thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-09-20 08:40:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'origin/master' into fix-grouping-for-grouping-sets

Browse Source
This commit is contained in:
Dmitry Novik 2023-03-14 12:01:51 +00:00
commit ae3d30a736
511 changed files with 5916 additions and 2544 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -14,7 +14,7 @@ curl https://clickhouse.com/ | sh
* [Tutorial](https://clickhouse.com/docs/en/getting_started/tutorial/) shows how to set up and query a small ClickHouse cluster.
* [Documentation](https://clickhouse.com/docs/en/) provides more in-depth information.
* [YouTube channel](https://www.youtube.com/c/ClickHouseDB) has a lot of content about ClickHouse in video format.
* [Slack](https://join.slack.com/t/clickhousedb/shared_invite/zt-1gh9ds7f4-PgDhJAaF8ad5RbWBAAjzFg) and [Telegram](https://telegram.me/clickhouse_en) allow chatting with ClickHouse users in real-time.
* [Slack](https://clickhouse.com/slack) and [Telegram](https://telegram.me/clickhouse_en) allow chatting with ClickHouse users in real-time.
* [Blog](https://clickhouse.com/blog/) contains various ClickHouse-related articles, as well as announcements and reports about events.
* [Code Browser (Woboq)](https://clickhouse.com/codebrowser/ClickHouse/index.html) with syntax highlight and navigation.
* [Code Browser (github.dev)](https://github.dev/ClickHouse/ClickHouse) with syntax highlight, powered by github.dev.

214
base/base/hex.h Normal file
View File

@ -0,0 +1,214 @@
#pragma once
#include <cstring>
#include "types.h"
/// Maps 0..15 to 0..9A..F or 0..9a..f correspondingly.
constexpr inline std::string_view hex_digit_to_char_uppercase_table = "0123456789ABCDEF";
constexpr inline std::string_view hex_digit_to_char_lowercase_table = "0123456789abcdef";
constexpr char hexDigitUppercase(unsigned char c)
{
return hex_digit_to_char_uppercase_table[c];
}
constexpr char hexDigitLowercase(unsigned char c)
{
return hex_digit_to_char_lowercase_table[c];
}
/// Maps 0..255 to 00..FF or 00..ff correspondingly
constexpr inline std::string_view hex_byte_to_char_uppercase_table = //
"000102030405060708090A0B0C0D0E0F"
"101112131415161718191A1B1C1D1E1F"
"202122232425262728292A2B2C2D2E2F"
"303132333435363738393A3B3C3D3E3F"
"404142434445464748494A4B4C4D4E4F"
"505152535455565758595A5B5C5D5E5F"
"606162636465666768696A6B6C6D6E6F"
"707172737475767778797A7B7C7D7E7F"
"808182838485868788898A8B8C8D8E8F"
"909192939495969798999A9B9C9D9E9F"
"A0A1A2A3A4A5A6A7A8A9AAABACADAEAF"
"B0B1B2B3B4B5B6B7B8B9BABBBCBDBEBF"
"C0C1C2C3C4C5C6C7C8C9CACBCCCDCECF"
"D0D1D2D3D4D5D6D7D8D9DADBDCDDDEDF"
"E0E1E2E3E4E5E6E7E8E9EAEBECEDEEEF"
"F0F1F2F3F4F5F6F7F8F9FAFBFCFDFEFF";
constexpr inline std::string_view hex_byte_to_char_lowercase_table = //
"000102030405060708090a0b0c0d0e0f"
"101112131415161718191a1b1c1d1e1f"
"202122232425262728292a2b2c2d2e2f"
"303132333435363738393a3b3c3d3e3f"
"404142434445464748494a4b4c4d4e4f"
"505152535455565758595a5b5c5d5e5f"
"606162636465666768696a6b6c6d6e6f"
"707172737475767778797a7b7c7d7e7f"
"808182838485868788898a8b8c8d8e8f"
"909192939495969798999a9b9c9d9e9f"
"a0a1a2a3a4a5a6a7a8a9aaabacadaeaf"
"b0b1b2b3b4b5b6b7b8b9babbbcbdbebf"
"c0c1c2c3c4c5c6c7c8c9cacbcccdcecf"
"d0d1d2d3d4d5d6d7d8d9dadbdcdddedf"
"e0e1e2e3e4e5e6e7e8e9eaebecedeeef"
"f0f1f2f3f4f5f6f7f8f9fafbfcfdfeff";
inline void writeHexByteUppercase(UInt8 byte, void * out)
{
memcpy(out, &hex_byte_to_char_uppercase_table[static_cast<size_t>(byte) * 2], 2);
}
inline void writeHexByteLowercase(UInt8 byte, void * out)
{
memcpy(out, &hex_byte_to_char_lowercase_table[static_cast<size_t>(byte) * 2], 2);
}
constexpr inline std::string_view bin_byte_to_char_table = //
"0000000000000001000000100000001100000100000001010000011000000111"
"0000100000001001000010100000101100001100000011010000111000001111"
"0001000000010001000100100001001100010100000101010001011000010111"
"0001100000011001000110100001101100011100000111010001111000011111"
"0010000000100001001000100010001100100100001001010010011000100111"
"0010100000101001001010100010101100101100001011010010111000101111"
"0011000000110001001100100011001100110100001101010011011000110111"
"0011100000111001001110100011101100111100001111010011111000111111"
"0100000001000001010000100100001101000100010001010100011001000111"
"0100100001001001010010100100101101001100010011010100111001001111"
"0101000001010001010100100101001101010100010101010101011001010111"
"0101100001011001010110100101101101011100010111010101111001011111"
"0110000001100001011000100110001101100100011001010110011001100111"
"0110100001101001011010100110101101101100011011010110111001101111"
"0111000001110001011100100111001101110100011101010111011001110111"
"0111100001111001011110100111101101111100011111010111111001111111"
"1000000010000001100000101000001110000100100001011000011010000111"
"1000100010001001100010101000101110001100100011011000111010001111"
"1001000010010001100100101001001110010100100101011001011010010111"
"1001100010011001100110101001101110011100100111011001111010011111"
"1010000010100001101000101010001110100100101001011010011010100111"
"1010100010101001101010101010101110101100101011011010111010101111"
"1011000010110001101100101011001110110100101101011011011010110111"
"1011100010111001101110101011101110111100101111011011111010111111"
"1100000011000001110000101100001111000100110001011100011011000111"
"1100100011001001110010101100101111001100110011011100111011001111"
"1101000011010001110100101101001111010100110101011101011011010111"
"1101100011011001110110101101101111011100110111011101111011011111"
"1110000011100001111000101110001111100100111001011110011011100111"
"1110100011101001111010101110101111101100111011011110111011101111"
"1111000011110001111100101111001111110100111101011111011011110111"
"1111100011111001111110101111101111111100111111011111111011111111";
inline void writeBinByte(UInt8 byte, void * out)
{
memcpy(out, &bin_byte_to_char_table[static_cast<size_t>(byte) * 8], 8);
}
/// Produces hex representation of an unsigned int with leading zeros (for checksums)
template <typename TUInt>
inline void writeHexUIntImpl(TUInt uint_, char * out, std::string_view table)
{
union
{
TUInt value;
UInt8 uint8[sizeof(TUInt)];
};
value = uint_;
for (size_t i = 0; i < sizeof(TUInt); ++i)
{
if constexpr (std::endian::native == std::endian::little)
memcpy(out + i * 2, &table[static_cast<size_t>(uint8[sizeof(TUInt) - 1 - i]) * 2], 2);
else
memcpy(out + i * 2, &table[static_cast<size_t>(uint8[i]) * 2], 2);
}
}
template <typename TUInt>
inline void writeHexUIntUppercase(TUInt uint_, char * out)
{
writeHexUIntImpl(uint_, out, hex_byte_to_char_uppercase_table);
}
template <typename TUInt>
inline void writeHexUIntLowercase(TUInt uint_, char * out)
{
writeHexUIntImpl(uint_, out, hex_byte_to_char_lowercase_table);
}
template <typename TUInt>
std::string getHexUIntUppercase(TUInt uint_)
{
std::string res(sizeof(TUInt) * 2, '\0');
writeHexUIntUppercase(uint_, res.data());
return res;
}
template <typename TUInt>
std::string getHexUIntLowercase(TUInt uint_)
{
std::string res(sizeof(TUInt) * 2, '\0');
writeHexUIntLowercase(uint_, res.data());
return res;
}
/// Maps 0..9, A..F, a..f to 0..15. Other chars are mapped to implementation specific value.
constexpr inline std::string_view hex_char_to_digit_table
= {"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\xff\xff\xff\xff\xff\xff" //0-9
"\xff\x0a\x0b\x0c\x0d\x0e\x0f\xff\xff\xff\xff\xff\xff\xff\xff\xff" //A-Z
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\x0a\x0b\x0c\x0d\x0e\x0f\xff\xff\xff\xff\xff\xff\xff\xff\xff" //a-z
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff",
256};
constexpr UInt8 unhex(char c)
{
return hex_char_to_digit_table[static_cast<UInt8>(c)];
}
constexpr UInt8 unhex2(const char * data)
{
return static_cast<UInt8>(unhex(data[0])) * 0x10 + static_cast<UInt8>(unhex(data[1]));
}
constexpr UInt16 unhex4(const char * data)
{
return static_cast<UInt16>(unhex(data[0])) * 0x1000 + static_cast<UInt16>(unhex(data[1])) * 0x100
+ static_cast<UInt16>(unhex(data[2])) * 0x10 + static_cast<UInt16>(unhex(data[3]));
}
template <typename TUInt>
constexpr TUInt unhexUInt(const char * data)
{
TUInt res = 0;
if constexpr ((sizeof(TUInt) <= 8) || ((sizeof(TUInt) % 8) != 0))
{
for (size_t i = 0; i < sizeof(TUInt) * 2; ++i, ++data)
{
res <<= 4;
res += unhex(*data);
}
}
else
{
for (size_t i = 0; i < sizeof(TUInt) / 8; ++i, data += 16)
{
res <<= 64;
res += unhexUInt<UInt64>(data);
}
}
return res;
}

13
base/base/interpolate.h Normal file
View File

@ -0,0 +1,13 @@
#pragma once
#include <cassert>
#include <cmath>
/** Linear interpolation in logarithmic coordinates.
* Exponential interpolation is related to linear interpolation
* exactly in same way as geometric mean is related to arithmetic mean.
*/
constexpr double interpolateExponential(double min, double max, double ratio)
{
assert(min > 0 && ratio >= 0 && ratio <= 1);
return min * std::pow(max / min, ratio);
}

2
base/poco/Foundation/include/Poco/Message.h
View File

@ -105,6 +105,8 @@ public:
const std::string & getText() const;
/// Returns the text of the message.
void appendText(const std::string & text);
void setPriority(Priority prio);
/// Sets the priority of the message.

6
base/poco/Foundation/src/Message.cpp
View File

@ -157,6 +157,12 @@ void Message::setText(const std::string& text)
}
void Message::appendText(const std::string & text)
{
_text.append(text);
}
void Message::setPriority(Priority prio)
{
_prio = prio;

20
contrib/arrow-cmake/CMakeLists.txt
View File

@ -115,6 +115,13 @@ configure_file("${ORC_SOURCE_SRC_DIR}/Adaptor.hh.in" "${ORC_BUILD_INCLUDE_DIR}/A
# ARROW_ORC + adapters/orc/CMakefiles
set(ORC_SRCS
"${CMAKE_CURRENT_BINARY_DIR}/orc_proto.pb.h"
"${ORC_SOURCE_SRC_DIR}/sargs/ExpressionTree.cc"
"${ORC_SOURCE_SRC_DIR}/sargs/Literal.cc"
"${ORC_SOURCE_SRC_DIR}/sargs/PredicateLeaf.cc"
"${ORC_SOURCE_SRC_DIR}/sargs/SargsApplier.cc"
"${ORC_SOURCE_SRC_DIR}/sargs/SearchArgument.cc"
"${ORC_SOURCE_SRC_DIR}/sargs/TruthValue.cc"
"${ORC_SOURCE_SRC_DIR}/Exceptions.cc"
"${ORC_SOURCE_SRC_DIR}/OrcFile.cc"
"${ORC_SOURCE_SRC_DIR}/Reader.cc"
@ -129,13 +136,20 @@ set(ORC_SRCS
"${ORC_SOURCE_SRC_DIR}/MemoryPool.cc"
"${ORC_SOURCE_SRC_DIR}/RLE.cc"
"${ORC_SOURCE_SRC_DIR}/RLEv1.cc"
"${ORC_SOURCE_SRC_DIR}/RLEv2.cc"
"${ORC_SOURCE_SRC_DIR}/RleDecoderV2.cc"
"${ORC_SOURCE_SRC_DIR}/RleEncoderV2.cc"
"${ORC_SOURCE_SRC_DIR}/RLEV2Util.cc"
"${ORC_SOURCE_SRC_DIR}/Statistics.cc"
"${ORC_SOURCE_SRC_DIR}/StripeStream.cc"
"${ORC_SOURCE_SRC_DIR}/Timezone.cc"
"${ORC_SOURCE_SRC_DIR}/TypeImpl.cc"
"${ORC_SOURCE_SRC_DIR}/Vector.cc"
"${ORC_SOURCE_SRC_DIR}/Writer.cc"
"${ORC_SOURCE_SRC_DIR}/Adaptor.cc"
"${ORC_SOURCE_SRC_DIR}/BloomFilter.cc"
"${ORC_SOURCE_SRC_DIR}/Murmur3.cc"
"${ORC_SOURCE_SRC_DIR}/BlockBuffer.cc"
"${ORC_SOURCE_SRC_DIR}/wrap/orc-proto-wrapper.cc"
"${ORC_SOURCE_SRC_DIR}/io/InputStream.cc"
"${ORC_SOURCE_SRC_DIR}/io/OutputStream.cc"
"${ORC_ADDITION_SOURCE_DIR}/orc_proto.pb.cc"
@ -358,6 +372,9 @@ SET(ARROW_SRCS "${LIBRARY_DIR}/util/compression_zlib.cc" ${ARROW_SRCS})
add_definitions(-DARROW_WITH_ZSTD)
SET(ARROW_SRCS "${LIBRARY_DIR}/util/compression_zstd.cc" ${ARROW_SRCS})
add_definitions(-DARROW_WITH_BROTLI)
SET(ARROW_SRCS "${LIBRARY_DIR}/util/compression_brotli.cc" ${ARROW_SRCS})
add_library(_arrow ${ARROW_SRCS})
@ -372,6 +389,7 @@ target_link_libraries(_arrow PRIVATE
ch_contrib::snappy
ch_contrib::zlib
ch_contrib::zstd
ch_contrib::brotli
)
target_link_libraries(_arrow PUBLIC _orc)

2
contrib/orc vendored

@ -1 +1 @@
Subproject commit f9a393ed2433a60034795284f82d093b348f2102
Subproject commit c5d7755ba0b9a95631c8daea4d094101f26ec761

2
docker/keeper/Dockerfile
View File

@ -29,7 +29,7 @@ RUN arch=${TARGETARCH:-amd64} \
esac
ARG REPOSITORY="https://s3.amazonaws.com/clickhouse-builds/22.4/31c367d3cd3aefd316778601ff6565119fe36682/package_release"
ARG VERSION="23.2.3.17"
ARG VERSION="23.2.4.12"
ARG PACKAGES="clickhouse-keeper"
# user/group precreated explicitly with fixed uid/gid on purpose.

2
docker/server/Dockerfile.alpine
View File

@ -33,7 +33,7 @@ RUN arch=${TARGETARCH:-amd64} \
# lts / testing / prestable / etc
ARG REPO_CHANNEL="stable"
ARG REPOSITORY="https://packages.clickhouse.com/tgz/${REPO_CHANNEL}"
ARG VERSION="23.2.3.17"
ARG VERSION="23.2.4.12"
ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static"
# user/group precreated explicitly with fixed uid/gid on purpose.

