls: if url.endswith("/clickhouse"): build_url = url break else: - raise Exception("Cannot find the clickhouse binary among build results") + raise ValueError("Cannot find the clickhouse binary among build results") logging.info("Got build url %s", build_url) diff --git a/tests/ci/ssh.py b/tests/ci/ssh.py index 275f26fd65f..321826fcf44 100644 --- a/tests/ci/ssh.py +++ b/tests/ci/ssh.py @@ -1,11 +1,11 @@ #!/usr/bin/env python3 -import shutil +import logging import os +import shutil +import signal import subprocess import tempfile -import logging -import signal class SSHAgent: @@ -21,7 +21,7 @@ class SSHAgent: def start(self): if shutil.which("ssh-agent") is None: - raise Exception("ssh-agent binary is not available") + raise RuntimeError("ssh-agent binary is not available") self._env_backup["SSH_AUTH_SOCK"] = os.environ.get("SSH_AUTH_SOCK") self._env_backup["SSH_OPTIONS"] = os.environ.get("SSH_OPTIONS") @@ -54,7 +54,7 @@ class SSHAgent: def remove(self, key_pub): if key_pub not in self._keys: - raise Exception(f"Private key not found, public part: {key_pub}") + raise ValueError(f"Private key not found, public part: {key_pub}") if self._keys[key_pub] > 1: self._keys[key_pub] -= 1 @@ -107,7 +107,7 @@ class SSHAgent: if p.returncode: message = stderr.strip() + b"\n" + stdout.strip() - raise Exception(message.strip().decode()) + raise RuntimeError(message.strip().decode()) return stdout @@ -115,9 +115,9 @@ class SSHAgent: class SSHKey: def __init__(self, key_name=None, key_value=None): if key_name is None and key_value is None: - raise Exception("Either key_name or key_value must be specified") + raise ValueError("Either key_name or key_value must be specified") if key_name is not None and key_value is not None: - raise Exception("key_name or key_value must be specified") + raise ValueError("key_name or key_value must be specified") if key_name is not None: self.key = os.getenv(key_name) else: diff --git a/tests/ci/stress.py b/tests/ci/stress.py index 7d582e683e0..7ccc058f79f 100755 --- a/tests/ci/stress.py +++ b/tests/ci/stress.py @@ -1,14 +1,14 @@ #!/usr/bin/env python3 # -*- coding: utf-8 -*- """This script is used in docker images for stress tests and upgrade tests""" -from multiprocessing import cpu_count -from pathlib import Path -from subprocess import Popen, call, check_output, STDOUT, PIPE -from typing import List import argparse import logging import random import time +from multiprocessing import cpu_count +from pathlib import Path +from subprocess import PIPE, STDOUT, Popen, call, check_output +from typing import List def get_options(i: int, upgrade_check: bool) -> str: @@ -90,12 +90,13 @@ def run_func_test( ] pipes = [] for i, path in enumerate(output_paths): - with open(path, "w") as op: + with open(path, "w", encoding="utf-8") as op: full_command = ( f"{cmd} {get_options(i, upgrade_check)} {global_time_limit_option} " f"{skip_tests_option} {upgrade_check_option}" ) logging.info("Run func tests '%s'", full_command) + # pylint:disable-next=consider-using-with pipes.append(Popen(full_command, shell=True, stdout=op, stderr=op)) time.sleep(0.5) return pipes @@ -204,6 +205,7 @@ def prepare_for_hung_check(drop_databases: bool) -> bool: continue command = make_query_command(f"DETACH DATABASE {db}") # we don't wait for drop + # pylint:disable-next=consider-using-with Popen(command, shell=True) break except Exception as ex: @@ -212,7 +214,7 @@ def prepare_for_hung_check(drop_databases: bool) -> bool: ) time.sleep(i) else: - raise Exception( + raise RuntimeError( "Cannot drop databases after stress tests. Probably server consumed " "too much memory and cannot execute simple queries" ) @@ -293,7 +295,9 @@ def main(): args = parse_args() if args.drop_databases and not args.hung_check: - raise Exception("--drop-databases only used in hung check (--hung-check)") + raise argparse.ArgumentTypeError( + "--drop-databases only used in hung check (--hung-check)" + ) # FIXME Hung check with ubsan is temporarily disabled due to # https://github.com/ClickHouse/ClickHouse/issues/45372 @@ -359,15 +363,17 @@ def main(): ] ) hung_check_log = args.output_folder / "hung_check.log" # type: Path - tee = Popen(["/usr/bin/tee", hung_check_log], stdin=PIPE) - res = call(cmd, shell=True, stdout=tee.stdin, stderr=STDOUT, timeout=600) - if tee.stdin is not None: - tee.stdin.close() + with Popen(["/usr/bin/tee", hung_check_log], stdin=PIPE) as tee: + res = call(cmd, shell=True, stdout=tee.stdin, stderr=STDOUT, timeout=600) + if tee.stdin is not None: + tee.stdin.close() if res != 0 and have_long_running_queries and not suppress_hung_check: logging.info("Hung check failed with exit code %d", res) else: hung_check_status = "No queries hung\tOK\t\\N\t\n" - with open(args.output_folder / "test_results.tsv", "w+") as results: + with open( + args.output_folder / "test_results.tsv", "w+", encoding="utf-8" + ) as results: results.write(hung_check_status) hung_check_log.unlink() diff --git a/tests/ci/stress_check.py b/tests/ci/stress_check.py index 49c1515c69f..0f2c0cdd222 100644 --- a/tests/ci/stress_check.py +++ b/tests/ci/stress_check.py @@ -95,7 +95,7 @@ def process_results( results_path = result_directory / "test_results.tsv" test_results = read_test_results(results_path, True) if len(test_results) == 0: - raise Exception("Empty results") + raise ValueError("Empty results") except Exception as e: return ( ERROR, diff --git a/tests/ci/style_check.py b/tests/ci/style_check.py index 0c7160aeea4..d0565e136d3 100644 --- a/tests/ci/style_check.py +++ b/tests/ci/style_check.py @@ -1,5 +1,6 @@ #!/usr/bin/env python3 import argparse +from concurrent.futures import ProcessPoolExecutor import csv import logging import os @@ -43,7 +44,7 @@ def process_result( results_path = result_directory / "test_results.tsv" test_results = read_test_results(results_path) if len(test_results) == 0: - raise Exception("Empty results") + raise ValueError("Empty results") return state, description, test_results, additional_files except Exception: @@ -119,7 +120,7 @@ def checkout_last_ref(pr_info: PRInfo) -> None: def main(): logging.basicConfig(level=logging.INFO) logging.getLogger("git_helper").setLevel(logging.DEBUG) - args = parse_args() + # args = parse_args() stopwatch = Stopwatch() @@ -127,28 +128,46 @@ def main(): temp_path = Path(TEMP_PATH) temp_path.mkdir(parents=True, exist_ok=True) - pr_info = PRInfo() + # pr_info = PRInfo() IMAGE_NAME = "clickhouse/style-test" image = pull_image(get_docker_image(IMAGE_NAME)) - cmd = ( + cmd_1 = ( f"docker run -u $(id -u ${{USER}}):$(id -g ${{USER}}) --cap-add=SYS_PTRACE " f"--volume={repo_path}:/ClickHouse --volume={temp_path}:/test_output " - f"{image}" + f"--entrypoint= -w/ClickHouse/utils/check-style " + f"{image} ./check_cpp_docs.sh" ) + cmd_2 = ( + f"docker run -u $(id -u ${{USER}}):$(id -g ${{USER}}) --cap-add=SYS_PTRACE " + f"--volume={repo_path}:/ClickHouse --volume={temp_path}:/test_output " + f"--entrypoint= -w/ClickHouse/utils/check-style " + f"{image} ./check_py.sh" + ) + logging.info("Is going to run the command: %s", cmd_1) + logging.info("Is going to run the command: %s", cmd_2) - if args.push: - checkout_head(pr_info) + with ProcessPoolExecutor(max_workers=2) as executor: + # Submit commands for execution in parallel + future1 = executor.submit(subprocess.run, cmd_1, shell=True) + future2 = executor.submit(subprocess.run, cmd_2, shell=True) + # Wait for both commands to complete + _ = future1.result() + _ = future2.result() + + # if args.push: + # checkout_head(pr_info) - logging.info("Is going to run the command: %s", cmd) subprocess.check_call( - cmd, + f"python3 ../../utils/check-style/process_style_check_result.py --in-results-dir {temp_path} " + f"--out-results-file {temp_path}/test_results.tsv --out-status-file {temp_path}/check_status.tsv || " + f'echo -e "failure\tCannot parse results" > {temp_path}/check_status.tsv', shell=True, ) - if args.push: - commit_push_staged(pr_info) - checkout_last_ref(pr_info) + # if args.push: + # commit_push_staged(pr_info) + # checkout_last_ref(pr_info) state, description, test_results, additional_files = process_result(temp_path) diff --git a/tests/ci/team_keys_lambda/app.py b/tests/ci/team_keys_lambda/app.py index f562fbe101d..c056808a9b4 100644 --- a/tests/ci/team_keys_lambda/app.py +++ b/tests/ci/team_keys_lambda/app.py @@ -2,13 +2,12 @@ import argparse import json - from datetime import datetime from queue import Queue from threading import Thread -import requests # type: ignore import boto3 # type: ignore +import requests class Keys(set): @@ -34,7 +33,7 @@ class Worker(Thread): m = self.queue.get() if m == "": break - response = requests.get(f"https://github.com/{m}.keys") + response = requests.get(f"https://github.com/{m}.keys", timeout=30) self.results.add(f"# {m}\n{response.text}\n") self.queue.task_done() @@ -45,7 +44,9 @@ def get_org_team_members(token: str, org: str, team_slug: str) -> set: "Accept": "application/vnd.github.v3+json", } response = requests.get( - f"https://api.github.com/orgs/{org}/teams/{team_slug}/members", headers=headers + f"https://api.github.com/orgs/{org}/teams/{team_slug}/members", + headers=headers, + timeout=30, ) response.raise_for_status() data = response.json() diff --git a/tests/ci/tee_popen.py b/tests/ci/tee_popen.py index a50532aea20..13db50df53f 100644 --- a/tests/ci/tee_popen.py +++ b/tests/ci/tee_popen.py @@ -1,14 +1,14 @@ #!/usr/bin/env python3 -from io import TextIOWrapper -from pathlib import Path -from subprocess import Popen, PIPE, STDOUT -from threading import Thread -from time import sleep -from typing import Optional, Union import logging import os import sys +from io import TextIOWrapper +from pathlib import Path +from subprocess import PIPE, STDOUT, Popen +from threading import Thread +from time import sleep +from typing import Optional, Union # Very simple tee logic implementation. You can specify a shell command, output @@ -98,5 +98,6 @@ class TeePopen: @property def log_file(self) -> TextIOWrapper: if self._log_file is None: + # pylint:disable-next=consider-using-with self._log_file = open(self._log_file_name, "w", encoding="utf-8") return self._log_file diff --git a/tests/ci/terminate_runner_lambda/app.py b/tests/ci/terminate_runner_lambda/app.py index 010f7dd6734..ab3e33d9e11 100644 --- a/tests/ci/terminate_runner_lambda/app.py +++ b/tests/ci/terminate_runner_lambda/app.py @@ -8,8 +8,7 @@ from dataclasses import dataclass from typing import Any, Dict, List import boto3 # type: ignore - -from lambda_shared import RunnerDescriptions, list_runners, cached_value_is_valid +from lambda_shared import RunnerDescriptions, cached_value_is_valid, list_runners from lambda_shared.token import get_access_token_by_key_app, get_cached_access_token @@ -134,7 +133,7 @@ def main(access_token: str, event: dict) -> Dict[str, List[str]]: candidates = instances_by_zone[zone] total_to_kill += num_to_kill if num_to_kill > len(candidates): - raise Exception( + raise RuntimeError( f"Required to kill {num_to_kill}, but have only {len(candidates)}" f" candidates in AV {zone}" ) @@ -196,6 +195,7 @@ def main(access_token: str, event: dict) -> Dict[str, List[str]]: def handler(event: dict, context: Any) -> Dict[str, List[str]]: + _ = context return main(get_cached_access_token(), event) @@ -226,7 +226,7 @@ if __name__ == "__main__": if args.private_key: private_key = args.private_key else: - with open(args.private_key_path, "r") as key_file: + with open(args.private_key_path, "r", encoding="utf-8") as key_file: private_key = key_file.read() token = get_access_token_by_key_app(private_key, args.app_id) diff --git a/tests/ci/test_ci_config.py b/tests/ci/test_ci_config.py index 49d49d9c328..badbc4c5dcf 100644 --- a/tests/ci/test_ci_config.py +++ b/tests/ci/test_ci_config.py @@ -1,15 +1,30 @@ #!/usr/bin/env python3 import unittest +from ci_config import CIStages, JobNames, CI_CONFIG, Runners class TestCIConfig(unittest.TestCase): - def test_no_errors_in_ci_config(self): - raised = None - try: - from ci_config import ( # pylint: disable=import-outside-toplevel - CI_CONFIG as _, - ) - except Exception as exc: - raised = exc - self.assertIsNone(raised, f"CI_CONFIG import raised error {raised}") + def test_runner_config(self): + """check runner is provided w/o exception""" + for job in JobNames: + runner = CI_CONFIG.get_runner_type(job) + self.assertIn(runner, Runners) + + def test_job_stage_config(self): + """check runner is provided w/o exception""" + for job in JobNames: + stage = CI_CONFIG.get_job_ci_stage(job) + if job in [ + JobNames.STYLE_CHECK, + JobNames.FAST_TEST, + JobNames.JEPSEN_KEEPER, + JobNames.BUILD_CHECK, + JobNames.BUILD_CHECK_SPECIAL, + ]: + assert ( + stage == CIStages.NA + ), "These jobs are not in CI stages, must be NA" + else: + assert stage != CIStages.NA, f"stage not found for [{job}]" + self.assertIn(stage, CIStages) diff --git a/tests/ci/unit_tests_check.py b/tests/ci/unit_tests_check.py index 41c52d53020..2c2862d926a 100644 --- a/tests/ci/unit_tests_check.py +++ b/tests/ci/unit_tests_check.py @@ -23,7 +23,7 @@ def get_test_name(line): for element in elements: if "(" not in element and ")" not in element: return element - raise Exception(f"No test name in line '{line}'") + raise ValueError(f"No test name in line '{line}'") def process_results( @@ -176,6 +176,7 @@ def main(): run_command = ( f"docker run --cap-add=SYS_PTRACE --volume={tests_binary}:/unit_tests_dbms " + "--security-opt seccomp=unconfined " # required to issue io_uring sys-calls f"--volume={test_output}:/test_output {docker_image}" ) diff --git a/tests/ci/workflow_approve_rerun_lambda/app.py b/tests/ci/workflow_approve_rerun_lambda/app.py index 5e68f2d4b53..8ed47a54ab1 100644 --- a/tests/ci/workflow_approve_rerun_lambda/app.py +++ b/tests/ci/workflow_approve_rerun_lambda/app.py @@ -6,7 +6,7 @@ import time from collections import namedtuple from urllib.parse import quote -import requests # type: ignore +import requests from lambda_shared.pr import TRUSTED_CONTRIBUTORS from lambda_shared.token import get_cached_access_token @@ -90,26 +90,29 @@ def is_trusted_contributor(pr_user_login, pr_user_orgs): def _exec_get_with_retry(url, token): headers = {"Authorization": f"token {token}"} + e = Exception() for i in range(MAX_RETRY): try: - response = requests.get(url, headers=headers) + response = requests.get(url, headers=headers, timeout=30) response.raise_for_status() return response.json() except Exception as ex: print("Got exception executing request", ex) + e = ex time.sleep(i + 1) - raise Exception("Cannot execute GET request with retries") + raise requests.HTTPError("Cannot execute GET request with retries") from e def _exec_post_with_retry(url, token, data=None): headers = {"Authorization": f"token {token}"} + e = Exception() for i in range(MAX_RETRY): try: if data: - response = requests.post(url, headers=headers, json=data) + response = requests.post(url, headers=headers, json=data, timeout=30) else: - response = requests.post(url, headers=headers) + response = requests.post(url, headers=headers, timeout=30) if response.status_code == 403: data = response.json() if ( @@ -123,9 +126,10 @@ def _exec_post_with_retry(url, token, data=None): return response.json() except Exception as ex: print("Got exception executing request", ex) + e = ex time.sleep(i + 1) - raise Exception("Cannot execute POST request with retry") + raise requests.HTTPError("Cannot execute POST request with retry") from e def _get_pull_requests_from(repo_url, owner, branch, token): diff --git a/tests/clickhouse-test b/tests/clickhouse-test index d44c80bc410..ce0feadf050 100755 --- a/tests/clickhouse-test +++ b/tests/clickhouse-test @@ -5,47 +5,46 @@ # pylint: disable=too-many-lines # pylint: disable=anomalous-backslash-in-string -import enum -from queue import Full -import shutil -import itertools -import sys -import os -import os.path -import glob -import signal -import re import copy -import traceback -import math +import enum +import glob # Not requests, to avoid requiring extra dependency. import http.client -import urllib.parse +import itertools import json +import math +import multiprocessing +import os +import os.path +import platform +import random +import re +import shutil +import signal +import socket +import string +import subprocess +import sys +import traceback +import urllib.parse # for crc32 import zlib - from argparse import ArgumentParser -from typing import Tuple, Union, Optional, Dict, Set, List -import subprocess -from subprocess import Popen -from subprocess import PIPE +from contextlib import closing from datetime import datetime, timedelta -from time import time, sleep from errno import ESRCH +from queue import Full +from subprocess import PIPE, Popen +from time import sleep, time +from typing import Dict, List, Optional, Set, Tuple, Union try: import termcolor # type: ignore except ImportError: termcolor = None -import random -import string -import multiprocessing -import socket -from contextlib import closing USE_JINJA = True try: @@ -69,7 +68,7 @@ TEST_FILE_EXTENSIONS = [".sql", ".sql.j2", ".sh", ".py", ".expect"] VERSION_PATTERN = r"^((\d+\.)?(\d+\.)?(\d+\.)?\d+)$" -def stringhash(s): +def stringhash(s: str) -> int: # default hash() function consistent # only during process invocation https://stackoverflow.com/a/42089311 return zlib.crc32(s.encode("utf-8")) @@ -93,6 +92,10 @@ def trim_for_log(s): return "\n".join(lines) +class TestException(Exception): + pass + + class HTTPError(Exception): def __init__(self, message=None, code=None): self.message = message @@ -249,7 +252,7 @@ def get_db_engine(args, database_name): def get_create_database_settings(args, testcase_args): - create_database_settings = dict() + create_database_settings = {} if testcase_args: create_database_settings["log_comment"] = testcase_args.testcase_basename if args.db_engine == "Ordinary": @@ -573,27 +576,6 @@ def get_localzone(): return os.getenv("TZ", "/".join(os.readlink("/etc/localtime").split("/")[-2:])) -def supports_io_uring(): - return not subprocess.call( - [ - args.binary, - "-q", - "select * from file('/dev/null', 'LineAsString')", - "--storage_file_read_method", - "io_uring", - ], - stdout=subprocess.DEVNULL, - stderr=subprocess.DEVNULL, - ) - - -def get_local_filesystem_methods(): - methods = ["read", "pread", "mmap", "pread_threadpool"] - if supports_io_uring(): - methods.append("io_uring") - return methods - - class SettingsRandomizer: settings = { "max_insert_threads": lambda: 0 @@ -634,7 +616,10 @@ class SettingsRandomizer: 0.2, 0.5, 1, 10 * 1024 * 1024 * 1024 ), "local_filesystem_read_method": lambda: random.choice( - get_local_filesystem_methods() + # Allow to use uring only when running on Linux + ["read", "pread", "mmap", "pread_threadpool", "io_uring"] + if platform.system().lower() == "linux" + else ["read", "pread", "mmap", "pread_threadpool"] ), "remote_filesystem_read_method": lambda: random.choice(["read", "threadpool"]), "local_filesystem_read_prefetch": lambda: random.randint(0, 1), @@ -1203,7 +1188,7 @@ class TestCase: ) if result_is_different: - diff_proc = Popen( + with Popen( [ "diff", "-U", @@ -1214,23 +1199,23 @@ class TestCase: encoding="latin-1", stdout=PIPE, universal_newlines=True, - ) - if self.show_whitespaces_in_diff: - sed_proc = Popen( - ["sed", "-e", "s/[ \t]\\+$/&$/g"], - stdin=diff_proc.stdout, - stdout=PIPE, - ) - diff_proc.stdout.close() # Allow diff to receive a SIGPIPE if cat exits. - diff = sed_proc.communicate()[0].decode("utf-8", errors="ignore") - else: - diff = diff_proc.communicate()[0] + ) as diff_proc: + if self.show_whitespaces_in_diff: + with Popen( + ["sed", "-e", "s/[ \t]\\+$/&$/g"], + stdin=diff_proc.stdout, + stdout=PIPE, + ) as sed_proc: + diff = sed_proc.communicate()[0].decode( + "utf-8", errors="ignore" + ) + else: + diff = diff_proc.communicate()[0] if diff.startswith("Binary files "): diff += "Content of stdout:\n===================\n" - file = open(self.stdout_file, "rb") - diff += str(file.read()) - file.close() + with open(self.stdout_file, "rb") as file: + diff += str(file.read()) diff += "===================" description += f"\n{diff}\n" if debug_log: @@ -1393,6 +1378,7 @@ class TestCase: command = pattern.format(**params) + # pylint:disable-next=consider-using-with; TODO: fix proc = Popen(command, shell=True, env=os.environ, start_new_session=True) while ( @@ -1559,7 +1545,7 @@ class TestCase: ) if len(leftover_tables) != 0: - raise Exception( + raise TestException( f"The test should cleanup its tables ({leftover_tables}), otherwise it is inconvenient for running it locally." ) @@ -1642,7 +1628,7 @@ class TestSuite: ): return "#" else: - raise Exception(f"Unknown file_extension: {filename}") + raise TestException(f"Unknown file_extension: {filename}") def parse_tags_from_line(line, comment_sign) -> Set[str]: if not line.startswith(comment_sign): @@ -1703,17 +1689,22 @@ class TestSuite: self.suite_tmp_path: str = suite_tmp_path self.suite: str = suite - filter_func = lambda x: True # noqa: ignore E731 - if args.run_by_hash_num is not None and args.run_by_hash_total is not None: if args.run_by_hash_num > args.run_by_hash_total: - raise Exception( + raise TestException( f"Incorrect run by hash, value {args.run_by_hash_num} bigger than total {args.run_by_hash_total}" ) - filter_func = ( - lambda x: stringhash(x) % args.run_by_hash_total == args.run_by_hash_num - ) + def filter_func(x: str) -> bool: + return bool( + stringhash(x) % args.run_by_hash_total == args.run_by_hash_num + ) + + else: + + def filter_func(x: str) -> bool: + _ = x + return True self.all_tests: List[str] = self.get_tests_list( self.tests_in_suite_key_func, filter_func @@ -2435,7 +2426,7 @@ def main(args): pid = get_server_pid() print("Got server pid", pid) print_stacktraces() - raise Exception(msg) + raise TestException(msg) args.build_flags = collect_build_flags(args) args.changed_merge_tree_settings = collect_changed_merge_tree_settings(args) @@ -2639,7 +2630,7 @@ def find_binary(name): if os.access(bin_path, os.X_OK): return bin_path - raise Exception(f"{name} was not found in PATH") + raise TestException(f"{name} was not found in PATH") def find_clickhouse_command(binary, command): diff --git a/tests/config/config.d/azure_storage_conf.xml b/tests/config/config.d/azure_storage_conf.xml index f42bb8e3cf8..412d40111a7 100644 --- a/tests/config/config.d/azure_storage_conf.xml +++ b/tests/config/config.d/azure_storage_conf.xml @@ -2,7 +2,8 @@ - azure_blob_storage + object_storage + azure http://localhost:10000/devstoreaccount1 cont false diff --git a/tests/config/config.d/storage_conf.xml b/tests/config/config.d/storage_conf.xml index 1429dfff724..00d8cb3aea5 100644 --- a/tests/config/config.d/storage_conf.xml +++ b/tests/config/config.d/storage_conf.xml @@ -37,11 +37,13 @@ - local_blob_storage + object_storage + local local_disk/ - local_blob_storage + object_storage + local_blob_storage local_disk_2/ diff --git a/tests/integration/ci-runner.py b/tests/integration/ci-runner.py index 08dd9ba276b..e7f691d2237 100755 --- a/tests/integration/ci-runner.py +++ b/tests/integration/ci-runner.py @@ -11,9 +11,11 @@ import shlex import shutil import string import subprocess +import sys import time import zlib # for crc32 from collections import defaultdict +from itertools import chain from integration_test_images import IMAGES @@ -102,7 +104,7 @@ def get_counters(fname): "SKIPPED": set([]), } - with open(fname, "r") as out: + with open(fname, "r", encoding="utf-8") as out: for line in out: line = line.strip() # Example of log: @@ -118,7 +120,7 @@ def get_counters(fname): # [gw0] [ 7%] ERROR test_mysql_protocol/test.py::test_golang_client # ^^^^^^^^^^^^^ if line.strip().startswith("["): - line = re.sub("^\[[^\[\]]*\] \[[^\[\]]*\] ", "", line) + line = re.sub(r"^\[[^\[\]]*\] \[[^\[\]]*\] ", "", line) line_arr = line.split(" ") if len(line_arr) < 2: @@ -160,7 +162,7 @@ def get_counters(fname): def parse_test_times(fname): read = False description_output = [] - with open(fname, "r") as out: + with open(fname, "r", encoding="utf-8") as out: for line in out: if read and "==" in line: break @@ -196,7 +198,7 @@ def clear_ip_tables_and_restart_daemons(): shell=True, ) except subprocess.CalledProcessError as err: - logging.info("docker kill excepted: " + str(err)) + logging.info("docker kill excepted: %s", str(err)) try: logging.info("Removing all docker containers") @@ -205,7 +207,7 @@ def clear_ip_tables_and_restart_daemons(): shell=True, ) except subprocess.CalledProcessError as err: - logging.info("docker rm excepted: " + str(err)) + logging.info("docker rm excepted: %s", str(err)) # don't restart docker if it's disabled if os.environ.get("CLICKHOUSE_TESTS_RUNNER_RESTART_DOCKER", "1") == "1": @@ -213,7 +215,7 @@ def clear_ip_tables_and_restart_daemons(): logging.info("Stopping docker daemon") subprocess.check_output("service docker stop", shell=True) except subprocess.CalledProcessError as err: - logging.info("docker stop excepted: " + str(err)) + logging.info("docker stop excepted: %s", str(err)) try: for i in range(200): @@ -226,9 +228,9 @@ def clear_ip_tables_and_restart_daemons(): time.sleep(0.5) logging.info("Waiting docker to start, current %s", str(err)) else: - raise Exception("Docker daemon doesn't responding") + raise RuntimeError("Docker daemon doesn't responding") except subprocess.CalledProcessError as err: - logging.info("Can't reload docker: " + str(err)) + logging.info("Can't reload docker: %s", str(err)) iptables_iter = 0 try: @@ -276,13 +278,14 @@ class ClickhouseIntegrationTestsRunner: def base_path(self): return os.path.join(str(self.result_path), "../") - def should_skip_tests(self): + @staticmethod + def should_skip_tests(): return [] def get_image_with_version(self, name): if name in self.image_versions: return name + ":" + self.image_versions[name] - logging.warn( + logging.warning( "Cannot find image %s in params list %s", name, self.image_versions ) if ":" not in name: @@ -292,7 +295,7 @@ class ClickhouseIntegrationTestsRunner: def get_image_version(self, name: str): if name in self.image_versions: return self.image_versions[name] - logging.warn( + logging.warning( "Cannot find image %s in params list %s", name, self.image_versions ) return "latest" @@ -304,13 +307,9 @@ class ClickhouseIntegrationTestsRunner: image_cmd = self._get_runner_image_cmd(repo_path) cmd = ( - "cd {repo_path}/tests/integration && " - "timeout --signal=KILL 1h ./runner {runner_opts} {image_cmd} --pre-pull --command '{command}' ".format( - repo_path=repo_path, - runner_opts=self._get_runner_opts(), - image_cmd=image_cmd, - command=r""" echo Pre Pull finished """, - ) + f"cd {repo_path}/tests/integration && " + f"timeout --signal=KILL 1h ./runner {self._get_runner_opts()} {image_cmd} " + "--pre-pull --command ' echo Pre Pull finished ' " ) for i in range(5): @@ -322,14 +321,15 @@ class ClickhouseIntegrationTestsRunner: ) return except subprocess.CalledProcessError as err: - logging.info("docker-compose pull failed: " + str(err)) + logging.info("docker-compose pull failed: %s", str(err)) continue logging.error("Pulling images failed for 5 attempts. Will fail the worker.") # We pass specific retcode to to ci/integration_test_check.py to skip status reporting and restart job - exit(13) + sys.exit(13) - def _can_run_with(self, path, opt): - with open(path, "r") as script: + @staticmethod + def _can_run_with(path, opt): + with open(path, "r", encoding="utf-8") as script: for line in script: if opt in line: return True @@ -349,19 +349,23 @@ class ClickhouseIntegrationTestsRunner: logging.info("Package found in %s", full_path) log_name = "install_" + f + ".log" log_path = os.path.join(str(self.path()), log_name) - with open(log_path, "w") as log: - cmd = "dpkg -x {} .".format(full_path) + with open(log_path, "w", encoding="utf-8") as log: + cmd = f"dpkg -x {full_path} ." logging.info("Executing installation cmd %s", cmd) - retcode = subprocess.Popen( + with subprocess.Popen( cmd, shell=True, stderr=log, stdout=log - ).wait() - if retcode == 0: - logging.info("Installation of %s successfull", full_path) - else: - raise Exception("Installation of %s failed", full_path) + ) as proc: + if proc.wait() == 0: + logging.info( + "Installation of %s successfull", full_path + ) + else: + raise RuntimeError( + f"Installation of {full_path} failed" + ) break else: - raise Exception("Package with {} not found".format(package)) + raise FileNotFoundError(f"Package with {package} not found") # logging.info("Unstripping binary") # logging.info( # "Unstring %s", @@ -387,11 +391,11 @@ class ClickhouseIntegrationTestsRunner: os.getenv("CLICKHOUSE_TESTS_LIBRARY_BRIDGE_BIN_PATH"), ) - def _compress_logs(self, dir, relpaths, result_path): + @staticmethod + def _compress_logs(directory, relpaths, result_path): retcode = subprocess.call( # STYLE_CHECK_ALLOW_SUBPROCESS_CHECK_CALL - "tar --use-compress-program='zstd --threads=0' -cf {} -C {} {}".format( - result_path, dir, " ".join(relpaths) - ), + f"tar --use-compress-program='zstd --threads=0' -cf {result_path} -C " + f"{directory} {' '.join(relpaths)}", shell=True, ) # tar return 1 when the files are changed on compressing, we ignore it @@ -443,26 +447,25 @@ class ClickhouseIntegrationTestsRunner: return list(sorted(all_tests)) - def _get_parallel_tests_skip_list(self, repo_path): - skip_list_file_path = "{}/tests/integration/parallel_skip.json".format( - repo_path - ) + @staticmethod + def _get_parallel_tests_skip_list(repo_path): + skip_list_file_path = f"{repo_path}/tests/integration/parallel_skip.json" if ( not os.path.isfile(skip_list_file_path) or os.path.getsize(skip_list_file_path) == 0 ): - raise Exception( - "There is something wrong with getting all tests list: file '{}' is empty or does not exist.".format( - skip_list_file_path - ) + raise ValueError( + "There is something wrong with getting all tests list: " + f"file '{skip_list_file_path}' is empty or does not exist." ) skip_list_tests = [] - with open(skip_list_file_path, "r") as skip_list_file: + with open(skip_list_file_path, "r", encoding="utf-8") as skip_list_file: skip_list_tests = json.load(skip_list_file) return list(sorted(skip_list_tests)) - def group_test_by_file(self, tests): + @staticmethod + def group_test_by_file(tests): result = {} for test in tests: test_file = test.split("::")[0] @@ -471,7 +474,8 @@ class ClickhouseIntegrationTestsRunner: result[test_file].append(test) return result - def _update_counters(self, main_counters, current_counters, broken_tests): + @staticmethod + def _update_counters(main_counters, current_counters, broken_tests): for test in current_counters["PASSED"]: if test not in main_counters["PASSED"]: if test in main_counters["FAILED"]: @@ -511,21 +515,23 @@ class ClickhouseIntegrationTestsRunner: logging.info( "Can run with custom docker image version %s", runner_version ) - image_cmd += " --docker-image-version={} ".format(runner_version) + image_cmd += f" --docker-image-version={runner_version} " else: if self._can_run_with( os.path.join(repo_path, "tests/integration", "runner"), "--docker-compose-images-tags", ): - image_cmd += "--docker-compose-images-tags={} ".format( - self.get_image_with_version(img) + image_cmd += ( + "--docker-compose-images-tags=" + f"{self.get_image_with_version(img)} " ) else: image_cmd = "" logging.info("Cannot run with custom docker image version :(") return image_cmd - def _find_test_data_dirs(self, repo_path, test_names): + @staticmethod + def _find_test_data_dirs(repo_path, test_names): relpaths = {} for test_name in test_names: if "/" in test_name: @@ -543,7 +549,8 @@ class ClickhouseIntegrationTestsRunner: relpaths[relpath] = mtime return relpaths - def _get_test_data_dirs_difference(self, new_snapshot, old_snapshot): + @staticmethod + def _get_test_data_dirs_difference(new_snapshot, old_snapshot): res = set() for path in new_snapshot: if (path not in old_snapshot) or (old_snapshot[path] != new_snapshot[path]): @@ -569,7 +576,7 @@ class ClickhouseIntegrationTestsRunner: broken_tests, ) except Exception as e: - logging.info("Failed to run {}:\n{}".format(str(test_group), str(e))) + logging.info("Failed to run %s:\n%s", test_group, e) counters = { "ERROR": [], "PASSED": [], @@ -630,31 +637,27 @@ class ClickhouseIntegrationTestsRunner: info_path = os.path.join(repo_path, "tests/integration", info_basename) test_cmd = " ".join([shlex.quote(test) for test in sorted(test_names)]) - parallel_cmd = ( - " --parallel {} ".format(num_workers) if num_workers > 0 else "" - ) + parallel_cmd = f" --parallel {num_workers} " if num_workers > 0 else "" # -r -- show extra test summary: # -f -- (f)ailed # -E -- (E)rror # -p -- (p)assed # -s -- (s)kipped - cmd = "cd {}/tests/integration && timeout --signal=KILL 1h ./runner {} {} -t {} {} -- -rfEps --run-id={} --color=no --durations=0 {} | tee {}".format( - repo_path, - self._get_runner_opts(), - image_cmd, - test_cmd, - parallel_cmd, - i, - _get_deselect_option(self.should_skip_tests()), - info_path, + cmd = ( + f"cd {repo_path}/tests/integration && " + f"timeout --signal=KILL 1h ./runner {self._get_runner_opts()} " + f"{image_cmd} -t {test_cmd} {parallel_cmd} -- -rfEps --run-id={i} " + f"--color=no --durations=0 {_get_deselect_option(self.should_skip_tests())} " + f"| tee {info_path}" ) log_basename = test_group_str + "_" + str(i) + ".log" log_path = os.path.join(repo_path, "tests/integration", log_basename) - with open(log_path, "w") as log: + with open(log_path, "w", encoding="utf-8") as log: logging.info("Executing cmd: %s", cmd) # ignore retcode, since it meaningful due to pipe to tee - subprocess.Popen(cmd, shell=True, stderr=log, stdout=log).wait() + with subprocess.Popen(cmd, shell=True, stderr=log, stdout=log) as proc: + proc.wait() extra_logs_names = [log_basename] log_result_path = os.path.join( @@ -745,11 +748,14 @@ class ClickhouseIntegrationTestsRunner: # want to mark them as error so we filter by '::'. for test in tests_in_group: if ( - test not in counters["PASSED"] - and test not in counters["ERROR"] - and test not in counters["SKIPPED"] - and test not in counters["FAILED"] - and test not in counters["BROKEN"] + test + not in chain( + counters["PASSED"], + counters["ERROR"], + counters["SKIPPED"], + counters["FAILED"], + counters["BROKEN"], + ) and "::" in test ): counters["ERROR"].append(test) @@ -814,7 +820,7 @@ class ClickhouseIntegrationTestsRunner: ( c + " (✕" + str(final_retry) + ")", text_state, - "{:.2f}".format(tests_times[c]), + f"{tests_times[c]:.2f}", ) for c in counters[state] ] @@ -836,7 +842,7 @@ class ClickhouseIntegrationTestsRunner: self._install_clickhouse(build_path) logging.info("Pulling images") - runner._pre_pull_images(repo_path) + self._pre_pull_images(repo_path) logging.info( "Dump iptables before run %s", @@ -909,11 +915,15 @@ class ClickhouseIntegrationTestsRunner: logging.info("Shuffling test groups") random.shuffle(items_to_run) - broken_tests = list() + broken_tests = [] if self.use_analyzer: - with open(f"{repo_path}/tests/analyzer_integration_broken_tests.txt") as f: + with open( + f"{repo_path}/tests/analyzer_integration_broken_tests.txt", + "r", + encoding="utf-8", + ) as f: broken_tests = f.read().splitlines() - logging.info(f"Broken tests in the list: {len(broken_tests)}") + logging.info("Broken tests in the list: %s", len(broken_tests)) for group, tests in items_to_run: logging.info("Running test group %s containing %s tests", group, len(tests)) @@ -965,12 +975,12 @@ class ClickhouseIntegrationTestsRunner: else: text_state = state test_result += [ - (c, text_state, "{:.2f}".format(tests_times[c]), tests_log_paths[c]) + (c, text_state, f"{tests_times[c]:.2f}", tests_log_paths[c]) for c in counters[state] ] failed_sum = len(counters["FAILED"]) + len(counters["ERROR"]) - status_text = "fail: {}, passed: {}".format(failed_sum, len(counters["PASSED"])) + status_text = f"fail: {failed_sum}, passed: {len(counters['PASSED'])}" if self.soft_deadline_time < time.time(): status_text = "Timeout, " + status_text @@ -987,10 +997,10 @@ class ClickhouseIntegrationTestsRunner: def write_results(results_file, status_file, results, status): - with open(results_file, "w") as f: + with open(results_file, "w", encoding="utf-8") as f: out = csv.writer(f, delimiter="\t") out.writerows(results) - with open(status_file, "w") as f: + with open(status_file, "w", encoding="utf-8") as f: out = csv.writer(f, delimiter="\t") out.writerow(status) @@ -1003,7 +1013,8 @@ if __name__ == "__main__": result_path = os.environ.get("CLICKHOUSE_TESTS_RESULT_PATH") params_path = os.environ.get("CLICKHOUSE_TESTS_JSON_PARAMS_PATH") - params = json.loads(open(params_path, "r").read()) + with open(params_path, "r", encoding="utf-8") as jfd: + params = json.loads(jfd.read()) runner = ClickhouseIntegrationTestsRunner(result_path, params) logging.info("Running tests") diff --git a/tests/integration/conftest.py b/tests/integration/conftest.py index 6e76270c607..a4b18ff523a 100644 --- a/tests/integration/conftest.py +++ b/tests/integration/conftest.py @@ -1,10 +1,11 @@ -from helpers.cluster import run_and_check -import pytest +#!/usr/bin/env python3 + import logging import os -from helpers.test_tools import TSV -from helpers.network import _NetworkManager +import pytest # pylint:disable=import-error; for style check +from helpers.cluster import run_and_check +from helpers.network import _NetworkManager # This is a workaround for a problem with logging in pytest [1]. # @@ -32,32 +33,35 @@ def tune_local_port_range(): def cleanup_environment(): try: if int(os.environ.get("PYTEST_CLEANUP_CONTAINERS", 0)) == 1: - logging.debug(f"Cleaning all iptables rules") + logging.debug("Cleaning all iptables rules") _NetworkManager.clean_all_user_iptables_rules() result = run_and_check(["docker ps | wc -l"], shell=True) if int(result) > 1: if int(os.environ.get("PYTEST_CLEANUP_CONTAINERS", 0)) != 1: logging.warning( - f"Docker containters({int(result)}) are running before tests run. They can be left from previous pytest run and cause test failures.\n" - "You can set env PYTEST_CLEANUP_CONTAINERS=1 or use runner with --cleanup-containers argument to enable automatic containers cleanup." + "Docker containters(%s) are running before tests run. " + "They can be left from previous pytest run and cause test failures.\n" + "You can set env PYTEST_CLEANUP_CONTAINERS=1 or use runner with " + "--cleanup-containers argument to enable automatic containers cleanup.", + int(result), ) else: logging.debug("Trying to kill unstopped containers...") run_and_check( - [f"docker kill $(docker container list --all --quiet)"], + ["docker kill $(docker container list --all --quiet)"], shell=True, nothrow=True, ) run_and_check( - [f"docker rm $docker container list --all --quiet)"], + ["docker rm $docker container list --all --quiet)"], shell=True, nothrow=True, ) logging.debug("Unstopped containers killed") r = run_and_check(["docker-compose", "ps", "--services", "--all"]) - logging.debug(f"Docker ps before start:{r.stdout}") + logging.debug("Docker ps before start:%s", r.stdout) else: - logging.debug(f"No running containers") + logging.debug("No running containers") logging.debug("Pruning Docker networks") run_and_check( @@ -66,8 +70,7 @@ def cleanup_environment(): nothrow=True, ) except Exception as e: - logging.exception(f"cleanup_environment:{str(e)}") - pass + logging.exception("cleanup_environment:%s", e) yield diff --git a/tests/integration/runner b/tests/integration/runner index b1193b5b471..f1d5198f545 100755 --- a/tests/integration/runner +++ b/tests/integration/runner @@ -39,9 +39,7 @@ def check_args_and_update_paths(args): else: CLICKHOUSE_ROOT = args.clickhouse_root else: - logging.info( - "ClickHouse root is not set. Will use %s" % (DEFAULT_CLICKHOUSE_ROOT) - ) + logging.info("ClickHouse root is not set. Will use %s", DEFAULT_CLICKHOUSE_ROOT) CLICKHOUSE_ROOT = DEFAULT_CLICKHOUSE_ROOT if not os.path.isabs(args.binary): @@ -74,9 +72,7 @@ def check_args_and_update_paths(args): args.base_configs_dir = os.path.abspath( os.path.join(CLICKHOUSE_ROOT, CONFIG_DIR_IN_REPO) ) - logging.info( - "Base configs dir is not set. Will use %s" % (args.base_configs_dir) - ) + logging.info("Base configs dir is not set. Will use %s", args.base_configs_dir) if args.cases_dir: if not os.path.isabs(args.cases_dir): @@ -87,7 +83,7 @@ def check_args_and_update_paths(args): args.cases_dir = os.path.abspath( os.path.join(CLICKHOUSE_ROOT, INTEGRATION_DIR_IN_REPO) ) - logging.info("Cases dir is not set. Will use %s" % (args.cases_dir)) + logging.info("Cases dir is not set. Will use %s", args.cases_dir) if args.utils_dir: if not os.path.isabs(args.utils_dir): @@ -98,12 +94,13 @@ def check_args_and_update_paths(args): args.utils_dir = os.path.abspath( os.path.join(CLICKHOUSE_ROOT, UTILS_DIR_IN_REPO) ) - logging.info("utils dir is not set. Will use %s" % (args.utils_dir)) + logging.info("utils dir is not set. Will use %s", args.utils_dir) logging.info( - "base_configs_dir: {}, binary: {}, cases_dir: {} ".format( - args.base_configs_dir, args.binary, args.cases_dir - ) + "base_configs_dir: %s, binary: %s, cases_dir: %s ", + args.base_configs_dir, + args.binary, + args.cases_dir, ) for path in [ @@ -115,7 +112,7 @@ def check_args_and_update_paths(args): CLICKHOUSE_ROOT, ]: if not os.path.exists(path): - raise Exception("Path {} doesn't exist".format(path)) + raise FileNotFoundError(f"Path {path} doesn't exist") if args.dockerd_volume: if not os.path.isabs(args.dockerd_volume): @@ -126,21 +123,22 @@ def check_args_and_update_paths(args): if (not os.path.exists(os.path.join(args.base_configs_dir, "config.xml"))) and ( not os.path.exists(os.path.join(args.base_configs_dir, "config.yaml")) ): - raise Exception( - "No config.xml or config.yaml in {}".format(args.base_configs_dir) + raise FileNotFoundError( + f"No config.xml or config.yaml in {args.base_configs_dir}" ) if (not os.path.exists(os.path.join(args.base_configs_dir, "users.xml"))) and ( not os.path.exists(os.path.join(args.base_configs_dir, "users.yaml")) ): - raise Exception( - "No users.xml or users.yaml in {}".format(args.base_configs_dir) + raise FileNotFoundError( + f"No users.xml or users.yaml in {args.base_configs_dir}" ) def docker_kill_handler_handler(signum, frame): + _, _ = signum, frame subprocess.check_call( - "docker ps --all --quiet --filter name={name}".format(name=CONTAINER_NAME), + f"docker ps --all --quiet --filter name={CONTAINER_NAME}", shell=True, ) raise KeyboardInterrupt("Killed by Ctrl+C") @@ -318,7 +316,7 @@ if __name__ == "__main__": parallel_args = "" if args.parallel: parallel_args += "--dist=loadfile" - parallel_args += " -n {}".format(args.parallel) + parallel_args += f" -n {args.parallel}".format() rand_args = "" # if not args.no_random: @@ -326,7 +324,7 @@ if __name__ == "__main__": net = "" if args.network: - net = "--net={}".format(args.network) + net = f"--net={args.network}" elif not args.disable_net_host: net = "--net=host" @@ -350,9 +348,7 @@ if __name__ == "__main__": dockerd_internal_volume = "--tmpfs /var/lib/docker -e DOCKER_RAMDISK=true" elif args.dockerd_volume: dockerd_internal_volume = ( - "--mount type=bind,source={},target=/var/lib/docker".format( - args.dockerd_volume - ) + f"--mount type=bind,source={args.dockerd_volume},target=/var/lib/docker" ) else: try: diff --git a/tests/integration/test_attach_without_fetching/test.py b/tests/integration/test_attach_without_fetching/test.py index b430387e0f1..67352e2dcbe 100644 --- a/tests/integration/test_attach_without_fetching/test.py +++ b/tests/integration/test_attach_without_fetching/test.py @@ -56,6 +56,7 @@ def check_data(nodes, detached_parts): node.query_with_retry("SYSTEM SYNC REPLICA test") + for node in nodes: print("> Checking data integrity for", node.name) for i in range(10): diff --git a/tests/integration/test_backup_restore_new/test_cancel_backup.py b/tests/integration/test_backup_restore_new/test_cancel_backup.py index 6016bac9197..cce23a7e932 100644 --- a/tests/integration/test_backup_restore_new/test_cancel_backup.py +++ b/tests/integration/test_backup_restore_new/test_cancel_backup.py @@ -177,7 +177,7 @@ def cancel_restore(restore_id): def test_cancel_backup(): # We use partitioning so backups would contain more files. node.query( - "CREATE TABLE tbl (x UInt64) ENGINE=MergeTree() ORDER BY tuple() PARTITION BY x%5" + "CREATE TABLE tbl (x UInt64) ENGINE=MergeTree() ORDER BY tuple() PARTITION BY x%20" ) node.query(f"INSERT INTO tbl SELECT number FROM numbers(500)") diff --git a/tests/integration/test_checking_s3_blobs_paranoid/configs/setting.xml b/tests/integration/test_checking_s3_blobs_paranoid/configs/setting.xml index 23ab57f9330..d94ef68d9c4 100644 --- a/tests/integration/test_checking_s3_blobs_paranoid/configs/setting.xml +++ b/tests/integration/test_checking_s3_blobs_paranoid/configs/setting.xml @@ -5,6 +5,7 @@ 1 1 + 10000 diff --git a/tests/integration/test_checking_s3_blobs_paranoid/configs/storage_conf.xml b/tests/integration/test_checking_s3_blobs_paranoid/configs/storage_conf.xml index 7b1f503ed55..84f7f9f1b6d 100644 --- a/tests/integration/test_checking_s3_blobs_paranoid/configs/storage_conf.xml +++ b/tests/integration/test_checking_s3_blobs_paranoid/configs/storage_conf.xml @@ -19,6 +19,7 @@ minio minio123 1 + 10000 diff --git a/tests/integration/test_disabled_access_control_improvements/test_row_policy.py b/tests/integration/test_disabled_access_control_improvements/test_row_policy.py index b620e88e7eb..c09a80cea06 100644 --- a/tests/integration/test_disabled_access_control_improvements/test_row_policy.py +++ b/tests/integration/test_disabled_access_control_improvements/test_row_policy.py @@ -41,7 +41,7 @@ def started_cluster(): CREATE TABLE mydb.filtered_table2 (a UInt8, b UInt8, c UInt8, d UInt8) ENGINE MergeTree ORDER BY a; INSERT INTO mydb.filtered_table2 values (0, 0, 0, 0), (1, 2, 3, 4), (4, 3, 2, 1), (0, 0, 6, 0); - CREATE TABLE mydb.filtered_table3 (a UInt8, b UInt8, c UInt16 ALIAS a + b) ENGINE MergeTree ORDER BY a; + CREATE TABLE mydb.filtered_table3 (a UInt8, b UInt8, bb ALIAS b + 1, c UInt16 ALIAS a + bb - 1) ENGINE MergeTree ORDER BY a; INSERT INTO mydb.filtered_table3 values (0, 0), (0, 1), (1, 0), (1, 1); CREATE TABLE mydb.`.filtered_table4` (a UInt8, b UInt8, c UInt16 ALIAS a + b) ENGINE MergeTree ORDER BY a; diff --git a/tests/integration/test_disk_over_web_server/test.py b/tests/integration/test_disk_over_web_server/test.py index e84209a03a1..dbcd7cc3c21 100644 --- a/tests/integration/test_disk_over_web_server/test.py +++ b/tests/integration/test_disk_over_web_server/test.py @@ -278,7 +278,7 @@ def test_unavailable_server(cluster): "Caught exception while loading metadata.*Connection refused" ) assert node2.contains_in_log( - "HTTP request to \`http://nginx:8080/test1/.*\` failed at try 1/10 with bytes read: 0/unknown. Error: Connection refused." + "Failed to make request to 'http://nginx:8080/test1/.*'. Error: 'Connection refused'. Failed at try 10/10." ) finally: node2.exec_in_container( diff --git a/tests/integration/test_dns_cache/test.py b/tests/integration/test_dns_cache/test.py index 9c1c9797383..a6db26c8575 100644 --- a/tests/integration/test_dns_cache/test.py +++ b/tests/integration/test_dns_cache/test.py @@ -46,6 +46,7 @@ def cluster_without_dns_cache_update(): except Exception as ex: print(ex) + raise finally: cluster.shutdown() @@ -61,6 +62,7 @@ def test_ip_change_drop_dns_cache(cluster_without_dns_cache_update): node2.set_hosts([("2001:3984:3989::1:1111", "node1")]) # drop DNS cache node2.query("SYSTEM DROP DNS CACHE") + node2.query("SYSTEM DROP CONNECTIONS CACHE") # First we check, that normal replication works node1.query( @@ -86,6 +88,7 @@ def test_ip_change_drop_dns_cache(cluster_without_dns_cache_update): # drop DNS cache node2.query("SYSTEM DROP DNS CACHE") + node2.query("SYSTEM DROP CONNECTIONS CACHE") # Data is downloaded assert_eq_with_retry(node2, "SELECT count(*) from test_table_drop", "6") @@ -124,6 +127,7 @@ def cluster_with_dns_cache_update(): except Exception as ex: print(ex) + raise finally: cluster.shutdown() @@ -267,6 +271,11 @@ def test_user_access_ip_change(cluster_with_dns_cache_update, node): privileged=True, user="root", ) + node.exec_in_container( + ["bash", "-c", 'clickhouse client -q "SYSTEM DROP CONNECTIONS CACHE"'], + privileged=True, + user="root", + ) retry_count = 1 assert_eq_with_retry( @@ -296,7 +305,8 @@ def test_host_is_drop_from_cache_after_consecutive_failures( # Note that the list of hosts in variable since lost_host will be there too (and it's dropped and added back) # dns_update_short -> dns_max_consecutive_failures set to 6 assert node4.wait_for_log_line( - "Code: 198. DB::Exception: Not found address of host: InvalidHostThatDoesNotExist." + regexp="Code: 198. DB::NetException: Not found address of host: InvalidHostThatDoesNotExist.", + look_behind_lines=300, ) assert node4.wait_for_log_line( "Cached hosts not found:.*InvalidHostThatDoesNotExist**", diff --git a/tests/integration/test_http_failover/test.py b/tests/integration/test_http_failover/test.py index 41b55ef635c..5920fd980ce 100644 --- a/tests/integration/test_http_failover/test.py +++ b/tests/integration/test_http_failover/test.py @@ -56,9 +56,10 @@ def dst_node_addrs(started_cluster, request): yield - # Clear static DNS entries + # Clear static DNS entries and all keep alive connections src_node.set_hosts([]) src_node.query("SYSTEM DROP DNS CACHE") + src_node.query("SYSTEM DROP CONNECTIONS CACHE") @pytest.mark.parametrize( @@ -77,7 +78,8 @@ def dst_node_addrs(started_cluster, request): def test_url_destination_host_with_multiple_addrs(dst_node_addrs, expectation): with expectation: result = src_node.query( - "SELECT * FROM url('http://dst_node:8123/?query=SELECT+42', TSV, 'column1 UInt32')" + "SELECT * FROM url('http://dst_node:8123/?query=SELECT+42', TSV, 'column1 UInt32')", + settings={"http_max_tries": "3"}, ) assert result == "42\n" diff --git a/tests/integration/test_keeper_four_word_command/test.py b/tests/integration/test_keeper_four_word_command/test.py index 84dd2a2fd93..44b2b50673a 100644 --- a/tests/integration/test_keeper_four_word_command/test.py +++ b/tests/integration/test_keeper_four_word_command/test.py @@ -266,6 +266,7 @@ def test_cmd_conf(started_cluster): assert result["heart_beat_interval_ms"] == "500" assert result["election_timeout_lower_bound_ms"] == "1000" assert result["election_timeout_upper_bound_ms"] == "2000" + assert result["leadership_expiry_ms"] == "0" assert result["reserved_log_items"] == "100000" assert result["snapshot_distance"] == "75" diff --git a/tests/integration/test_odbc_interaction/test_exiled.py b/tests/integration/test_odbc_interaction/test_exiled.py deleted file mode 100644 index bdc819b8221..00000000000 --- a/tests/integration/test_odbc_interaction/test_exiled.py +++ /dev/null @@ -1,111 +0,0 @@ -import time -import logging -import pytest - -from helpers.cluster import ClickHouseCluster, assert_eq_with_retry -from test_odbc_interaction.test import ( - create_mysql_db, - create_mysql_table, - get_mysql_conn, - skip_test_msan, -) - - -cluster = ClickHouseCluster(__file__) -node1 = cluster.add_instance( - "node1", - with_odbc_drivers=True, - main_configs=["configs/openssl.xml", "configs/odbc_logging.xml"], - stay_alive=True, - dictionaries=["configs/dictionaries/sqlite3_odbc_hashed_dictionary.xml"], -) - - -@pytest.fixture(scope="module") -def started_cluster(): - try: - cluster.start() - - sqlite_db = node1.odbc_drivers["SQLite3"]["Database"] - logging.debug(f"sqlite data received: {sqlite_db}") - node1.exec_in_container( - [ - "sqlite3", - sqlite_db, - "CREATE TABLE t2(id INTEGER PRIMARY KEY ASC, X INTEGER, Y, Z);", - ], - privileged=True, - user="root", - ) - - node1.exec_in_container( - ["sqlite3", sqlite_db, "INSERT INTO t2 values(1, 1, 2, 3);"], - privileged=True, - user="root", - ) - - node1.query("SYSTEM RELOAD DICTIONARY sqlite3_odbc_hashed") - - yield cluster - except Exception as ex: - logging.exception(ex) - raise ex - finally: - cluster.shutdown() - - -# This test kills ClickHouse server and ODBC bridge and in worst scenario -# may cause group test crashes. Thus, this test is executed in a separate "module" -# with separate environment. -def test_bridge_dies_with_parent(started_cluster): - skip_test_msan(node1) - - if node1.is_built_with_address_sanitizer(): - # TODO: Leak sanitizer falsely reports about a leak of 16 bytes in clickhouse-odbc-bridge in this test and - # that's linked somehow with that we have replaced getauxval() in glibc-compatibility. - # The leak sanitizer calls getauxval() for its own purposes, and our replaced version doesn't seem to be equivalent in that case. - pytest.skip( - "Leak sanitizer falsely reports about a leak of 16 bytes in clickhouse-odbc-bridge" - ) - - assert_eq_with_retry( - node1, "select dictGetUInt8('sqlite3_odbc_hashed', 'Z', toUInt64(1))", "3" - ) - - clickhouse_pid = node1.get_process_pid("clickhouse server") - bridge_pid = node1.get_process_pid("odbc-bridge") - assert clickhouse_pid is not None - assert bridge_pid is not None - - try: - node1.exec_in_container( - ["kill", str(clickhouse_pid)], privileged=True, user="root" - ) - except: - pass - - for _ in range(30): - time.sleep(1) - clickhouse_pid = node1.get_process_pid("clickhouse server") - if clickhouse_pid is None: - break - - for _ in range(30): - time.sleep(1) # just for sure, that odbc-bridge