2
docker/server/Dockerfile.ubuntu
View File

@ -22,7 +22,7 @@ RUN sed -i "s|http://archive.ubuntu.com|${apt_archive}|g" /etc/apt/sources.list
ARG REPO_CHANNEL="stable"
ARG REPOSITORY="deb https://packages.clickhouse.com/deb ${REPO_CHANNEL} main"
ARG VERSION="23.2.3.17"
ARG VERSION="23.2.4.12"
ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static"
# set non-empty deb_location_url url to create a docker image

11
docker/test/upgrade/run.sh
View File

@ -60,6 +60,13 @@ install_packages previous_release_package_folder
export USE_S3_STORAGE_FOR_MERGE_TREE=1
# Previous version may not be ready for fault injections
export ZOOKEEPER_FAULT_INJECTION=0
# force_sync=false doesn't work correctly on some older versions
sudo cat /etc/clickhouse-server/config.d/keeper_port.xml \
| sed "s|<force_sync>false</force_sync>|<force_sync>true</force_sync>|" \
> /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
configure
# But we still need default disk because some tables loaded only into it
@ -161,7 +168,9 @@ rg -Fav -e "Code: 236. DB::Exception: Cancelled merging parts" \
-e "Authentication failed" \
-e "Cannot flush" \
-e "Container already exists" \
/var/log/clickhouse-server/clickhouse-server.upgrade.log | zgrep -Fa "<Error>" > /test_output/upgrade_error_messages.txt \
clickhouse-server.upgrade.log \
| grep -av -e "_repl_01111_.*Mapping for table with UUID" \
| zgrep -Fa "<Error>" > /test_output/upgrade_error_messages.txt \
&& echo -e "Error message in clickhouse-server.log (see upgrade_error_messages.txt)$FAIL$(head_escaped /test_output/upgrade_error_messages.txt)" \
>> /test_output/test_results.tsv \
|| echo -e "No Error messages after server upgrade$OK" >> /test_output/test_results.tsv

29
docs/changelogs/v22.12.5.34-stable.md Normal file
View File

@ -0,0 +1,29 @@
---
sidebar_position: 1
sidebar_label: 2023
---
# 2023 Changelog
### ClickHouse release v22.12.5.34-stable (b82d6401ca1) FIXME as compared to v22.12.4.76-stable (cb5772db805)
#### Improvement
* Backported in [#46983](https://github.com/ClickHouse/ClickHouse/issues/46983): - Apply `ALTER TABLE table_name ON CLUSTER cluster MOVE PARTITION|PART partition_expr TO DISK|VOLUME 'disk_name'` to all replicas. Because `ALTER TABLE t MOVE` is not replicated. [#46402](https://github.com/ClickHouse/ClickHouse/pull/46402) ([lizhuoyu5](https://github.com/lzydmxy)).
#### Bug Fix (user-visible misbehavior in official stable or prestable release)
* Backported in [#45729](https://github.com/ClickHouse/ClickHouse/issues/45729): Fix key description when encountering duplicate primary keys. This can happen in projections. See [#45590](https://github.com/ClickHouse/ClickHouse/issues/45590) for details. [#45686](https://github.com/ClickHouse/ClickHouse/pull/45686) ([Amos Bird](https://github.com/amosbird)).
* Backported in [#46398](https://github.com/ClickHouse/ClickHouse/issues/46398): Fix `SYSTEM UNFREEZE` queries failing with the exception `CANNOT_PARSE_INPUT_ASSERTION_FAILED`. [#46325](https://github.com/ClickHouse/ClickHouse/pull/46325) ([Aleksei Filatov](https://github.com/aalexfvk)).
* Backported in [#46903](https://github.com/ClickHouse/ClickHouse/issues/46903): - Fix incorrect alias recursion in QueryNormalizer. [#46609](https://github.com/ClickHouse/ClickHouse/pull/46609) ([Raúl Marín](https://github.com/Algunenano)).
* Backported in [#47210](https://github.com/ClickHouse/ClickHouse/issues/47210): `INSERT` queries through native TCP protocol and HTTP protocol were not canceled correctly in some cases. It could lead to a partially applied query if a client canceled the query, or if a client died or, in rare cases, on network errors. As a result, it could lead to not working deduplication. Fixes [#27667](https://github.com/ClickHouse/ClickHouse/issues/27667) and [#45377](https://github.com/ClickHouse/ClickHouse/issues/45377). [#46681](https://github.com/ClickHouse/ClickHouse/pull/46681) ([Alexander Tokmakov](https://github.com/tavplubix)).
* Backported in [#47157](https://github.com/ClickHouse/ClickHouse/issues/47157): - Fix arithmetic operations in aggregate optimization with `min` and `max`. [#46705](https://github.com/ClickHouse/ClickHouse/pull/46705) ([Duc Canh Le](https://github.com/canhld94)).
* Backported in [#46881](https://github.com/ClickHouse/ClickHouse/issues/46881): Fix MSan report in the `maxIntersections` function. This closes [#43126](https://github.com/ClickHouse/ClickHouse/issues/43126). [#46847](https://github.com/ClickHouse/ClickHouse/pull/46847) ([Alexey Milovidov](https://github.com/alexey-milovidov)).
* Backported in [#47359](https://github.com/ClickHouse/ClickHouse/issues/47359): Fix possible deadlock on distributed query cancellation. [#47161](https://github.com/ClickHouse/ClickHouse/pull/47161) ([Kruglov Pavel](https://github.com/Avogar)).
#### NOT FOR CHANGELOG / INSIGNIFICANT
* Use /etc/default/clickhouse in systemd too [#47003](https://github.com/ClickHouse/ClickHouse/pull/47003) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).
* Update typing for a new PyGithub version [#47123](https://github.com/ClickHouse/ClickHouse/pull/47123) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).
* Follow-up to [#46681](https://github.com/ClickHouse/ClickHouse/issues/46681) [#47284](https://github.com/ClickHouse/ClickHouse/pull/47284) ([Alexander Tokmakov](https://github.com/tavplubix)).
* Add a manual trigger for release workflow [#47302](https://github.com/ClickHouse/ClickHouse/pull/47302) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).

28
docs/changelogs/v22.8.15.23-lts.md Normal file
View File

@ -0,0 +1,28 @@
---
sidebar_position: 1
sidebar_label: 2023
---
# 2023 Changelog
### ClickHouse release v22.8.15.23-lts (d36fa168bbf) FIXME as compared to v22.8.14.53-lts (4ea67c40077)
#### Improvement
* Backported in [#46981](https://github.com/ClickHouse/ClickHouse/issues/46981): - Apply `ALTER TABLE table_name ON CLUSTER cluster MOVE PARTITION|PART partition_expr TO DISK|VOLUME 'disk_name'` to all replicas. Because `ALTER TABLE t MOVE` is not replicated. [#46402](https://github.com/ClickHouse/ClickHouse/pull/46402) ([lizhuoyu5](https://github.com/lzydmxy)).
#### Bug Fix
* Backported in [#47336](https://github.com/ClickHouse/ClickHouse/issues/47336): Sometimes after changing a role that could be not reflected on the access rights of a user who uses that role. This PR fixes that. [#46772](https://github.com/ClickHouse/ClickHouse/pull/46772) ([Vitaly Baranov](https://github.com/vitlibar)).
#### Bug Fix (user-visible misbehavior in official stable or prestable release)
* Backported in [#46901](https://github.com/ClickHouse/ClickHouse/issues/46901): - Fix incorrect alias recursion in QueryNormalizer. [#46609](https://github.com/ClickHouse/ClickHouse/pull/46609) ([Raúl Marín](https://github.com/Algunenano)).
* Backported in [#47156](https://github.com/ClickHouse/ClickHouse/issues/47156): - Fix arithmetic operations in aggregate optimization with `min` and `max`. [#46705](https://github.com/ClickHouse/ClickHouse/pull/46705) ([Duc Canh Le](https://github.com/canhld94)).
* Backported in [#46987](https://github.com/ClickHouse/ClickHouse/issues/46987): Fix result of LIKE predicates which translate to substring searches and contain quoted non-LIKE metacharacters. [#46875](https://github.com/ClickHouse/ClickHouse/pull/46875) ([Robert Schulze](https://github.com/rschu1ze)).
* Backported in [#47357](https://github.com/ClickHouse/ClickHouse/issues/47357): Fix possible deadlock on distributed query cancellation. [#47161](https://github.com/ClickHouse/ClickHouse/pull/47161) ([Kruglov Pavel](https://github.com/Avogar)).
#### NOT FOR CHANGELOG / INSIGNIFICANT
* Reduce updates of Mergeable Check [#46781](https://github.com/ClickHouse/ClickHouse/pull/46781) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).
* Update typing for a new PyGithub version [#47123](https://github.com/ClickHouse/ClickHouse/pull/47123) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).
* Add a manual trigger for release workflow [#47302](https://github.com/ClickHouse/ClickHouse/pull/47302) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).

28
docs/changelogs/v23.1.5.24-stable.md Normal file
View File

@ -0,0 +1,28 @@
---
sidebar_position: 1
sidebar_label: 2023
---
# 2023 Changelog
### ClickHouse release v23.1.5.24-stable (0e51b53ba99) FIXME as compared to v23.1.4.58-stable (9ed562163a5)
#### Build/Testing/Packaging Improvement
* Backported in [#47060](https://github.com/ClickHouse/ClickHouse/issues/47060): Fix error during server startup on old distros (e.g. Amazon Linux 2) and on ARM that glibc 2.28 symbols are not found. [#47008](https://github.com/ClickHouse/ClickHouse/pull/47008) ([Robert Schulze](https://github.com/rschu1ze)).
#### Bug Fix (user-visible misbehavior in official stable or prestable release)
* Backported in [#46401](https://github.com/ClickHouse/ClickHouse/issues/46401): Fix `SYSTEM UNFREEZE` queries failing with the exception `CANNOT_PARSE_INPUT_ASSERTION_FAILED`. [#46325](https://github.com/ClickHouse/ClickHouse/pull/46325) ([Aleksei Filatov](https://github.com/aalexfvk)).
* Backported in [#46905](https://github.com/ClickHouse/ClickHouse/issues/46905): - Fix incorrect alias recursion in QueryNormalizer. [#46609](https://github.com/ClickHouse/ClickHouse/pull/46609) ([Raúl Marín](https://github.com/Algunenano)).
* Backported in [#47211](https://github.com/ClickHouse/ClickHouse/issues/47211): `INSERT` queries through native TCP protocol and HTTP protocol were not canceled correctly in some cases. It could lead to a partially applied query if a client canceled the query, or if a client died or, in rare cases, on network errors. As a result, it could lead to not working deduplication. Fixes [#27667](https://github.com/ClickHouse/ClickHouse/issues/27667) and [#45377](https://github.com/ClickHouse/ClickHouse/issues/45377). [#46681](https://github.com/ClickHouse/ClickHouse/pull/46681) ([Alexander Tokmakov](https://github.com/tavplubix)).
* Backported in [#47118](https://github.com/ClickHouse/ClickHouse/issues/47118): - Fix arithmetic operations in aggregate optimization with `min` and `max`. [#46705](https://github.com/ClickHouse/ClickHouse/pull/46705) ([Duc Canh Le](https://github.com/canhld94)).
* Backported in [#46883](https://github.com/ClickHouse/ClickHouse/issues/46883): Fix MSan report in the `maxIntersections` function. This closes [#43126](https://github.com/ClickHouse/ClickHouse/issues/43126). [#46847](https://github.com/ClickHouse/ClickHouse/pull/46847) ([Alexey Milovidov](https://github.com/alexey-milovidov)).
* Backported in [#47361](https://github.com/ClickHouse/ClickHouse/issues/47361): Fix possible deadlock on distributed query cancellation. [#47161](https://github.com/ClickHouse/ClickHouse/pull/47161) ([Kruglov Pavel](https://github.com/Avogar)).
#### NOT FOR CHANGELOG / INSIGNIFICANT
* Use /etc/default/clickhouse in systemd too [#47003](https://github.com/ClickHouse/ClickHouse/pull/47003) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).
* Update typing for a new PyGithub version [#47123](https://github.com/ClickHouse/ClickHouse/pull/47123) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).
* Follow-up to [#46681](https://github.com/ClickHouse/ClickHouse/issues/46681) [#47284](https://github.com/ClickHouse/ClickHouse/pull/47284) ([Alexander Tokmakov](https://github.com/tavplubix)).
* Add a manual trigger for release workflow [#47302](https://github.com/ClickHouse/ClickHouse/pull/47302) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).

20
docs/changelogs/v23.2.4.12-stable.md Normal file
View File

@ -0,0 +1,20 @@
---
sidebar_position: 1
sidebar_label: 2023
---
# 2023 Changelog
### ClickHouse release v23.2.4.12-stable (8fe866cb035) FIXME as compared to v23.2.3.17-stable (dec18bf7281)
#### Bug Fix (user-visible misbehavior in official stable or prestable release)
* Backported in [#47277](https://github.com/ClickHouse/ClickHouse/issues/47277): Fix IPv4/IPv6 serialization/deserialization in binary formats that was broken in https://github.com/ClickHouse/ClickHouse/pull/43221. Closes [#46522](https://github.com/ClickHouse/ClickHouse/issues/46522). [#46616](https://github.com/ClickHouse/ClickHouse/pull/46616) ([Kruglov Pavel](https://github.com/Avogar)).
* Backported in [#47212](https://github.com/ClickHouse/ClickHouse/issues/47212): `INSERT` queries through native TCP protocol and HTTP protocol were not canceled correctly in some cases. It could lead to a partially applied query if a client canceled the query, or if a client died or, in rare cases, on network errors. As a result, it could lead to not working deduplication. Fixes [#27667](https://github.com/ClickHouse/ClickHouse/issues/27667) and [#45377](https://github.com/ClickHouse/ClickHouse/issues/45377). [#46681](https://github.com/ClickHouse/ClickHouse/pull/46681) ([Alexander Tokmakov](https://github.com/tavplubix)).
* Backported in [#47363](https://github.com/ClickHouse/ClickHouse/issues/47363): Fix possible deadlock on distributed query cancellation. [#47161](https://github.com/ClickHouse/ClickHouse/pull/47161) ([Kruglov Pavel](https://github.com/Avogar)).
#### NOT FOR CHANGELOG / INSIGNIFICANT
* Follow-up to [#46681](https://github.com/ClickHouse/ClickHouse/issues/46681) [#47284](https://github.com/ClickHouse/ClickHouse/pull/47284) ([Alexander Tokmakov](https://github.com/tavplubix)).
* Add a manual trigger for release workflow [#47302](https://github.com/ClickHouse/ClickHouse/pull/47302) ([Mikhail f. Shiryaev](https://github.com/Felixoid)).

123
docs/en/development/build-cross-s390x.md Normal file
View File

@ -0,0 +1,123 @@
---
slug: /en/development/build-cross-s390x
sidebar_position: 69
title: How to Build, Run and Debug ClickHouse on Linux for s390x (zLinux)
sidebar_label: Build on Linux for s390x (zLinux)
---
As of writing (2023/3/10) building for s390x considered to be experimental. Not all features can be enabled, has broken features and is currently under active development.
## Building
As s390x does not support boringssl, it uses OpenSSL and has two related build options.
- By default, the s390x build will dynamically link to OpenSSL libraries. It will build OpenSSL shared objects, so it's not necessary to install OpenSSL beforehand. (This option is recommended in all cases.)
- Another option is to build OpenSSL in-tree. In this case two build flags need to be supplied to cmake
```bash
-DENABLE_OPENSSL_DYNAMIC=0 -DENABLE_OPENSSL=1
```
These instructions assume that the host machine is x86_64 and has all the tooling required to build natively based on the [build instructions](../development/build.md). It also assumes that the host is Ubuntu 22.04 but the following instructions should also work on Ubuntu 20.04.
In addition to installing the tooling used to build natively, the following additional packages need to be installed:
```bash
apt-get install binutils-s390x-linux-gnu libc6-dev-s390x-cross gcc-s390x-linux-gnu binfmt-support qemu-user-static
```
If you wish to cross compile rust code install the rust cross compile target for s390x:
```bash
rustup target add s390x-unknown-linux-gnu
```
To build for s390x:
```bash
cmake -DCMAKE_TOOLCHAIN_FILE=cmake/linux/toolchain-s390x.cmake ..
ninja
```
## Running
Once built, the binary can be run with, eg.:
```bash
qemu-s390x-static -L /usr/s390x-linux-gnu ./clickhouse
```
## Debugging
Install LLDB:
```bash
apt-get install lldb-15
```
To Debug a s390x executable, run clickhouse using QEMU in debug mode:
```bash
qemu-s390x-static -g 31338 -L /usr/s390x-linux-gnu ./clickhouse
```
In another shell run LLDB and attach, replace `<Clickhouse Parent Directory>` and `<build directory>` with the values corresponding to your environment.
```bash
lldb-15
(lldb) target create ./clickhouse
Current executable set to '/<Clickhouse Parent Directory>/ClickHouse/<build directory>/programs/clickhouse' (s390x).
(lldb) settings set target.source-map <build directory> /<Clickhouse Parent Directory>/ClickHouse
(lldb) gdb-remote 31338
Process 1 stopped
* thread #1, stop reason = signal SIGTRAP
frame #0: 0x0000004020e74cd0
-> 0x4020e74cd0: lgr %r2, %r15
0x4020e74cd4: aghi %r15, -160
0x4020e74cd8: xc 0(8,%r15), 0(%r15)
0x4020e74cde: brasl %r14, 275429939040
(lldb) b main
Breakpoint 1: 9 locations.
(lldb) c
Process 1 resuming
Process 1 stopped
* thread #1, stop reason = breakpoint 1.1
frame #0: 0x0000004005cd9fc0 clickhouse`main(argc_=1, argv_=0x0000004020e594a8) at main.cpp:450:17
447 #if !defined(FUZZING_MODE)
448 int main(int argc_, char ** argv_)
449 {
-> 450 inside_main = true;
451 SCOPE_EXIT({ inside_main = false; });
452
453 /// PHDR cache is required for query profiler to work reliably
```
## Visual Studio Code integration
- (CodeLLDB extension)[https://github.com/vadimcn/vscode-lldb] is required for visual debugging, the (Command Variable)[https://github.com/rioj7/command-variable] extension can help dynamic launches if using (cmake variants)[https://github.com/microsoft/vscode-cmake-tools/blob/main/docs/variants.md].
- Make sure to set the backend to your llvm installation eg. `"lldb.library": "/usr/lib/x86_64-linux-gnu/liblldb-15.so"`
- Launcher:
```json
{
"version": "0.2.0",
"configurations": [
{
"name": "Debug",
"type": "lldb",
"request": "custom",
"targetCreateCommands": ["target create ${command:cmake.launchTargetDirectory}/clickhouse"],
"processCreateCommands": ["settings set target.source-map ${input:targetdir} ${workspaceFolder}", "gdb-remote 31338"],
"sourceMap": { "${input:targetdir}": "${workspaceFolder}" },
}
],
"inputs": [
{
"id": "targetdir",
"type": "command",
"command": "extension.commandvariable.transform",
"args": {
"text": "${command:cmake.launchTargetDirectory}",
"find": ".*/([^/]+)/[^/]+$",
"replace": "$1"
}
}
]
}
```
- Make sure to run the clickhouse executable in debug mode prior to launch. (It is also possible to create a `preLaunchTask` that automates this)