caught signal - bridge_pid = node1.get_process_pid("odbc-bridge") - if bridge_pid is None: - break - - if bridge_pid: - out = node1.exec_in_container( - ["gdb", "-p", str(bridge_pid), "--ex", "thread apply all bt", "--ex", "q"], - privileged=True, - user="root", - ) - logging.debug(f"Bridge is running, gdb output:\n{out}") - - try: - assert clickhouse_pid is None - assert bridge_pid is None - finally: - node1.start_clickhouse(20) diff --git a/tests/integration/test_redirect_url_storage/test.py b/tests/integration/test_redirect_url_storage/test.py index 17a9a03008e..033f02d7bde 100644 --- a/tests/integration/test_redirect_url_storage/test.py +++ b/tests/integration/test_redirect_url_storage/test.py @@ -150,7 +150,7 @@ def test_url_reconnect(started_cluster): def select(): global result result = node1.query( - "select sum(cityHash64(id)) from url('http://hdfs1:50075/webhdfs/v1/storage_big?op=OPEN&namenoderpcaddress=hdfs1:9000&offset=0', 'TSV', 'id Int32') settings http_max_tries = 10, http_retry_max_backoff_ms=1000" + "select sum(cityHash64(id)) from url('http://hdfs1:50075/webhdfs/v1/storage_big?op=OPEN&namenoderpcaddress=hdfs1:9000&offset=0', 'TSV', 'id Int32') settings http_max_tries=10, http_retry_max_backoff_ms=1000, http_make_head_request=false" ) assert int(result) == 6581218782194912115 diff --git a/tests/integration/test_replicated_database/test.py b/tests/integration/test_replicated_database/test.py index b47f86a843d..bd80ac25468 100644 --- a/tests/integration/test_replicated_database/test.py +++ b/tests/integration/test_replicated_database/test.py @@ -508,12 +508,9 @@ def test_alters_from_different_replicas(started_cluster): dummy_node.stop_clickhouse(kill=True) settings = {"distributed_ddl_task_timeout": 5} - assert ( - "There are 1 unfinished hosts (0 of them are currently executing the task" - in competing_node.query_and_get_error( - "ALTER TABLE alters_from_different_replicas.concurrent_test ADD COLUMN Added0 UInt32;", - settings=settings, - ) + assert "is not finished on 1 of 3 hosts" in competing_node.query_and_get_error( + "ALTER TABLE alters_from_different_replicas.concurrent_test ADD COLUMN Added0 UInt32;", + settings=settings, ) settings = { "distributed_ddl_task_timeout": 5, diff --git a/tests/integration/test_replicated_database_cluster_groups/test.py b/tests/integration/test_replicated_database_cluster_groups/test.py index 647626d8014..91361c1850b 100644 --- a/tests/integration/test_replicated_database_cluster_groups/test.py +++ b/tests/integration/test_replicated_database_cluster_groups/test.py @@ -95,12 +95,9 @@ def test_cluster_groups(started_cluster): # Exception main_node_2.stop_clickhouse() settings = {"distributed_ddl_task_timeout": 5} - assert ( - "There are 1 unfinished hosts (0 of them are currently executing the task)" - in main_node_1.query_and_get_error( - "CREATE TABLE cluster_groups.table_2 (d Date, k UInt64) ENGINE=ReplicatedMergeTree ORDER BY k PARTITION BY toYYYYMM(d);", - settings=settings, - ) + assert "is not finished on 1 of 2 hosts" in main_node_1.query_and_get_error( + "CREATE TABLE cluster_groups.table_2 (d Date, k UInt64) ENGINE=ReplicatedMergeTree ORDER BY k PARTITION BY toYYYYMM(d);", + settings=settings, ) # 3. After start both groups are synced diff --git a/tests/integration/test_replicated_merge_tree_wait_on_shutdown/test.py b/tests/integration/test_replicated_merge_tree_wait_on_shutdown/test.py index 67dd03098e9..995afedf415 100644 --- a/tests/integration/test_replicated_merge_tree_wait_on_shutdown/test.py +++ b/tests/integration/test_replicated_merge_tree_wait_on_shutdown/test.py @@ -40,6 +40,10 @@ def test_shutdown_and_wait(start_cluster): f"CREATE TABLE test_table (value UInt64) ENGINE=ReplicatedMergeTree('/test/table', 'r{i}') ORDER BY tuple()" ) + # we stop merges on node1 to make node2 fetch all 51 origin parts from node1 + # and not to fetch a smaller set of merged covering parts + node1.query("SYSTEM STOP MERGES test_table") + node1.query("INSERT INTO test_table VALUES (0)") node2.query("SYSTEM SYNC REPLICA test_table") diff --git a/tests/integration/test_row_policy/test.py b/tests/integration/test_row_policy/test.py index 98653bf6106..8260be78e82 100644 --- a/tests/integration/test_row_policy/test.py +++ b/tests/integration/test_row_policy/test.py @@ -60,7 +60,7 @@ def started_cluster(): CREATE TABLE mydb.filtered_table2 (a UInt8, b UInt8, c UInt8, d UInt8) ENGINE MergeTree ORDER BY a; INSERT INTO mydb.filtered_table2 values (0, 0, 0, 0), (1, 2, 3, 4), (4, 3, 2, 1), (0, 0, 6, 0); - CREATE TABLE mydb.filtered_table3 (a UInt8, b UInt8, c UInt16 ALIAS a + b) ENGINE MergeTree ORDER BY a; + CREATE TABLE mydb.filtered_table3 (a UInt8, b UInt8, bb ALIAS b + 1, c UInt16 ALIAS a + bb - 1) ENGINE MergeTree ORDER BY a; INSERT INTO mydb.filtered_table3 values (0, 0), (0, 1), (1, 0), (1, 1); CREATE TABLE mydb.`.filtered_table4` (a UInt8, b UInt8, c UInt16 ALIAS a + b) ENGINE MergeTree ORDER BY a; @@ -113,6 +113,7 @@ def test_smoke(): assert node.query("SELECT a FROM mydb.filtered_table3") == TSV([[0], [1]]) assert node.query("SELECT b FROM mydb.filtered_table3") == TSV([[1], [0]]) + assert node.query("SELECT bb FROM mydb.filtered_table3") == TSV([[2], [1]]) assert node.query("SELECT c FROM mydb.filtered_table3") == TSV([[1], [1]]) assert node.query("SELECT a + b FROM mydb.filtered_table3") == TSV([[1], [1]]) assert node.query("SELECT a FROM mydb.filtered_table3 WHERE c = 1") == TSV( diff --git a/tests/integration/test_s3_table_functions/test.py b/tests/integration/test_s3_table_functions/test.py index a6def175136..ff62d1a9eac 100644 --- a/tests/integration/test_s3_table_functions/test.py +++ b/tests/integration/test_s3_table_functions/test.py @@ -80,6 +80,7 @@ def test_s3_table_functions_timeouts(started_cluster): Test with timeout limit of 1200ms. This should raise an Exception and pass. """ + with PartitionManager() as pm: pm.add_network_delay(node, 1200) diff --git a/tests/integration/test_storage_numbers/test.py b/tests/integration/test_storage_numbers/test.py index 61fe8719ea2..cbd7793fd8c 100644 --- a/tests/integration/test_storage_numbers/test.py +++ b/tests/integration/test_storage_numbers/test.py @@ -242,3 +242,19 @@ def test_overflow(started_cluster): ) assert response == "(18446744073709551614),(18446744073709551615),(0),(1),(2)" check_read_rows("test_overflow", 5) + + +def test_non_number_filter(started_cluster): + response = node.query( + "SELECT toString(number) as a FROM numbers(3) WHERE a = '1' FORMAT Values", + query_id="test_non_number_filter", + ) + assert response == "('1')" + check_read_rows("test_non_number_filter", 3) + + response = node.query( + "SELECT toString(number) as a FROM numbers(1, 4) WHERE a = '1' FORMAT Values SETTINGS max_block_size = 3", + query_id="test_non_number_filter2", + ) + assert response == "('1')" + check_read_rows("test_non_number_filter2", 4) diff --git a/tests/integration/test_storage_s3/s3_mocks/unstable_server.py b/tests/integration/test_storage_s3/s3_mocks/unstable_server.py index 5ef781bdc9e..70d49b7c1b0 100644 --- a/tests/integration/test_storage_s3/s3_mocks/unstable_server.py +++ b/tests/integration/test_storage_s3/s3_mocks/unstable_server.py @@ -118,5 +118,5 @@ class RequestHandler(http.server.BaseHTTPRequestHandler): self.wfile.write(b"OK") -httpd = http.server.HTTPServer(("0.0.0.0", int(sys.argv[1])), RequestHandler) +httpd = http.server.ThreadingHTTPServer(("0.0.0.0", int(sys.argv[1])), RequestHandler) httpd.serve_forever() diff --git a/tests/performance/multiif.xml b/tests/performance/multiif.xml index ad56ab3f5f2..0c2d95cc553 100644 --- a/tests/performance/multiif.xml +++ b/tests/performance/multiif.xml @@ -5,4 +5,12 @@ select count(1) from test_multiif_t where multiIf(d > 2, d-2, d > 1, d-1, d >0, d, 0) > 1 SETTINGS max_threads=1 DROP TABLE IF EXISTS test_multiif_t + + + + SELECT count() FROM zeros(10000000) WHERE NOT ignore(multiIf( rand(1) % 2 = 0, materialize(1::Nullable(Decimal256(3))), rand(2) % 2 = 0, materialize(2::Nullable(Decimal256(3))), rand(3) % 2 = 0, materialize(3::Nullable(Decimal256(3))), rand(4) % 2 = 0, materialize(4::Nullable(Decimal256(3))), rand(5) % 2 = 0, materialize(5::Nullable(Decimal256(3))), materialize(6::Nullable(Decimal256(3))))) + SELECT count() FROM zeros(10000000) WHERE NOT ignore(multiIf( rand(1) % 2 = 0, materialize(1::Decimal256(3)), rand(2) % 2 = 0, materialize(2::Decimal256(3)), rand(3) % 2 = 0, materialize(3::Decimal256(3)), rand(4) % 2 = 0, materialize(4::Decimal256(3)), rand(5) % 2 = 0, materialize(5::Decimal256(3)), materialize(6::Decimal256(3)))) + + + SELECT count() FROM zeros(10000000) WHERE NOT ignore(multiIf(rand() % 2 = 0, rand()+2, rand() % 3 = 0, rand()+3, rand() % 4 = 0, rand()+4, rand() % 5 = 0, rand() + 5, rand() % 6 = 0, rand() + 6, rand())) diff --git a/tests/queries/0_stateless/00597_push_down_predicate_long.reference b/tests/queries/0_stateless/00597_push_down_predicate_long.reference index 04b0432c0a9..2c46edc98bf 100644 --- a/tests/queries/0_stateless/00597_push_down_predicate_long.reference +++ b/tests/queries/0_stateless/00597_push_down_predicate_long.reference @@ -114,7 +114,7 @@ FROM ( SELECT 1 AS id, - identity(_CAST(1, \'Nullable(UInt8)\')) AS subquery + __scalarSubqueryResult(_CAST(1, \'Nullable(UInt8)\')) AS subquery WHERE subquery = 1 ) WHERE subquery = 1 diff --git a/tests/queries/0_stateless/00646_url_engine.python b/tests/queries/0_stateless/00646_url_engine.python index dc0fdd1a71d..931d18a3f80 100644 --- a/tests/queries/0_stateless/00646_url_engine.python +++ b/tests/queries/0_stateless/00646_url_engine.python @@ -12,6 +12,7 @@ import urllib.request import subprocess from io import StringIO from http.server import BaseHTTPRequestHandler, HTTPServer +from socketserver import ThreadingMixIn def is_ipv6(host): @@ -145,11 +146,19 @@ class HTTPServerV6(HTTPServer): address_family = socket.AF_INET6 +class ThreadedHTTPServer(ThreadingMixIn, HTTPServer): + pass + + +class ThreadedHTTPServerV6(ThreadingMixIn, HTTPServerV6): + pass + + def start_server(): if IS_IPV6: - httpd = HTTPServerV6(HTTP_SERVER_ADDRESS, CSVHTTPServer) + httpd = ThreadedHTTPServerV6(HTTP_SERVER_ADDRESS, CSVHTTPServer) else: - httpd = HTTPServer(HTTP_SERVER_ADDRESS, CSVHTTPServer) + httpd = ThreadedHTTPServer(HTTP_SERVER_ADDRESS, CSVHTTPServer) t = threading.Thread(target=httpd.serve_forever) return t, httpd diff --git a/tests/queries/0_stateless/00849_multiple_comma_join_2.sql b/tests/queries/0_stateless/00849_multiple_comma_join_2.sql index 51bf5a2ede1..6530f691087 100644 --- a/tests/queries/0_stateless/00849_multiple_comma_join_2.sql +++ b/tests/queries/0_stateless/00849_multiple_comma_join_2.sql @@ -62,49 +62,49 @@ SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explai --- EXPLAIN QUERY TREE SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2 WHERE t1.a = t2.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2 WHERE t1.a = t2.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2 WHERE t1.b = t2.b SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2 WHERE t1.b = t2.b) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3 WHERE t1.a = t2.a AND t1.a = t3.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3 WHERE t1.a = t2.a AND t1.a = t3.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3 WHERE t1.b = t2.b AND t1.b = t3.b SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3 WHERE t1.b = t2.b AND t1.b = t3.b) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t1.a = t2.a AND t1.a = t3.a AND t1.a = t4.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t1.a = t2.a AND t1.a = t3.a AND t1.a = t4.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t1.b = t2.b AND t1.b = t3.b AND t1.b = t4.b SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t1.b = t2.b AND t1.b = t3.b AND t1.b = t4.b) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t2.a = t1.a AND t2.a = t3.a AND t2.a = t4.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t2.a = t1.a AND t2.a = t3.a AND t2.a = t4.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t3.a = t1.a AND t3.a = t2.a AND t3.a = t4.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t3.a = t1.a AND t3.a = t2.a AND t3.a = t4.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t4.a = t1.a AND t4.a = t2.a AND t4.a = t3.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t4.a = t1.a AND t4.a = t2.a AND t4.a = t3.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t1.a = t2.a AND t2.a = t3.a AND t3.a = t4.a SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 WHERE t1.a = t2.a AND t2.a = t3.a AND t3.a = t4.a) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4 SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2, t3, t4) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1 CROSS JOIN t2 CROSS JOIN t3 CROSS JOIN t4 SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1 CROSS JOIN t2 CROSS JOIN t3 CROSS JOIN t4) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2 CROSS JOIN t3 SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1, t2 CROSS JOIN t3) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1 JOIN t2 USING a CROSS JOIN t3 SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1 JOIN t2 USING a CROSS JOIN t3) SETTINGS allow_experimental_analyzer = 1; SELECT countIf(explain like '%COMMA%' OR explain like '%CROSS%'), countIf(explain like '%INNER%') FROM ( - EXPLAIN QUERY TREE SELECT t1.a FROM t1 JOIN t2 ON t1.a = t2.a CROSS JOIN t3 SETTINGS allow_experimental_analyzer = 1); + EXPLAIN QUERY TREE SELECT t1.a FROM t1 JOIN t2 ON t1.a = t2.a CROSS JOIN t3) SETTINGS allow_experimental_analyzer = 1; INSERT INTO t1 values (1,1), (2,2), (3,3), (4,4); INSERT INTO t2 values (1,1), (1, Null); diff --git a/tests/queries/0_stateless/01029_early_constant_folding.reference b/tests/queries/0_stateless/01029_early_constant_folding.reference index abcb2ddc6a7..4df5414ba4a 100644 --- a/tests/queries/0_stateless/01029_early_constant_folding.reference +++ b/tests/queries/0_stateless/01029_early_constant_folding.reference @@ -2,7 +2,7 @@ SELECT 1 WHERE 0 SELECT 1 SELECT 1 -WHERE (1 IN (0, 2)) AND (2 = (identity(_CAST(2, \'Nullable(UInt8)\')) AS subquery)) +WHERE (1 IN (0, 2)) AND (2 = (__scalarSubqueryResult(_CAST(2, \'Nullable(UInt8)\')) AS subquery)) SELECT 1 WHERE 1 IN (( SELECT arrayJoin([1, 2, 3]) diff --git a/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.reference b/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.reference index b9a66a1e1a9..2151328d8b7 100644 --- a/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.reference +++ b/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.reference @@ -3,7 +3,7 @@ Received exception from server: Code: 57. Error: Received from localhost:9000. Error: There was an error on [localhost:9000]: Code: 57. Error: Table default.none already exists. (TABLE_ALREADY_EXISTS) (query: create table none on cluster test_shard_localhost (n int) engine=Memory;) Received exception from server: -Code: 159. Error: Received from localhost:9000. Error: Watching task is executing longer than distributed_ddl_task_timeout (=1) seconds. There are 1 unfinished hosts (0 of them are currently executing the task), they are going to execute the query in background. (TIMEOUT_EXCEEDED) +Code: 159. Error: Received from localhost:9000. Error: Distributed DDL task is not finished on 1 of 2 hosts (0 of them are currently executing the task, 0 are inactive). They are going to execute the query in background. Was waiting for seconds, which is longer than distributed_ddl_task_timeout. (TIMEOUT_EXCEEDED) (query: drop table if exists none on cluster test_unavailable_shard;) throw localhost 9000 0 0 0 @@ -12,7 +12,7 @@ Code: 57. Error: Received from localhost:9000. Error: There was an error on [loc (query: create table throw on cluster test_shard_localhost (n int) engine=Memory format Null;) localhost 9000 0 1 0 Received exception from server: -Code: 159. Error: Received from localhost:9000. Error: Watching task is executing longer than distributed_ddl_task_timeout (=1) seconds. There are 1 unfinished hosts (0 of them are currently executing the task), they are going to execute the query in background. (TIMEOUT_EXCEEDED) +Code: 159. Error: Received from localhost:9000. Error: Distributed DDL task is not finished on 1 of 2 hosts (0 of them are currently executing the task, 0 are inactive). They are going to execute the query in background. Was waiting for seconds, which is longer than distributed_ddl_task_timeout. (TIMEOUT_EXCEEDED) (query: drop table if exists throw on cluster test_unavailable_shard;) null_status_on_timeout localhost 9000 0 0 0 diff --git a/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.sh b/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.sh index 12e142adda9..f17e85da60a 100755 --- a/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.sh +++ b/tests/queries/0_stateless/01175_distributed_ddl_output_mode_long.sh @@ -54,14 +54,14 @@ $CLIENT --distributed_ddl_output_mode=none -q "create table none on cluster test $CLIENT --distributed_ddl_output_mode=none -q "create table none on cluster test_shard_localhost (n int) engine=Memory;" 2>&1 | sed "s/DB::Exception/Error/g" | sed "s/ (version.*)//" # Timeout -run_until_out_contains 'There are 1 unfinished hosts' $CLICKHOUSE_CLIENT_WITH_SETTINGS --distributed_ddl_output_mode=none -q "drop table if exists none on cluster test_unavailable_shard;" 2>&1 | sed "s/DB::Exception/Error/g" | sed "s/ (version.*)//" | sed "s/Watching task .* is executing longer/Watching task is executing longer/" +run_until_out_contains 'not finished on 1 ' $CLICKHOUSE_CLIENT_WITH_SETTINGS --distributed_ddl_output_mode=none -q "drop table if exists none on cluster test_unavailable_shard;" 2>&1 | sed "s/DB::Exception/Error/g" | sed "s/ (version.*)//" | sed "s/Distributed DDL task .* is not finished/Distributed DDL task is not finished/" | sed "s/for .* seconds/for seconds/" $CLIENT --distributed_ddl_output_mode=throw -q "select value from system.settings where name='distributed_ddl_output_mode';" $CLIENT --distributed_ddl_output_mode=throw -q "create table throw on cluster test_shard_localhost (n int) engine=Memory;" $CLIENT --distributed_ddl_output_mode=throw -q "create table throw on cluster test_shard_localhost (n int) engine=Memory format Null;" 2>&1 | sed "s/DB::Exception/Error/g" | sed "s/ (version.*)//" -run_until_out_contains 'There are 1 unfinished hosts' $CLICKHOUSE_CLIENT_WITH_SETTINGS --distributed_ddl_output_mode=throw -q "drop table if exists throw on cluster test_unavailable_shard;" 2>&1 | sed "s/DB::Exception/Error/g" | sed "s/ (version.*)//" | sed "s/Watching task .* is executing longer/Watching task is executing longer/" +run_until_out_contains 'not finished on 1 ' $CLICKHOUSE_CLIENT_WITH_SETTINGS --distributed_ddl_output_mode=throw -q "drop table if exists throw on cluster test_unavailable_shard;" 2>&1 | sed "s/DB::Exception/Error/g" | sed "s/ (version.*)//" | sed "s/Distributed DDL task .* is not finished/Distributed DDL task is not finished/" | sed "s/for .* seconds/for seconds/" $CLIENT --distributed_ddl_output_mode=null_status_on_timeout -q "select value from system.settings where name='distributed_ddl_output_mode';" diff --git a/tests/queries/0_stateless/01271_show_privileges.reference b/tests/queries/0_stateless/01271_show_privileges.reference index 7af299c6728..b18ae8a99be 100644 --- a/tests/queries/0_stateless/01271_show_privileges.reference +++ b/tests/queries/0_stateless/01271_show_privileges.reference @@ -105,6 +105,7 @@ NAMED COLLECTION ADMIN ['NAMED COLLECTION CONTROL'] NAMED_COLLECTION ALL SET DEFINER [] USER_NAME ALL SYSTEM SHUTDOWN ['SYSTEM KILL','SHUTDOWN'] GLOBAL SYSTEM SYSTEM DROP DNS CACHE ['SYSTEM DROP DNS','DROP DNS CACHE','DROP DNS'] GLOBAL SYSTEM DROP CACHE +SYSTEM DROP CONNECTIONS CACHE ['SYSTEM DROP CONNECTIONS CACHE','DROP CONNECTIONS CACHE'] GLOBAL SYSTEM DROP CACHE SYSTEM DROP MARK CACHE ['SYSTEM DROP MARK','DROP MARK CACHE','DROP MARKS'] GLOBAL SYSTEM DROP CACHE SYSTEM DROP UNCOMPRESSED CACHE ['SYSTEM DROP UNCOMPRESSED','DROP UNCOMPRESSED CACHE','DROP UNCOMPRESSED'] GLOBAL SYSTEM DROP CACHE SYSTEM DROP MMAP CACHE ['SYSTEM DROP MMAP','DROP MMAP CACHE','DROP MMAP'] GLOBAL SYSTEM DROP CACHE diff --git a/tests/queries/0_stateless/01281_parseDateTime64BestEffort.sql b/tests/queries/0_stateless/01281_parseDateTime64BestEffort.sql index c1cec6ea212..808eaf291d5 100644 --- a/tests/queries/0_stateless/01281_parseDateTime64BestEffort.sql +++ b/tests/queries/0_stateless/01281_parseDateTime64BestEffort.sql @@ -7,7 +7,7 @@ SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184Z', 'bar'); -- {ser SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184Z', 3, 4); -- {serverError 43} -- invalid timezone parameter SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184Z', 3, 'baz'); -- {serverError BAD_ARGUMENTS} -- unknown timezone -SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184Z', materialize(3), 4); -- {serverError 44} -- non-const precision +SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184Z', materialize(3), 4); -- {serverError 43, 44} -- non-const precision SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184Z', 3, materialize('UTC')); -- {serverError 44} -- non-const timezone SELECT parseDateTime64BestEffort('2020-05-14T03:37:03.253184012345678910111213141516171819Z', 3, 'UTC'); -- {serverError 6} diff --git a/tests/queries/0_stateless/01293_show_settings.reference b/tests/queries/0_stateless/01293_show_settings.reference index f053387d1c5..187f55697e4 100644 --- a/tests/queries/0_stateless/01293_show_settings.reference +++ b/tests/queries/0_stateless/01293_show_settings.reference @@ -3,6 +3,7 @@ connect_timeout Seconds 10 connect_timeout_with_failover_ms Milliseconds 2000 connect_timeout_with_failover_secure_ms Milliseconds 3000 external_storage_connect_timeout_sec UInt64 10 +s3_connect_timeout_ms UInt64 1000 filesystem_prefetch_max_memory_usage UInt64 1073741824 max_untracked_memory UInt64 1048576 memory_profiler_step UInt64 1048576 diff --git a/tests/queries/0_stateless/01611_constant_folding_subqueries.reference b/tests/queries/0_stateless/01611_constant_folding_subqueries.reference index c3df2314112..327a4694aa8 100644 --- a/tests/queries/0_stateless/01611_constant_folding_subqueries.reference +++ b/tests/queries/0_stateless/01611_constant_folding_subqueries.reference @@ -5,7 +5,13 @@ SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n) FO 1,10 EXPLAIN SYNTAX SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n); SELECT - identity(_CAST(0, \'Nullable(UInt64)\')) AS n, + __scalarSubqueryResult(_CAST(0, \'Nullable(UInt64)\')) AS n, toUInt64(10 / n) SELECT * FROM (WITH (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n) as q SELECT * FROM system.one WHERE q > 0); 0 +SELECT * FROM (SELECT (SELECT '\d[a-z]') AS n, extractAll('5abc', assumeNotNull(n))) FORMAT CSV; +"\d[a-z]","['5a']" +EXPLAIN SYNTAX SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n); +SELECT + __scalarSubqueryResult(_CAST(0, \'Nullable(UInt64)\')) AS n, + toUInt64(10 / n) diff --git a/tests/queries/0_stateless/01611_constant_folding_subqueries.sql b/tests/queries/0_stateless/01611_constant_folding_subqueries.sql index 59f057d1ec5..b30fb43f621 100644 --- a/tests/queries/0_stateless/01611_constant_folding_subqueries.sql +++ b/tests/queries/0_stateless/01611_constant_folding_subqueries.sql @@ -3,3 +3,6 @@ SELECT * FROM (SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUI SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n) FORMAT CSV; EXPLAIN SYNTAX SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n); SELECT * FROM (WITH (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n) as q SELECT * FROM system.one WHERE q > 0); + +SELECT * FROM (SELECT (SELECT '\d[a-z]') AS n, extractAll('5abc', assumeNotNull(n))) FORMAT CSV; +EXPLAIN SYNTAX SELECT (SELECT * FROM system.numbers LIMIT 1 OFFSET 1) AS n, toUInt64(10 / n); diff --git a/tests/queries/0_stateless/01848_partition_value_column.sql b/tests/queries/0_stateless/01848_partition_value_column.sql index 28d842af3e9..de5e766c92c 100644 --- a/tests/queries/0_stateless/01848_partition_value_column.sql +++ b/tests/queries/0_stateless/01848_partition_value_column.sql @@ -14,8 +14,8 @@ select count() from tbl where _partition_value.3 = 4 settings max_rows_to_read = create table tbl2(i int) engine MergeTree order by i; insert into tbl2 values (1); -select _partition_value from tbl2; -- { serverError 16 } -select _partition_value from tbl2 group by 1; -- { serverError 16 } +select _partition_value from tbl2; -- { serverError UNKNOWN_IDENTIFIER } +select _partition_value from tbl2 group by 1; -- { serverError UNKNOWN_IDENTIFIER } drop table tbl; drop table tbl2; diff --git a/tests/queries/0_stateless/02116_tuple_element.sql b/tests/queries/0_stateless/02116_tuple_element.sql index 97f6c049705..64d9b9db331 100644 --- a/tests/queries/0_stateless/02116_tuple_element.sql +++ b/tests/queries/0_stateless/02116_tuple_element.sql @@ -19,7 +19,7 @@ SELECT tupleElement(t1) FROM t_tuple_element; -- { serverError NUMBER_OF_ARGUMEN SELECT tupleElement(t1, 'b') FROM t_tuple_element; -- { serverError NOT_FOUND_COLUMN_IN_BLOCK, UNKNOWN_IDENTIFIER } SELECT tupleElement(t1, 0) FROM t_tuple_element; -- { serverError ILLEGAL_INDEX, NOT_FOUND_COLUMN_IN_BLOCK } SELECT tupleElement(t1, 3) FROM t_tuple_element; -- { serverError ILLEGAL_INDEX, NOT_FOUND_COLUMN_IN_BLOCK } -SELECT tupleElement(t1, materialize('a')) FROM t_tuple_element; -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT tupleElement(t1, materialize('a')) FROM t_tuple_element; -- { serverError ILLEGAL_TYPE_OF_ARGUMENT, NOT_FOUND_COLUMN_IN_BLOCK } SELECT t2.1 FROM t_tuple_element; EXPLAIN SYNTAX SELECT t2.1 FROM t_tuple_element; @@ -31,7 +31,7 @@ SELECT tupleElement(t2) FROM t_tuple_element; -- { serverError NUMBER_OF_ARGUMEN SELECT tupleElement(t2, 'a') FROM t_tuple_element; -- { serverError NOT_FOUND_COLUMN_IN_BLOCK, UNKNOWN_IDENTIFIER } SELECT tupleElement(t2, 0) FROM t_tuple_element; -- { serverError ILLEGAL_INDEX, NOT_FOUND_COLUMN_IN_BLOCK } SELECT tupleElement(t2, 3) FROM t_tuple_element; -- { serverError ILLEGAL_INDEX, NOT_FOUND_COLUMN_IN_BLOCK } -SELECT tupleElement(t2, materialize(1)) FROM t_tuple_element; -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT tupleElement(t2, materialize(1)) FROM t_tuple_element; -- { serverError ILLEGAL_TYPE_OF_ARGUMENT, NOT_FOUND_COLUMN_IN_BLOCK } DROP TABLE t_tuple_element; diff --git a/tests/queries/0_stateless/02205_HTTP_user_agent.python b/tests/queries/0_stateless/02205_HTTP_user_agent.python index d8f8a32b6db..83089741bf2 100644 --- a/tests/queries/0_stateless/02205_HTTP_user_agent.python +++ b/tests/queries/0_stateless/02205_HTTP_user_agent.python @@ -1,6 +1,7 @@ #!/usr/bin/env python3 from http.server import SimpleHTTPRequestHandler, HTTPServer +from socketserver import ThreadingMixIn import socket import sys import threading @@ -116,11 +117,19 @@ class HTTPServerV6(HTTPServer): address_family = socket.AF_INET6 +class ThreadedHTTPServer(ThreadingMixIn, HTTPServer): + pass + + +class ThreadedHTTPServerV6(ThreadingMixIn, HTTPServerV6): + pass + + def start_server(requests_amount): if IS_IPV6: - httpd = HTTPServerV6(HTTP_SERVER_ADDRESS, HttpProcessor) + httpd = ThreadedHTTPServerV6(HTTP_SERVER_ADDRESS, HttpProcessor) else: - httpd = HTTPServer(HTTP_SERVER_ADDRESS, HttpProcessor) + httpd = ThreadedHTTPServer(HTTP_SERVER_ADDRESS, HttpProcessor) def real_func(): for i in range(requests_amount): diff --git a/tests/queries/0_stateless/02227_union_match_by_name.reference b/tests/queries/0_stateless/02227_union_match_by_name.reference index c28035fab49..d726ae86de7 100644 --- a/tests/queries/0_stateless/02227_union_match_by_name.reference +++ b/tests/queries/0_stateless/02227_union_match_by_name.reference @@ -36,7 +36,7 @@ Header: avgWeighted(x, y) Nullable(Float64) Header: x Nullable(Nothing) y UInt8 Expression (Projection) - Header: NULL_Nullable(Nothing) Nullable(Nothing) + Header: _CAST(NULL_Nullable(Nothing), \'Nullable(Nothing)\'_String) Nullable(Nothing) 1_UInt8 UInt8 Expression (Change column names to column identifiers) Header: __table5.dummy UInt8 diff --git a/tests/queries/0_stateless/02233_HTTP_ranged.python b/tests/queries/0_stateless/02233_HTTP_ranged.python index 66ef3304098..5d06e4824b1 100644 --- a/tests/queries/0_stateless/02233_HTTP_ranged.python +++ b/tests/queries/0_stateless/02233_HTTP_ranged.python @@ -1,6 +1,7 @@ #!/usr/bin/env python3 from http.server import BaseHTTPRequestHandler, HTTPServer +from socketserver import ThreadingMixIn import socket import sys import re @@ -206,13 +207,22 @@ class HTTPServerV6(HTTPServer): address_family = socket.AF_INET6 +class ThreadedHTTPServer(ThreadingMixIn, HTTPServer): + pass + + +class ThreadedHTTPServerV6(ThreadingMixIn, HTTPServerV6): + pass + + def start_server(): if IS_IPV6: - httpd = HTTPServerV6(HTTP_SERVER_ADDRESS, HttpProcessor) + httpd = ThreadedHTTPServerV6(HTTP_SERVER_ADDRESS, HttpProcessor) else: - httpd = HTTPServer(HTTP_SERVER_ADDRESS, HttpProcessor) + httpd = ThreadedHTTPServer(HTTP_SERVER_ADDRESS, HttpProcessor) t = threading.Thread(target=httpd.serve_forever) + t.start() return t, httpd @@ -235,8 +245,6 @@ def run_test(allow_range, settings, check_retries=False): HttpProcessor.responses_to_get = ["500", "200", "206"] retries_num = len(HttpProcessor.responses_to_get) - t, httpd = start_server() - t.start() test_select(settings) download_buffer_size = settings["max_download_buffer_size"] @@ -261,12 +269,12 @@ def run_test(allow_range, settings, check_retries=False): if HttpProcessor.range_used: raise Exception("HTTP Range used while not supported") - httpd.shutdown() - t.join() print("PASSED") def main(): + t, httpd = start_server() + settings = {"max_download_buffer_size": 20} # Test Accept-Ranges=False @@ -285,10 +293,15 @@ def main(): settings["max_download_threads"] = 2 run_test(allow_range=True, settings=settings, check_retries=True) + httpd.shutdown() + t.join() + if __name__ == "__main__": try: main() + sys.stdout.flush() + os._exit(0) except Exception as ex: exc_type, exc_value, exc_traceback = sys.exc_info() traceback.print_tb(exc_traceback, file=sys.stderr) diff --git a/tests/queries/0_stateless/02235_add_part_offset_virtual_column.reference b/tests/queries/0_stateless/02235_add_part_offset_virtual_column.reference index 2455f50b7f2..e88abb35ab4 100644 --- a/tests/queries/0_stateless/02235_add_part_offset_virtual_column.reference +++ b/tests/queries/0_stateless/02235_add_part_offset_virtual_column.reference @@ -50,6 +50,8 @@ SOME GRANULES FILTERED OUT 100002 foo PREWHERE 301408 164953047376 164953047376 +335872 166463369216 166463369216 +301407 164952947376 164952947376 42 10042 20042 diff --git a/tests/queries/0_stateless/02235_add_part_offset_virtual_column.sql b/tests/queries/0_stateless/02235_add_part_offset_virtual_column.sql index 73ae6eb499f..5af6565c03d 100644 --- a/tests/queries/0_stateless/02235_add_part_offset_virtual_column.sql +++ b/tests/queries/0_stateless/02235_add_part_offset_virtual_column.sql @@ -52,7 +52,7 @@ SELECT _part_offset, foo FROM t_1 where granule == 0 AND _part_offset >= 100000 SELECT 'PREWHERE'; SELECT count(*), sum(_part_offset), sum(order_0) from t_1 prewhere granule == 0 where _part_offset >= 100000; -SELECT count(*), sum(_part_offset), sum(order_0) from t_1 prewhere _part != '' where granule == 0; -- { serverError 10, 16 } -SELECT count(*), sum(_part_offset), sum(order_0) from t_1 prewhere _part_offset > 100000 where granule == 0; -- { serverError 10, 16 } +SELECT count(*), sum(_part_offset), sum(order_0) from t_1 prewhere _part != '' where granule == 0; +SELECT count(*), sum(_part_offset), sum(order_0) from t_1 prewhere _part_offset > 100000 where granule == 0; SELECT _part_offset FROM t_1 PREWHERE order_0 % 10000 == 42 ORDER BY order_0 LIMIT 3; SELECT _part_offset, foo FROM t_1 PREWHERE order_0 % 10000 == 42 ORDER BY order_0 LIMIT 3; diff --git a/tests/queries/0_stateless/02265_test_dns_profile_events.sh b/tests/queries/0_stateless/02265_test_dns_profile_events.sh index 756a761a0ae..50fa6ba2cda 100755 --- a/tests/queries/0_stateless/02265_test_dns_profile_events.sh +++ b/tests/queries/0_stateless/02265_test_dns_profile_events.sh @@ -1,5 +1,5 @@ #!/usr/bin/env bash -# Tags: no-parallel +# Tags: no-parallel, no-fasttest CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) # shellcheck source=../shell_config.sh diff --git a/tests/queries/0_stateless/02294_system_certificates.reference b/tests/queries/0_stateless/02294_system_certificates.reference index 4655f996c40..432df9110a2 100644 --- a/tests/queries/0_stateless/02294_system_certificates.reference +++ b/tests/queries/0_stateless/02294_system_certificates.reference @@ -1,10 +1,10 @@ -version Int32 -serial_number Nullable(String) -signature_algo Nullable(String) -issuer Nullable(String) -not_before Nullable(String) -not_after Nullable(String) -subject Nullable(String) -pkey_algo Nullable(String) -path String -default UInt8 +version Int32 Version of the certificate. Values are 0 for v1, 1 for v2, 2 for v3. +serial_number Nullable(String) Serial Number of the certificate assigned by the issuer. +signature_algo Nullable(String) Signature Algorithm - an algorithm used by the issuer to sign this certificate. +issuer Nullable(String) Issuer - an unique identifier for the Certificate Authority issuing this certificate. +not_before Nullable(String) The beginning of the time window when this certificate is valid. +not_after Nullable(String) The end of the time window when this certificate is valid. +subject Nullable(String) Subject - identifies the owner of the public key. +pkey_algo Nullable(String) Public Key Algorithm defines the algorithm the public key can be used with. +path String Path to the file or directory containing this certificate. +default UInt8 Certificate is in the default certificate location. diff --git a/tests/queries/0_stateless/02326_settings_changes_system_table.reference b/tests/queries/0_stateless/02326_settings_changes_system_table.reference index 1c8c4fa1880..946b2727d30 100644 --- a/tests/queries/0_stateless/02326_settings_changes_system_table.reference +++ b/tests/queries/0_stateless/02326_settings_changes_system_table.reference @@ -1,3 +1,3 @@ -version String -changes Array(Tuple(\n name String,\n previous_value String,\n new_value String,\n reason String)) +version String The ClickHouse server version. +changes Array(Tuple(\n name String,\n previous_value String,\n new_value String,\n reason String)) The list of changes in settings which changed the behaviour of ClickHouse. 22.5 [('memory_overcommit_ratio_denominator','0','1073741824','Enable memory overcommit feature by default'),('memory_overcommit_ratio_denominator_for_user','0','1073741824','Enable memory overcommit feature by default')] diff --git a/tests/queries/0_stateless/02366_kql_mvexpand.sql b/tests/queries/0_stateless/02366_kql_mvexpand.sql index e7798609646..af336a19638 100644 --- a/tests/queries/0_stateless/02366_kql_mvexpand.sql +++ b/tests/queries/0_stateless/02366_kql_mvexpand.sql @@ -33,3 +33,7 @@ print '-- mv_expand_test_table | mv-expand with_itemindex=index c,d to typeof(bo mv_expand_test_table | mv-expand with_itemindex=index c,d to typeof(bool); print '-- mv_expand_test_table | mv-expand c to typeof(bool) --'; mv_expand_test_table | mv-expand c to typeof(bool); +SET max_query_size = 28; +SET dialect='kusto'; +mv_expand_test_table | mv-expand c, d; -- { serverError SYNTAX_ERROR } +SET max_query_size=262144; diff --git a/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference b/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference index 379eea4dbbb..cd776611857 100644 --- a/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference +++ b/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference @@ -62,6 +62,7 @@ __bitBoolMaskOr __bitSwapLastTwo __bitWrapperFunc __getScalar +__scalarSubqueryResult abs accurateCast accurateCastOrDefault diff --git a/tests/queries/0_stateless/02447_drop_database_replica.reference b/tests/queries/0_stateless/02447_drop_database_replica.reference index 1af3ee244f1..bd3b689ca3c 100644 --- a/tests/queries/0_stateless/02447_drop_database_replica.reference +++ b/tests/queries/0_stateless/02447_drop_database_replica.reference @@ -13,6 +13,7 @@ t rdb_default 1 1 s1 r1 1 2 2 +2 s1 r1 OK 2 0 s1 r2 QUEUED 2 0 s2 r1 QUEUED 2 0 @@ -24,4 +25,5 @@ rdb_default 1 2 s1 r2 0 t t2 t3 +t4 rdb_default_4 1 1 s1 r1 1 diff --git a/tests/queries/0_stateless/02447_drop_database_replica.sh b/tests/queries/0_stateless/02447_drop_database_replica.sh index fb89db5045b..5c4604bc8cd 100755 --- a/tests/queries/0_stateless/02447_drop_database_replica.sh +++ b/tests/queries/0_stateless/02447_drop_database_replica.sh @@ -33,8 +33,9 @@ $CLICKHOUSE_CLIENT -q "select cluster, shard_num, replica_num, database_shard_na $CLICKHOUSE_CLIENT -q "system drop database replica 's1|r1' from database $db2" 2>&1| grep -Fac "is active, cannot drop it" # Also check that it doesn't exceed distributed_ddl_task_timeout waiting for inactive replicas -timeout 60s $CLICKHOUSE_CLIENT --distributed_ddl_task_timeout=1000 --distributed_ddl_output_mode=throw_only_active -q "create table $db.t2 (n int) engine=Log" 2>&1| grep -Fac "TIMEOUT_EXCEEDED" -timeout 60s $CLICKHOUSE_CLIENT --distributed_ddl_task_timeout=1000 --distributed_ddl_output_mode=null_status_on_timeout_only_active -q "create table $db.t3 (n int) engine=Log" | sort +timeout 60s $CLICKHOUSE_CLIENT --distributed_ddl_task_timeout=1000 --distributed_ddl_output_mode=none_only_active -q "create table $db.t2 (n int) engine=Log" 2>&1| grep -Fac "TIMEOUT_EXCEEDED" +timeout 60s $CLICKHOUSE_CLIENT --distributed_ddl_task_timeout=1000 --distributed_ddl_output_mode=throw_only_active -q "create table $db.t3 (n int) engine=Log" 2>&1| grep -Fac "TIMEOUT_EXCEEDED" +timeout 60s $CLICKHOUSE_CLIENT --distributed_ddl_task_timeout=1000 --distributed_ddl_output_mode=null_status_on_timeout_only_active -q "create table $db.t4 (n int) engine=Log" | sort $CLICKHOUSE_CLIENT -q "detach database $db3" $CLICKHOUSE_CLIENT -q "system drop database replica 'r1' from shard 's2' from database $db" diff --git a/tests/queries/0_stateless/02514_analyzer_drop_join_on.reference b/tests/queries/0_stateless/02514_analyzer_drop_join_on.reference index a5a71560d00..2c62e278050 100644 --- a/tests/queries/0_stateless/02514_analyzer_drop_join_on.reference +++ b/tests/queries/0_stateless/02514_analyzer_drop_join_on.reference @@ -55,33 +55,33 @@ Header: a2 String Header: __table1.a2 String __table1.k UInt64 __table4.d2 String - Expression ((Actions for left table alias column keys + DROP unused columns after JOIN)) + Expression (DROP unused columns after JOIN) Header: __table1.a2 String __table1.k UInt64 Join (JOIN FillRightFirst) Header: __table1.a2 String __table1.k UInt64 - Expression ((Actions for left table alias column keys + DROP unused columns after JOIN)) + Expression (DROP unused columns after JOIN) Header: __table1.a2 String __table1.k UInt64 Join (JOIN FillRightFirst) Header: __table1.a2 String __table1.k UInt64 - Expression ((Actions for left table alias column keys + Change column names to column identifiers)) + Expression (Change column names to column identifiers) Header: __table1.a2 String __table1.k UInt64 ReadFromMemoryStorage Header: a2 String k UInt64 - Expression ((Actions for right table alias column keys + Change column names to column identifiers)) + Expression (Change column names to column identifiers) Header: __table2.k UInt64 ReadFromMemoryStorage Header: k UInt64 - Expression ((Actions for right table alias column keys + Change column names to column identifiers)) + Expression (Change column names to column identifiers) Header: __table3.k UInt64 ReadFromMemoryStorage Header: k UInt64 - Expression ((Actions for right table alias column keys + Change column names to column identifiers)) + Expression (Change column names to column identifiers) Header: __table4.d2 String __table4.k UInt64 ReadFromMemoryStorage diff --git a/tests/queries/0_stateless/02534_join_prewhere_bug_44062.reference b/tests/queries/0_stateless/02534_join_prewhere_bug.reference similarity index 88% rename from tests/queries/0_stateless/02534_join_prewhere_bug_44062.reference rename to tests/queries/0_stateless/02534_join_prewhere_bug.reference index aaef17371d8..115ea994de1 100644 --- a/tests/queries/0_stateless/02534_join_prewhere_bug_44062.reference +++ b/tests/queries/0_stateless/02534_join_prewhere_bug.reference @@ -34,5 +34,12 @@ ORDER BY test2.col1 ; 5600 123 123 5601 321 -32 +SELECT col2, col2 + 1 FROM test1 +FULL OUTER JOIN test2 USING (col1) +PREWHERE (col2 * 2) :: UInt8 +; +123 124 +-32 -31 +-30 -29 DROP TABLE IF EXISTS test1; DROP TABLE IF EXISTS test2; diff --git a/tests/queries/0_stateless/02534_join_prewhere_bug_44062.sql b/tests/queries/0_stateless/02534_join_prewhere_bug.sql similarity index 92% rename from tests/queries/0_stateless/02534_join_prewhere_bug_44062.sql rename to tests/queries/0_stateless/02534_join_prewhere_bug.sql index 073f81e4ff3..016c92597ec 100644 --- a/tests/queries/0_stateless/02534_join_prewhere_bug_44062.sql +++ b/tests/queries/0_stateless/02534_join_prewhere_bug.sql @@ -42,5 +42,10 @@ WHERE test2.col1 IS NOT NULL ORDER BY test2.col1 ; +SELECT col2, col2 + 1 FROM test1 +FULL OUTER JOIN test2 USING (col1) +PREWHERE (col2 * 2) :: UInt8 +; + DROP TABLE IF EXISTS test1; DROP TABLE IF EXISTS test2; diff --git a/tests/queries/0_stateless/02675_profile_events_from_query_log_and_client.sh b/tests/queries/0_stateless/02675_profile_events_from_query_log_and_client.sh index 288f1129b53..e346d9893a7 100755 --- a/tests/queries/0_stateless/02675_profile_events_from_query_log_and_client.sh +++ b/tests/queries/0_stateless/02675_profile_events_from_query_log_and_client.sh @@ -9,7 +9,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) echo "INSERT TO S3" $CLICKHOUSE_CLIENT --print-profile-events --profile-events-delay-ms=-1 -nq " INSERT INTO TABLE FUNCTION s3('http://localhost:11111/test/profile_events.csv', 'test', 'testtest', 'CSV', 'number UInt64') SELECT number FROM numbers(1000000) SETTINGS s3_max_single_part_upload_size = 10, s3_truncate_on_insert = 1; -" 2>&1 | grep -o -e '\ \[\ .*\ \]\ S3.*:\ .*\ ' | grep -v 'Microseconds' | sort +" 2>&1 | grep -o -e '\ \[\ .*\ \]\ S3.*:\ .*\ ' | grep -v 'Microseconds' | grep -v 'S3DiskConnections' | grep -v 'S3DiskAddresses' | sort echo "CHECK WITH query_log" $CLICKHOUSE_CLIENT -nq " diff --git a/tests/queries/0_stateless/02722_database_filesystem.reference b/tests/queries/0_stateless/02722_database_filesystem.reference index 17f84cfc49c..2aa7e74befb 100644 --- a/tests/queries/0_stateless/02722_database_filesystem.reference +++ b/tests/queries/0_stateless/02722_database_filesystem.reference @@ -23,4 +23,3 @@ OK OK OK OK -OK diff --git a/tests/queries/0_stateless/02722_database_filesystem.sh b/tests/queries/0_stateless/02722_database_filesystem.sh index f3af7abcbb3..0c75c15fc69 100755 --- a/tests/queries/0_stateless/02722_database_filesystem.sh +++ b/tests/queries/0_stateless/02722_database_filesystem.sh @@ -81,8 +81,6 @@ CREATE DATABASE test2 ENGINE = Filesystem('relative_unknown_dir'); # FILE_DOESNT_EXIST: unknown file ${CLICKHOUSE_CLIENT} --query "SELECT COUNT(*) FROM test1.\`tmp2.csv\`;" 2>&1 | tr '\n' ' ' | grep -oF -e "UNKNOWN_TABLE" -e "FILE_DOESNT_EXIST" > /dev/null && echo "OK" || echo 'FAIL' ||: -# BAD_ARGUMENTS: Cannot determine the file format by it's extension -${CLICKHOUSE_CLIENT} --query "SELECT COUNT(*) FROM test1.\`${unique_name}/tmp.myext\`;" 2>&1 | tr '\n' ' ' | grep -oF -e "UNKNOWN_TABLE" -e "BAD_ARGUMENTS" > /dev/null && echo "OK" || echo 'FAIL' ||: # Clean ${CLICKHOUSE_CLIENT} --query "DROP DATABASE test1;" rm -rd $tmp_dir diff --git a/tests/queries/0_stateless/02785_date_predicate_optimizations_ast_query_tree_rewrite.reference b/tests/queries/0_stateless/02785_date_predicate_optimizations_ast_query_tree_rewrite.reference index 63658890119..fca48238778 100644 --- a/tests/queries/0_stateless/02785_date_predicate_optimizations_ast_query_tree_rewrite.reference +++ b/tests/queries/0_stateless/02785_date_predicate_optimizations_ast_query_tree_rewrite.reference @@ -24,21 +24,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -66,21 +66,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -244,21 +244,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 15, constant_value: \'1998-01-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 14, constant_value: \'1998-01-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -289,34 +289,34 @@ QUERY id: 0 FUNCTION id: 14, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 15, nodes: 2 - COLUMN id: 16, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 17, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 18, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 12, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 16, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 17, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - FUNCTION id: 20, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + FUNCTION id: 19, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 21, nodes: 2 - COLUMN id: 22, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 23, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 24, function_name: less, function_type: ordinary, result_type: UInt8 + LIST id: 20, nodes: 2 + COLUMN id: 12, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 21, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 22, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 25, nodes: 2 - COLUMN id: 26, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 27, constant_value: \'1995-01-01\', constant_value_type: String - FUNCTION id: 28, function_name: and, function_type: ordinary, result_type: UInt8 + LIST id: 23, nodes: 2 + COLUMN id: 12, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 24, constant_value: \'1995-01-01\', constant_value_type: String + FUNCTION id: 25, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 29, nodes: 2 - FUNCTION id: 30, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 26, nodes: 2 + FUNCTION id: 27, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 31, nodes: 2 - COLUMN id: 32, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 33, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 34, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 28, nodes: 2 + COLUMN id: 29, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 30, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 31, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 35, nodes: 2 - COLUMN id: 32, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 36, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 32, nodes: 2 + COLUMN id: 29, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 33, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1, @@ -346,26 +346,26 @@ QUERY id: 0 FUNCTION id: 11, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 12, nodes: 2 - COLUMN id: 13, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 14, constant_value: \'1993-01-01\', constant_value_type: String - FUNCTION id: 15, function_name: less, function_type: ordinary, result_type: UInt8 + COLUMN id: 6, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 13, constant_value: \'1993-01-01\', constant_value_type: String + FUNCTION id: 14, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 16, nodes: 2 - COLUMN id: 17, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 18, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 19, function_name: and, function_type: ordinary, result_type: UInt8 + LIST id: 15, nodes: 2 + COLUMN id: 6, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 16, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 17, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 20, nodes: 2 - FUNCTION id: 21, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + FUNCTION id: 19, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 22, nodes: 2 - COLUMN id: 23, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 25, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 20, nodes: 2 + COLUMN id: 21, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 22, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 23, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 26, nodes: 2 - COLUMN id: 23, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 27, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 24, nodes: 2 + COLUMN id: 21, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 25, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -425,22 +425,22 @@ QUERY id: 0 FUNCTION id: 10, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 11, nodes: 2 - COLUMN id: 12, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 13, constant_value: \'1994-01-01\', constant_value_type: String + COLUMN id: 8, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 12, constant_value: \'1994-01-01\', constant_value_type: String WHERE - FUNCTION id: 14, function_name: and, function_type: ordinary, result_type: UInt8 + FUNCTION id: 13, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 15, nodes: 2 - FUNCTION id: 16, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 14, nodes: 2 + FUNCTION id: 15, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - COLUMN id: 18, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 19, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 20, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + COLUMN id: 17, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 18, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 19, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 21, nodes: 2 - COLUMN id: 18, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 22, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 20, nodes: 2 + COLUMN id: 17, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 21, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -479,8 +479,8 @@ QUERY id: 0 FUNCTION id: 19, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 20, nodes: 2 - COLUMN id: 21, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 22, constant_value: \'1994-01-01\', constant_value_type: String + COLUMN id: 17, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 21, constant_value: \'1994-01-01\', constant_value_type: String SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -582,21 +582,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -624,21 +624,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 15, constant_value: \'1992-04-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 14, constant_value: \'1992-04-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -666,21 +666,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 15, constant_value: \'1992-04-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 14, constant_value: \'1992-04-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date_t @@ -847,26 +847,26 @@ QUERY id: 0 FUNCTION id: 14, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 15, nodes: 2 - COLUMN id: 16, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 17, constant_value: \'1993-01-01\', constant_value_type: String - FUNCTION id: 18, function_name: less, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 16, constant_value: \'1993-01-01\', constant_value_type: String + FUNCTION id: 17, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 21, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 22, function_name: and, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 10, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 19, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 20, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - FUNCTION id: 24, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 21, nodes: 2 + FUNCTION id: 22, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 25, nodes: 2 - COLUMN id: 26, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 27, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 28, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 23, nodes: 2 + COLUMN id: 24, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 25, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 26, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 29, nodes: 2 - COLUMN id: 26, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 30, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 27, nodes: 2 + COLUMN id: 24, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 28, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM datetime_t @@ -894,21 +894,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: DateTime, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: DateTime, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM datetime_t @@ -936,21 +936,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: DateTime, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: DateTime, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date32_t @@ -978,21 +978,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date32, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date32, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM date32_t @@ -1020,21 +1020,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: Date32, source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: Date32, source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM datetime64_t @@ -1062,21 +1062,21 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: DateTime64(3), source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: DateTime64(3), source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 SELECT value1 FROM datetime64_t @@ -1104,19 +1104,19 @@ QUERY id: 0 FUNCTION id: 12, function_name: less, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 13, nodes: 2 - COLUMN id: 14, column_name: date1, result_type: DateTime64(3), source_id: 3 - CONSTANT id: 15, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String - FUNCTION id: 16, function_name: and, function_type: ordinary, result_type: UInt8 + COLUMN id: 10, column_name: date1, result_type: DateTime64(3), source_id: 3 + CONSTANT id: 14, constant_value: \'1994-01-01 00:00:00\', constant_value_type: String + FUNCTION id: 15, function_name: and, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 17, nodes: 2 - FUNCTION id: 18, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 16, nodes: 2 + FUNCTION id: 17, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 21, constant_value: UInt64_1, constant_value_type: UInt8 - FUNCTION id: 22, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 20, constant_value: UInt64_1, constant_value_type: UInt8 + FUNCTION id: 21, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 23, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - CONSTANT id: 24, constant_value: UInt64_3, constant_value_type: UInt8 + LIST id: 22, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + CONSTANT id: 23, constant_value: UInt64_3, constant_value_type: UInt8 SETTINGS allow_experimental_analyzer=1 diff --git a/tests/queries/0_stateless/02789_reading_from_s3_with_connection_pool.sh b/tests/queries/0_stateless/02789_reading_from_s3_with_connection_pool.sh index ce90157d004..e4a1de9a2ec 100755 --- a/tests/queries/0_stateless/02789_reading_from_s3_with_connection_pool.sh +++ b/tests/queries/0_stateless/02789_reading_from_s3_with_connection_pool.sh @@ -19,8 +19,8 @@ query_id=$(${CLICKHOUSE_CLIENT} --query "select queryID() from ($query) limit 1" ${CLICKHOUSE_CLIENT} --query "SYSTEM FLUSH LOGS" ${CLICKHOUSE_CLIENT} -nm --query " WITH - ProfileEvents['ReadBufferFromS3ResetSessions'] AS reset, - ProfileEvents['ReadBufferFromS3PreservedSessions'] AS preserved + ProfileEvents['DiskConnectionsReset'] AS reset, + ProfileEvents['DiskConnectionsPreserved'] AS preserved SELECT preserved > reset FROM system.query_log WHERE type = 'QueryFinish' @@ -51,7 +51,7 @@ select queryID() from( " 2>&1) ${CLICKHOUSE_CLIENT} --query "SYSTEM FLUSH LOGS" ${CLICKHOUSE_CLIENT} -nm --query " -SELECT ProfileEvents['ReadWriteBufferFromHTTPPreservedSessions'] > 0 +SELECT ProfileEvents['StorageConnectionsPreserved'] > 0 FROM system.query_log WHERE type = 'QueryFinish' AND current_database = currentDatabase() diff --git a/tests/queries/0_stateless/02813_seriesDecomposeSTL.sql b/tests/queries/0_stateless/02813_seriesDecomposeSTL.sql index 929d0474e09..496267f2476 100644 --- a/tests/queries/0_stateless/02813_seriesDecomposeSTL.sql +++ b/tests/queries/0_stateless/02813_seriesDecomposeSTL.sql @@ -1,6 +1,3 @@ --- Tags: no-cpu-aarch64 --- Tag no-cpu-aarch64: values generated are slighly different on aarch64 - DROP TABLE IF EXISTS tb2; CREATE TABLE tb2 (`period` UInt32, `ts` Array(Float64)) ENGINE = Memory; diff --git a/tests/queries/0_stateless/02833_url_without_path_encoding.sh b/tests/queries/0_stateless/02833_url_without_path_encoding.sh index b71586099cf..eb845c6b45b 100755 --- a/tests/queries/0_stateless/02833_url_without_path_encoding.sh +++ b/tests/queries/0_stateless/02833_url_without_path_encoding.sh @@ -8,5 +8,5 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) $CLICKHOUSE_CLIENT -q "select count() from url('http://localhost:11111/test%2Fa.tsv') settings enable_url_encoding=1" # Grep 'test%2Fa.tsv' to ensure that path wasn't encoded/decoded -$CLICKHOUSE_CLIENT -q "select count() from url('http://localhost:11111/test%2Fa.tsv') settings enable_url_encoding=0" 2>&1 | grep -o "test%2Fa.tsv" -m1 - +$CLICKHOUSE_CLIENT -q "select count() from url('http://localhost:11111/test%2Fa.tsv') settings enable_url_encoding=0" 2>&1 | \ + grep -o "test%2Fa.tsv" -m1 | head -n 1 diff --git a/tests/queries/0_stateless/02867_page_cache.reference b/tests/queries/0_stateless/02867_page_cache.reference index 5502059508a..c3d6484a175 100644 --- a/tests/queries/0_stateless/02867_page_cache.reference +++ b/tests/queries/0_stateless/02867_page_cache.reference @@ -1,23 +1,21 @@ -54975576145920 +cold read 54975576145920 PageCacheBytesUnpinnedRoundedToHugePages 1 PageCacheBytesUnpinnedRoundedToPages 1 PageCacheChunkMisses 1 ReadBufferFromS3Bytes 1 -54975576145920 +repeat read 1 54975576145920 PageCacheBytesUnpinnedRoundedToHugePages 1 PageCacheBytesUnpinnedRoundedToPages 1 PageCacheChunkDataHits 1 -54975576145920 +dropped and bypassed cache 54975576145920 +PageCacheChunkMisses 1 +ReadBufferFromS3Bytes 1 +repeat read 2 54975576145920 PageCacheBytesUnpinnedRoundedToHugePages 1 PageCacheBytesUnpinnedRoundedToPages 1 PageCacheChunkMisses 1 ReadBufferFromS3Bytes 1 -54975576145920 -PageCacheBytesUnpinnedRoundedToHugePages 1 -PageCacheBytesUnpinnedRoundedToPages 1 -PageCacheChunkMisses 1 -ReadBufferFromS3Bytes 1 -54975576145920 +repeat read 3 54975576145920 PageCacheBytesUnpinnedRoundedToHugePages 1 PageCacheBytesUnpinnedRoundedToPages 1 PageCacheChunkDataHits 1 diff --git a/tests/queries/0_stateless/02867_page_cache.sql b/tests/queries/0_stateless/02867_page_cache.sql index 8765b30ebc3..f1882de4af6 100644 --- a/tests/queries/0_stateless/02867_page_cache.sql +++ b/tests/queries/0_stateless/02867_page_cache.sql @@ -46,7 +46,7 @@ insert into events_snapshot select * from system.events; -- Cold read, should miss cache. (Populating cache on write is not implemented yet.) -select sum(k) from page_cache_03055; +select 'cold read', sum(k) from page_cache_03055; select * from events_diff where event not in ('PageCacheChunkDataHits'); truncate table events_snapshot; @@ -54,7 +54,7 @@ insert into events_snapshot select * from system.events; -- Repeat read, should hit cache. -select sum(k) from page_cache_03055; +select 'repeat read 1', sum(k) from page_cache_03055; select * from events_diff; truncate table events_snapshot; @@ -64,16 +64,17 @@ insert into events_snapshot select * from system.events; system drop page cache; -select sum(k) from page_cache_03055 settings read_from_page_cache_if_exists_otherwise_bypass_cache = 1; +select 'dropped and bypassed cache', sum(k) from page_cache_03055 settings read_from_page_cache_if_exists_otherwise_bypass_cache = 1; -- Data could be read multiple times because we're not writing to cache. -select event, if(event in ('PageCacheChunkMisses', 'ReadBufferFromS3Bytes'), diff >= 1, diff) from events_diff where event not in ('PageCacheChunkDataHits'); +-- (Not checking PageCacheBytesUnpinned* because it's unreliable in this case because of an intentional race condition, see PageCache::evictChunk.) +select event, if(event in ('PageCacheChunkMisses', 'ReadBufferFromS3Bytes'), diff >= 1, diff) from events_diff where event not in ('PageCacheChunkDataHits', 'PageCacheBytesUnpinnedRoundedToPages', 'PageCacheBytesUnpinnedRoundedToHugePages'); truncate table events_snapshot; insert into events_snapshot select * from system.events; -- Repeat read, should still miss, but populate cache. -select sum(k) from page_cache_03055; +select 'repeat read 2', sum(k) from page_cache_03055; select * from events_diff where event not in ('PageCacheChunkDataHits'); truncate table events_snapshot; @@ -81,7 +82,7 @@ insert into events_snapshot select * from system.events; -- Read again, hit the cache. -select sum(k) from page_cache_03055 settings read_from_page_cache_if_exists_otherwise_bypass_cache = 1; +select 'repeat read 3', sum(k) from page_cache_03055 settings read_from_page_cache_if_exists_otherwise_bypass_cache = 1; select * from events_diff; truncate table events_snapshot; diff --git a/tests/queries/0_stateless/02890_describe_table_options.reference b/tests/queries/0_stateless/02890_describe_table_options.reference index 5d99df36bb4..ff58202ae49 100644 --- a/tests/queries/0_stateless/02890_describe_table_options.reference +++ b/tests/queries/0_stateless/02890_describe_table_options.reference @@ -2,237 +2,197 @@ SET describe_compact_output = 0, describe_include_virtual_columns = 0, describe_include_subcolumns = 0; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name─┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┠-│ id │ UInt64 │ │ │ index column │ │ │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ -└──────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┘ +┌─name─┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┠+│ id │ UInt64 │ │ │ index column │ │ │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ +└──────┴───────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name─┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┠-│ id │ UInt64 │ │ │ index column │ │ │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ -└──────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┘ +┌─name─┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┠+│ id │ UInt64 │ │ │ index column │ │ │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ +└──────┴───────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┘ SET describe_compact_output = 0, describe_include_virtual_columns = 0, describe_include_subcolumns = 1; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name──────┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┠-│ id │ UInt64 │ │ │ index column │ │ │ 0 │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ -│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ -│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ -│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ -└───────────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴──────────────┘ +┌─name──────┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┠+│ id │ UInt64 │ │ │ index column │ │ │ 0 │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ +│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ +│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ +│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ +└───────────┴───────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴──────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name──────┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┠-│ id │ UInt64 │ │ │ index column │ │ │ 0 │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ -│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ -│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ -│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ -└───────────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴──────────────┘ +┌─name──────┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┠+│ id │ UInt64 │ │ │ index column │ │ │ 0 │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ +│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ +│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ +│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ +└───────────┴───────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴──────────────┘ SET describe_compact_output = 0, describe_include_virtual_columns = 1, describe_include_subcolumns = 0; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name─────────────┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_virtual─┠-│ id │ UInt64 │ │ │ index column │ │ │ 0 │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ -│ _part │ LowCardinality(String) │ │ │ │ │ │ 1 │ -│ _part_index │ UInt64 │ │ │ │ │ │ 1 │ -│ _part_uuid │ UUID │ │ │ │ │ │ 1 │ -│ _partition_id │ LowCardinality(String) │ │ │ │ │ │ 1 │ -│ _partition_value │ UInt8 │ │ │ │ │ │ 1 │ -│ _sample_factor │ Float64 │ │ │ │ │ │ 1 │ -│ _part_offset │ UInt64 │ │ │ │ │ │ 1 │ -│ _row_exists │ UInt8 │ │ │ │ │ │ 1 │ -│ _block_number │ UInt64 │ │ │ │ │ │ 1 │ -└──────────────────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment─────────────────────────────────────────────────────────────────────────────────┬─codec_expression─┬─ttl_expression─┬─is_virtual─┠+│ id │ UInt64 │ │ │ index column │ │ │ 0 │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ +│ _part │ LowCardinality(String) │ │ │ Name of part │ │ │ 1 │ +│ _part_index │ UInt64 │ │ │ Sequential index of the part in the query result │ │ │ 1 │ +│ _part_uuid │ UUID │ │ │ Unique part identifier (if enabled MergeTree setting assign_part_uuids) │ │ │ 1 │ +│ _partition_id │ LowCardinality(String) │ │ │ Name of partition │ │ │ 1 │ +│ _sample_factor │ Float64 │ │ │ Sample factor (from the query) │ │ │ 1 │ +│ _part_offset │ UInt64 │ │ │ Number of row in the part │ │ │ 1 │ +│ _row_exists │ UInt8 │ │ │ Persisted mask created by lightweight delete that show whether row exists or is deleted │ │ │ 1 │ +│ _block_number │ UInt64 │ │ │ Persisted original number of block that was assigned at insert │ Delta, LZ4 │ │ 1 │ +└────────────────┴───────────────────────────┴──────────────┴────────────────────┴─────────────────────────────────────────────────────────────────────────────────────────┴──────────────────┴────────────────┴────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name───────────┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_virtual─┠-│ id │ UInt64 │ │ │ index column │ │ │ 0 │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ -│ _table │ LowCardinality(String) │ │ │ │ │ │ 1 │ -│ _part │ LowCardinality(String) │ │ │ │ │ │ 1 │ -│ _part_index │ UInt64 │ │ │ │ │ │ 1 │ -│ _part_uuid │ UUID │ │ │ │ │ │ 1 │ -│ _partition_id │ LowCardinality(String) │ │ │ │ │ │ 1 │ -│ _sample_factor │ Float64 │ │ │ │ │ │ 1 │ -│ _part_offset │ UInt64 │ │ │ │ │ │ 1 │ -│ _row_exists │ UInt8 │ │ │ │ │ │ 1 │ -│ _block_number │ UInt64 │ │ │ │ │ │ 1 │ -│ _shard_num │ UInt32 │ │ │ │ │ │ 1 │ -└────────────────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment─────────────────────────────────────────────────────────────────────────────────┬─codec_expression─┬─ttl_expression─┬─is_virtual─┠+│ id │ UInt64 │ │ │ index column │ │ │ 0 │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ +│ _part │ LowCardinality(String) │ │ │ Name of part │ │ │ 1 │ +│ _part_index │ UInt64 │ │ │ Sequential index of the part in the query result │ │ │ 1 │ +│ _part_uuid │ UUID │ │ │ Unique part identifier (if enabled MergeTree setting assign_part_uuids) │ │ │ 1 │ +│ _partition_id │ LowCardinality(String) │ │ │ Name of partition │ │ │ 1 │ +│ _sample_factor │ Float64 │ │ │ Sample factor (from the query) │ │ │ 1 │ +│ _part_offset │ UInt64 │ │ │ Number of row in the part │ │ │ 1 │ +│ _row_exists │ UInt8 │ │ │ Persisted mask created by lightweight delete that show whether row exists or is deleted │ │ │ 1 │ +│ _block_number │ UInt64 │ │ │ Persisted original number of block that was assigned at insert │ Delta, LZ4 │ │ 1 │ +│ _shard_num │ UInt32 │ │ │ Deprecated. Use function shardNum instead │ │ │ 1 │ +└────────────────┴───────────────────────────┴──────────────┴────────────────────┴─────────────────────────────────────────────────────────────────────────────────────────┴──────────────────┴────────────────┴────────────┘ SET describe_compact_output = 0, describe_include_virtual_columns = 1, describe_include_subcolumns = 1; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name─────────────┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┬─is_virtual─┠-│ id │ UInt64 │ │ │ index column │ │ │ 0 │ 0 │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ 0 │ -│ _part │ LowCardinality(String) │ │ │ │ │ │ 0 │ 1 │ -│ _part_index │ UInt64 │ │ │ │ │ │ 0 │ 1 │ -│ _part_uuid │ UUID │ │ │ │ │ │ 0 │ 1 │ -│ _partition_id │ LowCardinality(String) │ │ │ │ │ │ 0 │ 1 │ -│ _partition_value │ UInt8 │ │ │ │ │ │ 0 │ 1 │ -│ _sample_factor │ Float64 │ │ │ │ │ │ 0 │ 1 │ -│ _part_offset │ UInt64 │ │ │ │ │ │ 0 │ 1 │ -│ _row_exists │ UInt8 │ │ │ │ │ │ 0 │ 1 │ -│ _block_number │ UInt64 │ │ │ │ │ │ 0 │ 1 │ -│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ 0 │ -│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ -│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ -└──────────────────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴──────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment─────────────────────────────────────────────────────────────────────────────────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┬─is_virtual─┠+│ id │ UInt64 │ │ │ index column │ │ │ 0 │ 0 │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ 0 │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ 0 │ +│ _part │ LowCardinality(String) │ │ │ Name of part │ │ │ 0 │ 1 │ +│ _part_index │ UInt64 │ │ │ Sequential index of the part in the query result │ │ │ 0 │ 1 │ +│ _part_uuid │ UUID │ │ │ Unique part identifier (if enabled MergeTree setting assign_part_uuids) │ │ │ 0 │ 1 │ +│ _partition_id │ LowCardinality(String) │ │ │ Name of partition │ │ │ 0 │ 1 │ +│ _sample_factor │ Float64 │ │ │ Sample factor (from the query) │ │ │ 0 │ 1 │ +│ _part_offset │ UInt64 │ │ │ Number of row in the part │ │ │ 0 │ 1 │ +│ _row_exists │ UInt8 │ │ │ Persisted mask created by lightweight delete that show whether row exists or is deleted │ │ │ 0 │ 1 │ +│ _block_number │ UInt64 │ │ │ Persisted original number of block that was assigned at insert │ Delta, LZ4 │ │ 0 │ 1 │ +│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ 0 │ +│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ +│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ +└────────────────┴───────────────────────────┴──────────────┴────────────────────┴─────────────────────────────────────────────────────────────────────────────────────────┴──────────────────┴────────────────┴──────────────┴────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name───────────┬─type─────────────────────────────┬─default_type─┬─default_expression─┬─comment──────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┬─is_virtual─┠-│ id │ UInt64 │ │ │ index column │ │ │ 0 │ 0 │ -│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ 0 │ -│ _table │ LowCardinality(String) │ │ │ │ │ │ 0 │ 1 │ -│ _part │ LowCardinality(String) │ │ │ │ │ │ 0 │ 1 │ -│ _part_index │ UInt64 │ │ │ │ │ │ 0 │ 1 │ -│ _part_uuid │ UUID │ │ │ │ │ │ 0 │ 1 │ -│ _partition_id │ LowCardinality(String) │ │ │ │ │ │ 0 │ 1 │ -│ _sample_factor │ Float64 │ │ │ │ │ │ 0 │ 1 │ -│ _part_offset │ UInt64 │ │ │ │ │ │ 0 │ 1 │ -│ _row_exists │ UInt8 │ │ │ │ │ │ 0 │ 1 │ -│ _block_number │ UInt64 │ │ │ │ │ │ 0 │ 1 │ -│ _shard_num │ UInt32 │ │ │ │ │ │ 0 │ 1 │ -│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ 0 │ -│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ -│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ -└────────────────┴──────────────────────────────────┴──────────────┴────────────────────┴──────────────┴──────────────────┴────────────────┴──────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─default_type─┬─default_expression─┬─comment─────────────────────────────────────────────────────────────────────────────────┬─codec_expression─┬─ttl_expression─┬─is_subcolumn─┬─is_virtual─┠+│ id │ UInt64 │ │ │ index column │ │ │ 0 │ 0 │ +│ arr │ Array(UInt64) │ DEFAULT │ [10, 20] │ │ ZSTD(1) │ │ 0 │ 0 │ +│ t │ Tuple(a String, b UInt64) │ DEFAULT │ ('foo', 0) │ │ ZSTD(1) │ │ 0 │ 0 │ +│ _part │ LowCardinality(String) │ │ │ Name of part │ │ │ 0 │ 1 │ +│ _part_index │ UInt64 │ │ │ Sequential index of the part in the query result │ │ │ 0 │ 1 │ +│ _part_uuid │ UUID │ │ │ Unique part identifier (if enabled MergeTree setting assign_part_uuids) │ │ │ 0 │ 1 │ +│ _partition_id │ LowCardinality(String) │ │ │ Name of partition │ │ │ 0 │ 1 │ +│ _sample_factor │ Float64 │ │ │ Sample factor (from the query) │ │ │ 0 │ 1 │ +│ _part_offset │ UInt64 │ │ │ Number of row in the part │ │ │ 0 │ 1 │ +│ _row_exists │ UInt8 │ │ │ Persisted mask created by lightweight delete that show whether row exists or is deleted │ │ │ 0 │ 1 │ +│ _block_number │ UInt64 │ │ │ Persisted original number of block that was assigned at insert │ Delta, LZ4 │ │ 0 │ 1 │ +│ _shard_num │ UInt32 │ │ │ Deprecated. Use function shardNum instead │ │ │ 0 │ 1 │ +│ arr.size0 │ UInt64 │ │ │ │ │ │ 1 │ 0 │ +│ t.a │ String │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ +│ t.b │ UInt64 │ │ │ │ ZSTD(1) │ │ 1 │ 0 │ +└────────────────┴───────────────────────────┴──────────────┴────────────────────┴─────────────────────────────────────────────────────────────────────────────────────────┴──────────────────┴────────────────┴──────────────┴────────────┘ SET describe_compact_output = 1, describe_include_virtual_columns = 0, describe_include_subcolumns = 0; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name─┬─type─────────────────────────────┠-│ id │ UInt64 │ -│ arr │ Array(UInt64) │ -│ t │ Tuple( - a String, - b UInt64) │ -└──────┴──────────────────────────────────┘ +┌─name─┬─type──────────────────────┠+│ id │ UInt64 │ +│ arr │ Array(UInt64) │ +│ t │ Tuple(a String, b UInt64) │ +└──────┴───────────────────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name─┬─type─────────────────────────────┠-│ id │ UInt64 │ -│ arr │ Array(UInt64) │ -│ t │ Tuple( - a String, - b UInt64) │ -└──────┴──────────────────────────────────┘ +┌─name─┬─type──────────────────────┠+│ id │ UInt64 │ +│ arr │ Array(UInt64) │ +│ t │ Tuple(a String, b UInt64) │ +└──────┴───────────────────────────┘ SET describe_compact_output = 1, describe_include_virtual_columns = 0, describe_include_subcolumns = 1; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name──────┬─type─────────────────────────────┬─is_subcolumn─┠-│ id │ UInt64 │ 0 │ -│ arr │ Array(UInt64) │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ 0 │ -│ arr.size0 │ UInt64 │ 1 │ -│ t.a │ String │ 1 │ -│ t.b │ UInt64 │ 1 │ -└───────────┴──────────────────────────────────┴──────────────┘ +┌─name──────┬─type──────────────────────┬─is_subcolumn─┠+│ id │ UInt64 │ 0 │ +│ arr │ Array(UInt64) │ 0 │ +│ t │ Tuple(a String, b UInt64) │ 0 │ +│ arr.size0 │ UInt64 │ 1 │ +│ t.a │ String │ 1 │ +│ t.b │ UInt64 │ 1 │ +└───────────┴───────────────────────────┴──────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name──────┬─type─────────────────────────────┬─is_subcolumn─┠-│ id │ UInt64 │ 0 │ -│ arr │ Array(UInt64) │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ 0 │ -│ arr.size0 │ UInt64 │ 1 │ -│ t.a │ String │ 1 │ -│ t.b │ UInt64 │ 1 │ -└───────────┴──────────────────────────────────┴──────────────┘ +┌─name──────┬─type──────────────────────┬─is_subcolumn─┠+│ id │ UInt64 │ 0 │ +│ arr │ Array(UInt64) │ 0 │ +│ t │ Tuple(a String, b UInt64) │ 0 │ +│ arr.size0 │ UInt64 │ 1 │ +│ t.a │ String │ 1 │ +│ t.b │ UInt64 │ 1 │ +└───────────┴───────────────────────────┴──────────────┘ SET describe_compact_output = 1, describe_include_virtual_columns = 1, describe_include_subcolumns = 0; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name─────────────┬─type─────────────────────────────┬─is_virtual─┠-│ id │ UInt64 │ 0 │ -│ arr │ Array(UInt64) │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ 0 │ -│ _part │ LowCardinality(String) │ 1 │ -│ _part_index │ UInt64 │ 1 │ -│ _part_uuid │ UUID │ 1 │ -│ _partition_id │ LowCardinality(String) │ 1 │ -│ _partition_value │ UInt8 │ 1 │ -│ _sample_factor │ Float64 │ 1 │ -│ _part_offset │ UInt64 │ 1 │ -│ _row_exists │ UInt8 │ 1 │ -│ _block_number │ UInt64 │ 1 │ -└──────────────────┴──────────────────────────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─is_virtual─┠+│ id │ UInt64 │ 0 │ +│ arr │ Array(UInt64) │ 0 │ +│ t │ Tuple(a String, b UInt64) │ 0 │ +│ _part │ LowCardinality(String) │ 1 │ +│ _part_index │ UInt64 │ 1 │ +│ _part_uuid │ UUID │ 1 │ +│ _partition_id │ LowCardinality(String) │ 1 │ +│ _sample_factor │ Float64 │ 1 │ +│ _part_offset │ UInt64 │ 1 │ +│ _row_exists │ UInt8 │ 1 │ +│ _block_number │ UInt64 │ 1 │ +└────────────────┴───────────────────────────┴────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name───────────┬─type─────────────────────────────┬─is_virtual─┠-│ id │ UInt64 │ 0 │ -│ arr │ Array(UInt64) │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ 0 │ -│ _table │ LowCardinality(String) │ 1 │ -│ _part │ LowCardinality(String) │ 1 │ -│ _part_index │ UInt64 │ 1 │ -│ _part_uuid │ UUID │ 1 │ -│ _partition_id │ LowCardinality(String) │ 1 │ -│ _sample_factor │ Float64 │ 1 │ -│ _part_offset │ UInt64 │ 1 │ -│ _row_exists │ UInt8 │ 1 │ -│ _block_number │ UInt64 │ 1 │ -│ _shard_num │ UInt32 │ 1 │ -└────────────────┴──────────────────────────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─is_virtual─┠+│ id │ UInt64 │ 0 │ +│ arr │ Array(UInt64) │ 0 │ +│ t │ Tuple(a String, b UInt64) │ 0 │ +│ _part │ LowCardinality(String) │ 1 │ +│ _part_index │ UInt64 │ 1 │ +│ _part_uuid │ UUID │ 1 │ +│ _partition_id │ LowCardinality(String) │ 1 │ +│ _sample_factor │ Float64 │ 1 │ +│ _part_offset │ UInt64 │ 1 │ +│ _row_exists │ UInt8 │ 1 │ +│ _block_number │ UInt64 │ 1 │ +│ _shard_num │ UInt32 │ 1 │ +└────────────────┴───────────────────────────┴────────────┘ SET describe_compact_output = 1, describe_include_virtual_columns = 1, describe_include_subcolumns = 1; DESCRIBE TABLE t_describe_options FORMAT PrettyCompactNoEscapes; -┌─name─────────────┬─type─────────────────────────────┬─is_subcolumn─┬─is_virtual─┠-│ id │ UInt64 │ 0 │ 0 │ -│ arr │ Array(UInt64) │ 0 │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ 0 │ 0 │ -│ _part │ LowCardinality(String) │ 0 │ 1 │ -│ _part_index │ UInt64 │ 0 │ 1 │ -│ _part_uuid │ UUID │ 0 │ 1 │ -│ _partition_id │ LowCardinality(String) │ 0 │ 1 │ -│ _partition_value │ UInt8 │ 0 │ 1 │ -│ _sample_factor │ Float64 │ 0 │ 1 │ -│ _part_offset │ UInt64 │ 0 │ 1 │ -│ _row_exists │ UInt8 │ 0 │ 1 │ -│ _block_number │ UInt64 │ 0 │ 1 │ -│ arr.size0 │ UInt64 │ 1 │ 0 │ -│ t.a │ String │ 1 │ 0 │ -│ t.b │ UInt64 │ 1 │ 0 │ -└──────────────────┴──────────────────────────────────┴──────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─is_subcolumn─┬─is_virtual─┠+│ id │ UInt64 │ 0 │ 0 │ +│ arr │ Array(UInt64) │ 0 │ 0 │ +│ t │ Tuple(a String, b UInt64) │ 0 │ 0 │ +│ _part │ LowCardinality(String) │ 0 │ 1 │ +│ _part_index │ UInt64 │ 0 │ 1 │ +│ _part_uuid │ UUID │ 0 │ 1 │ +│ _partition_id │ LowCardinality(String) │ 0 │ 1 │ +│ _sample_factor │ Float64 │ 0 │ 1 │ +│ _part_offset │ UInt64 │ 0 │ 1 │ +│ _row_exists │ UInt8 │ 0 │ 1 │ +│ _block_number │ UInt64 │ 0 │ 1 │ +│ arr.size0 │ UInt64 │ 1 │ 0 │ +│ t.a │ String │ 1 │ 0 │ +│ t.b │ UInt64 │ 1 │ 0 │ +└────────────────┴───────────────────────────┴──────────────┴────────────┘ DESCRIBE remote(default, currentDatabase(), t_describe_options) FORMAT