2
docs/en/development/contrib.md
View File

@ -1,6 +1,6 @@
---
slug: /en/development/contrib
sidebar_position: 71
sidebar_position: 72
sidebar_label: Third-Party Libraries
description: A list of third-party libraries used
---

2
docs/en/development/style.md
View File

@ -1,6 +1,6 @@
---
slug: /en/development/style
sidebar_position: 69
sidebar_position: 70
sidebar_label: C++ Guide
description: A list of recommendations regarding coding style, naming convention, formatting and more
---

5
docs/en/development/tests.md
View File

@ -1,6 +1,6 @@
---
slug: /en/development/tests
sidebar_position: 70
sidebar_position: 71
sidebar_label: Testing
title: ClickHouse Testing
description: Most of ClickHouse features can be tested with functional tests and they are mandatory to use for every change in ClickHouse code that can be tested that way.
@ -31,6 +31,9 @@ folder and run the following command:
PATH=$PATH:<path to clickhouse-client> tests/clickhouse-test 01428_hash_set_nan_key
```
Test results (`stderr` and `stdout`) are written to files `01428_hash_set_nan_key.[stderr|stdout]` which
are located near the test file itself (so for `queries/0_stateless/foo.sql` output will be in `queries/0_stateless/foo.stdout`).
For more options, see `tests/clickhouse-test --help`. You can simply run all tests or run subset of tests filtered by substring in test name: `./clickhouse-test substring`. There are also options to run tests in parallel or in randomized order.
### Adding a New Test

88
docs/en/getting-started/example-datasets/nyc-taxi.md
View File

File diff suppressed because one or more lines are too long

3
docs/en/interfaces/formats.md
View File

@ -1981,6 +1981,7 @@ To exchange data with Hadoop, you can use [HDFS table engine](/docs/en/engines/t
- [input_format_parquet_skip_columns_with_unsupported_types_in_schema_inference](/docs/en/operations/settings/settings-formats.md/#input_format_parquet_skip_columns_with_unsupported_types_in_schema_inference) - allow skipping columns with unsupported types while schema inference for Parquet format. Default value - `false`.
- [output_format_parquet_fixed_string_as_fixed_byte_array](/docs/en/operations/settings/settings-formats.md/#output_format_parquet_fixed_string_as_fixed_byte_array) - use Parquet FIXED_LENGTH_BYTE_ARRAY type instead of Binary/String for FixedString columns. Default value - `true`.
- [output_format_parquet_version](/docs/en/operations/settings/settings-formats.md/#output_format_parquet_version) - The version of Parquet format used in output format. Default value - `2.latest`.
- [output_format_parquet_compression_method](/docs/en/operations/settings/settings-formats.md/#output_format_parquet_compression_method) - compression method used in output Parquet format. Default value - `snappy`.
## Arrow {#data-format-arrow}
@ -2051,6 +2052,7 @@ $ clickhouse-client --query="SELECT * FROM {some_table} FORMAT Arrow" > {filenam
- [input_format_arrow_allow_missing_columns](/docs/en/operations/settings/settings-formats.md/#input_format_arrow_allow_missing_columns) - allow missing columns while reading Arrow data. Default value - `false`.
- [input_format_arrow_skip_columns_with_unsupported_types_in_schema_inference](/docs/en/operations/settings/settings-formats.md/#input_format_arrow_skip_columns_with_unsupported_types_in_schema_inference) - allow skipping columns with unsupported types while schema inference for Arrow format. Default value - `false`.
- [output_format_arrow_fixed_string_as_fixed_byte_array](/docs/en/operations/settings/settings-formats.md/#output_format_arrow_fixed_string_as_fixed_byte_array) - use Arrow FIXED_SIZE_BINARY type instead of Binary/String for FixedString columns. Default value - `true`.
- [output_format_arrow_compression_method](/docs/en/operations/settings/settings-formats.md/#output_format_arrow_compression_method) - compression method used in output Arrow format. Default value - `none`.
## ArrowStream {#data-format-arrow-stream}
@ -2107,6 +2109,7 @@ $ clickhouse-client --query="SELECT * FROM {some_table} FORMAT ORC" > {filename.
### Arrow format settings {#parquet-format-settings}
- [output_format_arrow_string_as_string](/docs/en/operations/settings/settings-formats.md/#output_format_arrow_string_as_string) - use Arrow String type instead of Binary for String columns. Default value - `false`.
- [output_format_orc_compression_method](/docs/en/operations/settings/settings-formats.md/#output_format_orc_compression_method) - compression method used in output ORC format. Default value - `none`.
- [input_format_arrow_import_nested](/docs/en/operations/settings/settings-formats.md/#input_format_arrow_import_nested) - allow inserting array of structs into Nested table in Arrow input format. Default value - `false`.
- [input_format_arrow_case_insensitive_column_matching](/docs/en/operations/settings/settings-formats.md/#input_format_arrow_case_insensitive_column_matching) - ignore case when matching Arrow columns with ClickHouse columns. Default value - `false`.
- [input_format_arrow_allow_missing_columns](/docs/en/operations/settings/settings-formats.md/#input_format_arrow_allow_missing_columns) - allow missing columns while reading Arrow data. Default value - `false`.

80
docs/en/interfaces/schema-inference.md
View File

@ -117,7 +117,7 @@ clickhouse-local --file='hobbies.jsonl' --table='hobbies' --query='SELECT * FROM
4 47 Brayan ['movies','skydiving']
```
# Using structure from insertion table {#using-structure-from-insertion-table}
## Using structure from insertion table {#using-structure-from-insertion-table}
When table functions `file/s3/url/hdfs` are used to insert data into a table,
there is an option to use the structure from the insertion table instead of extracting it from the data.
@ -222,7 +222,7 @@ INSERT INTO hobbies4 SELECT id, empty(hobbies) ? NULL : hobbies[1] FROM file(hob
In this case, there are some operations performed on the column `hobbies` in the `SELECT` query to insert it into the table, so ClickHouse cannot use the structure from the insertion table, and schema inference will be used.
# Schema inference cache {#schema-inference-cache}
## Schema inference cache {#schema-inference-cache}
For most input formats schema inference reads some data to determine its structure and this process can take some time.
To prevent inferring the same schema every time ClickHouse read the data from the same file, the inferred schema is cached and when accessing the same file again, ClickHouse will use the schema from the cache.
@ -326,14 +326,14 @@ SELECT count() FROM system.schema_inference_cache WHERE storage='S3'
└─────────┘
```
# Text formats {#text-formats}
## Text formats {#text-formats}
For text formats, ClickHouse reads the data row by row, extracts column values according to the format,
and then uses some recursive parsers and heuristics to determine the type for each value. The maximum number of rows read from the data in schema inference
is controlled by the setting `input_format_max_rows_to_read_for_schema_inference` with default value 25000.
By default, all inferred types are [Nullable](../sql-reference/data-types/nullable.md), but you can change this by setting `schema_inference_make_columns_nullable` (see examples in the [settings](#settings-for-text-formats) section).
## JSON formats {#json-formats}
### JSON formats {#json-formats}
In JSON formats ClickHouse parses values according to the JSON specification and then tries to find the most appropriate data type for them.
@ -464,9 +464,9 @@ most likely this column contains only Nulls or empty Arrays/Maps.
...
```
### JSON settings {#json-settings}
#### JSON settings {#json-settings}
#### input_format_json_read_objects_as_strings
##### input_format_json_read_objects_as_strings
Enabling this setting allows reading nested JSON objects as strings.
This setting can be used to read nested JSON objects without using JSON object type.
@ -486,7 +486,7 @@ DESC format(JSONEachRow, $$
└──────┴──────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
#### input_format_json_try_infer_numbers_from_strings
##### input_format_json_try_infer_numbers_from_strings
Enabling this setting allows inferring numbers from string values.
@ -507,7 +507,7 @@ DESC format(JSONEachRow, $$
└───────┴─────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
#### input_format_json_read_numbers_as_strings
##### input_format_json_read_numbers_as_strings
Enabling this setting allows reading numeric values as strings.
@ -528,7 +528,7 @@ DESC format(JSONEachRow, $$
└───────┴──────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
#### input_format_json_read_bools_as_numbers
##### input_format_json_read_bools_as_numbers
Enabling this setting allows reading Bool values as numbers.
@ -549,7 +549,7 @@ DESC format(JSONEachRow, $$
└───────┴─────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## CSV {#csv}
### CSV {#csv}
In CSV format ClickHouse extracts column values from the row according to delimiters. ClickHouse expects all types except numbers and strings to be enclosed in double quotes. If the value is in double quotes, ClickHouse tries to parse
the data inside quotes using the recursive parser and then tries to find the most appropriate data type for it. If the value is not in double quotes, ClickHouse tries to parse it as a number,
@ -726,7 +726,7 @@ $$)
└──────────────┴───────────────┘
```
## TSV/TSKV {#tsv-tskv}
### TSV/TSKV {#tsv-tskv}
In TSV/TSKV formats ClickHouse extracts column value from the row according to tabular delimiters and then parses extracted value using
the recursive parser to determine the most appropriate type. If the type cannot be determined, ClickHouse treats this value as String.
@ -1019,7 +1019,7 @@ DESC format(TSV, '[1,2,3] 42.42 Hello World!')
└──────┴──────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## CustomSeparated {#custom-separated}
### CustomSeparated {#custom-separated}
In CustomSeparated format ClickHouse first extracts all column values from the row according to specified delimiters and then tries to infer
the data type for each value according to escaping rule.
@ -1080,7 +1080,7 @@ $$)
└────────┴───────────────┴────────────┘
```
## Template {#template}
### Template {#template}
In Template format ClickHouse first extracts all column values from the row according to the specified template and then tries to infer the
data type for each value according to its escaping rule.
@ -1120,7 +1120,7 @@ $$)
└──────────┴────────────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## Regexp {#regexp}
### Regexp {#regexp}
Similar to Template, in Regexp format ClickHouse first extracts all column values from the row according to specified regular expression and then tries to infer
data type for each value according to the specified escaping rule.
@ -1142,9 +1142,9 @@ Line: value_1=2, value_2="Some string 2", value_3="[4, 5, NULL]"$$)
└──────┴────────────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## Settings for text formats {settings-for-text-formats}
### Settings for text formats {#settings-for-text-formats}
### input_format_max_rows_to_read_for_schema_inference
#### input_format_max_rows_to_read_for_schema_inference
This setting controls the maximum number of rows to be read while schema inference.
The more rows are read, the more time is spent on schema inference, but the greater the chance to
@ -1152,7 +1152,7 @@ correctly determine the types (especially when the data contains a lot of nulls)
Default value: `25000`.
### column_names_for_schema_inference
#### column_names_for_schema_inference
The list of column names to use in schema inference for formats without explicit column names. Specified names will be used instead of default `c1,c2,c3,...`. The format: `column1,column2,column3,...`.
@ -1169,7 +1169,7 @@ DESC format(TSV, 'Hello, World! 42 [1, 2, 3]') settings column_names_for_schema_
└──────┴────────────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
### schema_inference_hints
#### schema_inference_hints
The list of column names and types to use in schema inference instead of automatically determined types. The format: 'column_name1 column_type1, column_name2 column_type2, ...'.
This setting can be used to specify the types of columns that could not be determined automatically or for optimizing the schema.
@ -1189,7 +1189,7 @@ DESC format(JSONEachRow, '{"id" : 1, "age" : 25, "name" : "Josh", "status" : nul
└─────────┴─────────────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
### schema_inference_make_columns_nullable
#### schema_inference_make_columns_nullable
Controls making inferred types `Nullable` in schema inference for formats without information about nullability.
If the setting is enabled, all inferred type will be `Nullable`, if disabled, the inferred type will be `Nullable` only if the column contains `NULL` in a sample that is parsed during schema inference.
@ -1232,7 +1232,7 @@ DESC format(JSONEachRow, $$
└─────────┴──────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
### input_format_try_infer_integers
#### input_format_try_infer_integers
If enabled, ClickHouse will try to infer integers instead of floats in schema inference for text formats.
If all numbers in the column from sample data are integers, the result type will be `Int64`, if at least one number is float, the result type will be `Float64`.
@ -1289,7 +1289,7 @@ DESC format(JSONEachRow, $$
└────────┴───────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
### input_format_try_infer_datetimes
#### input_format_try_infer_datetimes
If enabled, ClickHouse will try to infer type `DateTime64` from string fields in schema inference for text formats.
If all fields from a column in sample data were successfully parsed as datetimes, the result type will be `DateTime64(9)`,
@ -1337,7 +1337,7 @@ DESC format(JSONEachRow, $$
Note: Parsing datetimes during schema inference respect setting [date_time_input_format](/docs/en/operations/settings/settings-formats.md#date_time_input_format)
### input_format_try_infer_dates
#### input_format_try_infer_dates
If enabled, ClickHouse will try to infer type `Date` from string fields in schema inference for text formats.
If all fields from a column in sample data were successfully parsed as dates, the result type will be `Date`,
@ -1383,14 +1383,14 @@ DESC format(JSONEachRow, $$
└──────┴──────────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
# Self describing formats {#self-describing-formats}
## Self describing formats {#self-describing-formats}
Self-describing formats contain information about the structure of the data in the data itself,
it can be some header with a description, a binary type tree, or some kind of table.
To automatically infer a schema from files in such formats, ClickHouse reads a part of the data containing
information about the types and converts it into a schema of the ClickHouse table.
## Formats with -WithNamesAndTypes suffix {#formats-with-names-and-types}
### Formats with -WithNamesAndTypes suffix {#formats-with-names-and-types}
ClickHouse supports some text formats with the suffix -WithNamesAndTypes. This suffix means that the data contains two additional rows with column names and types before the actual data.
While schema inference for such formats, ClickHouse reads the first two rows and extracts column names and types.
@ -1412,7 +1412,7 @@ $$)
└──────┴──────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## JSON formats with metadata {#json-with-metadata}
### JSON formats with metadata {#json-with-metadata}
Some JSON input formats ([JSON](formats.md#json), [JSONCompact](formats.md#json-compact), [JSONColumnsWithMetadata](formats.md#jsoncolumnswithmetadata)) contain metadata with column names and types.
In schema inference for such formats, ClickHouse reads this metadata.
@ -1465,7 +1465,7 @@ $$)
└──────┴──────────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## Avro {#avro}
### Avro {#avro}
In Avro format ClickHouse reads its schema from the data and converts it to ClickHouse schema using the following type matches:
@ -1485,7 +1485,7 @@ In Avro format ClickHouse reads its schema from the data and converts it to Clic
Other Avro types are not supported.
## Parquet {#parquet}
### Parquet {#parquet}
In Parquet format ClickHouse reads its schema from the data and converts it to ClickHouse schema using the following type matches:
@ -1513,7 +1513,7 @@ In Parquet format ClickHouse reads its schema from the data and converts it to C
Other Parquet types are not supported. By default, all inferred types are inside `Nullable`, but it can be changed using the setting `schema_inference_make_columns_nullable`.
## Arrow {#arrow}
### Arrow {#arrow}
In Arrow format ClickHouse reads its schema from the data and converts it to ClickHouse schema using the following type matches:
@ -1541,7 +1541,7 @@ In Arrow format ClickHouse reads its schema from the data and converts it to Cli
Other Arrow types are not supported. By default, all inferred types are inside `Nullable`, but it can be changed using the setting `schema_inference_make_columns_nullable`.
## ORC {#orc}
### ORC {#orc}
In ORC format ClickHouse reads its schema from the data and converts it to ClickHouse schema using the following type matches:
@ -1564,17 +1564,17 @@ In ORC format ClickHouse reads its schema from the data and converts it to Click
Other ORC types are not supported. By default, all inferred types are inside `Nullable`, but it can be changed using the setting `schema_inference_make_columns_nullable`.
## Native {#native}
### Native {#native}
Native format is used inside ClickHouse and contains the schema in the data.
In schema inference, ClickHouse reads the schema from the data without any transformations.
# Formats with external schema {#formats-with-external-schema}
## Formats with external schema {#formats-with-external-schema}
Such formats require a schema describing the data in a separate file in a specific schema language.
To automatically infer a schema from files in such formats, ClickHouse reads external schema from a separate file and transforms it to a ClickHouse table schema.
# Protobuf {#protobuf}
### Protobuf {#protobuf}
In schema inference for Protobuf format ClickHouse uses the following type matches:
@ -1592,7 +1592,7 @@ In schema inference for Protobuf format ClickHouse uses the following type match
| `repeated T` | [Array(T)](../sql-reference/data-types/array.md) |
| `message`, `group` | [Tuple](../sql-reference/data-types/tuple.md) |
# CapnProto {#capnproto}
### CapnProto {#capnproto}
In schema inference for CapnProto format ClickHouse uses the following type matches:
@ -1615,13 +1615,13 @@ In schema inference for CapnProto format ClickHouse uses the following type matc
| `struct` | [Tuple](../sql-reference/data-types/tuple.md) |
| `union(T, Void)`, `union(Void, T)` | [Nullable(T)](../sql-reference/data-types/nullable.md) |
# Strong-typed binary formats {#strong-typed-binary-formats}
## Strong-typed binary formats {#strong-typed-binary-formats}
In such formats, each serialized value contains information about its type (and possibly about its name), but there is no information about the whole table.
In schema inference for such formats, ClickHouse reads data row by row (up to `input_format_max_rows_to_read_for_schema_inference` rows) and extracts
the type (and possibly name) for each value from the data and then converts these types to ClickHouse types.
## MsgPack {msgpack}
### MsgPack {#msgpack}
In MsgPack format there is no delimiter between rows, to use schema inference for this format you should specify the number of columns in the table
using the setting `input_format_msgpack_number_of_columns`. ClickHouse uses the following type matches:
@ -1641,7 +1641,7 @@ using the setting `input_format_msgpack_number_of_columns`. ClickHouse uses the
By default, all inferred types are inside `Nullable`, but it can be changed using the setting `schema_inference_make_columns_nullable`.
## BSONEachRow {#bsoneachrow}
### BSONEachRow {#bsoneachrow}
In BSONEachRow each row of data is presented as a BSON document. In schema inference ClickHouse reads BSON documents one by one and extracts
values, names, and types from the data and then transforms these types to ClickHouse types using the following type matches:
@ -1661,11 +1661,11 @@ values, names, and types from the data and then transforms these types to ClickH
By default, all inferred types are inside `Nullable`, but it can be changed using the setting `schema_inference_make_columns_nullable`.
# Formats with constant schema {#formats-with-constant-schema}
## Formats with constant schema {#formats-with-constant-schema}
Data in such formats always have the same schema.
## LineAsString {#line-as-string}
### LineAsString {#line-as-string}
In this format, ClickHouse reads the whole line from the data into a single column with `String` data type. The inferred type for this format is always `String` and the column name is `line`.
@ -1680,7 +1680,7 @@ DESC format(LineAsString, 'Hello\nworld!')
└──────┴────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## JSONAsString {#json-as-string}
### JSONAsString {#json-as-string}
In this format, ClickHouse reads the whole JSON object from the data into a single column with `String` data type. The inferred type for this format is always `String` and the column name is `json`.
@ -1695,7 +1695,7 @@ DESC format(JSONAsString, '{"x" : 42, "y" : "Hello, World!"}')
└──────┴────────┴──────────────┴────────────────────┴─────────┴──────────────────┴────────────────┘
```
## JSONAsObject {#json-as-object}
### JSONAsObject {#json-as-object}
In this format, ClickHouse reads the whole JSON object from the data into a single column with `Object('json')` data type. Inferred type for this format is always `String` and the column name is `json`.

2
docs/en/operations/server-configuration-parameters/settings.md
View File

@ -1319,7 +1319,7 @@ Settings:
``` xml
<prometheus>
<endpoint>/metrics</endpoint>
<port>8001</port>
<port>9363</port>
<metrics>true</metrics>
<events>true</events>
<asynchronous_metrics>true</asynchronous_metrics>

30
docs/en/operations/settings/settings-formats.md
View File

@ -1014,6 +1014,12 @@ Use Arrow FIXED_SIZE_BINARY type instead of Binary/String for FixedString column
Enabled by default.
### output_format_arrow_compression_method {#output_format_arrow_compression_method}
Compression method used in output Arrow format. Supported codecs: `lz4_frame`, `zstd`, `none` (uncompressed)
Default value: `none`.
## ORC format settings {#orc-format-settings}
### input_format_orc_import_nested {#input_format_orc_import_nested}
@ -1057,6 +1063,12 @@ Use ORC String type instead of Binary for String columns.
Disabled by default.
### output_format_orc_compression_method {#output_format_orc_compression_method}
Compression method used in output ORC format. Supported codecs: `lz4`, `snappy`, `zlib`, `zstd`, `none` (uncompressed)
Default value: `none`.
## Parquet format settings {#parquet-format-settings}
### input_format_parquet_import_nested {#input_format_parquet_import_nested}
@ -1112,6 +1124,12 @@ The version of Parquet format used in output format. Supported versions: `1.0`,
Default value: `2.latest`.
### output_format_parquet_compression_method {#output_format_parquet_compression_method}
Compression method used in output Parquet format. Supported codecs: `snappy`, `lz4`, `brotli`, `zstd`, `gzip`, `none` (uncompressed)
Default value: `snappy`.
## Hive format settings {#hive-format-settings}
### input_format_hive_text_fields_delimiter {#input_format_hive_text_fields_delimiter}
@ -1474,7 +1492,7 @@ Default value: `65505`.
The name of table that will be used in the output INSERT statement.
Default value: `'table''`.
Default value: `table`.
### output_format_sql_insert_include_column_names {#output_format_sql_insert_include_column_names}
@ -1514,4 +1532,12 @@ Disabled by default.
The maximum allowed size for String in RowBinary format. It prevents allocating large amount of memory in case of corrupted data. 0 means there is no limit.
Default value: `1GiB`
Default value: `1GiB`.
## Native format settings {#native-format-settings}
### input_format_native_allow_types_conversion {#input_format_native_allow_types_conversion}
Allow types conversion in Native input format between columns from input data and requested columns.
Enabled by default.

45
docs/en/operations/settings/settings.md
View File

@ -1248,7 +1248,9 @@ Possible values:
Default value: 1.
:::warning
Disable this setting if you use [max_parallel_replicas](#settings-max_parallel_replicas).
Disable this setting if you use [max_parallel_replicas](#settings-max_parallel_replicas) without [parallel_replicas_custom_key](#settings-parallel_replicas_custom_key).
If [parallel_replicas_custom_key](#settings-parallel_replicas_custom_key) is set, disable this setting only if it's used on a cluster with multiple shards containing multiple replicas.
If it's used on a cluster with a single shard and multiple replicas, disabling this setting will have negative effects.
:::
## totals_mode {#totals-mode}
@ -1273,16 +1275,47 @@ Default value: `1`.
**Additional Info**
This setting is useful for replicated tables with a sampling key. A query may be processed faster if it is executed on several servers in parallel. But the query performance may degrade in the following cases:
This options will produce different results depending on the settings used.
:::warning
This setting will produce incorrect results when joins or subqueries are involved, and all tables don't meet certain requirements. See [Distributed Subqueries and max_parallel_replicas](../../sql-reference/operators/in.md/#max_parallel_replica-subqueries) for more details.
:::
### Parallel processing using `SAMPLE` key
A query may be processed faster if it is executed on several servers in parallel. But the query performance may degrade in the following cases:
- The position of the sampling key in the partitioning key does not allow efficient range scans.
- Adding a sampling key to the table makes filtering by other columns less efficient.
- The sampling key is an expression that is expensive to calculate.
- The cluster latency distribution has a long tail, so that querying more servers increases the query overall latency.
:::warning
This setting will produce incorrect results when joins or subqueries are involved, and all tables don't meet certain requirements. See [Distributed Subqueries and max_parallel_replicas](../../sql-reference/operators/in.md/#max_parallel_replica-subqueries) for more details.
:::
### Parallel processing using [parallel_replicas_custom_key](#settings-parallel_replicas_custom_key)
This setting is useful for any replicated table.
## parallel_replicas_custom_key {#settings-parallel_replicas_custom_key}
An arbitrary integer expression that can be used to split work between replicas for a specific table.
The value can be any integer expression.
A query may be processed faster if it is executed on several servers in parallel but it depends on the used [parallel_replicas_custom_key](#settings-parallel_replicas_custom_key)
and [parallel_replicas_custom_key_filter_type](#settings-parallel_replicas_custom_key_filter_type).
Simple expressions using primary keys are preferred.
If the setting is used on a cluster that consists of a single shard with multiple replicas, those replicas will be converted into virtual shards.
Otherwise, it will behave same as for `SAMPLE` key, it will use multiple replicas of each shard.
## parallel_replicas_custom_key_filter_type {#settings-parallel_replicas_custom_key_filter_type}
How to use `parallel_replicas_custom_key` expression for splitting work between replicas.
Possible values:
- `default` — Use the default implementation using modulo operation on the `parallel_replicas_custom_key`.
- `range` — Split the entire value space of the expression in the ranges. This type of filtering is useful if values of `parallel_replicas_custom_key` are uniformly spread across the entire integer space, e.g. hash values.
Default value: `default`.
## compile_expressions {#compile-expressions}
@ -1515,7 +1548,7 @@ Enables or disables asynchronous inserts. This makes sense only for insertion ov
If enabled, the data is combined into batches before the insertion into tables, so it is possible to do small and frequent insertions into ClickHouse (up to 15000 queries per second) without buffer tables.
The data is inserted either after the [async_insert_max_data_size](#async-insert-max-data-size) is exceeded or after [async_insert_busy_timeout_ms](#async-insert-busy-timeout-ms) milliseconds since the first `INSERT` query. If the [async_insert_stale_timeout_ms](#async-insert-stale-timeout-ms) is set to a non-zero value, the data is inserted after `async_insert_stale_timeout_ms` milliseconds since the last query.
The data is inserted either after the [async_insert_max_data_size](#async-insert-max-data-size) is exceeded or after [async_insert_busy_timeout_ms](#async-insert-busy-timeout-ms) milliseconds since the first `INSERT` query. If the [async_insert_stale_timeout_ms](#async-insert-stale-timeout-ms) is set to a non-zero value, the data is inserted after `async_insert_stale_timeout_ms` milliseconds since the last query. Also the buffer will be flushed to disk if at least [async_insert_max_query_number](#async-insert-max-query-number) async insert queries per block were received. This last setting takes effect only if [async_insert_deduplicate](#async-insert-deduplicate) is enabled.
If [wait_for_async_insert](#wait-for-async-insert) is enabled, every client will wait for the data to be processed and flushed to the table. Otherwise, the query would be processed almost instantly, even if the data is not inserted.

8
docs/en/operations/system-tables/opentelemetry_span_log.md
View File

@ -15,6 +15,13 @@ Columns:
- `operation_name` ([String](../../sql-reference/data-types/string.md)) — The name of the operation.
- `kind` ([Enum8](../../sql-reference/data-types/enum.md)) — The [SpanKind](https://opentelemetry.io/docs/reference/specification/trace/api/#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.
- `start_time_us` ([UInt64](../../sql-reference/data-types/int-uint.md)) — The start time of the `trace span` (in microseconds).
- `finish_time_us` ([UInt64](../../sql-reference/data-types/int-uint.md)) — The finish time of the `trace span` (in microseconds).
@ -42,6 +49,7 @@ trace_id: cdab0847-0d62-61d5-4d38-dd65b19a1914
span_id: 701487461015578150
parent_span_id: 2991972114672045096
operation_name: DB::Block DB::InterpreterSelectQuery::getSampleBlockImpl()
kind: INTERNAL
start_time_us: 1612374594529090
finish_time_us: 1612374594529108
finish_date: 2021-02-03

27
docs/en/operations/utilities/clickhouse-local.md
View File

@ -6,29 +6,26 @@ sidebar_label: clickhouse-local
# clickhouse-local
The `clickhouse-local` program enables you to perform fast processing on local files, without having to deploy and configure the ClickHouse server.
The `clickhouse-local` program enables you to perform fast processing on local files, without having to deploy and configure the ClickHouse server. It accepts data that represent tables and queries them using [ClickHouse SQL dialect](../../sql-reference/). `clickhouse-local` uses the same core as ClickHouse server, so it supports most of the features and the same set of formats and table engines.
Accepts data that represent tables and queries them using [ClickHouse SQL dialect](../../sql-reference/).
`clickhouse-local` uses the same core as ClickHouse server, so it supports most of the features and the same set of formats and table engines.
By default `clickhouse-local` does not have access to data on the same host, but it supports loading server configuration using `--config-file` argument.
:::warning
It is not recommended to load production server configuration into `clickhouse-local` because data can be damaged in case of human error.
:::
For temporary data, a unique temporary data directory is created by default.
By default `clickhouse-local` has access to data on the same host, and it does not depend on the server's configuration. It also supports loading server configuration using `--config-file` argument. For temporary data, a unique temporary data directory is created by default.
## Usage {#usage}
Basic usage:
Basic usage (Linux):
``` bash
$ clickhouse-local --structure "table_structure" --input-format "format_of_incoming_data" \
--query "query"
$ clickhouse-local --structure "table_structure" --input-format "format_of_incoming_data" --query "query"
```
Basic usage (Mac):
``` bash
$ ./clickhouse local --structure "table_structure" --input-format "format_of_incoming_data" --query "query"
```
Also supported on Windows through WSL2.
Arguments:
- `-S`, `--structure` — table structure for input data.

61
docs/en/sql-reference/aggregate-functions/reference/exponentialmovingaverage.md
View File