PrettyCompactNoEscapes; -┌─name───────────┬─type─────────────────────────────┬─is_subcolumn─┬─is_virtual─┠-│ id │ UInt64 │ 0 │ 0 │ -│ arr │ Array(UInt64) │ 0 │ 0 │ -│ t │ Tuple( - a String, - b UInt64) │ 0 │ 0 │ -│ _table │ LowCardinality(String) │ 0 │ 1 │ -│ _part │ LowCardinality(String) │ 0 │ 1 │ -│ _part_index │ UInt64 │ 0 │ 1 │ -│ _part_uuid │ UUID │ 0 │ 1 │ -│ _partition_id │ LowCardinality(String) │ 0 │ 1 │ -│ _sample_factor │ Float64 │ 0 │ 1 │ -│ _part_offset │ UInt64 │ 0 │ 1 │ -│ _row_exists │ UInt8 │ 0 │ 1 │ -│ _block_number │ UInt64 │ 0 │ 1 │ -│ _shard_num │ UInt32 │ 0 │ 1 │ -│ arr.size0 │ UInt64 │ 1 │ 0 │ -│ t.a │ String │ 1 │ 0 │ -│ t.b │ UInt64 │ 1 │ 0 │ -└────────────────┴──────────────────────────────────┴──────────────┴────────────┘ +┌─name───────────┬─type──────────────────────┬─is_subcolumn─┬─is_virtual─┠+│ id │ UInt64 │ 0 │ 0 │ +│ arr │ Array(UInt64) │ 0 │ 0 │ +│ t │ Tuple(a String, b UInt64) │ 0 │ 0 │ +│ _part │ LowCardinality(String) │ 0 │ 1 │ +│ _part_index │ UInt64 │ 0 │ 1 │ +│ _part_uuid │ UUID │ 0 │ 1 │ +│ _partition_id │ LowCardinality(String) │ 0 │ 1 │ +│ _sample_factor │ Float64 │ 0 │ 1 │ +│ _part_offset │ UInt64 │ 0 │ 1 │ +│ _row_exists │ UInt8 │ 0 │ 1 │ +│ _block_number │ UInt64 │ 0 │ 1 │ +│ _shard_num │ UInt32 │ 0 │ 1 │ +│ arr.size0 │ UInt64 │ 1 │ 0 │ +│ t.a │ String │ 1 │ 0 │ +│ t.b │ UInt64 │ 1 │ 0 │ +└────────────────┴───────────────────────────┴──────────────┴────────────┘ diff --git a/tests/queries/0_stateless/02890_describe_table_options.sql b/tests/queries/0_stateless/02890_describe_table_options.sql index 236100148ff..63806c7ee3d 100644 --- a/tests/queries/0_stateless/02890_describe_table_options.sql +++ b/tests/queries/0_stateless/02890_describe_table_options.sql @@ -1,5 +1,7 @@ DROP TABLE IF EXISTS t_describe_options; +SET print_pretty_type_names = 0; + CREATE TABLE t_describe_options ( id UInt64 COMMENT 'index column', arr Array(UInt64) DEFAULT [10, 20] CODEC(ZSTD), diff --git a/tests/queries/0_stateless/02908_many_requests_to_system_replicas.reference b/tests/queries/0_stateless/02908_many_requests_to_system_replicas.reference index af0e50ec332..f1ca07ef408 100644 --- a/tests/queries/0_stateless/02908_many_requests_to_system_replicas.reference +++ b/tests/queries/0_stateless/02908_many_requests_to_system_replicas.reference @@ -1,5 +1,5 @@ Creating 300 tables -Making making 200 requests to system.replicas +Making 200 requests to system.replicas Query system.replicas while waiting for other concurrent requests to finish 0 900 diff --git a/tests/queries/0_stateless/02908_many_requests_to_system_replicas.sh b/tests/queries/0_stateless/02908_many_requests_to_system_replicas.sh index f93175529c0..d3eed891ab9 100755 --- a/tests/queries/0_stateless/02908_many_requests_to_system_replicas.sh +++ b/tests/queries/0_stateless/02908_many_requests_to_system_replicas.sh @@ -33,7 +33,7 @@ done wait; -echo "Making making $CONCURRENCY requests to system.replicas" +echo "Making $CONCURRENCY requests to system.replicas" for i in `seq 1 $CONCURRENCY`; do diff --git a/tests/queries/0_stateless/02911_support_alias_column_in_indices.reference b/tests/queries/0_stateless/02911_support_alias_column_in_indices.reference index 883966ce6b5..b867a31dcc3 100644 --- a/tests/queries/0_stateless/02911_support_alias_column_in_indices.reference +++ b/tests/queries/0_stateless/02911_support_alias_column_in_indices.reference @@ -14,13 +14,13 @@ Expression ((Projection + Before ORDER BY)) Parts: 1/1 Granules: 1/1 Expression ((Project names + Projection)) - Filter ((WHERE + Change column names to column identifiers)) + Filter ((WHERE + (Change column names to column identifiers + Compute alias columns))) ReadFromMergeTree (02911_support_alias_column_in_indices.test1) Indexes: PrimaryKey Keys: c - Condition: (_CAST(plus(c, \'UInt64\'), 1) in [11, +Inf)) + Condition: (plus(c, 1) in [11, +Inf)) Parts: 1/2 Granules: 1/2 Skip @@ -44,12 +44,17 @@ Expression ((Projection + Before ORDER BY)) Parts: 1/1 Granules: 1/1 Expression ((Project names + Projection)) - Filter ((WHERE + Change column names to column identifiers)) + Filter ((WHERE + (Change column names to column identifiers + Compute alias columns))) ReadFromMergeTree (02911_support_alias_column_in_indices.test2) Indexes: PrimaryKey Keys: c - Condition: (_CAST(plus(_CAST(plus(c, \'UInt64\'), 1), \'UInt64\'), 1) in [16, +Inf)) + Condition: (plus(plus(c, 1), 1) in [16, +Inf)) Parts: 1/2 Granules: 1/2 + Skip + Name: i + Description: minmax GRANULARITY 1 + Parts: 1/1 + Granules: 1/1 diff --git a/tests/queries/0_stateless/02922_deduplication_with_zero_copy.reference b/tests/queries/0_stateless/02922_deduplication_with_zero_copy.reference index e69de29bb2d..29376f82128 100644 --- a/tests/queries/0_stateless/02922_deduplication_with_zero_copy.reference +++ b/tests/queries/0_stateless/02922_deduplication_with_zero_copy.reference @@ -0,0 +1,6 @@ +zero_copy: + +tables: +table +zero_copy: +tables: diff --git a/tests/queries/0_stateless/02922_deduplication_with_zero_copy.sh b/tests/queries/0_stateless/02922_deduplication_with_zero_copy.sh index e9deed2a7cd..bb013dccb65 100755 --- a/tests/queries/0_stateless/02922_deduplication_with_zero_copy.sh +++ b/tests/queries/0_stateless/02922_deduplication_with_zero_copy.sh @@ -3,7 +3,7 @@ set -e -CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +CURDIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" # shellcheck source=../shell_config.sh . "$CURDIR"/../shell_config.sh @@ -27,6 +27,7 @@ create table r2 (n int) allow_remote_fs_zero_copy_replication=1; " + function get_shared_locks() { table_shared_id="$1" @@ -42,11 +43,12 @@ function get_shared_locks() done } + function filter_temporary_locks() { while read -r lock do - owner=$($CLICKHOUSE_KEEPER_CLIENT -q "get_stat ${lock}" | grep 'ephemeralOwner' | sed 's/.*= //') + owner="$($CLICKHOUSE_KEEPER_CLIENT -q "get_stat ${lock}" | grep 'ephemeralOwner' | sed 's/.*= //')" if [[ "${owner}" -eq "0" ]] then echo "${lock}" @@ -63,10 +65,11 @@ function insert_duplicates() { wait $CLICKHOUSE_CLIENT -nm -q " +system sync replica r1; system sync replica r2; " - count=$($CLICKHOUSE_CLIENT -q "select count() from r2;") + count="$($CLICKHOUSE_CLIENT -q "select count() from r2;")" [[ "${count}" -eq "1" ]] } @@ -75,7 +78,7 @@ function loop() { set -e - table_shared_id=$($CLICKHOUSE_KEEPER_CLIENT -q "get /test/02922/${CLICKHOUSE_DATABASE}/table/table_shared_id") + table_shared_id="$1" while : do @@ -89,8 +92,8 @@ system sync replica r2; " done - persistent_locks="$(get_shared_locks ${table_shared_id} | filter_temporary_locks)" - num=$(echo "${persistent_locks}" | wc -w) + persistent_locks="$(get_shared_locks "${table_shared_id}" | filter_temporary_locks)" + num="$(echo "${persistent_locks}" | wc -w)" if [[ "${num}" -ne "2" ]] then @@ -101,24 +104,41 @@ system sync replica r2; } + export -f query_with_retry export -f filter_temporary_locks export -f insert_duplicates export -f get_shared_locks export -f loop +table_shared_id="$($CLICKHOUSE_KEEPER_CLIENT -q "get /test/02922/${CLICKHOUSE_DATABASE}/table/table_shared_id")" exit_code=0 -timeout 60 bash -c loop || exit_code="${?}" +timeout 40 bash -c "loop '${table_shared_id}'" || exit_code="${?}" if [[ "${exit_code}" -ne "124" ]] then echo "timeout expected, but loop exited with code: ${exit_code}." echo "the error is found if loop ends with 0." + echo "table_shared_id=${table_shared_id}" exit 1 fi -$CLICKHOUSE_CLIENT -nm -q " -drop table r1; -drop table r2; -" +function list_keeper_nodes() { + table_shared_id=$1 + + echo "zero_copy:" + $CLICKHOUSE_KEEPER_CLIENT -q "ls /clickhouse/zero_copy/zero_copy_s3" | grep -o "${table_shared_id}" | \ + sed "s/${table_shared_id}//g" || : + + echo "tables:" + $CLICKHOUSE_KEEPER_CLIENT -q "ls /test/02922/${CLICKHOUSE_DATABASE}" | grep -o "table" || : +} + +list_keeper_nodes "${table_shared_id}" + +$CLICKHOUSE_CLIENT -nm -q "drop table r1;" --send_logs_level="error" & +$CLICKHOUSE_CLIENT -nm -q "drop table r2;" --send_logs_level="error" & +wait + +list_keeper_nodes "${table_shared_id}" diff --git a/tests/queries/0_stateless/02932_refreshable_materialized_views.reference b/tests/queries/0_stateless/02932_refreshable_materialized_views.reference index b52d0847ff9..aa76806da9d 100644 --- a/tests/queries/0_stateless/02932_refreshable_materialized_views.reference +++ b/tests/queries/0_stateless/02932_refreshable_materialized_views.reference @@ -1,8 +1,8 @@ <1: created view> a [] 1 -CREATE MATERIALIZED VIEW default.a\nREFRESH AFTER 1 SECOND\n(\n `x` UInt64\n)\nENGINE = Memory\nAS SELECT number AS x\nFROM numbers(2)\nUNION ALL\nSELECT rand64() AS x +CREATE MATERIALIZED VIEW default.a\nREFRESH AFTER 2 SECOND\n(\n `x` UInt64\n)\nENGINE = Memory\nAS SELECT number AS x\nFROM numbers(2)\nUNION ALL\nSELECT rand64() AS x <2: refreshed> 3 1 1 -<3: time difference at least> 500 -<4: next refresh in> 1 +<3: time difference at least> 1000 +<4: next refresh in> 2 <4.5: altered> Scheduled Finished 2052-01-01 00:00:00 CREATE MATERIALIZED VIEW default.a\nREFRESH EVERY 2 YEAR\n(\n `x` Int16\n)\nENGINE = Memory\nAS SELECT x * 2 AS x\nFROM default.src <5: no refresh> 3 diff --git a/tests/queries/0_stateless/02932_refreshable_materialized_views.sh b/tests/queries/0_stateless/02932_refreshable_materialized_views.sh index 8daea063fc5..89942e25b67 100755 --- a/tests/queries/0_stateless/02932_refreshable_materialized_views.sh +++ b/tests/queries/0_stateless/02932_refreshable_materialized_views.sh @@ -18,7 +18,7 @@ $CLICKHOUSE_CLIENT -nq "create view refreshes as select * from system.view_refre # Basic refreshing. $CLICKHOUSE_CLIENT -nq " create materialized view a - refresh after 1 second + refresh after 2 second engine Memory empty as select number as x from numbers(2) union all select rand64() as x" @@ -29,6 +29,7 @@ while [ "`$CLICKHOUSE_CLIENT -nq "select last_refresh_result from refreshes -- $ do sleep 0.1 done +start_time="`$CLICKHOUSE_CLIENT -nq "select reinterpret(now64(), 'Int64')"`" # Check table contents. $CLICKHOUSE_CLIENT -nq "select '<2: refreshed>', count(), sum(x=0), sum(x=1) from a" # Wait for table contents to change. @@ -39,7 +40,6 @@ do [ "$res2" == "$res1" ] || break sleep 0.1 done -time2="`$CLICKHOUSE_CLIENT -nq "select reinterpret(now64(), 'Int64')"`" # Wait for another change. while : do @@ -47,11 +47,11 @@ do [ "$res3" == "$res2" ] || break sleep 0.1 done -# Check that the two changes were at least 500ms apart, in particular that we're not refreshing +# Check that the two changes were at least 1 second apart, in particular that we're not refreshing # like crazy. This is potentially flaky, but we need at least one test that uses non-mocked timer # to make sure the clock+timer code works at all. If it turns out flaky, increase refresh period above. $CLICKHOUSE_CLIENT -nq " - select '<3: time difference at least>', min2(reinterpret(now64(), 'Int64') - $time2, 500); + select '<3: time difference at least>', min2(reinterpret(now64(), 'Int64') - $start_time, 1000); select '<4: next refresh in>', next_refresh_time-last_refresh_time from refreshes;" # Create a source table from which views will read. @@ -61,7 +61,7 @@ $CLICKHOUSE_CLIENT -nq " # Switch to fake clock, change refresh schedule, change query. $CLICKHOUSE_CLIENT -nq " system test view a set fake time '2050-01-01 00:00:01';" -while [ "`$CLICKHOUSE_CLIENT -nq "select status, last_refresh_time, next_refresh_time from refreshes -- $LINENO" | xargs`" != 'Scheduled 2050-01-01 00:00:01 2050-01-01 00:00:02' ] +while [ "`$CLICKHOUSE_CLIENT -nq "select status, last_refresh_time, next_refresh_time from refreshes -- $LINENO" | xargs`" != 'Scheduled 2050-01-01 00:00:01 2050-01-01 00:00:03' ] do sleep 0.1 done diff --git a/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.reference b/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.reference index 4d33751c699..8ae3cdf8f3a 100644 --- a/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.reference +++ b/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.reference @@ -2,7 +2,12 @@ 2 test2 8 3 test3 8 4 test4 1985 +2 test2 8 +3 test3 8 +4 test4 1985 --- 1 test1 42 +1 test1 42 --- 3 test3 +3 test3 diff --git a/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.sql b/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.sql index 53b8a761cda..ab6e1532299 100644 --- a/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.sql +++ b/tests/queries/0_stateless/02949_parallel_replicas_in_subquery.sql @@ -8,19 +8,23 @@ INSERT INTO merge_tree_in_subqueries VALUES(5, 'test5', 0); SET max_parallel_replicas=3, cluster_for_parallel_replicas='test_cluster_one_shard_three_replicas_localhost', parallel_replicas_for_non_replicated_merge_tree=1; -SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 0) SETTINGS allow_experimental_parallel_reading_from_replicas=2; -- { serverError SUPPORT_IS_DISABLED } +SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 0) SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=0; -- { serverError SUPPORT_IS_DISABLED } +SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 0) SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=1; SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 0) SETTINGS allow_experimental_parallel_reading_from_replicas=1; SELECT '---'; -SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 2, 3) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=2; -- { serverError SUPPORT_IS_DISABLED }; +SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 2, 3) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=0; -- { serverError SUPPORT_IS_DISABLED }; +SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 2, 3) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=1; SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT * FROM system.numbers LIMIT 2, 3) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=1; SELECT '---'; -SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT 1) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=2; -- { serverError SUPPORT_IS_DISABLED }; +SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT 1) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=0; -- { serverError SUPPORT_IS_DISABLED }; +SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT 1) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=1; SELECT * FROM merge_tree_in_subqueries WHERE id IN (SELECT 1) ORDER BY id SETTINGS allow_experimental_parallel_reading_from_replicas=1; -- IN with tuples is allowed SELECT '---'; -SELECT id, name FROM merge_tree_in_subqueries WHERE (id, name) IN (3, 'test3') SETTINGS allow_experimental_parallel_reading_from_replicas=2; +SELECT id, name FROM merge_tree_in_subqueries WHERE (id, name) IN (3, 'test3') SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=0; +SELECT id, name FROM merge_tree_in_subqueries WHERE (id, name) IN (3, 'test3') SETTINGS allow_experimental_parallel_reading_from_replicas=2, parallel_replicas_allow_in_with_subquery=1; DROP TABLE IF EXISTS merge_tree_in_subqueries; diff --git a/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.reference b/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.reference index d3a002c4fd4..413c79828c7 100644 --- a/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.reference +++ b/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.reference @@ -2,3 +2,6 @@ 2023-10-09 10:11:12.001 2023-10-09 10:11:12.000 2023-10-09 10:11:12.000 +2023-10-09 00:00:00.000000 +2023-10-09 00:00:00.000 +2023-10-09 00:00:00 diff --git a/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.sql b/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.sql index 178f21a9e63..15753d4532c 100644 --- a/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.sql +++ b/tests/queries/0_stateless/02956_fix_to_start_of_milli_microsecond.sql @@ -1,4 +1,7 @@ SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.000999', 6), toIntervalMillisecond(1)); SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.000500', 6), toIntervalMillisecond(1)); SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.000499', 6), toIntervalMillisecond(1)); -SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.000999', 6), toIntervalMillisecond(10)); \ No newline at end of file +SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.000999', 6), toIntervalMillisecond(10)); +select toStartOfInterval(toDateTime64('2023-10-09 00:01:34', 9), toIntervalMicrosecond(100000000)); +select toStartOfInterval(toDateTime64('2023-10-09 00:01:34', 9), toIntervalMillisecond(100000)); +select toStartOfInterval(toDateTime64('2023-10-09 00:01:34', 9), toIntervalSecond(100)); \ No newline at end of file diff --git a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference index 6b1fdfd42a2..100e4e500cd 100644 --- a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference +++ b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference @@ -237,7 +237,7 @@ sub2 as (select y, z from tab2 where y != 4), sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), sub4 as (select z, a from tab3 where z != 8), sub5 as (select z, a, x, y, r.y, ll.z from sub4 rr any right join sub3 ll on ll.z = rr.z) -select * from sub5 order by x SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1;-- { echoOn } +select * from sub5 order by x SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; Expression Sorting Expression @@ -250,6 +250,93 @@ Expression ReadFromRemoteParallelReplicas Expression ReadFromRemoteParallelReplicas +-- +-- Subqueries for IN allowed +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; +0 0 0 0 0 0 +1 1 0 0 0 0 +3 3 0 0 0 0 +4 4 0 0 0 0 +5 5 0 0 0 0 +6 6 6 6 0 0 +7 7 0 0 0 0 +8 8 8 8 0 0 +9 9 0 0 0 0 +10 10 10 10 0 0 +11 11 0 0 0 0 +12 12 12 12 12 12 +13 13 0 0 0 0 +14 14 14 14 0 0 +15 15 0 0 0 0 +explain description=0 +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; +Expression + Sorting + Expression + ReadFromRemoteParallelReplicas +-- +-- Subqueries for IN are not allowed +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1, parallel_replicas_allow_in_with_subquery=0; +0 0 0 0 0 0 +1 1 0 0 0 0 +3 3 0 0 0 0 +4 4 0 0 0 0 +5 5 0 0 0 0 +6 6 6 6 0 0 +7 7 0 0 0 0 +8 8 8 8 0 0 +9 9 0 0 0 0 +10 10 10 10 0 0 +11 11 0 0 0 0 +12 12 12 12 12 12 +13 13 0 0 0 0 +14 14 14 14 0 0 +15 15 0 0 0 0 +explain description=0 +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1, parallel_replicas_allow_in_with_subquery=0;-- { echoOn } +Expression + Sorting + Expression + Join + Expression + Join + Expression + CreatingSets + Expression + Expression + ReadFromMergeTree + CreatingSet + Expression + Filter + ReadFromSystemNumbers + Expression + ReadFromRemoteParallelReplicas + Expression + ReadFromRemoteParallelReplicas set parallel_replicas_prefer_local_join = 1; -- A query with only INNER/LEFT joins is fully send to replicas. JOIN is executed in GLOBAL mode. select x, y, r.y, z, rr.z, a from (select l.x, l.y, r.y, r.z as z from (select x, y from tab1 where x != 2) l any left join (select y, z from tab2 where y != 4) r on l.y = r.y) ll any left join (select z, a from tab3 where z != 8) rr on ll.z = rr.z order by x SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; @@ -500,3 +587,90 @@ Expression ReadFromRemoteParallelReplicas Expression ReadFromRemoteParallelReplicas +-- +-- Subqueries for IN allowed +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; +0 0 0 0 0 0 +1 1 0 0 0 0 +3 3 0 0 0 0 +4 4 0 0 0 0 +5 5 0 0 0 0 +6 6 6 6 0 0 +7 7 0 0 0 0 +8 8 8 8 0 0 +9 9 0 0 0 0 +10 10 10 10 0 0 +11 11 0 0 0 0 +12 12 12 12 12 12 +13 13 0 0 0 0 +14 14 14 14 0 0 +15 15 0 0 0 0 +explain description=0 +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; +Expression + Sorting + Expression + ReadFromRemoteParallelReplicas +-- +-- Subqueries for IN are not allowed +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1, parallel_replicas_allow_in_with_subquery=0; +0 0 0 0 0 0 +1 1 0 0 0 0 +3 3 0 0 0 0 +4 4 0 0 0 0 +5 5 0 0 0 0 +6 6 6 6 0 0 +7 7 0 0 0 0 +8 8 8 8 0 0 +9 9 0 0 0 0 +10 10 10 10 0 0 +11 11 0 0 0 0 +12 12 12 12 12 12 +13 13 0 0 0 0 +14 14 14 14 0 0 +15 15 0 0 0 0 +explain description=0 +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1, parallel_replicas_allow_in_with_subquery=0; +Expression + Sorting + Expression + Join + Expression + Join + Expression + CreatingSets + Expression + Expression + ReadFromMergeTree + CreatingSet + Expression + Filter + ReadFromSystemNumbers + Expression + ReadFromRemoteParallelReplicas + Expression + ReadFromRemoteParallelReplicas diff --git a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2 b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2 index 7d2766d52f8..54505b147a3 100644 --- a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2 +++ b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2 @@ -126,4 +126,42 @@ sub4 as (select z, a from tab3 where z != 8), sub5 as (select z, a, x, y, r.y, ll.z from sub4 rr any right join sub3 ll on ll.z = rr.z) select * from sub5 order by x SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; +-- +-- Subqueries for IN allowed +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; + +explain description=0 +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1; + +-- +-- Subqueries for IN are not allowed +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1, parallel_replicas_allow_in_with_subquery=0; + +explain description=0 +with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)), +sub2 as (select y, z from tab2 where y != 4), +sub3 as (select l.x, l.y, r.y, r.z as z from sub1 l any left join sub2 r on l.y = r.y), +sub4 as (select z, a from tab3 where z != 8), +sub5 as (select x, y, r.y, z, rr.z, a from sub3 ll any left join sub4 rr on ll.z = rr.z) +select * from sub5 order by x +SETTINGS allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', allow_experimental_analyzer=1, parallel_replicas_allow_in_with_subquery=0; + {%- endfor %} diff --git a/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference index b6d1ff865e5..47e9b86237a 100644 --- a/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference +++ b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference @@ -1,2 +1,3 @@ c1 Nullable(String) c1 Nullable(Float64) +x Nullable(Float64) diff --git a/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql index 2a281e898f1..4f78855f5ce 100644 --- a/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql +++ b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql @@ -1,2 +1,5 @@ DESC format(CSV, '1E20\n1.1E20') settings input_format_try_infer_exponent_floats = 0; DESC format(CSV, '1E20\n1.1E20') settings input_format_try_infer_exponent_floats = 1; +-- This setting should not take affect on JSON formats +DESC format(JSONEachRow, '{"x" : 1.1e20}') settings input_format_try_infer_exponent_floats = 0; + diff --git a/tests/queries/0_stateless/02992_all_columns_should_have_comment.reference b/tests/queries/0_stateless/02992_all_columns_should_have_comment.