@ -11,15 +11,15 @@ sidebar_title: exponentialMovingAverage
**Syntax**
```sql
exponentialMovingAverage(x)(value, timestamp)
exponentialMovingAverage(x)(value, timeunit)
```
Each `value` corresponds to the determinate `timestamp`. The half-life `x` is the time lag at which the exponential weights decay by one-half. The function returns a weighted average: the older the time point, the less weight the corresponding value is considered to be.
Each `value` corresponds to the determinate `timeunit`. The half-life `x` is the time lag at which the exponential weights decay by one-half. The function returns a weighted average: the older the time point, the less weight the corresponding value is considered to be.
**Arguments**
- `value` — Value. [Integer](../../../sql-reference/data-types/int-uint.md), [Float](../../../sql-reference/data-types/float.md) or [Decimal](../../../sql-reference/data-types/decimal.md).
- `timestamp` — Timestamp. [Integer](../../../sql-reference/data-types/int-uint.md), [Float](../../../sql-reference/data-types/float.md) or [Decimal](../../../sql-reference/data-types/decimal.md).
- `timeunit` — Timeunit. [Integer](../../../sql-reference/data-types/int-uint.md), [Float](../../../sql-reference/data-types/float.md) or [Decimal](../../../sql-reference/data-types/decimal.md). Timeunit is not timestamp (seconds), it's -- an index of the time interval. Can be calculated using [intDiv](../../functions/arithmetic-functions/#intdiva-b).
**Parameters**
@ -148,3 +148,58 @@ Result:
│ 1 │ 49 │ 0.825 │ █████████████████████████████████████████▎│
└───────┴──────┴──────────────────────┴────────────────────────────────────────────┘
```
```sql
CREATE TABLE data
ENGINE = Memory AS
SELECT
10 AS value,
toDateTime('2020-01-01') + (3600 * number) AS time
FROM numbers_mt(10);
-- Calculate timeunit using intDiv
SELECT
value,
time,
exponentialMovingAverage(1)(value, intDiv(toUInt32(time), 3600)) OVER (ORDER BY time ASC) AS res,
intDiv(toUInt32(time), 3600) AS timeunit
FROM data
ORDER BY time ASC;
┌─value─┬────────────────time─┬─────────res─┬─timeunit─┐
│ 10 │ 2020-01-01 00:00:00 │ 5 │ 438288 │
│ 10 │ 2020-01-01 01:00:00 │ 7.5 │ 438289 │
│ 10 │ 2020-01-01 02:00:00 │ 8.75 │ 438290 │
│ 10 │ 2020-01-01 03:00:00 │ 9.375 │ 438291 │
│ 10 │ 2020-01-01 04:00:00 │ 9.6875 │ 438292 │
│ 10 │ 2020-01-01 05:00:00 │ 9.84375 │ 438293 │
│ 10 │ 2020-01-01 06:00:00 │ 9.921875 │ 438294 │
│ 10 │ 2020-01-01 07:00:00 │ 9.9609375 │ 438295 │
│ 10 │ 2020-01-01 08:00:00 │ 9.98046875 │ 438296 │
│ 10 │ 2020-01-01 09:00:00 │ 9.990234375 │ 438297 │
└───────┴─────────────────────┴─────────────┴──────────┘
-- Calculate timeunit using toRelativeHourNum
SELECT
value,
time,
exponentialMovingAverage(1)(value, toRelativeHourNum(time)) OVER (ORDER BY time ASC) AS res,
toRelativeHourNum(time) AS timeunit
FROM data
ORDER BY time ASC;
┌─value─┬────────────────time─┬─────────res─┬─timeunit─┐
│ 10 │ 2020-01-01 00:00:00 │ 5 │ 438288 │
│ 10 │ 2020-01-01 01:00:00 │ 7.5 │ 438289 │
│ 10 │ 2020-01-01 02:00:00 │ 8.75 │ 438290 │
│ 10 │ 2020-01-01 03:00:00 │ 9.375 │ 438291 │
│ 10 │ 2020-01-01 04:00:00 │ 9.6875 │ 438292 │
│ 10 │ 2020-01-01 05:00:00 │ 9.84375 │ 438293 │
│ 10 │ 2020-01-01 06:00:00 │ 9.921875 │ 438294 │
│ 10 │ 2020-01-01 07:00:00 │ 9.9609375 │ 438295 │
│ 10 │ 2020-01-01 08:00:00 │ 9.98046875 │ 438296 │
│ 10 │ 2020-01-01 09:00:00 │ 9.990234375 │ 438297 │
└───────┴─────────────────────┴─────────────┴──────────┘
```

11
docs/en/sql-reference/operators/in.md
View File

@ -233,8 +233,9 @@ If `some_predicate` is not selective enough, it will return large amount of data
### Distributed Subqueries and max_parallel_replicas
When max_parallel_replicas is greater than 1, distributed queries are further transformed. For example, the following:
When [max_parallel_replicas](#settings-max_parallel_replicas) is greater than 1, distributed queries are further transformed.
For example, the following:
```sql
SELECT CounterID, count() FROM distributed_table_1 WHERE UserID IN (SELECT UserID FROM local_table_2 WHERE CounterID < 100)
SETTINGS max_parallel_replicas=3
@ -247,8 +248,12 @@ SELECT CounterID, count() FROM local_table_1 WHERE UserID IN (SELECT UserID FROM
SETTINGS parallel_replicas_count=3, parallel_replicas_offset=M
```
where M is between 1 and 3 depending on which replica the local query is executing on. These settings affect every MergeTree-family table in the query and have the same effect as applying `SAMPLE 1/3 OFFSET (M-1)/3` on each table.
where M is between 1 and 3 depending on which replica the local query is executing on.
Therefore adding the max_parallel_replicas setting will only produce correct results if both tables have the same replication scheme and are sampled by UserID or a subkey of it. In particular, if local_table_2 does not have a sampling key, incorrect results will be produced. The same rule applies to JOIN.
These settings affect every MergeTree-family table in the query and have the same effect as applying `SAMPLE 1/3 OFFSET (M-1)/3` on each table.
Therefore adding the [max_parallel_replicas](#settings-max_parallel_replicas) setting will only produce correct results if both tables have the same replication scheme and are sampled by UserID or a subkey of it. In particular, if local_table_2 does not have a sampling key, incorrect results will be produced. The same rule applies to JOIN.
One workaround if local_table_2 does not meet the requirements, is to use `GLOBAL IN` or `GLOBAL JOIN`.
If a table doesn't have a sampling key, more flexible options for [parallel_replicas_custom_key](#settings-parallel_replicas_custom_key) can be used that can produce different and more optimal behaviour.

33
docs/en/sql-reference/statements/create/table.md
View File

@ -112,23 +112,21 @@ See also [data_type_default_nullable](../../../operations/settings/settings.md#d
## Default Values {#default_values}
The column description can specify an expression for a default value, in one of the following ways: `DEFAULT expr`, `MATERIALIZED expr`, `ALIAS expr`.
The column description can specify a default value expression in the form of `DEFAULT expr`, `MATERIALIZED expr`, or `ALIAS expr`. Example: `URLDomain String DEFAULT domain(URL)`.
Example: `URLDomain String DEFAULT domain(URL)`.
The expression `expr` is optional. If it is omitted, the column type must be specified explicitly and the default value will be `0` for numeric columns, `''` (the empty string) for string columns, `[]` (the empty array) for array columns, `1970-01-01` for date columns, or `NULL` for nullable columns.
If an expression for the default value is not defined, the default values will be set to zeros for numbers, empty strings for strings, empty arrays for arrays, and `1970-01-01` for dates or zero unix timestamp for DateTime, NULL for Nullable.
The column type of a default value column can be omitted in which case it is infered from `expr`'s type. For example the type of column `EventDate DEFAULT toDate(EventTime)` will be date.
If the default expression is defined, the column type is optional. If there isn’t an explicitly defined type, the default expression type is used. Example: `EventDate DEFAULT toDate(EventTime)` – the ‘Date’ type will be used for the ‘EventDate’ column.
If both a data type and a default value expression are specified, an implicit type casting function inserted which converts the expression to the specified type. Example: `Hits UInt32 DEFAULT 0` is internally represented as `Hits UInt32 DEFAULT toUInt32(0)`.
If the data type and default expression are defined explicitly, this expression will be cast to the specified type using type casting functions. Example: `Hits UInt32 DEFAULT 0` means the same thing as `Hits UInt32 DEFAULT toUInt32(0)`.
Default expressions may be defined as an arbitrary expression from table constants and columns. When creating and changing the table structure, it checks that expressions do not contain loops. For INSERT, it checks that expressions are resolvable – that all columns they can be calculated from have been passed.
A default value expression `expr` may reference arbitrary table columns and constants. ClickHouse checks that changes of the table structure do not introduce loops in the expression calculation. For INSERT, it checks that expressions are resolvable – that all columns they can be calculated from have been passed.
### DEFAULT
`DEFAULT expr`
Normal default value. If the INSERT query does not specify the corresponding column, it will be filled in by computing the corresponding expression.
Normal default value. If the value of such a column is not specified in an INSERT query, it is computed from `expr`.
Example:
@ -154,9 +152,9 @@ SELECT * FROM test;
`MATERIALIZED expr`
Materialized expression. Such a column can’t be specified for INSERT, because it is always calculated.
For an INSERT without a list of columns, these columns are not considered.
In addition, this column is not substituted when using an asterisk in a SELECT query. This is to preserve the invariant that the dump obtained using `SELECT *` can be inserted back into the table using INSERT without specifying the list of columns.
Materialized expression. Values of such columns are always calculated, they cannot be specified in INSERT queries.
Also, default value columns of this type are not included in the result of `SELECT *`. This is to preserve the invariant that the result of a `SELECT *` can always be inserted back into the table using `INSERT`. This behavior can be disabled with setting `asterisk_include_materialized_columns`.
Example:
@ -192,8 +190,9 @@ SELECT * FROM test SETTINGS asterisk_include_materialized_columns=1;
`EPHEMERAL [expr]`
Ephemeral column. Such a column isn't stored in the table and cannot be SELECTed, but can be referenced in the defaults of CREATE statement. If `expr` is omitted type for column is required.
INSERT without list of columns will skip such column, so SELECT/INSERT invariant is preserved - the dump obtained using `SELECT *` can be inserted back into the table using INSERT without specifying the list of columns.
Ephemeral column. Columns of this type are not stored in the table and it is not possible to SELECT from them. The only purpose of ephemeral columns is to build default value expressions of other columns from them.
An insert without explicitly specified columns will skip columns of this type. This is to preserve the invariant that the result of a `SELECT *` can always be inserted back into the table using `INSERT`.
Example:
@ -205,7 +204,7 @@ CREATE OR REPLACE TABLE test
hexed FixedString(4) DEFAULT unhex(unhexed)
)
ENGINE = MergeTree
ORDER BY id
ORDER BY id;
INSERT INTO test (id, unhexed) Values (1, '5a90b714');
@ -227,9 +226,9 @@ hex(hexed): 5A90B714
`ALIAS expr`
Synonym. Such a column isn’t stored in the table at all.
Its values can’t be inserted in a table, and it is not substituted when using an asterisk in a SELECT query.
It can be used in SELECTs if the alias is expanded during query parsing.
Calculated columns (synonym). Column of this type are not stored in the table and it is not possible to INSERT values into them.
When SELECT queries explicitly reference columns of this type, the value is computed at query time from `expr`. By default, `SELECT *` excludes ALIAS columns. This behavior can be disabled with setting `asteriks_include_alias_columns`.
When using the ALTER query to add new columns, old data for these columns is not written. Instead, when reading old data that does not have values for the new columns, expressions are computed on the fly by default. However, if running the expressions requires different columns that are not indicated in the query, these columns will additionally be read, but only for the blocks of data that need it.

6
docs/en/sql-reference/statements/create/view.md
View File

@ -70,6 +70,12 @@ A materialized view is implemented as follows: when inserting data to the table
Materialized views in ClickHouse use **column names** instead of column order during insertion into destination table. If some column names are not present in the `SELECT` query result, ClickHouse uses a default value, even if the column is not [Nullable](../../data-types/nullable.md). A safe practice would be to add aliases for every column when using Materialized views.
Materialized views in ClickHouse are implemented more like insert triggers. If there’s some aggregation in the view query, it’s applied only to the batch of freshly inserted data. Any changes to existing data of source table (like update, delete, drop partition, etc.) does not change the materialized view.
Materialized views in ClickHouse do not have deterministic behaviour in case of errors. This means that blocks that had been already written will be preserved in the destination table, but all blocks after error will not.
By default if pushing to one of views fails, then the INSERT query will fail too, and some blocks may not be written to the destination table. This can be changed using `materialized_views_ignore_errors` setting (you should set it for `INSERT` query), if you will set `materialized_views_ignore_errors=true`, then any errors while pushing to views will be ignored and all blocks will be written to the destination table.
Also note, that `materialized_views_ignore_errors` set to `true` by default for `system.*_log` tables.
:::
If you specify `POPULATE`, the existing table data is inserted into the view when creating it, as if making a `CREATE TABLE ... AS SELECT ...` . Otherwise, the query contains only the data inserted in the table after creating the view. We **do not recommend** using `POPULATE`, since data inserted in the table during the view creation will not be inserted in it.

2
docs/ru/engines/table-engines/mergetree-family/collapsingmergetree.md
View File

@ -89,7 +89,7 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster]
└─────────────────────┴───────────┴──────────┴──────┘
```
Первая строка отменяет предыдущее состояние объекта (пользователя). Она должен повторять все поля из ключа сортировки для отменённого состояния за исключением `Sign`.
Первая строка отменяет предыдущее состояние объекта (пользователя). Она должна повторять все поля из ключа сортировки для отменённого состояния за исключением `Sign`.
Вторая строка содержит текущее состояние.

6
docs/ru/engines/table-engines/mergetree-family/mergetree.md
View File

@ -584,7 +584,7 @@ TTL d + INTERVAL 1 MONTH GROUP BY k1, k2 SET x = max(x), y = min(y);
Данные с истекшим `TTL` удаляются, когда ClickHouse мёржит куски данных.
Когда ClickHouse видит, что некоторые данные устарели, он выполняет внеплановые мёржи. Для управление частотой подобных мёржей, можно задать настройку `merge_with_ttl_timeout`. Если её значение слишком низкое, придется выполнять много внеплановых мёржей, которые могут начать потреблять значительную долю ресурсов сервера.
Когда ClickHouse видит, что некоторые данные устарели, он выполняет внеплановые мёржи. Для управления частотой подобных мёржей, можно задать настройку `merge_with_ttl_timeout`. Если её значение слишком низкое, придется выполнять много внеплановых мёржей, которые могут начать потреблять значительную долю ресурсов сервера.
Если вы выполните запрос `SELECT` между слияниями вы можете получить устаревшие данные. Чтобы избежать этого используйте запрос [OPTIMIZE](../../../engines/table-engines/mergetree-family/mergetree.md#misc_operations-optimize) перед `SELECT`.
@ -679,7 +679,7 @@ TTL d + INTERVAL 1 MONTH GROUP BY k1, k2 SET x = max(x), y = min(y);
- `policy_name_N` — название политики. Названия политик должны быть уникальны.
- `volume_name_N` — название тома. Названия томов должны быть уникальны.
- `disk` — диск, находящийся внутри тома.
- `max_data_part_size_bytes` — максимальный размер куска данных, который может находится на любом из дисков этого тома. Если в результате слияния размер куска ожидается больше, чем max_data_part_size_bytes, то этот кусок будет записан в следующий том. В основном эта функция позволяет хранить новые / мелкие куски на горячем (SSD) томе и перемещать их на холодный (HDD) том, когда они достигают большого размера. Не используйте этот параметр, если политика имеет только один том.
- `max_data_part_size_bytes` — максимальный размер куска данных, который может находиться на любом из дисков этого тома. Если в результате слияния размер куска ожидается больше, чем max_data_part_size_bytes, то этот кусок будет записан в следующий том. В основном эта функция позволяет хранить новые / мелкие куски на горячем (SSD) томе и перемещать их на холодный (HDD) том, когда они достигают большого размера. Не используйте этот параметр, если политика имеет только один том.
- `move_factor` — доля доступного свободного места на томе, если места становится меньше, то данные начнут перемещение на следующий том, если он есть (по умолчанию 0.1). Для перемещения куски сортируются по размеру от большего к меньшему (по убыванию) и выбираются куски, совокупный размер которых достаточен для соблюдения условия `move_factor`, если совокупный размер всех партов недостаточен, будут перемещены все парты.
- `prefer_not_to_merge` — Отключает слияние кусков данных, хранящихся на данном томе. Если данная настройка включена, то слияние данных, хранящихся на данном томе, не допускается. Это позволяет контролировать работу ClickHouse с медленными дисками.
@ -730,7 +730,7 @@ TTL d + INTERVAL 1 MONTH GROUP BY k1, k2 SET x = max(x), y = min(y);
В приведенном примере, политика `hdd_in_order` реализует прицип [round-robin](https://ru.wikipedia.org/wiki/Round-robin_(%D0%B0%D0%BB%D0%B3%D0%BE%D1%80%D0%B8%D1%82%D0%BC)). Так как в политике есть всего один том (`single`), то все записи производятся на его диски по круговому циклу. Такая политика может быть полезна при наличии в системе нескольких похожих дисков, но при этом не сконфигурирован RAID. Учтите, что каждый отдельный диск ненадёжен и чтобы не потерять важные данные это необходимо скомпенсировать за счет хранения данных в трёх копиях.
Если система содержит диски различных типов, то может пригодиться политика `moving_from_ssd_to_hdd`. В томе `hot` находится один SSD-диск (`fast_ssd`), а также задается ограничение на максимальный размер куска, который может храниться на этом томе (1GB). Все куски такой таблицы больше 1GB будут записываться сразу на том `cold`, в котором содержится один HDD-диск `disk1`. Также, при заполнении диска `fast_ssd` более чем на 80% данные будут переносится на диск `disk1` фоновым процессом.
Если система содержит диски различных типов, то может пригодиться политика `moving_from_ssd_to_hdd`. В томе `hot` находится один SSD-диск (`fast_ssd`), а также задается ограничение на максимальный размер куска, который может храниться на этом томе (1GB). Все куски такой таблицы больше 1GB будут записываться сразу на том `cold`, в котором содержится один HDD-диск `disk1`. Также при заполнении диска `fast_ssd` более чем на 80% данные будут переноситься на диск `disk1` фоновым процессом.
Порядок томов в политиках хранения важен, при достижении условий на переполнение тома данные переносятся на следующий. Порядок дисков в томах так же важен, данные пишутся по очереди на каждый из них.

1
docs/ru/interfaces/third-party/client-libraries.md vendored
View File

@ -8,6 +8,7 @@ sidebar_label: "Клиентские библиотеки от сторонни
:::danger "Disclaimer"
Яндекс не поддерживает перечисленные ниже библиотеки и не проводит тщательного тестирования для проверки их качества.
:::
- Python:
- [infi.clickhouse_orm](https://github.com/Infinidat/infi.clickhouse_orm)

9
docs/ru/interfaces/third-party/gui.md vendored
View File

@ -177,19 +177,20 @@ sidebar_label: "Визуальные интерфейсы от сторонни
### Yandex DataLens {#yandex-datalens}
[Yandex DataLens](https://cloud.yandex.ru/services/datalens) — cервис визуализации и анализа данных.
[Yandex DataLens](https://datalens.yandex.ru) — cервис визуализации и анализа данных.
Основные возможности:
- Широкий выбор инструментов визуализации, от простых столбчатых диаграмм до сложных дашбордов.
- Возможность опубликовать дашборды на широкую аудиторию.
- Поддержка множества источников данных, включая ClickHouse.
- Хранение материализованных данных в кластере ClickHouse DataLens.
Для небольших проектов DataLens [доступен бесплатно](https://cloud.yandex.ru/docs/datalens/pricing), в том числе и для коммерческого использования.
DataLens [доступен бесплатно](https://cloud.yandex.ru/docs/datalens/pricing), в том числе и для коммерческого использования.
- [Знакомство с DataLens]((https://youtu.be/57ngi_6BINE).
- [Чат сообщества DataLens](https://t.me/YandexDataLens)
- [Документация DataLens](https://cloud.yandex.ru/docs/datalens/).
- [Пособие по визуализации данных из ClickHouse](https://cloud.yandex.ru/docs/solutions/datalens/data-from-ch-visualization).
- [Сценарий по визуализации данных из ClickHouse](https://cloud.yandex.ru/docs/solutions/datalens/data-from-ch-visualization).
### Holistics Software {#holistics-software}

12
docs/ru/operations/clickhouse-keeper.md
View File

@ -325,21 +325,21 @@ clickhouse-keeper-converter --zookeeper-logs-dir /var/lib/zookeeper/version-2 --
Например, для кластера из 3 нод, алгоритм кворума продолжает работать при отказе не более чем одной ноды.
Конфигурация кластера может быть изменена динамически с некоторыми ограничениями.
Переконфигурация также использует Raft, поэтому для добавление новой ноды кластера или исключения старой ноды из него требуется достижения кворума в рамках текущей конфигурации кластера.
Переконфигурация также использует Raft, поэтому для добавления новой ноды кластера или исключения старой ноды требуется достижение кворума в рамках текущей конфигурации кластера.
Если в вашем кластере произошел отказ большего числа нод, чем допускает Raft для вашей текущей конфигурации и у вас нет возможности восстановить их работоспособность, Raft перестанет работать и не позволит изменить конфигурацию стандартным механизмом.
Тем не менее ClickHousr Keeper имеет возможность запуститься в режиме восстановления, который позволяет переконфигурировать класте используя только одну ноду кластера.
Тем не менее ClickHouse Keeper имеет возможность запуститься в режиме восстановления, который позволяет переконфигурировать кластер используя только одну ноду кластера.
Этот механизм может использоваться только как крайняя мера, когда вы не можете восстановить существующие ноды кластера или запустить новый сервер с тем же идентификатором.
Важно:
- Удостоверьтесь, что отказавшие ноды не смогут в дальнейшем подключиться к кластеру в будущем.
- Не запускайте новые ноды, пока не завешите процедуру ниже.
- Не запускайте новые ноды, пока не завершите процедуру ниже.
После того, как выполнили действия выше выполните следующие шаги.
1. Выберете одну ноду Keeper, которая станет новым лидером. Учтите, что данные которые с этой ноды будут испольщзованы всем кластером, поэтому рекомендуется выбрать ноду с наиболее актуальным состоянием.
1. Выберете одну ноду Keeper, которая станет новым лидером. Учтите, что данные с этой ноды будут использованы всем кластером, поэтому рекомендуется выбрать ноду с наиболее актуальным состоянием.
2. Перед дальнейшими действиям сделайте резервную копию данных из директорий `log_storage_path` и `snapshot_storage_path`.
3. Измените настройки на всех нодах кластера, которые вы собираетесь использовать.
4. Отправьте команду `rcvr` на ноду, которую вы выбрали или остановите ее и запустите заново с аргументом `--force-recovery`. Это переведет ноду в режим восстановления.
4. Отправьте команду `rcvr` на ноду, которую вы выбрали, или остановите ее и запустите заново с аргументом `--force-recovery`. Это переведет ноду в режим восстановления.
5. Запускайте остальные ноды кластера по одной и проверяйте, что команда `mntr` возвращает `follower` в выводе состояния `zk_server_state` перед тем, как запустить следующую ноду.
6. Пока нода работает в режиме восстановления, лидер будет возвращать ошибку на запрос `mntr` пока кворум не будет достигнут с помощью новых нод. Любые запросы от клиентов и постедователей будут возвращать ошибку.
6. Пока нода работает в режиме восстановления, лидер будет возвращать ошибку на запрос `mntr` пока кворум не будет достигнут с помощью новых нод. Любые запросы от клиентов и последователей будут возвращать ошибку.
7. После достижения кворума лидер перейдет в нормальный режим работы и станет обрабатывать все запросы через Raft. Удостоверьтесь, что запрос `mntr` возвращает `leader` в выводе состояния `zk_server_state`.

1
docs/ru/operations/opentelemetry.md
View File

@ -10,6 +10,7 @@ ClickHouse поддерживает [OpenTelemetry](https://opentelemetry.io/)
:::danger "Предупреждение"
Поддержка стандарта экспериментальная и будет со временем меняться.
:::
## Обеспечение поддержки контекста трассировки в ClickHouse

3
docs/ru/operations/server-configuration-parameters/settings.md
View File

@ -26,6 +26,7 @@ ClickHouse перезагружает встроенные словари с з
:::danger "Внимание"
Лучше не использовать, если вы только начали работать с ClickHouse.
:::
Общий вид конфигурации:
@ -1064,6 +1065,7 @@ ClickHouse использует потоки из глобального пул
:::danger "Обратите внимание"
Завершающий слеш обязателен.
:::
**Пример**
@ -1330,6 +1332,7 @@ TCP порт для защищённого обмена данными с кли
:::danger "Обратите внимание"
Завершающий слеш обязателен.
:::
**Пример**

2
docs/ru/operations/storing-data.md
View File

@ -82,7 +82,7 @@ sidebar_label: "Хранение данных на внешних дисках"
- `type` — `encrypted`. Иначе зашифрованный диск создан не будет.
- `disk` — тип диска для хранения данных.
- `key` — ключ для шифрования и расшифровки. Тип: [Uint64](../sql-reference/data-types/int-uint.md). Вы можете использовать параметр `key_hex` для шифрования в шестнадцатеричной форме.
- `key` — ключ для шифрования и расшифровки. Тип: [UInt64](../sql-reference/data-types/int-uint.md). Вы можете использовать параметр `key_hex` для шифрования в шестнадцатеричной форме.
Вы можете указать несколько ключей, используя атрибут `id` (смотрите пример выше).
Необязательные параметры:

1
docs/ru/sql-reference/data-types/geo.md
View File

@ -10,6 +10,7 @@ ClickHouse поддерживает типы данных для отображ
:::danger "Предупреждение"
Сейчас использование типов данных для работы с географическими структурами является экспериментальной возможностью. Чтобы использовать эти типы данных, включите настройку `allow_experimental_geo_types = 1`.
:::
**См. также**
- [Хранение географических структур данных](https://ru.wikipedia.org/wiki/GeoJSON).

1
docs/ru/sql-reference/data-types/special-data-types/interval.md
View File

@ -10,6 +10,7 @@ sidebar_label: Interval
:::danger "Внимание"
Нельзя использовать типы данных `Interval` для хранения данных в таблице.
:::
Структура:

2
docs/ru/sql-reference/data-types/tuple.md
View File

@ -34,7 +34,7 @@ SELECT tuple(1,'a') AS x, toTypeName(x)
## Особенности работы с типами данных {#osobennosti-raboty-s-tipami-dannykh}
При создании кортежа «на лету» ClickHouse автоматически определяет тип каждого аргументов как минимальный из типов, который может сохранить значение аргумента. Если аргумент — [NULL](../../sql-reference/data-types/tuple.md#null-literal), то тип элемента кортежа — [Nullable](nullable.md).
При создании кортежа «на лету» ClickHouse автоматически определяет тип всех аргументов как минимальный из типов, который может сохранить значение аргумента. Если аргумент — [NULL](../../sql-reference/data-types/tuple.md#null-literal), то тип элемента кортежа — [Nullable](nullable.md).
Пример автоматического определения типа данных:

4
docs/ru/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-polygon.md
View File

@ -61,7 +61,7 @@ LAYOUT(POLYGON(STORE_POLYGON_KEY_COLUMN 1))
- Мультиполигон. Представляет из себя массив полигонов. Каждый полигон задается двумерным массивом точек — первый элемент этого массива задает внешнюю границу полигона,
последующие элементы могут задавать дырки, вырезаемые из него.
Точки могут задаваться массивом или кортежем из своих координат. В текущей реализации поддерживается только двумерные точки.
Точки могут задаваться массивом или кортежем из своих координат. В текущей реализации поддерживаются только двумерные точки.
Пользователь может [загружать свои собственные данные](../../../sql-reference/dictionaries/external-dictionaries/external-dicts-dict-sources.md) во всех поддерживаемых ClickHouse форматах.
@ -80,7 +80,7 @@ LAYOUT(POLYGON(STORE_POLYGON_KEY_COLUMN 1))
- `POLYGON`. Синоним к `POLYGON_INDEX_CELL`.
Запросы к словарю осуществляются с помощью стандартных [функций](../../../sql-reference/functions/ext-dict-functions.md) для работы со внешними словарями.
Важным отличием является то, что здесь ключами будут являются точки, для которых хочется найти содержащий их полигон.
Важным отличием является то, что здесь ключами являются точки, для которых хочется найти содержащий их полигон.
**Пример**

1
docs/ru/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-structure.md
View File

@ -59,6 +59,7 @@ ClickHouse поддерживает следующие виды ключей:
:::danger "Обратите внимание"
Ключ не надо дополнительно описывать в атрибутах.
:::
### Числовой ключ {#ext_dict-numeric-key}

2
docs/ru/sql-reference/dictionaries/external-dictionaries/external-dicts.md
View File

@ -14,7 +14,7 @@ ClickHouse:
- Периодически обновляет их и динамически подгружает отсутствующие значения.
- Позволяет создавать внешние словари с помощью xml-файлов или [DDL-запросов](../../statements/create/dictionary.md#create-dictionary-query).
Конфигурация внешних словарей может находится в одном или нескольких xml-файлах. Путь к конфигурации указывается в параметре [dictionaries_config](../../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-dictionaries_config).
Конфигурация внешних словарей может находиться в одном или нескольких xml-файлах. Путь к конфигурации указывается в параметре [dictionaries_config](../../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-dictionaries_config).
Словари могут загружаться при старте сервера или при первом использовании, в зависимости от настройки [dictionaries_lazy_load](../../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-dictionaries_lazy_load).

2
docs/ru/sql-reference/functions/introspection.md
View File

@ -22,7 +22,7 @@ sidebar_label: "Функции интроспекции"
ClickHouse сохраняет отчеты профилировщика в [журнал трассировки](../../operations/system-tables/trace_log.md#system_tables-trace_log) в системной таблице. Убедитесь, что таблица и профилировщик настроены правильно.
## addresssToLine {#addresstoline}
## addressToLine {#addresstoline}
Преобразует адрес виртуальной памяти внутри процесса сервера ClickHouse в имя файла и номер строки в исходном коде ClickHouse.

1
docs/ru/sql-reference/operators/exists.md
View File

@ -9,6 +9,7 @@ slug: /ru/sql-reference/operators/exists
:::danger "Предупреждение"
Ссылки на таблицы или столбцы основного запроса не поддерживаются в подзапросе.
:::
**Синтаксис**

12
docs/ru/sql-reference/operators/in.md
View File

@ -38,9 +38,9 @@ SELECT '1' IN (SELECT 1);
└──────────────────────┘
```
Если в качестве правой части оператора указано имя таблицы (например, `UserID IN users`), то это эквивалентно подзапросу `UserID IN (SELECT * FROM users)`. Это используется при работе с внешними данными, отправляемым вместе с запросом. Например, вместе с запросом может быть отправлено множество идентификаторов посетителей, загруженное во временную таблицу users, по которому следует выполнить фильтрацию.
Если в качестве правой части оператора указано имя таблицы (например, `UserID IN users`), то это эквивалентно подзапросу `UserID IN (SELECT * FROM users)`. Это используется при работе с внешними данными, отправляемыми вместе с запросом. Например, вместе с запросом может быть отправлено множество идентификаторов посетителей, загруженное во временную таблицу users, по которому следует выполнить фильтрацию.
Если в качестве правой части оператора, указано имя таблицы, имеющий движок Set (подготовленное множество, постоянно находящееся в оперативке), то множество не будет создаваться заново при каждом запросе.
Если в качестве правой части оператора, указано имя таблицы, имеющей движок Set (подготовленное множество, постоянно находящееся в оперативке), то множество не будет создаваться заново при каждом запросе.
В подзапросе может быть указано более одного столбца для фильтрации кортежей.
Пример:
@ -49,9 +49,9 @@ SELECT '1' IN (SELECT 1);
SELECT (CounterID, UserID) IN (SELECT CounterID, UserID FROM ...) FROM ...
```
Типы столбцов слева и справа оператора IN, должны совпадать.
Типы столбцов слева и справа оператора IN должны совпадать.
Оператор IN и подзапрос могут встречаться в любой части запроса, в том числе в агрегатных и лямбда функциях.
Оператор IN и подзапрос могут встречаться в любой части запроса, в том числе в агрегатных и лямбда-функциях.
Пример:
``` sql
@ -122,7 +122,7 @@ FROM t_null
Существует два варианта IN-ов с подзапросами (аналогично для JOIN-ов): обычный `IN` / `JOIN` и `GLOBAL IN` / `GLOBAL JOIN`. Они отличаются способом выполнения при распределённой обработке запроса.
:::note "Attention"
:::note "Внимание"
Помните, что алгоритмы, описанные ниже, могут работать иначе в зависимости от [настройки](../../operations/settings/settings.md) `distributed_product_mode`.
:::
При использовании обычного IN-а, запрос отправляется на удалённые серверы, и на каждом из них выполняются подзапросы в секциях `IN` / `JOIN`.
@ -228,7 +228,7 @@ SELECT CounterID, count() FROM distributed_table_1 WHERE UserID IN (SELECT UserI
SETTINGS max_parallel_replicas=3
```
преобразуются на каждом сервере в
преобразуется на каждом сервере в
```sql
SELECT CounterID, count() FROM local_table_1 WHERE UserID IN (SELECT UserID FROM local_table_2 WHERE CounterID < 100)

1
docs/ru/sql-reference/operators/index.md
View File

@ -263,6 +263,7 @@ SELECT toDateTime('2014-10-26 00:00:00', 'Europe/Moscow') AS time, time + 60 * 6
│ 2014-10-26 00:00:00 │ 2014-10-26 23:00:00 │ 2014-10-27 00:00:00 │
└─────────────────────┴─────────────────────┴─────────────────────┘
```
:::
**Смотрите также**

2
docs/ru/sql-reference/statements/alter/view.md
View File

@ -6,7 +6,7 @@ sidebar_label: VIEW
# Выражение ALTER TABLE … MODIFY QUERY {#alter-modify-query}
Вы можеие изменить запрос `SELECT`, который был задан при создании [материализованного представления](../create/view.md#materialized), с помощью запроса 'ALTER TABLE … MODIFY QUERY'. Используйте его если при создании материализованного представления не использовалась секция `TO [db.]name`. Настройка `allow_experimental_alter_materialized_view_structure` должна быть включена.
Вы можете изменить запрос `SELECT`, который был задан при создании [материализованного представления](../create/view.md#materialized), с помощью запроса 'ALTER TABLE … MODIFY QUERY'. Используйте его если при создании материализованного представления не использовалась секция `TO [db.]name`. Настройка `allow_experimental_alter_materialized_view_structure` должна быть включена.
Если при создании материализованного представления использовалась конструкция `TO [db.]name`, то для изменения отсоедините представление с помощью [DETACH](../detach.md), измените таблицу с помощью [ALTER TABLE](index.md), а затем снова присоедините запрос с помощью [ATTACH](../attach.md).

1
docs/ru/sql-reference/statements/optimize.md
View File

@ -10,6 +10,7 @@ sidebar_label: OPTIMIZE
:::danger "Внимание"
`OPTIMIZE` не устраняет причину появления ошибки `Too many parts`.
:::
**Синтаксис**

13
programs/copier/ClusterCopier.cpp
View File

@ -1757,8 +1757,7 @@ void ClusterCopier::dropParticularPartitionPieceFromAllHelpingTables(const TaskT
LOG_INFO(log, "All helping tables dropped partition {}", partition_name);
}
String ClusterCopier::getRemoteCreateTable(
const DatabaseAndTableName & table, Connection & connection, const Settings & settings)
String ClusterCopier::getRemoteCreateTable(const DatabaseAndTableName & table, Connection & connection, const Settings & settings)
{
auto remote_context = Context::createCopy(context);
remote_context->setSettings(settings);
@ -1777,8 +1776,10 @@ ASTPtr ClusterCopier::getCreateTableForPullShard(const ConnectionTimeouts & time
{
/// Fetch and parse (possibly) new definition
auto connection_entry = task_shard.info.pool->get(timeouts, &task_cluster->settings_pull, true);
String create_query_pull_str
= getRemoteCreateTable(task_shard.task_table.table_pull, *connection_entry, task_cluster->settings_pull);
String create_query_pull_str = getRemoteCreateTable(
task_shard.task_table.table_pull,
*connection_entry,
task_cluster->settings_pull);
ParserCreateQuery parser_create_query;
const auto & settings = getContext()->getSettingsRef();
@ -1867,8 +1868,8 @@ std::set<String> ClusterCopier::getShardPartitions(const ConnectionTimeouts & ti
String query;
{
WriteBufferFromOwnString wb;
wb << "SELECT DISTINCT " << partition_name << " AS partition FROM"
<< " " << getQuotedTable(task_shard.table_read_shard) << " ORDER BY partition DESC";
wb << "SELECT " << partition_name << " AS partition FROM "
<< getQuotedTable(task_shard.table_read_shard) << " GROUP BY partition ORDER BY partition DESC";
query = wb.str();
}

2
programs/install/Install.cpp
View File

@ -20,7 +20,7 @@
#include <Common/formatReadable.h>
#include <Common/Config/ConfigProcessor.h>
#include <Common/OpenSSLHelpers.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Common/getResource.h>
#include <base/sleep.h>
#include <IO/ReadBufferFromFileDescriptor.h>

3
programs/server/Server.cpp
View File

@ -67,7 +67,6 @@
#include <TableFunctions/registerTableFunctions.h>
#include <Formats/registerFormats.h>
#include <Storages/registerStorages.h>
#include <QueryPipeline/ConnectionCollector.h>
#include <Dictionaries/registerDictionaries.h>
#include <Disks/registerDisks.h>
#include <IO/Resource/registerSchedulerNodes.h>
@ -816,8 +815,6 @@ try
}
);
ConnectionCollector::init(global_context, server_settings.max_threads_for_connection_collector);
bool has_zookeeper = config().has("zookeeper");
zkutil::ZooKeeperNodeCache main_config_zk_node_cache([&] { return global_context->getZooKeeper(); });