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02992_all_columns_should_have_comment.sql b/tests/queries/0_stateless/02992_all_columns_should_have_comment.sql new file mode 100644 index 00000000000..b984c12e285 --- /dev/null +++ b/tests/queries/0_stateless/02992_all_columns_should_have_comment.sql @@ -0,0 +1,4 @@ +SYSTEM FLUSH LOGS; +SELECT 'Column ' || name || ' from table ' || concat(database, '.', table) || ' should have a comment' +FROM system.columns +WHERE (database = 'system') AND (comment = '') AND (table NOT ILIKE '%_log_%') AND (table NOT IN ('numbers', 'numbers_mt', 'one')) AND (default_kind != 'ALIAS'); diff --git a/tests/queries/0_stateless/02998_analyzer_secret_args_tree_node.reference b/tests/queries/0_stateless/02998_analyzer_secret_args_tree_node.reference new file mode 100644 index 00000000000..ffd2f68990b --- /dev/null +++ b/tests/queries/0_stateless/02998_analyzer_secret_args_tree_node.reference @@ -0,0 +1,71 @@ +-- { echoOn } +SET allow_experimental_analyzer = 1; +EXPLAIN QUERY TREE SELECT encrypt('aes-256-ofb', (SELECT 'qwerty'), '12345678901234567890123456789012'), encrypt('aes-256-ofb', (SELECT 'asdf'), '12345678901234567890123456789012'); +QUERY id: 0 + PROJECTION COLUMNS + encrypt(\'aes-256-ofb\', [HIDDEN id: 1], [HIDDEN id: 2]) Nullable(String) + encrypt(\'aes-256-ofb\', [HIDDEN id: 3], [HIDDEN id: 2]) Nullable(String) + PROJECTION + LIST id: 1, nodes: 2 + CONSTANT id: 2, constant_value: \'\\nãì&\', constant_value_type: Nullable(String) + EXPRESSION + FUNCTION id: 3, function_name: encrypt, function_type: ordinary, result_type: Nullable(String) + ARGUMENTS + LIST id: 4, nodes: 3 + CONSTANT id: 5, constant_value: \'aes-256-ofb\', constant_value_type: String + CONSTANT id: 6, constant_value: [HIDDEN id: 1], constant_value_type: Nullable(String) + CONSTANT id: 7, constant_value: [HIDDEN id: 2], constant_value_type: String + CONSTANT id: 8, constant_value: \'çø\', constant_value_type: Nullable(String) + EXPRESSION + FUNCTION id: 9, function_name: encrypt, function_type: ordinary, result_type: Nullable(String) + ARGUMENTS + LIST id: 10, nodes: 3 + CONSTANT id: 11, constant_value: \'aes-256-ofb\', constant_value_type: String + CONSTANT id: 12, constant_value: [HIDDEN id: 3], constant_value_type: Nullable(String) + CONSTANT id: 13, constant_value: [HIDDEN id: 2], constant_value_type: String + JOIN TREE + TABLE id: 14, alias: __table1, table_name: system.one +SET format_display_secrets_in_show_and_select = 1; +EXPLAIN QUERY TREE SELECT encrypt('aes-256-ofb', (SELECT 'qwerty'), '12345678901234567890123456789012'), encrypt('aes-256-ofb', (SELECT 'asdf'), '12345678901234567890123456789012'); +QUERY id: 0 + PROJECTION COLUMNS + encrypt(\'aes-256-ofb\', _subquery_1, \'12345678901234567890123456789012\') Nullable(String) + encrypt(\'aes-256-ofb\', _subquery_2, \'12345678901234567890123456789012\') Nullable(String) + PROJECTION + LIST id: 1, nodes: 2 + CONSTANT id: 2, constant_value: \'\\nãì&\', constant_value_type: Nullable(String) + EXPRESSION + FUNCTION id: 3, function_name: encrypt, function_type: ordinary, result_type: Nullable(String) + ARGUMENTS + LIST id: 4, nodes: 3 + CONSTANT id: 5, constant_value: \'aes-256-ofb\', constant_value_type: String + CONSTANT id: 6, constant_value: \'qwerty\', constant_value_type: Nullable(String) + EXPRESSION + QUERY id: 7, is_subquery: 1 + PROJECTION COLUMNS + \'qwerty\' String + PROJECTION + LIST id: 8, nodes: 1 + CONSTANT id: 9, constant_value: \'qwerty\', constant_value_type: String + JOIN TREE + TABLE id: 10, table_name: system.one + CONSTANT id: 11, constant_value: \'12345678901234567890123456789012\', constant_value_type: String + CONSTANT id: 12, constant_value: \'çø\', constant_value_type: Nullable(String) + EXPRESSION + FUNCTION id: 13, function_name: encrypt, function_type: ordinary, result_type: Nullable(String) + ARGUMENTS + LIST id: 14, nodes: 3 + CONSTANT id: 15, constant_value: \'aes-256-ofb\', constant_value_type: String + CONSTANT id: 16, constant_value: \'asdf\', constant_value_type: Nullable(String) + EXPRESSION + QUERY id: 17, is_subquery: 1 + PROJECTION COLUMNS + \'asdf\' String + PROJECTION + LIST id: 18, nodes: 1 + CONSTANT id: 19, constant_value: \'asdf\', constant_value_type: String + JOIN TREE + TABLE id: 20, table_name: system.one + CONSTANT id: 21, constant_value: \'12345678901234567890123456789012\', constant_value_type: String + JOIN TREE + TABLE id: 22, alias: __table1, table_name: system.one diff --git a/tests/queries/0_stateless/02998_analyzer_secret_args_tree_node.sql b/tests/queries/0_stateless/02998_analyzer_secret_args_tree_node.sql new file mode 100644 index 00000000000..f40b40b6c8c --- /dev/null +++ b/tests/queries/0_stateless/02998_analyzer_secret_args_tree_node.sql @@ -0,0 +1,12 @@ +-- Tags: no-fasttest +-- encrypt function doesn't exist in the fastest build + +-- { echoOn } +SET allow_experimental_analyzer = 1; + +EXPLAIN QUERY TREE SELECT encrypt('aes-256-ofb', (SELECT 'qwerty'), '12345678901234567890123456789012'), encrypt('aes-256-ofb', (SELECT 'asdf'), '12345678901234567890123456789012'); + +SET format_display_secrets_in_show_and_select = 1; + +EXPLAIN QUERY TREE SELECT encrypt('aes-256-ofb', (SELECT 'qwerty'), '12345678901234567890123456789012'), encrypt('aes-256-ofb', (SELECT 'asdf'), '12345678901234567890123456789012'); +-- { echoOff } diff --git a/tests/queries/0_stateless/02998_system_dns_cache_table.sh b/tests/queries/0_stateless/02998_system_dns_cache_table.sh index 41d2386fe9c..b74fc00ab3b 100755 --- a/tests/queries/0_stateless/02998_system_dns_cache_table.sh +++ b/tests/queries/0_stateless/02998_system_dns_cache_table.sh @@ -5,12 +5,14 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) . "$CURDIR"/../shell_config.sh # Retries are necessary because the DNS cache may be flushed before second statement is executed -i=0 retries=3 +i=0 +retries=5 while [[ $i -lt $retries ]]; do - ${CLICKHOUSE_CURL} -sS --fail --data "SELECT * FROM url('http://localhost:8123/ping', CSV, 'auto', headers())" "${CLICKHOUSE_URL}" | grep -oP -q 'Ok.' && \ + ${CLICKHOUSE_CURL} -sS --fail --data "SELECT * FROM url('http://localhost:8123/ping', CSV, 'auto', headers())" "${CLICKHOUSE_URL}" | grep -oP -q 'Ok.' || continue + RECORDS=$(${CLICKHOUSE_CURL} -sS --fail --data "SELECT hostname, ip_address, ip_family, (isNotNull(cached_at) AND cached_at > '1970-01-01 00:00:00') FROM system.dns_cache WHERE hostname = 'localhost' and ip_family = 'IPv4';" "${CLICKHOUSE_URL}") - if [ "${RECORDS}" != "" ]; then + if [[ -n "${RECORDS}" ]]; then echo "${RECORDS}" exit 0 fi diff --git a/tests/queries/0_stateless/02998_to_milliseconds.reference b/tests/queries/0_stateless/02998_to_milliseconds.reference new file mode 100644 index 00000000000..05139c19d1d --- /dev/null +++ b/tests/queries/0_stateless/02998_to_milliseconds.reference @@ -0,0 +1,8 @@ +2023-04-21 10:20:30 0 0 +2023-04-21 10:20:30 0 0 +2023-04-21 10:20:30.123 123 123 +2023-04-21 10:20:30.123456 123 123 +2023-04-21 10:20:30.123456789 123 123 +120 +2023-04-21 10:20:30 0 +2023-04-21 10:20:30 0 diff --git a/tests/queries/0_stateless/02998_to_milliseconds.sql b/tests/queries/0_stateless/02998_to_milliseconds.sql new file mode 100644 index 00000000000..f159f6aab50 --- /dev/null +++ b/tests/queries/0_stateless/02998_to_milliseconds.sql @@ -0,0 +1,17 @@ +-- Negative tests +SELECT toMillisecond(); -- { serverError NUMBER_OF_ARGUMENTS_DOESNT_MATCH } +SELECT toMillisecond('string'); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT toMillisecond(toDate('2024-02-28')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT toMillisecond(toDate32('2024-02-28')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } + +-- Tests with constant and non-constant arguments +SELECT toDateTime('2023-04-21 10:20:30') AS dt, toMillisecond(dt), toMillisecond(materialize(dt)); +SELECT toDateTime64('2023-04-21 10:20:30', 0) AS dt64, toMillisecond(dt64), toMillisecond(materialize(dt64)); +SELECT toDateTime64('2023-04-21 10:20:30.123', 3) AS dt64, toMillisecond(dt64), toMillisecond(materialize(dt64)); +SELECT toDateTime64('2023-04-21 10:20:30.123456', 6) AS dt64, toMillisecond(dt64), toMillisecond(materialize(dt64)); +SELECT toDateTime64('2023-04-21 10:20:30.123456789', 9) AS dt64, toMillisecond(dt64), toMillisecond(materialize(dt64)); + +-- Special cases +SELECT MILLISECOND(toDateTime64('2023-04-21 10:20:30.123456', 2)); -- Alias +SELECT toNullable(toDateTime('2023-04-21 10:20:30')) AS dt, toMillisecond(dt); -- Nullable +SELECT toLowCardinality(toDateTime('2023-04-21 10:20:30')) AS dt, toMillisecond(dt); -- LowCardinality diff --git a/tests/queries/0_stateless/02999_analyzer_preimage_null.reference b/tests/queries/0_stateless/02999_analyzer_preimage_null.reference index 6f9afedfd07..c5eb6b23d0d 100644 --- a/tests/queries/0_stateless/02999_analyzer_preimage_null.reference +++ b/tests/queries/0_stateless/02999_analyzer_preimage_null.reference @@ -108,14 +108,14 @@ QUERY id: 0 FUNCTION id: 14, function_name: greaterOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS LIST id: 15, nodes: 2 - COLUMN id: 16, column_name: date1, result_type: Date, source_id: 3 - CONSTANT id: 17, constant_value: \'1994-01-01\', constant_value_type: String - FUNCTION id: 18, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 + COLUMN id: 12, column_name: date1, result_type: Date, source_id: 3 + CONSTANT id: 16, constant_value: \'1994-01-01\', constant_value_type: String + FUNCTION id: 17, function_name: lessOrEquals, function_type: ordinary, result_type: UInt8 ARGUMENTS - LIST id: 19, nodes: 2 - COLUMN id: 20, column_name: id, result_type: UInt32, source_id: 3 - FUNCTION id: 21, function_name: toYear, function_type: ordinary, result_type: UInt16 + LIST id: 18, nodes: 2 + COLUMN id: 19, column_name: id, result_type: UInt32, source_id: 3 + FUNCTION id: 20, function_name: toYear, function_type: ordinary, result_type: UInt16 ARGUMENTS - LIST id: 22, nodes: 1 - COLUMN id: 23, column_name: date1, result_type: Date, source_id: 3 + LIST id: 21, nodes: 1 + COLUMN id: 12, column_name: date1, result_type: Date, source_id: 3 SETTINGS optimize_time_filter_with_preimage=1 diff --git a/tests/queries/0_stateless/02999_scalar_subqueries_bug_1.reference b/tests/queries/0_stateless/02999_scalar_subqueries_bug_1.reference new file mode 100644 index 00000000000..0740afe92c6 --- /dev/null +++ b/tests/queries/0_stateless/02999_scalar_subqueries_bug_1.reference @@ -0,0 +1,66 @@ +0 0 +0 0 +0 0 +0 0 +1 \N +1 \N +2 \N +2 \N +3 \N +3 \N +4 \N +4 \N +5 \N +5 \N +6 \N +6 \N +7 \N +7 \N +8 \N +8 \N +9 \N +9 \N +10 10 +10 10 +10 10 +10 10 +11 \N +11 \N +12 \N +12 \N +13 \N +13 \N +14 \N +14 \N +15 \N +15 \N +16 \N +16 \N +17 \N +17 \N +18 \N +18 \N +19 \N +19 \N +20 20 +20 20 +20 20 +20 20 +21 \N +21 \N +22 \N +22 \N +23 \N +23 \N +24 \N +24 \N +25 \N +25 \N +26 \N +26 \N +27 \N +27 \N +28 \N +28 \N +29 \N +29 \N diff --git a/tests/queries/0_stateless/02999_scalar_subqueries_bug_1.sql b/tests/queries/0_stateless/02999_scalar_subqueries_bug_1.sql new file mode 100644 index 00000000000..88bcdeb7f77 --- /dev/null +++ b/tests/queries/0_stateless/02999_scalar_subqueries_bug_1.sql @@ -0,0 +1,8 @@ +drop table if exists t_table_select; +CREATE TABLE t_table_select (id UInt32) ENGINE = MergeTree ORDER BY id; +INSERT INTO t_table_select (id) SELECT number FROM numbers(30); + +CREATE TEMPORARY TABLE t_test (x UInt32, y Nullable(UInt32)) AS SELECT a.id, b.id FROM remote('127.0.0.{1,2}', currentDatabase(), t_table_select) AS a GLOBAL LEFT JOIN (SELECT id FROM remote('127.0.0.{1,2}', currentDatabase(), t_table_select) AS b WHERE (b.id % 10) = 0) AS b ON b.id = a.id SETTINGS join_use_nulls = 1; + +select * from t_test order by x; + diff --git a/tests/queries/0_stateless/02999_scalar_subqueries_bug_2.reference b/tests/queries/0_stateless/02999_scalar_subqueries_bug_2.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02999_scalar_subqueries_bug_2.sql b/tests/queries/0_stateless/02999_scalar_subqueries_bug_2.sql new file mode 100644 index 00000000000..03ac91e401a --- /dev/null +++ b/tests/queries/0_stateless/02999_scalar_subqueries_bug_2.sql @@ -0,0 +1,18 @@ +drop table if exists source; +drop table if exists target1; +drop table if exists target2; +drop table if exists v_heavy; + + +create table source(type String) engine=MergeTree order by type; + +create view v_heavy as +with nums as (select number from numbers(1e5)) +select count(*) n from (select number from numbers(1e5) n1 cross join nums); + +create table target1(type String) engine=MergeTree order by type; +create table target2(type String) engine=MergeTree order by type; + +set max_execution_time=2; +-- we should not execute scalar subquery here +create materialized view vm_target2 to target2 as select * from source where type='two' and (select sum(sleepEachRow(0.1)) from numbers(30)); diff --git a/tests/queries/0_stateless/03000_minmax_index_first.reference b/tests/queries/0_stateless/03000_minmax_index_first.reference new file mode 100644 index 00000000000..7cf792d8ed4 --- /dev/null +++ b/tests/queries/0_stateless/03000_minmax_index_first.reference @@ -0,0 +1,2 @@ +Name: v_mm +Name: v_set diff --git a/tests/queries/0_stateless/03000_minmax_index_first.sql b/tests/queries/0_stateless/03000_minmax_index_first.sql new file mode 100644 index 00000000000..d978e6ea8fa --- /dev/null +++ b/tests/queries/0_stateless/03000_minmax_index_first.sql @@ -0,0 +1,18 @@ +DROP TABLE IF EXISTS skip_table; + +CREATE TABLE skip_table +( + k UInt64, + v UInt64, + INDEX v_set v TYPE set(100) GRANULARITY 2, -- set index is declared before minmax intentionally + INDEX v_mm v TYPE minmax GRANULARITY 2 +) +ENGINE = MergeTree +PRIMARY KEY k +SETTINGS index_granularity = 8192; + +INSERT INTO skip_table SELECT number, intDiv(number, 4096) FROM numbers(100000); + +SELECT trim(explain) FROM ( EXPLAIN indexes = 1 SELECT * FROM skip_table WHERE v = 125) WHERE explain like '%Name%'; + +DROP TABLE skip_table; diff --git a/tests/queries/0_stateless/03000_virtual_columns_in_prewhere.reference b/tests/queries/0_stateless/03000_virtual_columns_in_prewhere.reference new file mode 100644 index 00000000000..06b63ea6c2f --- /dev/null +++ b/tests/queries/0_stateless/03000_virtual_columns_in_prewhere.reference @@ -0,0 +1 @@ +0 0 0 diff --git a/tests/queries/0_stateless/03000_virtual_columns_in_prewhere.sql b/tests/queries/0_stateless/03000_virtual_columns_in_prewhere.sql new file mode 100644 index 00000000000..d57db9151b9 --- /dev/null +++ b/tests/queries/0_stateless/03000_virtual_columns_in_prewhere.sql @@ -0,0 +1,10 @@ +drop table if exists x; + +create table x (i int, j int, k int) engine MergeTree order by tuple() settings index_granularity=8192, index_granularity_bytes = '10Mi', min_bytes_for_wide_part=0, min_rows_for_wide_part=0, ratio_of_defaults_for_sparse_serialization=1; + +insert into x select number, number * 2, number * 3 from numbers(100000); + +-- One granule, (_part_offset (8 bytes) + (4 bytes)) * 8192 + (8 bytes) * 1 = 98312 +select * from x prewhere _part_offset = 0 settings max_bytes_to_read = 98312; + +drop table x; diff --git a/tests/queries/0_stateless/03001_insert_threads_deduplication.reference b/tests/queries/0_stateless/03001_insert_threads_deduplication.reference new file mode 100644 index 00000000000..b6d6006f84c --- /dev/null +++ b/tests/queries/0_stateless/03001_insert_threads_deduplication.reference @@ -0,0 +1,4 @@ +3 +6 +12 +18 diff --git a/tests/queries/0_stateless/03001_insert_threads_deduplication.sql b/tests/queries/0_stateless/03001_insert_threads_deduplication.sql new file mode 100644 index 00000000000..093d2b3185d --- /dev/null +++ b/tests/queries/0_stateless/03001_insert_threads_deduplication.sql @@ -0,0 +1,69 @@ +-- Tags: distributed + +DROP TABLE IF EXISTS landing SYNC; +DROP TABLE IF EXISTS landing_dist SYNC; +DROP TABLE IF EXISTS ds SYNC; + +CREATE TABLE landing +( + timestamp DateTime64(3), + status String, + id String +) +ENGINE = MergeTree() +ORDER BY timestamp; + +CREATE TABLE landing_dist +( + timestamp DateTime64(3), + status String, + id String +) +ENGINE = Distributed('test_cluster_two_shards', currentDatabase(), 'landing', rand()); + +SYSTEM STOP MERGES landing; -- Stopping merges to force 3 parts + +INSERT INTO landing (status, id, timestamp) SELECT * FROM generateRandom() LIMIT 1; +INSERT INTO landing (status, id, timestamp) SELECT * FROM generateRandom() LIMIT 1; +INSERT INTO landing (status, id, timestamp) SELECT * FROM generateRandom() LIMIT 1; + +CREATE TABLE ds +( + timestamp DateTime64(3), + status String, + id String +) +ENGINE = MergeTree() +ORDER BY timestamp +SETTINGS non_replicated_deduplication_window=1000; + +INSERT INTO ds SELECT * FROM landing +SETTINGS insert_deduplicate=1, insert_deduplication_token='token1', + max_insert_threads=5; + +SELECT count() FROM ds; + +INSERT INTO ds SELECT * FROM landing +SETTINGS insert_deduplicate=1, insert_deduplication_token='token2', + max_insert_threads=1; + +SELECT count() FROM ds; + +-- When reading from distributed table, 6 rows are going to be retrieved +-- due to the being using the two shards cluster + +INSERT INTO ds SELECT * FROM landing_dist +SETTINGS insert_deduplicate=1, insert_deduplication_token='token3', + max_insert_threads=5; + +SELECT count() FROM ds; + +INSERT INTO ds SELECT * FROM landing_dist +SETTINGS insert_deduplicate=1, insert_deduplication_token='token4', + max_insert_threads=1; + +SELECT count() FROM ds; + +DROP TABLE IF EXISTS landing SYNC; +DROP TABLE IF EXISTS landing_dist SYNC; +DROP TABLE IF EXISTS ds SYNC; diff --git a/tests/queries/0_stateless/03002_int_div_decimal_with_date_bug.reference b/tests/queries/0_stateless/03002_int_div_decimal_with_date_bug.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/03002_int_div_decimal_with_date_bug.sql b/tests/queries/0_stateless/03002_int_div_decimal_with_date_bug.sql new file mode 100644 index 00000000000..1668821200c --- /dev/null +++ b/tests/queries/0_stateless/03002_int_div_decimal_with_date_bug.sql @@ -0,0 +1,5 @@ +SELECT intDiv(CAST('1.0', 'Decimal256(3)'), today()); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT intDiv(CAST('1.0', 'Decimal256(3)'), toDate('2023-01-02')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT intDiv(CAST('1.0', 'Decimal256(2)'), toDate32('2023-01-02 12:12:12')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT intDiv(CAST('1.0', 'Decimal256(2)'), toDateTime('2023-01-02 12:12:12')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT intDiv(CAST('1.0', 'Decimal256(2)'), toDateTime64('2023-01-02 12:12:12.002', 3)); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } diff --git a/tests/queries/0_stateless/03003_analyzer_setting.reference b/tests/queries/0_stateless/03003_analyzer_setting.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/03003_analyzer_setting.sql b/tests/queries/0_stateless/03003_analyzer_setting.sql new file mode 100644 index 00000000000..2e5cab71277 --- /dev/null +++ b/tests/queries/0_stateless/03003_analyzer_setting.sql @@ -0,0 +1,9 @@ +CREATE TABLE test (dummy Int8) ENGINE = Distributed(test_cluster_two_shards, 'system', 'one'); + +SET allow_experimental_analyzer = 0; + +SELECT * FROM (SELECT * FROM test SETTINGS allow_experimental_analyzer = 1); -- { serverError INCORRECT_QUERY } + +SET allow_experimental_analyzer = 1; + +SELECT * FROM (SELECT * FROM test SETTINGS allow_experimental_analyzer = 0); -- { serverError INCORRECT_QUERY } diff --git a/tests/queries/0_stateless/03003_database_filesystem_format_detection.reference b/tests/queries/0_stateless/03003_database_filesystem_format_detection.reference new file mode 100644 index 00000000000..87d15e83593 --- /dev/null +++ b/tests/queries/0_stateless/03003_database_filesystem_format_detection.reference @@ -0,0 +1,2 @@ +a Nullable(Int64) +42 diff --git a/tests/queries/0_stateless/03003_database_filesystem_format_detection.sh b/tests/queries/0_stateless/03003_database_filesystem_format_detection.sh new file mode 100755 index 00000000000..8b476a1f82e --- /dev/null +++ b/tests/queries/0_stateless/03003_database_filesystem_format_detection.sh @@ -0,0 +1,15 @@ +#!/usr/bin/env bash + +CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +# shellcheck source=../shell_config.sh +. "$CURDIR"/../shell_config.sh + + +DATA_FILE=$CLICKHOUSE_TEST_UNIQUE_NAME.data + +echo '{"a" : 42}' > $DATA_FILE +$CLICKHOUSE_LOCAL -q "desc table \`$DATA_FILE\`" +$CLICKHOUSE_LOCAL -q "select * from \`$DATA_FILE\`" + +rm $DATA_FILE + diff --git a/tests/queries/0_stateless/03003_prql_panic.reference b/tests/queries/0_stateless/03003_prql_panic.reference new file mode 100644 index 00000000000..8e0782c44f2 --- /dev/null +++ b/tests/queries/0_stateless/03003_prql_panic.reference @@ -0,0 +1 @@ +SYNTAX_ERROR diff --git a/tests/queries/0_stateless/03003_prql_panic.sh b/tests/queries/0_stateless/03003_prql_panic.sh new file mode 100755 index 00000000000..01d7beed99b --- /dev/null +++ b/tests/queries/0_stateless/03003_prql_panic.sh @@ -0,0 +1,13 @@ +#!/usr/bin/env bash +# Tags: no-fasttest +# Requires Rust, which is not built for Fast Test. + +CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +# shellcheck source=../shell_config.sh +. "$CUR_DIR"/../shell_config.sh + +# Before [1] this causes a panic, but it will be fixed soon, so do not check +# for panic, but just for SYNTAX_ERROR. +# +# [1]: https://github.com/PRQL/prql/pull/4285 +$CLICKHOUSE_CLIENT --dialect prql -q "SELECT id FROM distributed_test_table GROUP BY x -> concat(concat(materialize(toNullable(NULL)))) LIMIT 3" |& grep -o -m1 SYNTAX_ERROR diff --git a/tests/queries/0_stateless/03004_json_named_tuples_inference_ambiguous_paths_as_string.reference b/tests/queries/0_stateless/03004_json_named_tuples_inference_ambiguous_paths_as_string.reference new file mode 100644 index 00000000000..0318b136ade --- /dev/null +++ b/tests/queries/0_stateless/03004_json_named_tuples_inference_ambiguous_paths_as_string.reference @@ -0,0 +1,3 @@ +obj Tuple(\n a Nullable(String)) +('42') +('{"b" : 42}') diff --git a/tests/queries/0_stateless/03004_json_named_tuples_inference_ambiguous_paths_as_string.sql b/tests/queries/0_stateless/03004_json_named_tuples_inference_ambiguous_paths_as_string.sql new file mode 100644 index 00000000000..4b986c94868 --- /dev/null +++ b/tests/queries/0_stateless/03004_json_named_tuples_inference_ambiguous_paths_as_string.sql @@ -0,0 +1,4 @@ +set input_format_json_use_string_type_for_ambiguous_paths_in_named_tuples_inference_from_objects=1; +desc format(JSONEachRow, '{"obj" : {"a" : 42}}, {"obj" : {"a" : {"b" : 42}}}'); +select * from format(JSONEachRow, '{"obj" : {"a" : 42}}, {"obj" : {"a" : {"b" : 42}}}'); + diff --git a/tests/queries/0_stateless/03006_analyzer_executable_table_function.reference b/tests/queries/0_stateless/03006_analyzer_executable_table_function.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/03006_analyzer_executable_table_function.sql b/tests/queries/0_stateless/03006_analyzer_executable_table_function.sql new file mode 100644 index 00000000000..4edbcc97f50 --- /dev/null +++ b/tests/queries/0_stateless/03006_analyzer_executable_table_function.sql @@ -0,0 +1,4 @@ +SELECT + toFixedString(toFixedString(toLowCardinality(toFixedString('--------------------', toNullable(20))), toLowCardinality(20)), 20), + * +FROM executable('data String', SETTINGS max_command_execution_time = 100); -- { serverError NUMBER_OF_ARGUMENTS_DOESNT_MATCH} diff --git a/tests/queries/0_stateless/03006_async_insert_deadlock_log.reference b/tests/queries/0_stateless/03006_async_insert_deadlock_log.reference new file mode 100644 index 00000000000..d00491fd7e5 --- /dev/null +++ b/tests/queries/0_stateless/03006_async_insert_deadlock_log.reference @@ -0,0 +1 @@ +1 diff --git a/tests/queries/0_stateless/03006_async_insert_deadlock_log.sh b/tests/queries/0_stateless/03006_async_insert_deadlock_log.sh new file mode 100755 index 00000000000..f01c34cdbda --- /dev/null +++ b/tests/queries/0_stateless/03006_async_insert_deadlock_log.sh @@ -0,0 +1,12 @@ +#!/usr/bin/env bash + +CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +# shellcheck source=../shell_config.sh +. "$CURDIR"/../shell_config.sh + +$CLICKHOUSE_CLIENT --query "CREATE TABLE t_async_insert_deadlock (a UInt64) ENGINE = Log" + +echo '{"a": 1}' | $CLICKHOUSE_CLIENT --async_insert 1 --wait_for_async_insert 1 --query "INSERT INTO t_async_insert_deadlock FORMAT JSONEachRow" + +$CLICKHOUSE_CLIENT --query "SELECT * FROM t_async_insert_deadlock ORDER BY a" +$CLICKHOUSE_CLIENT --query "DROP TABLE t_async_insert_deadlock" diff --git a/tests/queries/0_stateless/03006_buffer_overflow_join.reference b/tests/queries/0_stateless/03006_buffer_overflow_join.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/03006_buffer_overflow_join.sql b/tests/queries/0_stateless/03006_buffer_overflow_join.sql new file mode 100644 index 00000000000..8c1fa3cecc0 --- /dev/null +++ b/tests/queries/0_stateless/03006_buffer_overflow_join.sql @@ -0,0 +1,6 @@ +CREATE TABLE 03006_buffer_overflow_l (`a` String, `b` Tuple(String, String)) ENGINE = Memory; +INSERT INTO 03006_buffer_overflow_l SELECT * FROM generateRandom() limit 1000; +CREATE TABLE 03006_buffer_overflow_r (`a` LowCardinality(Nullable(String)), `c` Tuple(LowCardinality(String), LowCardinality(String))) ENGINE = Memory; +INSERT INTO 03006_buffer_overflow_r SELECT * FROM generateRandom() limit 1000; + +SELECT a FROM 03006_buffer_overflow_l RIGHT JOIN 03006_buffer_overflow_r USING (a) ORDER BY a ASC NULLS FIRST FORMAT Null; diff --git a/tests/queries/0_stateless/03006_mv_deduplication_throw_if_async_insert.reference b/tests/queries/0_stateless/03006_mv_deduplication_throw_if_async_insert.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/03006_mv_deduplication_throw_if_async_insert.sql b/tests/queries/0_stateless/03006_mv_deduplication_throw_if_async_insert.sql new file mode 100644 index 00000000000..808317c917e --- /dev/null +++ b/tests/queries/0_stateless/03006_mv_deduplication_throw_if_async_insert.sql @@ -0,0 +1,19 @@ +DROP TABLE IF EXISTS 02985_test; + +SET async_insert = 1; +SET deduplicate_blocks_in_dependent_materialized_views = 1; + +CREATE TABLE 03006_test +( + d Date, + value UInt64 +) +ENGINE = MergeTree ORDER BY tuple(); + +INSERT INTO 03006_test VALUES ('2024-03-05', 1), ('2024-03-05', 2), ('2024-03-05', 1); -- { serverError SUPPORT_IS_DISABLED } +INSERT INTO 03006_test SETTINGS compatibility='24.1' VALUES ('2024-03-05', 1), ('2024-03-05', 2), ('2024-03-05', 1); +INSERT INTO 03006_test SETTINGS async_insert=0 VALUES ('2024-03-05', 1), ('2024-03-05', 2), ('2024-03-05', 1); +INSERT INTO 03006_test SETTINGS deduplicate_blocks_in_dependent_materialized_views=0 VALUES ('2024-03-05', 1), ('2024-03-05', 2), ('2024-03-05', 1); +INSERT INTO 03006_test SETTINGS throw_if_deduplication_in_dependent_materialized_views_enabled_with_async_insert=0 VALUES ('2024-03-05', 1), ('2024-03-05', 2), ('2024-03-05', 1); + +DROP TABLE IF EXISTS 02985_test; diff --git a/tests/queries/0_stateless/03006_parallel_replicas_prewhere.reference b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql new file mode 100644 index 00000000000..4b84646c034 --- /dev/null +++ b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql @@ -0,0 +1,31 @@ +DROP POLICY IF EXISTS url_na_log_policy0 ON url_na_log; +DROP TABLE IF EXISTS url_na_log; + +CREATE TABLE url_na_log +( + `SiteId` UInt32, + `DateVisit` Date +) +ENGINE = MergeTree +PRIMARY KEY SiteId +ORDER BY (SiteId, DateVisit) +SETTINGS index_granularity = 1000, min_bytes_for_wide_part = 0; + +CREATE ROW POLICY url_na_log_policy0 ON url_na_log FOR SELECT USING (DateVisit < '2022-08-11') OR (DateVisit > '2022-08-19') TO default; + +INSERT INTO url_na_log +SETTINGS max_insert_block_size = 200000 +SELECT + 209, + CAST('2022-08-09', 'Date') + toIntervalDay(intDiv(number, 10000)) +FROM numbers(130000) +SETTINGS max_insert_block_size = 200000; + +EXPLAIN ESTIMATE +SELECT count() +FROM url_na_log +PREWHERE (DateVisit >= toFixedString('2022-08-10', 10)) AND (DateVisit <= '2022-08-20') +SETTINGS max_block_size = 1048576, max_threads = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', max_parallel_replicas = 3, parallel_replicas_min_number_of_rows_per_replica=10000; + +DROP POLICY url_na_log_policy0 ON url_na_log; +DROP TABLE url_na_log; diff --git a/tests/queries/0_stateless/03007_column_nullable_uninitialzed_value.reference b/tests/queries/0_stateless/03007_column_nullable_uninitialzed_value.reference new file mode 100644 index 00000000000..bead7ee1474 --- /dev/null +++ b/tests/queries/0_stateless/03007_column_nullable_uninitialzed_value.reference @@ -0,0 +1,3 @@ +\N 1000 + +\N 1000 diff --git a/tests/queries/0_stateless/03007_column_nullable_uninitialzed_value.sql b/tests/queries/0_stateless/03007_column_nullable_uninitialzed_value.sql new file mode 100644 index 00000000000..9479044e0e0 --- /dev/null +++ b/tests/queries/0_stateless/03007_column_nullable_uninitialzed_value.sql @@ -0,0 +1 @@ +SELECT count(NULL) IGNORE NULLS > avg(toDecimal32(NULL)) IGNORE NULLS, count() FROM numbers(1000) WITH TOTALS SETTINGS allow_experimental_analyzer = 1; diff --git a/tests/sqllogic/connection.py b/tests/sqllogic/connection.py index 2d5e1f8a9e9..169e0f0f440 100644 --- a/tests/sqllogic/connection.py +++ b/tests/sqllogic/connection.py @@ -1,18 +1,15 @@ #!/usr/bin/env python3 # -*- coding: utf-8 -*- -import datetime -import logging -import pyodbc -import sqlite3 -import traceback import enum +import logging import random +import sqlite3 import string from contextlib import contextmanager +import pyodbc # pylint:disable=import-error; for style check from exceptions import ProgramError - logger = logging.getLogger("connection") logger.setLevel(logging.DEBUG) @@ -22,9 +19,7 @@ class OdbcConnectingArgs: self._kwargs = kwargs def __str__(self): - conn_str = ";".join( - ["{}={}".format(x, y) for x, y in self._kwargs.items() if y] - ) + conn_str = ";".join([f"{x}={y}" for x, y in self._kwargs.items() if y]) return conn_str def update_database(self, database): @@ -49,6 +44,7 @@ class OdbcConnectingArgs: for kv in conn_str.split(";"): if kv: k, v = kv.split("=", 1) + # pylint:disable-next=protected-access args._kwargs[k] = v return args @@ -63,7 +59,10 @@ def default_clickhouse_odbc_conn_str(): OdbcConnectingArgs.create_from_kw( dsn="ClickHouse DSN (ANSI)", Timeout="300", - Url="http://localhost:8123/query?default_format=ODBCDriver2&default_table_engine=MergeTree&union_default_mode=DISTINCT&group_by_use_nulls=1&join_use_nulls=1&allow_create_index_without_type=1&create_index_ignore_unique=1", + Url="http://localhost:8123/query?default_format=ODBCDriver2&" + "default_table_engine=MergeTree&union_default_mode=DISTINCT&" + "group_by_use_nulls=1&join_use_nulls=1&allow_create_index_without_type=1&" + "create_index_ignore_unique=1", ) ) @@ -82,7 +81,7 @@ class KnownDBMS(str, enum.Enum): clickhouse = "ClickHouse" -class ConnectionWrap(object): +class ConnectionWrap: def __init__(self, connection=None, factory=None, factory_kwargs=None): self._factory = factory self._factory_kwargs = factory_kwargs @@ -126,7 +125,7 @@ class ConnectionWrap(object): f"SELECT name FROM system.tables WHERE database='{self.DATABASE_NAME}'" ) elif self.DBMS_NAME == KnownDBMS.sqlite.value: - list_query = f"SELECT name FROM sqlite_master WHERE type='table'" + list_query = "SELECT name FROM sqlite_master WHERE type='table'" else: logger.warning( "unable to drop all tables for unknown database: %s", self.DBMS_NAME @@ -154,7 +153,7 @@ class ConnectionWrap(object): self._use_database(database) logger.info( "currentDatabase : %s", - execute_request(f"SELECT currentDatabase()", self).get_result(), + execute_request("SELECT currentDatabase()", self).get_result(), ) @contextmanager @@ -174,7 +173,7 @@ class ConnectionWrap(object): def __exit__(self, *args): if hasattr(self._connection, "close"): - return self._connection.close() + self._connection.close() def setup_connection(engine, conn_str=None, make_debug_request=True): @@ -263,7 +262,7 @@ class ExecResult: def assert_no_exception(self): if self.has_exception(): raise ProgramError( - f"request doesn't have a result set, it has the exception", + "request doesn't have a result set, it has the exception", parent=self._exception, ) diff --git a/tests/sqllogic/exceptions.py b/tests/sqllogic/exceptions.py index 30c8983d80f..2e4da3fd78b 100644 --- a/tests/sqllogic/exceptions.py +++ b/tests/sqllogic/exceptions.py @@ -1,8 +1,6 @@ #!/usr/bin/env python3 # -*- coding: utf-8 -*- -from enum import Enum - class Error(Exception): def __init__( @@ -45,16 +43,8 @@ class Error(Exception): @property def reason(self): - return ", ".join( - ( - str(x) - for x in [ - super().__str__(), - "details: {}".format(self._details) if self._details else "", - ] - if x - ) - ) + details = f"details: {self._details}" if self._details else "" + return ", ".join((str(x) for x in [super().__str__(), details] if x)) def set_details(self, file=None, name=None, pos=None, request=None, details=None): if file is not None: @@ -88,16 +78,8 @@ class ErrorWithParent(Error): @property def reason(self): - return ", ".join( - ( - str(x) - for x in [ - super().reason, - "exception: {}".format(str(self._parent)) if self._parent else "", - ] - if x - ) - ) + exception = f"exception: {self._parent}" if self._parent else "" + return ", ".join((str(x) for x in [super().reason, exception] if x)) class ProgramError(ErrorWithParent): diff --git a/tests/sqllogic/runner.py b/tests/sqllogic/runner.py index 5f4baf8e59b..2e8e098a099 100755 --- a/tests/sqllogic/runner.py +++ b/tests/sqllogic/runner.py @@ -2,20 +2,25 @@ # -*- coding: utf-8 -*- import argparse -import enum -import os -import logging import csv +import enum import json +import logging import multiprocessing +import os from functools import reduce -from deepdiff import DeepDiff -from connection import setup_connection, Engines, default_clickhouse_odbc_conn_str -from test_runner import TestRunner, Status, RequestType +# isort: off +from deepdiff import DeepDiff # pylint:disable=import-error; for style check +# isort: on -LEVEL_NAMES = [x.lower() for x in logging._nameToLevel.keys() if x != logging.NOTSET] +from connection import Engines, default_clickhouse_odbc_conn_str, setup_connection +from test_runner import RequestType, Status, TestRunner + +LEVEL_NAMES = [ # pylint:disable-next=protected-access + l.lower() for l, n in logging._nameToLevel.items() if n != logging.NOTSET +] def setup_logger(args): @@ -41,7 +46,7 @@ def __write_check_status(status_row, out_dir): if len(status_row) > 140: status_row = status_row[0:135] + "..." check_status_path = os.path.join(out_dir, "check_status.tsv") - with open(check_status_path, "a") as stream: + with open(check_status_path, "a", encoding="utf-8") as stream: writer = csv.writer(stream, delimiter="\t", lineterminator="\n") writer.writerow(status_row) @@ -60,7 +65,7 @@ def __write_test_result( ): all_stages = reports.keys() test_results_path = os.path.join(out_dir, "test_results.tsv") - with open(test_results_path, "a") as stream: + with open(test_results_path, "a", encoding="utf-8") as stream: writer = csv.writer(stream, delimiter="\t", lineterminator="\n") for stage in all_stages: report = reports[stage] @@ -182,7 +187,7 @@ def mode_check_statements(parser): input_dir, f"check statements:: not a dir {input_dir}" ) - reports = dict() + reports = {} out_stages_dir = os.path.join(out_dir, f"{args.mode}-stages") @@ -242,7 +247,7 @@ def mode_check_complete(parser): input_dir, f"check statements:: not a dir {input_dir}" ) - reports = dict() + reports = {} out_stages_dir = os.path.join(out_dir, f"{args.mode}-stages") @@ -286,9 +291,9 @@ def make_actual_report(reports): return {stage: report.get_map() for stage, report in reports.items()} -def write_actual_report(actial, out_dir): - with open(os.path.join(out_dir, "actual_report.json"), "w") as f: - f.write(json.dumps(actial)) +def write_actual_report(actual, out_dir): + with open(os.path.join(out_dir, "actual_report.json"), "w", encoding="utf-8") as f: + f.write(json.dumps(actual)) def read_canonic_report(input_dir): @@ -296,13 +301,15 @@ def read_canonic_report(input_dir): if not os.path.exists(file): return {} - with open(os.path.join(input_dir, "canonic_report.json"), "r") as f: + with open( + os.path.join(input_dir, "canonic_report.json"), "r", encoding="utf-8" + ) as f: data = f.read() return json.loads(data) def write_canonic_report(canonic, out_dir): - with open(os.path.join(out_dir, "canonic_report.json"), "w") as f: + with open(os.path.join(out_dir, "canonic_report.json"), "w", encoding="utf-8") as f: f.write(json.dumps(canonic)) @@ -370,7 +377,7 @@ def mode_self_test(parser): if not os.path.isdir(out_dir): raise NotADirectoryError(out_dir, f"self test: not a dir {out_dir}") - reports = dict() + reports = {} out_stages_dir = os.path.join(out_dir, f"{args.mode}-stages") diff --git a/tests/sqllogic/test_parser.py b/tests/sqllogic/test_parser.py index cb1144d7dd9..648fa9f6bf6 100755 --- a/tests/sqllogic/test_parser.py +++ b/tests/sqllogic/test_parser.py @@ -2,24 +2,27 @@ # -*- coding: utf-8 -*- import logging -import os - -from itertools import chain from enum import Enum -from hashlib import md5 from functools import reduce +from hashlib import md5 +from itertools import chain + +# isort: off +# pylint:disable=import-error; for style check import sqlglot -from sqlglot.expressions import PrimaryKeyColumnConstraint, ColumnDef +from sqlglot.expressions import ColumnDef, PrimaryKeyColumnConstraint + +# pylint:enable=import-error; for style check +# isort: on from exceptions import ( - Error, - ProgramError, - ErrorWithParent, DataResultDiffer, + Error, + ErrorWithParent, + ProgramError, QueryExecutionError, ) - logger = logging.getLogger("parser") logger.setLevel(logging.DEBUG) @@ -248,6 +251,7 @@ class FileBlockBase: ) block.with_result(result) return block + raise ValueError(f"Unknown block_type {block_type}") def dump_to(self, output): if output is None: @@ -258,9 +262,6 @@ class FileBlockBase: class FileBlockComments(FileBlockBase): - def __init__(self, parser, start, end): - super().__init__(parser, start, end) - def get_block_type(self): return BlockType.comments @@ -469,20 +470,18 @@ class QueryResult: ( str(x) for x in [ - "rows: {}".format(self.rows) if self.rows else "", - "values_count: {}".format(self.values_count) - if self.values_count - else "", - "data_hash: {}".format(self.data_hash) if self.data_hash else "", - "exception: {}".format(self.exception) if self.exception else "", - "hash_threshold: {}".format(self.hash_threshold) + f"rows: {self.rows}" if self.rows else "", + f"values_count: {self.values_count}" if self.values_count else "", + f"data_hash: {self.data_hash}" if self.data_hash else "", + f"exception: {self.exception}" if self.exception else "", + f"hash_threshold: {self.hash_threshold}" if self.hash_threshold else "", ] if x ) ) - return "QueryResult({})".format(params) + return f"QueryResult({params})" def __iter__(self): if self.rows is not None: @@ -491,12 +490,10 @@ class QueryResult: if self.values_count <= self.hash_threshold: return iter(self.rows) if self.data_hash is not None: - return iter( - [["{} values hashing to {}".format(self.values_count, self.data_hash)]] - ) + return iter([[f"{self.values_count} values hashing to {self.data_hash}"]]) if self.exception is not None: - return iter([["exception: {}".format(self.exception)]]) - raise ProgramError("Query result is empty", details="{}".format(self.__str__())) + return iter([[f"exception: {self.exception}"]]) + raise ProgramError("Query result is empty", details=str(self)) @staticmethod def __value_count(rows): @@ -528,7 +525,7 @@ class QueryResult: for row in rows: res_row = [] for c, t in zip(row, types): - logger.debug(f"Builging row. c:{c} t:{t}") + logger.debug("Builging row. c:%s t:%s", c, t) if c is None: res_row.append("NULL") continue @@ -541,7 +538,7 @@ class QueryResult: elif t == "I": try: res_row.append(str(int(c))) - except ValueError as ex: + except ValueError: # raise QueryExecutionError( # f"Got non-integer result '{c}' for I type." # ) @@ -549,7 +546,7 @@ class QueryResult: except OverflowError as ex: raise QueryExecutionError( f"Got overflowed result '{c}' for I type." - ) + ) from ex elif t == "R": res_row.append(f"{c:.3f}") @@ -567,6 +564,7 @@ class QueryResult: values = list(chain(*rows)) values.sort() return [values] if values else [] + return [] @staticmethod def __calculate_hash(rows): @@ -595,9 +593,9 @@ class QueryResult: # do not print details to the test file # but print original exception if isinstance(e, ErrorWithParent): - message = "{}, original is: {}".format(e, e.get_parent()) + message = f"{e}, original is: {e.get_parent()}" else: - message = "{}".format(e) + message = str(e) return QueryResult(exception=message) @@ -616,9 +614,8 @@ class QueryResult: "canonic and actual results have different exceptions", details=f"canonic: {canonic.exception}, actual: {actual.exception}", ) - else: - # exceptions are the same - return + # exceptions are the same + return elif canonic.exception is not None: raise DataResultDiffer( "canonic result has exception and actual result doesn't", @@ -639,9 +636,8 @@ class QueryResult: if canonic.values_count != actual.values_count: raise DataResultDiffer( "canonic and actual results have different value count", - details="canonic values count {}, actual {}".format( - canonic.values_count, actual.values_count - ), + details=f"canonic values count {canonic.values_count}, " + f"actual {actual.values_count}", ) if canonic.data_hash != actual.data_hash: raise DataResultDiffer( @@ -653,9 +649,8 @@ class QueryResult: if canonic.values_count != actual.values_count: raise DataResultDiffer( "canonic and actual results have different value count", - details="canonic values count {}, actual {}".format( - canonic.values_count, actual.values_count - ), + details=f"canonic values count {canonic.values_count}, " + f"actual {actual.values_count}", ) if canonic.rows != actual.rows: raise DataResultDiffer( @@ -665,5 +660,5 @@ class QueryResult: raise ProgramError( "Unable to compare results", - details="actual {}, canonic {}".format(actual, canonic), + details=f"actual {actual}, canonic {canonic}", ) diff --git a/tests/sqllogic/test_runner.py b/tests/sqllogic/test_runner.py index baec0dc7924..8f2242a45b9 100644 --- a/tests/sqllogic/test_runner.py +++ b/tests/sqllogic/test_runner.py @@ -1,25 +1,23 @@ #!/usr/bin/env python3 # -*- coding: utf-8 -*- import enum -import logging -import os -import traceback import io import json +import logging +import os import test_parser +from connection import execute_request from exceptions import ( + DataResultDiffer, Error, ProgramError, - DataResultDiffer, - StatementExecutionError, - StatementSuccess, QueryExecutionError, QuerySuccess, SchemeResultDiffer, + StatementExecutionError, + StatementSuccess, ) -from connection import execute_request - logger = logging.getLogger("parser") logger.setLevel(logging.DEBUG) @@ -55,6 +53,7 @@ class Status(str, enum.Enum): class TestStatus: def __init__(self): + self.name = None self.status = None self.file = None self.position = None @@ -155,7 +154,7 @@ class SimpleStats: self.success += 1 def get_map(self): - result = dict() + result = {} result["success"] = self.success result["fail"] = self.fail return result @@ -187,7 +186,7 @@ class Stats: choose.update(status) def get_map(self): - result = dict() + result = {} result["statements"] = self.statements.get_map() result["queries"] = self.queries.get_map() result["total"] = self.total.get_map() @@ -205,7 +204,7 @@ class OneReport: self.test_name = test_name self.test_file = test_file self.stats = Stats() - self.requests = dict() # type: dict(int, TestStatus) + self.requests = {} def update(self, status): if not isinstance(status, TestStatus): @@ -218,11 +217,11 @@ class OneReport: return str(self.get_map()) def get_map(self): - result = dict() + result = {} result["test_name"] = self.test_name result["test_file"] = self.test_file result["stats"] = self.stats.get_map() - result["requests"] = dict() + result["requests"] = {} requests = result["requests"] for pos, status in self.requests.items(): requests[pos] = status.get_map() @@ -233,7 +232,7 @@ class Report: def __init__(self, dbms_name, input_dir=None): self.dbms_name = dbms_name self.stats = Stats() - self.tests = dict() # type: dict(str, OneReport) + self.tests = {} self.input_dir = input_dir self.output_dir = None @@ -256,7 +255,7 @@ class Report: self.output_dir = res_dir def get_map(self): - result = dict() + result = {} result["dbms_name"] = self.dbms_name result["stats"] = self.stats.get_map() result["input_dir"] = self.input_dir @@ -264,7 +263,7 @@ class Report: result["input_dir"] = self.input_dir if self.output_dir is not None: result["output_dir"] = self.output_dir - result["tests"] = dict() + result["tests"] = {} tests = result["tests"] for test_name, one_report in self.tests.items(): tests.update({test_name: one_report.get_map()}) @@ -297,8 +296,8 @@ class Report: def write_report(self, report_dir): report_path = os.path.join(report_dir, "report.json") - logger.info(f"create file {report_path}") - with open(report_path, "w") as stream: + logger.info("create file %s", report_path) + with open(report_path, "w", encoding="utf-8") as stream: stream.write(json.dumps(self.get_map(), indent=4)) @@ -434,38 +433,34 @@ class TestRunner: details=f"expected error: {expected_error}", parent=exec_res.get_exception(), ) - else: - clogger.debug("errors matched") - raise QuerySuccess() - else: - clogger.debug("missed error") - raise QueryExecutionError( - "query is expected to fail with error", - details="expected error: {}".format(expected_error), + clogger.debug("errors matched") + raise QuerySuccess() + clogger.debug("missed error") + raise QueryExecutionError( + "query is expected to fail with error", + details=f"expected error: {expected_error}", + ) + clogger.debug("success is expected") + if exec_res.has_exception(): + clogger.debug("had error") + if self.verify: + clogger.debug("verify mode") + canonic = test_parser.QueryResult.parse_it( + block.get_result(), 10 ) - else: - clogger.debug("success is expected") - if exec_res.has_exception(): - clogger.debug("had error") - if self.verify: - clogger.debug("verify mode") - canonic = test_parser.QueryResult.parse_it( - block.get_result(), 10 - ) - exception = QueryExecutionError( - "query execution failed with an exception", - parent=exec_res.get_exception(), - ) - actual = test_parser.QueryResult.as_exception(exception) - test_parser.QueryResult.assert_eq(canonic, actual) - block.with_result(actual) - raise QuerySuccess() - else: - clogger.debug("completion mode") - raise QueryExecutionError( - "query execution failed with an exception", - parent=exec_res.get_exception(), - ) + exception = QueryExecutionError( + "query execution failed with an exception", + parent=exec_res.get_exception(), + ) + actual = test_parser.QueryResult.as_exception(exception) + test_parser.QueryResult.assert_eq(canonic, actual) + block.with_result(actual) + raise QuerySuccess() + clogger.debug("completion mode") + raise QueryExecutionError( + "query execution failed with an exception", + parent=exec_res.get_exception(), + ) canonic_types = block.get_types() clogger.debug("canonic types %s", canonic_types) @@ -476,9 +471,8 @@ class TestRunner: if canonic_columns_count != actual_columns_count: raise SchemeResultDiffer( "canonic and actual columns count differ", - details="expected columns {}, actual columns {}".format( - canonic_columns_count, actual_columns_count - ), + details=f"expected columns {canonic_columns_count}, " + f"actual columns {actual_columns_count}", ) actual = test_parser.QueryResult.make_it( @@ -528,7 +522,7 @@ class TestRunner: self.report = Report(self.dbms_name, self._input_dir) if self.results is None: - self.results = dict() + self.results = {} if self.dbms_name == "ClickHouse" and test_name in [ "test/select5.test", @@ -536,7 +530,7 @@ class TestRunner: "test/evidence/slt_lang_replace.test", "test/evidence/slt_lang_droptrigger.test", ]: - logger.info(f"Let's skip test %s for ClickHouse", test_name) + logger.info("Let's skip test %s for ClickHouse", test_name) return with self.connection.with_one_test_scope(): @@ -565,7 +559,7 @@ class TestRunner: test_name = os.path.relpath(test_file, start=self._input_dir) logger.debug("open file %s", test_name) - with open(test_file, "r") as stream: + with open(test_file, "r", encoding="utf-8") as stream: self.run_one_test(stream, test_name, test_file) def run_all_tests_from_dir(self, input_dir): @@ -582,10 +576,10 @@ class TestRunner: for test_name, stream in self.results.items(): test_file = os.path.join(dir_path, test_name) - logger.info(f"create file {test_file}") + logger.info("create file %s", test_file) result_dir = os.path.dirname(test_file) os.makedirs(result_dir, exist_ok=True) - with open(test_file, "w") as output: + with open(test_file, "w", encoding="utf-8") as output: output.write(stream.getvalue()) def write_report(self, report_dir): diff --git a/utils/check-style/aspell-ignore/en/aspell-dict.txt b/utils/check-style/aspell-ignore/en/aspell-dict.txt index 3614bcb7452..8bf12247ef2 100644 --- a/utils/check-style/aspell-ignore/en/aspell-dict.txt +++ b/utils/check-style/aspell-ignore/en/aspell-dict.txt @@ -2545,6 +2545,7 @@ toRelativeSecondNum toRelativeWeekNum toRelativeYearNum toSecond +toMillisecond toStartOfDay toStartOfFifteenMinutes toStartOfFiveMinutes diff --git a/utils/check-style/check-pylint b/utils/check-style/check-pylint new file mode 100755 index 00000000000..7959a414023 --- /dev/null +++ b/utils/check-style/check-pylint @@ -0,0 +1,15 @@ +#!/usr/bin/env bash + + +LC_ALL="en_US.UTF-8" +ROOT_PATH=$(git rev-parse --show-toplevel) + +function xargs-pylint { + # $1 is number maximum arguments per pylint process + sort | awk '$2=="text/x-script.python" {print $1}' | \ + xargs -P "$(nproc)" -n "$1" pylint --rcfile="$ROOT_PATH/pyproject.toml" --persistent=no --score=n +} + +find "$ROOT_PATH/tests" -maxdepth 2 -type f -exec file -F' ' --mime-type {} + | xargs-pylint 50 +# Beware, there lambdas are checked. All of them contain `app`, and it causes brain-cucumber-zalgo +find "$ROOT_PATH/tests/ci" -mindepth 2 -type f -exec file -F' ' --mime-type {} + | xargs-pylint 1 diff --git a/utils/check-style/check-style b/utils/check-style/check-style index 3a966daea41..d7387c3f843 100755 --- a/utils/check-style/check-style +++ b/utils/check-style/check-style @@ -152,9 +152,6 @@ find $ROOT_PATH/{src,base,programs,utils} -name '*.xml' | grep -vP $EXCLUDE_DIRS | xargs xmllint --noout --nonet -# FIXME: for now only clickhouse-test -pylint --rcfile=$ROOT_PATH/.pylintrc --persistent=no --score=n $ROOT_PATH/tests/clickhouse-test $ROOT_PATH/tests/ci/*.py - find $ROOT_PATH -not -path $ROOT_PATH'/contrib*' \( -name '*.yaml' -or -name '*.yml' \) -type f | grep -vP $EXCLUDE_DIRS | xargs yamllint --config-file=$ROOT_PATH/.yamllint diff --git a/utils/check-style/check_cpp_docs.sh b/utils/check-style/check_cpp_docs.sh new file mode 100755 index 00000000000..7ad3cede758 --- /dev/null +++ b/utils/check-style/check_cpp_docs.sh @@ -0,0 +1,34 @@ +#!/bin/bash + +# yaml check is not the best one + +cd /ClickHouse/utils/check-style || echo -e "failure\tRepo not found" > /test_output/check_status.tsv + +# FIXME: 30 sec to wait +# echo "Check duplicates" | ts +# ./check-duplicate-includes.sh |& tee /test_output/duplicate_includes_output.txt + +echo "Check style" | ts +./check-style -n |& tee /test_output/style_output.txt +echo "Check typos" | ts +./check-typos |& tee /test_output/typos_output.txt +echo "Check docs spelling" | ts +./check-doc-aspell |& tee /test_output/docs_spelling_output.txt +echo "Check whitespaces" | ts +./check-whitespaces -n |& tee /test_output/whitespaces_output.txt +echo "Check workflows" | ts +./check-workflows |& tee /test_output/workflows_output.txt +echo "Check submodules" | ts +./check-submodules |& tee /test_output/submodules_output.txt +echo "Check style. Done" | ts + +# FIXME: 6 min to wait +# echo "Check shell scripts with shellcheck" | ts +# ./shellcheck-run.sh |& tee /test_output/shellcheck_output.txt + + +# FIXME: move out +# /process_style_check_result.py || echo -e "failure\tCannot parse results" > /test_output/check_status.tsv +# echo "Check help for changelog generator works" | ts +# cd ../changelog || exit 1 +# ./changelog.py -h 2>/dev/null 1>&2 diff --git a/utils/check-style/check_py.sh b/utils/check-style/check_py.sh new file mode 100755 index 00000000000..48c02013734 --- /dev/null +++ b/utils/check-style/check_py.sh @@ -0,0 +1,17 @@ +#!/bin/bash + +# yaml check is not the best one + +cd /ClickHouse/utils/check-style || echo -e "failure\tRepo not found" > /test_output/check_status.tsv + +# FIXME: 1 min to wait + head checkout +# echo "Check python formatting with black" | ts +# ./check-black -n |& tee /test_output/black_output.txt + +echo "Check pylint" | ts +./check-pylint -n |& tee /test_output/pylint_output.txt +echo "Check pylint. Done" | ts + +echo "Check python type hinting with mypy" | ts +./check-mypy -n |& tee /test_output/mypy_output.txt +echo "Check python type hinting with mypy. Done" | ts diff --git a/docker/test/style/process_style_check_result.py b/utils/check-style/process_style_check_result.py similarity index 96% rename from docker/test/style/process_style_check_result.py rename to utils/check-style/process_style_check_result.py index bc06df1af31..7980c01dd37 100755 --- a/docker/test/style/process_style_check_result.py +++ b/utils/check-style/process_style_check_result.py @@ -13,10 +13,11 @@ def process_result(result_folder): description = "" test_results = [] checks = ( - "duplicate includes", - "shellcheck", + #"duplicate includes", + #"shellcheck", "style", - "black", + "pylint", + #"black", "mypy", "typos", "whitespaces", diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 572ceddf590..e372e407ce1 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -25,6 +25,7 @@ v23.9.4.11-stable 2023-11-08 v23.9.3.12-stable 2023-10-31 v23.9.2.56-stable 2023-10-19 v23.9.1.1854-stable 2023-09-29 +v23.8.10.43-lts 2024-03-05 v23.8.9.54-lts 2024-01-05 v23.8.8.20-lts 2023-11-25 v23.8.7.24-lts 2023-11-17 @@ -54,6 +55,7 @@ v23.4.4.16-stable 2023-06-17 v23.4.3.48-stable 2023-06-12 v23.4.2.11-stable 2023-05-02 v23.4.1.1943-stable 2023-04-27 +v23.3.20.27-lts 2024-03-05 v23.3.19.32-lts 2024-01-05 v23.3.18.15-lts 2023-11-25 v23.3.17.13-lts 2023-11-17 diff --git a/utils/memcpy-bench/FastMemcpy.h b/utils/memcpy-bench/FastMemcpy.h index 85d09c5f53e..650a6761771 100644 --- a/utils/memcpy-bench/FastMemcpy.h +++ b/utils/memcpy-bench/FastMemcpy.h @@ -33,9 +33,11 @@ #endif #endif +/// NOLINTBEGIN(modernize-use-using) typedef __attribute__((__aligned__(1))) uint16_t uint16_unaligned_t; typedef __attribute__((__aligned__(1))) uint32_t uint32_unaligned_t; typedef __attribute__((__aligned__(1))) uint64_t uint64_unaligned_t; +/// NOLINTEND(modernize-use-using) //--------------------------------------------------------------------- // fast copy for different sizes @@ -98,7 +100,7 @@ __attribute__((__no_sanitize__("undefined"))) inline void *memcpy_tiny(void * __ unsigned char *dd = ((unsigned char*)dst) + size; const unsigned char *ss = ((const unsigned char*)src) + size; - switch (size) + switch (size) /// NOLINT(bugprone-switch-missing-default-case) { case 64: memcpy_sse2_64(dd - 64, ss - 64); @@ -652,7 +654,7 @@ __attribute__((__no_sanitize__("undefined"))) inline void *memcpy_tiny(void * __ //--------------------------------------------------------------------- // main routine //--------------------------------------------------------------------- -void* memcpy_fast_sse(void * __restrict destination, const void * __restrict source, size_t size) +void* memcpy_fast_sse(void * __restrict destination, const void * __restrict source, size_t size) /// NOLINT(misc-definitions-in-headers) { unsigned char *dst = (unsigned char*)destination; const unsigned char *src = (const unsigned char*)source; diff --git a/utils/memcpy-bench/FastMemcpy_Avx.h b/utils/memcpy-bench/FastMemcpy_Avx.h index ee7d4e19536..aecf8abbe03 100644 --- a/utils/memcpy-bench/FastMemcpy_Avx.h +++ b/utils/memcpy-bench/FastMemcpy_Avx.h @@ -103,7 +103,7 @@ static INLINE void *memcpy_tiny_avx(void * __restrict dst, const void * __restri unsigned char *dd = reinterpret_cast(dst) + size; const unsigned char *ss = reinterpret_cast(src) + size; - switch (size) + switch (size) /// NOLINT(bugprone-switch-missing-default-case) { case 128: memcpy_avx_128(dd - 128, ss - 128); [[fallthrough]]; case 0: break; @@ -371,7 +371,7 @@ static INLINE void *memcpy_tiny_avx(void * __restrict dst, const void * __restri //--------------------------------------------------------------------- // main routine //--------------------------------------------------------------------- -void* memcpy_fast_avx(void * __restrict destination, const void * __restrict source, size_t size) +void* memcpy_fast_avx(void * __restrict destination, const void * __restrict source, size_t size) /// NOLINT(misc-definitions-in-headers) { unsigned char *dst = reinterpret_cast(destination); const unsigned char *src = reinterpret_cast(source);