4
programs/server/config.xml
View File

@ -348,10 +348,6 @@
<background_distributed_schedule_pool_size>16</background_distributed_schedule_pool_size>
-->
<!-- Number of workers to recycle connections in background (see also drain_timeout).
If the pool is full, connection will be drained synchronously. -->
<!-- <max_threads_for_connection_collector>10</max_threads_for_connection_collector> -->
<!-- On memory constrained environments you may have to set this to value larger than 1.
-->
<max_server_memory_usage_to_ram_ratio>0.9</max_server_memory_usage_to_ram_ratio>

31
src/Analyzer/FunctionNode.cpp
View File

@ -2,18 +2,21 @@
#include <Common/SipHash.h>
#include <Common/FieldVisitorToString.h>
#include <DataTypes/IDataType.h>
#include <Analyzer/ConstantNode.h>
#include <IO/WriteBufferFromString.h>
#include <IO/Operators.h>
#include <DataTypes/IDataType.h>
#include <DataTypes/DataTypeSet.h>
#include <Parsers/ASTFunction.h>
#include <Functions/IFunction.h>
#include <AggregateFunctions/IAggregateFunction.h>
#include <Analyzer/Utils.h>
#include <Analyzer/ConstantNode.h>
#include <Analyzer/IdentifierNode.h>
namespace DB
@ -44,17 +47,29 @@ const DataTypes & FunctionNode::getArgumentTypes() const
ColumnsWithTypeAndName FunctionNode::getArgumentColumns() const
{
const auto & arguments = getArguments().getNodes();
size_t arguments_size = arguments.size();
ColumnsWithTypeAndName argument_columns;
argument_columns.reserve(arguments.size());
for (const auto & arg : arguments)
for (size_t i = 0; i < arguments_size; ++i)
{
ColumnWithTypeAndName argument;
argument.type = arg->getResultType();
if (auto * constant = arg->as<ConstantNode>())
argument.column = argument.type->createColumnConst(1, constant->getValue());
argument_columns.push_back(std::move(argument));
const auto & argument = arguments[i];
ColumnWithTypeAndName argument_column;
if (isNameOfInFunction(function_name) && i == 1)
argument_column.type = std::make_shared<DataTypeSet>();
else
argument_column.type = argument->getResultType();
auto * constant = argument->as<ConstantNode>();
if (constant && !isNotCreatable(argument_column.type))
argument_column.column = argument_column.type->createColumnConst(1, constant->getValue());
argument_columns.push_back(std::move(argument_column));
}
return argument_columns;
}

17
src/Analyzer/InDepthQueryTreeVisitor.h
View File

@ -99,8 +99,9 @@ class InDepthQueryTreeVisitorWithContext
public:
using VisitQueryTreeNodeType = std::conditional_t<const_visitor, const QueryTreeNodePtr, QueryTreeNodePtr>;
explicit InDepthQueryTreeVisitorWithContext(ContextPtr context)
explicit InDepthQueryTreeVisitorWithContext(ContextPtr context, size_t initial_subquery_depth = 0)
: current_context(std::move(context))
, subquery_depth(initial_subquery_depth)
{}
/// Return true if visitor should traverse tree top to bottom, false otherwise
@ -125,11 +126,17 @@ public:
return current_context->getSettingsRef();
}
size_t getSubqueryDepth() const
{
return subquery_depth;
}
void visit(VisitQueryTreeNodeType & query_tree_node)
{
auto current_scope_context_ptr = current_context;
SCOPE_EXIT(
current_context = std::move(current_scope_context_ptr);
--subquery_depth;
);
if (auto * query_node = query_tree_node->template as<QueryNode>())
@ -137,6 +144,8 @@ public:
else if (auto * union_node = query_tree_node->template as<UnionNode>())
current_context = union_node->getContext();
++subquery_depth;
bool traverse_top_to_bottom = getDerived().shouldTraverseTopToBottom();
if (!traverse_top_to_bottom)
visitChildren(query_tree_node);
@ -145,7 +154,12 @@ public:
if (traverse_top_to_bottom)
visitChildren(query_tree_node);
getDerived().leaveImpl(query_tree_node);
}
void leaveImpl(VisitQueryTreeNodeType & node [[maybe_unused]])
{}
private:
Derived & getDerived()
{
@ -172,6 +186,7 @@ private:
}
ContextPtr current_context;
size_t subquery_depth = 0;
};
template <typename Derived>

6
src/Analyzer/JoinNode.h
View File

@ -106,6 +106,12 @@ public:
return locality;
}
/// Set join locality
void setLocality(JoinLocality locality_value)
{
locality = locality_value;
}
/// Get join strictness
JoinStrictness getStrictness() const
{

2
src/Analyzer/Passes/AutoFinalOnQueryPass.cpp
View File

@ -42,7 +42,7 @@ private:
return;
const auto & storage = table_node ? table_node->getStorage() : table_function_node->getStorage();
bool is_final_supported = storage && storage->supportsFinal() && !storage->isRemote();
bool is_final_supported = storage && storage->supportsFinal();
if (!is_final_supported)
return;

8
src/Analyzer/Passes/LogicalExpressionOptimizer.cpp → src/Analyzer/Passes/LogicalExpressionOptimizerPass.cpp
View File

@ -7,8 +7,6 @@
#include <Analyzer/ConstantNode.h>
#include <Analyzer/HashUtils.h>
#include <DataTypes/DataTypeString.h>
namespace DB
{
@ -100,6 +98,9 @@ private:
}
}
if (and_operands.size() == function_node.getArguments().getNodes().size())
return;
if (and_operands.size() == 1)
{
/// AND operator can have UInt8 or bool as its type.
@ -207,6 +208,9 @@ private:
or_operands.push_back(std::move(in_function));
}
if (or_operands.size() == function_node.getArguments().getNodes().size())
return;
if (or_operands.size() == 1)
{
/// if the result type of operand is the same as the result type of OR

4
src/Analyzer/Passes/OptimizeGroupByFunctionKeysPass.cpp
View File

@ -69,8 +69,7 @@ private:
for (auto it = function_arguments.rbegin(); it != function_arguments.rend(); ++it)
candidates.push_back({ *it, is_deterministic });
// Using DFS we traverse function tree and try to find if it uses other keys as function arguments.
// TODO: Also process CONSTANT here. We can simplify GROUP BY x, x + 1 to GROUP BY x.
/// Using DFS we traverse function tree and try to find if it uses other keys as function arguments.
while (!candidates.empty())
{
auto [candidate, parents_are_only_deterministic] = candidates.back();
@ -108,6 +107,7 @@ private:
return false;
}
}
return true;
}

167
src/Analyzer/Passes/QueryAnalysisPass.cpp
View File

@ -193,13 +193,9 @@ namespace ErrorCodes
* lookup should not be continued, and exception must be thrown because if lookup continues identifier can be resolved from parent scope.
*
* TODO: Update exception messages
* TODO: JOIN TREE subquery constant columns
* TODO: Table identifiers with optional UUID.
* TODO: Lookup functions arrayReduce(sum, [1, 2, 3]);
* TODO: SELECT (compound_expression).*, (compound_expression).COLUMNS are not supported on parser level.
* TODO: SELECT a.b.c.*, a.b.c.COLUMNS. Qualified matcher where identifier size is greater than 2 are not supported on parser level.
* TODO: Support function identifier resolve from parent query scope, if lambda in parent scope does not capture any columns.
* TODO: Scalar subqueries cache.
*/
namespace
@ -701,7 +697,9 @@ struct IdentifierResolveScope
}
if (auto * union_node = scope_node->as<UnionNode>())
{
context = union_node->getContext();
}
else if (auto * query_node = scope_node->as<QueryNode>())
{
context = query_node->getContext();
@ -1336,6 +1334,9 @@ private:
/// Global resolve expression node to projection names map
std::unordered_map<QueryTreeNodePtr, ProjectionNames> resolved_expressions;
/// Global resolve expression node to tree size
std::unordered_map<QueryTreeNodePtr, size_t> node_to_tree_size;
/// Global scalar subquery to scalar value map
std::unordered_map<QueryTreeNodePtrWithHash, Block> scalar_subquery_to_scalar_value;
@ -1864,7 +1865,10 @@ void QueryAnalyzer::evaluateScalarSubqueryIfNeeded(QueryTreeNodePtr & node, Iden
Block scalar_block;
QueryTreeNodePtrWithHash node_with_hash(node);
auto node_without_alias = node->clone();
node_without_alias->removeAlias();
QueryTreeNodePtrWithHash node_with_hash(node_without_alias);
auto scalar_value_it = scalar_subquery_to_scalar_value.find(node_with_hash);
if (scalar_value_it != scalar_subquery_to_scalar_value.end())
@ -1954,21 +1958,7 @@ void QueryAnalyzer::evaluateScalarSubqueryIfNeeded(QueryTreeNodePtr & node, Iden
*
* Example: SELECT (SELECT 2 AS x, x)
*/
NameSet block_column_names;
size_t unique_column_name_counter = 1;
for (auto & column_with_type : block)
{
if (!block_column_names.contains(column_with_type.name))
{
block_column_names.insert(column_with_type.name);
continue;
}
column_with_type.name += '_';
column_with_type.name += std::to_string(unique_column_name_counter);
++unique_column_name_counter;
}
makeUniqueColumnNamesInBlock(block);
scalar_block.insert({
ColumnTuple::create(block.getColumns()),
@ -2348,7 +2338,13 @@ QueryTreeNodePtr QueryAnalyzer::tryResolveTableIdentifierFromDatabaseCatalog(con
storage_id = context->resolveStorageID(storage_id);
bool is_temporary_table = storage_id.getDatabaseName() == DatabaseCatalog::TEMPORARY_DATABASE;
auto storage = DatabaseCatalog::instance().tryGetTable(storage_id, context);
StoragePtr storage;
if (is_temporary_table)
storage = DatabaseCatalog::instance().getTable(storage_id, context);
else
storage = DatabaseCatalog::instance().tryGetTable(storage_id, context);
if (!storage)
return {};
@ -2914,7 +2910,10 @@ QueryTreeNodePtr QueryAnalyzer::tryResolveIdentifierFromTableExpression(const Id
break;
IdentifierLookup column_identifier_lookup = {qualified_identifier_with_removed_part, IdentifierLookupContext::EXPRESSION};
if (tryBindIdentifierToAliases(column_identifier_lookup, scope) ||
if (tryBindIdentifierToAliases(column_identifier_lookup, scope))
break;
if (table_expression_data.should_qualify_columns &&
tryBindIdentifierToTableExpressions(column_identifier_lookup, table_expression_node, scope))
break;
@ -3018,11 +3017,39 @@ QueryTreeNodePtr QueryAnalyzer::tryResolveIdentifierFromJoin(const IdentifierLoo
resolved_identifier = std::move(result_column_node);
}
else if (scope.joins_count == 1 && scope.context->getSettingsRef().single_join_prefer_left_table)
else if (left_resolved_identifier->isEqual(*right_resolved_identifier, IQueryTreeNode::CompareOptions{.compare_aliases = false}))
{
const auto & identifier_path_part = identifier_lookup.identifier.front();
auto * left_resolved_identifier_column = left_resolved_identifier->as<ColumnNode>();
auto * right_resolved_identifier_column = right_resolved_identifier->as<ColumnNode>();
if (left_resolved_identifier_column && right_resolved_identifier_column)
{
const auto & left_column_source_alias = left_resolved_identifier_column->getColumnSource()->getAlias();
const auto & right_column_source_alias = right_resolved_identifier_column->getColumnSource()->getAlias();
/** If column from right table was resolved using alias, we prefer column from right table.
*
* Example: SELECT dummy FROM system.one JOIN system.one AS A ON A.dummy = system.one.dummy;
*
* If alias is specified for left table, and alias is not specified for right table and identifier was resolved
* without using left table alias, we prefer column from right table.
*
* Example: SELECT dummy FROM system.one AS A JOIN system.one ON A.dummy = system.one.dummy;
*
* Otherwise we prefer column from left table.
*/
if (identifier_path_part == right_column_source_alias)
return right_resolved_identifier;
else if (!left_column_source_alias.empty() &&
right_column_source_alias.empty() &&
identifier_path_part != left_column_source_alias)
return right_resolved_identifier;
}
return left_resolved_identifier;
}
else if (left_resolved_identifier->isEqual(*right_resolved_identifier, IQueryTreeNode::CompareOptions{.compare_aliases = false}))
else if (scope.joins_count == 1 && scope.context->getSettingsRef().single_join_prefer_left_table)
{
return left_resolved_identifier;
}
@ -4466,6 +4493,7 @@ ProjectionNames QueryAnalyzer::resolveFunction(QueryTreeNodePtr & node, Identifi
bool is_special_function_dict_get = false;
bool is_special_function_join_get = false;
bool is_special_function_exists = false;
bool is_special_function_if = false;
if (!lambda_expression_untyped)
{
@ -4473,6 +4501,7 @@ ProjectionNames QueryAnalyzer::resolveFunction(QueryTreeNodePtr & node, Identifi
is_special_function_dict_get = functionIsDictGet(function_name);
is_special_function_join_get = functionIsJoinGet(function_name);
is_special_function_exists = function_name == "exists";
is_special_function_if = function_name == "if";
auto function_name_lowercase = Poco::toLower(function_name);
@ -4571,6 +4600,60 @@ ProjectionNames QueryAnalyzer::resolveFunction(QueryTreeNodePtr & node, Identifi
is_special_function_in = true;
}
if (is_special_function_if && !function_node_ptr->getArguments().getNodes().empty())
{
/** Handle special case with constant If function, even if some of the arguments are invalid.
*
* SELECT if(hasColumnInTable('system', 'numbers', 'not_existing_column'), not_existing_column, 5) FROM system.numbers;
*/
auto & if_function_arguments = function_node_ptr->getArguments().getNodes();
auto if_function_condition = if_function_arguments[0];
resolveExpressionNode(if_function_condition, scope, false /*allow_lambda_expression*/, false /*allow_table_expression*/);
auto constant_condition = tryExtractConstantFromConditionNode(if_function_condition);
if (constant_condition.has_value() && if_function_arguments.size() == 3)
{
QueryTreeNodePtr constant_if_result_node;
QueryTreeNodePtr possibly_invalid_argument_node;
if (*constant_condition)
{
possibly_invalid_argument_node = if_function_arguments[2];
constant_if_result_node = if_function_arguments[1];
}
else
{
possibly_invalid_argument_node = if_function_arguments[1];
constant_if_result_node = if_function_arguments[2];
}
bool apply_constant_if_optimization = false;
try
{
resolveExpressionNode(possibly_invalid_argument_node,
scope,
false /*allow_lambda_expression*/,
false /*allow_table_expression*/);
}
catch (...)
{
apply_constant_if_optimization = true;
}
if (apply_constant_if_optimization)
{
auto result_projection_names = resolveExpressionNode(constant_if_result_node,
scope,
false /*allow_lambda_expression*/,
false /*allow_table_expression*/);
node = std::move(constant_if_result_node);
return result_projection_names;
}
}
}
/// Resolve function arguments
bool allow_table_expressions = is_special_function_in;
@ -5059,7 +5142,7 @@ ProjectionNames QueryAnalyzer::resolveFunction(QueryTreeNodePtr & node, Identifi
/// Do not constant fold get scalar functions
bool disable_constant_folding = function_name == "__getScalar" || function_name == "shardNum" ||
function_name == "shardCount";
function_name == "shardCount" || function_name == "hostName";
/** If function is suitable for constant folding try to convert it to constant.
* Example: SELECT plus(1, 1);
@ -5085,7 +5168,8 @@ ProjectionNames QueryAnalyzer::resolveFunction(QueryTreeNodePtr & node, Identifi
/** Do not perform constant folding if there are aggregate or arrayJoin functions inside function.
* Example: SELECT toTypeName(sum(number)) FROM numbers(10);
*/
if (column && isColumnConst(*column) && (!hasAggregateFunctionNodes(node) && !hasFunctionNode(node, "arrayJoin")))
if (column && isColumnConst(*column) && !typeid_cast<const ColumnConst *>(column.get())->getDataColumn().isDummy() &&
(!hasAggregateFunctionNodes(node) && !hasFunctionNode(node, "arrayJoin")))
{
/// Replace function node with result constant node
Field column_constant_value;
@ -5433,9 +5517,9 @@ ProjectionNames QueryAnalyzer::resolveExpressionNode(QueryTreeNodePtr & node, Id
}
}
if (node
&& scope.nullable_group_by_keys.contains(node)
&& !scope.expressions_in_resolve_process_stack.hasAggregateFunction())
validateTreeSize(node, scope.context->getSettingsRef().max_expanded_ast_elements, node_to_tree_size);
if (scope.nullable_group_by_keys.contains(node) && !scope.expressions_in_resolve_process_stack.hasAggregateFunction())
{
node = node->clone();
node->convertToNullable();
@ -6592,6 +6676,17 @@ void QueryAnalyzer::resolveQuery(const QueryTreeNodePtr & query_node, Identifier
/// Resolve query node sections.
NamesAndTypes projection_columns;
if (!scope.group_by_use_nulls)
{
projection_columns = resolveProjectionExpressionNodeList(query_node_typed.getProjectionNode(), scope);
if (query_node_typed.getProjection().getNodes().empty())
throw Exception(ErrorCodes::EMPTY_LIST_OF_COLUMNS_QUERIED,
"Empty list of columns in projection. In scope {}",
scope.scope_node->formatASTForErrorMessage());
}
if (query_node_typed.hasWith())
resolveExpressionNodeList(query_node_typed.getWithNode(), scope, true /*allow_lambda_expression*/, false /*allow_table_expression*/);
@ -6686,11 +6781,14 @@ void QueryAnalyzer::resolveQuery(const QueryTreeNodePtr & query_node, Identifier
convertLimitOffsetExpression(query_node_typed.getOffset(), "OFFSET", scope);
}
auto projection_columns = resolveProjectionExpressionNodeList(query_node_typed.getProjectionNode(), scope);
if (scope.group_by_use_nulls)
{
projection_columns = resolveProjectionExpressionNodeList(query_node_typed.getProjectionNode(), scope);
if (query_node_typed.getProjection().getNodes().empty())
throw Exception(ErrorCodes::EMPTY_LIST_OF_COLUMNS_QUERIED,
"Empty list of columns in projection. In scope {}",
scope.scope_node->formatASTForErrorMessage());
}
/** Resolve nodes with duplicate aliases.
* Table expressions cannot have duplicate aliases.
@ -6757,6 +6855,15 @@ void QueryAnalyzer::resolveQuery(const QueryTreeNodePtr & query_node, Identifier
validateAggregates(query_node, { .group_by_use_nulls = scope.group_by_use_nulls });
for (const auto & column : projection_columns)
{
if (isNotCreatable(column.type))
throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT,
"Invalid projection column with type {}. In scope {}",
column.type->getName(),
scope.scope_node->formatASTForErrorMessage());
}
/** WITH section can be safely removed, because WITH section only can provide aliases to query expressions
* and CTE for other sections to use.
*

64
src/Analyzer/QueryTreeBuilder.cpp
View File

@ -355,21 +355,67 @@ QueryTreeNodePtr QueryTreeBuilder::buildSelectExpression(const ASTPtr & select_q
if (select_limit_by)
current_query_tree->getLimitByNode() = buildExpressionList(select_limit_by, current_context);
/// Combine limit expression with limit setting
/// Combine limit expression with limit and offset settings into final limit expression
/// The sequence of application is the following - offset expression, limit expression, offset setting, limit setting.
/// Since offset setting is applied after limit expression, but we want to transfer settings into expression
/// we must decrease limit expression by offset setting and then add offset setting to offset expression.
/// select_limit - limit expression
/// limit - limit setting
/// offset - offset setting
///
/// if select_limit
/// -- if offset >= select_limit (expr 0)
/// then (0) (0 rows)
/// -- else if limit > 0 (expr 1)
/// then min(select_limit - offset, limit) (expr 2)
/// -- else
/// then (select_limit - offset) (expr 3)
/// else if limit > 0
/// then limit
///
/// offset = offset + of_expr
auto select_limit = select_query_typed.limitLength();
if (select_limit && limit)
if (select_limit)
{
auto function_node = std::make_shared<FunctionNode>("least");
function_node->getArguments().getNodes().push_back(buildExpression(select_limit, current_context));
function_node->getArguments().getNodes().push_back(std::make_shared<ConstantNode>(limit));
/// Shortcut
if (offset == 0 && limit == 0)
{
current_query_tree->getLimit() = buildExpression(select_limit, current_context);
}
else
{
/// expr 3
auto expr_3 = std::make_shared<FunctionNode>("minus");
expr_3->getArguments().getNodes().push_back(buildExpression(select_limit, current_context));
expr_3->getArguments().getNodes().push_back(std::make_shared<ConstantNode>(offset));
/// expr 2
auto expr_2 = std::make_shared<FunctionNode>("least");
expr_2->getArguments().getNodes().push_back(expr_3->clone());
expr_2->getArguments().getNodes().push_back(std::make_shared<ConstantNode>(limit));
/// expr 0
auto expr_0 = std::make_shared<FunctionNode>("greaterOrEquals");
expr_0->getArguments().getNodes().push_back(std::make_shared<ConstantNode>(offset));
expr_0->getArguments().getNodes().push_back(buildExpression(select_limit, current_context));
/// expr 1
auto expr_1 = std::make_shared<ConstantNode>(limit > 0);
auto function_node = std::make_shared<FunctionNode>("multiIf");
function_node->getArguments().getNodes().push_back(expr_0);
function_node->getArguments().getNodes().push_back(std::make_shared<ConstantNode>(0));
function_node->getArguments().getNodes().push_back(expr_1);
function_node->getArguments().getNodes().push_back(expr_2);
function_node->getArguments().getNodes().push_back(expr_3);
current_query_tree->getLimit() = std::move(function_node);
}
else if (limit)
}
else if (limit > 0)
current_query_tree->getLimit() = std::make_shared<ConstantNode>(limit);
else if (select_limit)
current_query_tree->getLimit() = buildExpression(select_limit, current_context);
/// Combine offset expression with offset setting
/// Combine offset expression with offset setting into final offset expression
auto select_offset = select_query_typed.limitOffset();
if (select_offset && offset)
{

11
src/Analyzer/TableNode.cpp
View File

@ -60,13 +60,18 @@ bool TableNode::isEqualImpl(const IQueryTreeNode & rhs) const
}
void TableNode::updateTreeHashImpl(HashState & state) const
{
if (!temporary_table_name.empty())
{
state.update(temporary_table_name.size());
state.update(temporary_table_name);
}
else
{
auto full_name = storage_id.getFullNameNotQuoted();
state.update(full_name.size());
state.update(full_name);
state.update(temporary_table_name.size());
state.update(temporary_table_name);
}
if (table_expression_modifiers)
table_expression_modifiers->updateTreeHash(state);

92
src/Analyzer/Utils.cpp
View File

@ -8,6 +8,7 @@
#include <DataTypes/DataTypeString.h>
#include <DataTypes/DataTypeTuple.h>
#include <DataTypes/DataTypeArray.h>
#include <DataTypes/DataTypeLowCardinality.h>
#include <Functions/FunctionHelpers.h>
#include <Functions/FunctionFactory.h>
@ -32,6 +33,7 @@ namespace DB
namespace ErrorCodes
{
extern const int LOGICAL_ERROR;
extern const int BAD_ARGUMENTS;
}
bool isNodePartOfTree(const IQueryTreeNode * node, const IQueryTreeNode * root)
@ -79,6 +81,75 @@ bool isNameOfInFunction(const std::string & function_name)
return is_special_function_in;
}
bool isNameOfLocalInFunction(const std::string & function_name)
{
bool is_special_function_in = function_name == "in" ||
function_name == "notIn" ||
function_name == "nullIn" ||
function_name == "notNullIn" ||
function_name == "inIgnoreSet" ||
function_name == "notInIgnoreSet" ||
function_name == "nullInIgnoreSet" ||
function_name == "notNullInIgnoreSet";
return is_special_function_in;
}
bool isNameOfGlobalInFunction(const std::string & function_name)
{
bool is_special_function_in = function_name == "globalIn" ||
function_name == "globalNotIn" ||
function_name == "globalNullIn" ||
function_name == "globalNotNullIn" ||
function_name == "globalInIgnoreSet" ||
function_name == "globalNotInIgnoreSet" ||
function_name == "globalNullInIgnoreSet" ||
function_name == "globalNotNullInIgnoreSet";
return is_special_function_in;
}
std::string getGlobalInFunctionNameForLocalInFunctionName(const std::string & function_name)
{
if (function_name == "in")
return "globalIn";
else if (function_name == "notIn")
return "globalNotIn";
else if (function_name == "nullIn")
return "globalNullIn";
else if (function_name == "notNullIn")
return "globalNotNullIn";
else if (function_name == "inIgnoreSet")
return "globalInIgnoreSet";
else if (function_name == "notInIgnoreSet")
return "globalNotInIgnoreSet";
else if (function_name == "nullInIgnoreSet")
return "globalNullInIgnoreSet";
else if (function_name == "notNullInIgnoreSet")
return "globalNotNullInIgnoreSet";
throw Exception(ErrorCodes::BAD_ARGUMENTS, "Invalid local IN function name {}", function_name);
}
void makeUniqueColumnNamesInBlock(Block & block)
{
NameSet block_column_names;
size_t unique_column_name_counter = 1;
for (auto & column_with_type : block)
{
if (!block_column_names.contains(column_with_type.name))
{
block_column_names.insert(column_with_type.name);
continue;
}
column_with_type.name += '_';
column_with_type.name += std::to_string(unique_column_name_counter);
++unique_column_name_counter;
}
}
QueryTreeNodePtr buildCastFunction(const QueryTreeNodePtr & expression,
const DataTypePtr & type,
const ContextPtr & context,
@ -102,6 +173,27 @@ QueryTreeNodePtr buildCastFunction(const QueryTreeNodePtr & expression,
return cast_function_node;
}
std::optional<bool> tryExtractConstantFromConditionNode(const QueryTreeNodePtr & condition_node)
{
const auto * constant_node = condition_node->as<ConstantNode>();
if (!constant_node)
return {};
const auto & value = constant_node->getValue();
auto constant_type = constant_node->getResultType();
constant_type = removeNullable(removeLowCardinality(constant_type));
auto which_constant_type = WhichDataType(constant_type);
if (!which_constant_type.isUInt8() && !which_constant_type.isNothing())
return {};
if (value.isNull())
return false;
UInt8 predicate_value = value.safeGet<UInt8>();
return predicate_value > 0;
}
static ASTPtr convertIntoTableExpressionAST(const QueryTreeNodePtr & table_expression_node)
{
ASTPtr table_expression_node_ast;

15
src/Analyzer/Utils.h
View File

@ -13,6 +13,18 @@ bool isNodePartOfTree(const IQueryTreeNode * node, const IQueryTreeNode * root);
/// Returns true if function name is name of IN function or its variations, false otherwise
bool isNameOfInFunction(const std::string & function_name);
/// Returns true if function name is name of local IN function or its variations, false otherwise
bool isNameOfLocalInFunction(const std::string & function_name);
/// Returns true if function name is name of global IN function or its variations, false otherwise
bool isNameOfGlobalInFunction(const std::string & function_name);
/// Returns global IN function name for local IN function name
std::string getGlobalInFunctionNameForLocalInFunctionName(const std::string & function_name);
/// Add unique suffix to names of duplicate columns in block
void makeUniqueColumnNamesInBlock(Block & block);
/** Build cast function that cast expression into type.
* If resolve = true, then result cast function is resolved during build, otherwise
* result cast function is not resolved during build.
@ -22,6 +34,9 @@ QueryTreeNodePtr buildCastFunction(const QueryTreeNodePtr & expression,
const ContextPtr & context,
bool resolve = true);
/// Try extract boolean constant from condition node
std::optional<bool> tryExtractConstantFromConditionNode(const QueryTreeNodePtr & condition_node);
/** Add table expression in tables in select query children.
* If table expression node is not of identifier node, table node, query node, table function node, join node or array join node type throws logical error exception.
*/

49
src/Analyzer/ValidationUtils.cpp
View File

@ -17,6 +17,7 @@ namespace ErrorCodes
extern const int BAD_ARGUMENTS;
extern const int NOT_AN_AGGREGATE;
extern const int NOT_IMPLEMENTED;
extern const int BAD_ARGUMENTS;
}
class ValidateGroupByColumnsVisitor : public ConstInDepthQueryTreeVisitor<ValidateGroupByColumnsVisitor>
@ -284,4 +285,52 @@ void assertNoFunctionNodes(const QueryTreeNodePtr & node,
visitor.visit(node);
}
void validateTreeSize(const QueryTreeNodePtr & node,
size_t max_size,
std::unordered_map<QueryTreeNodePtr, size_t> & node_to_tree_size)
{
size_t tree_size = 0;
std::vector<std::pair<QueryTreeNodePtr, bool>> nodes_to_process;
nodes_to_process.emplace_back(node, false);
while (!nodes_to_process.empty())
{
const auto [node_to_process, processed_children] = nodes_to_process.back();
nodes_to_process.pop_back();
if (processed_children)
{
++tree_size;
node_to_tree_size.emplace(node_to_process, tree_size);
continue;
}
auto node_to_size_it = node_to_tree_size.find(node_to_process);
if (node_to_size_it != node_to_tree_size.end())
{
tree_size += node_to_size_it->second;
continue;
}
nodes_to_process.emplace_back(node_to_process, true);
for (const auto & node_to_process_child : node_to_process->getChildren())
{
if (!node_to_process_child)
continue;
nodes_to_process.emplace_back(node_to_process_child, false);
}
auto * constant_node = node_to_process->as<ConstantNode>();
if (constant_node && constant_node->hasSourceExpression())
nodes_to_process.emplace_back(constant_node->getSourceExpression(), false);
}
if (tree_size > max_size)
throw Exception(ErrorCodes::BAD_ARGUMENTS,
"Query tree is too big. Maximum: {}",
max_size);
}
}

9
src/Analyzer/ValidationUtils.h
View File

@ -7,7 +7,7 @@ namespace DB
struct ValidationParams
{
bool group_by_use_nulls;
bool group_by_use_nulls = false;
};
/** Validate aggregates in query node.
@ -31,4 +31,11 @@ void assertNoFunctionNodes(const QueryTreeNodePtr & node,
std::string_view exception_function_name,
std::string_view exception_place_message);
/** Validate tree size. If size of tree is greater than max size throws exception.
* Additionally for each node in tree, update node to tree size map.
*/
void validateTreeSize(const QueryTreeNodePtr & node,
size_t max_size,
std::unordered_map<QueryTreeNodePtr, size_t> & node_to_tree_size);
}

6
src/Analyzer/WindowNode.cpp
View File

@ -113,11 +113,17 @@ ASTPtr WindowNode::toASTImpl() const
window_definition->parent_window_name = parent_window_name;
if (hasPartitionBy())
{
window_definition->children.push_back(getPartitionByNode()->toAST());
window_definition->partition_by = window_definition->children.back();
}
if (hasOrderBy())
{
window_definition->children.push_back(getOrderByNode()->toAST());
window_definition->order_by = window_definition->children.back();
}
window_definition->frame_is_default = window_frame.is_default;
window_definition->frame_type = window_frame.type;

2
src/Backups/BackupCoordinationRemote.cpp
View File

@ -6,7 +6,7 @@
#include <IO/WriteHelpers.h>
#include <Common/ZooKeeper/KeeperException.h>
#include <Common/escapeForFileName.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Backups/BackupCoordinationStage.h>

2
src/Backups/BackupImpl.cpp
View File

@ -6,7 +6,7 @@
#include <Backups/BackupCoordinationLocal.h>
#include <Backups/BackupCoordinationRemote.h>
#include <Common/StringUtils/StringUtils.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Common/quoteString.h>
#include <Common/XMLUtils.h>
#include <Interpreters/Context.h>

2
src/Backups/IBackupCoordination.h
View File

@ -2,7 +2,7 @@
#include <optional>
#include <fmt/format.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Core/Types.h>

38
src/Client/ClientBase.cpp
View File

@ -1834,7 +1834,7 @@ bool ClientBase::executeMultiQuery(const String & all_queries_text)
{
/// disable logs if expects errors
TestHint test_hint(all_queries_text);
if (test_hint.clientError() || test_hint.serverError())
if (test_hint.hasClientErrors() || test_hint.hasServerErrors())
processTextAsSingleQuery("SET send_logs_level = 'fatal'");
}
@ -1876,17 +1876,17 @@ bool ClientBase::executeMultiQuery(const String & all_queries_text)
// the query ends because we failed to parse it, so we consume
// the entire line.
TestHint hint(String(this_query_begin, this_query_end - this_query_begin));
if (hint.serverError())
if (hint.hasServerErrors())
{
// Syntax errors are considered as client errors
current_exception->addMessage("\nExpected server error '{}'.", hint.serverError());
current_exception->addMessage("\nExpected server error: {}.", hint.serverErrors());
current_exception->rethrow();
}
if (hint.clientError() != current_exception->code())
if (!hint.hasExpectedClientError(current_exception->code()))
{
if (hint.clientError())
current_exception->addMessage("\nExpected client error: " + std::to_string(hint.clientError()));
if (hint.hasClientErrors())
current_exception->addMessage("\nExpected client error: {}.", hint.clientErrors());
current_exception->rethrow();
}
@ -1935,37 +1935,37 @@ bool ClientBase::executeMultiQuery(const String & all_queries_text)
bool error_matches_hint = true;
if (have_error)
{
if (test_hint.serverError())
if (test_hint.hasServerErrors())
{
if (!server_exception)
{
error_matches_hint = false;
fmt::print(stderr, "Expected server error code '{}' but got no server error (query: {}).\n",
test_hint.serverError(), full_query);
test_hint.serverErrors(), full_query);
}
else if (server_exception->code() != test_hint.serverError())
else if (!test_hint.hasExpectedServerError(server_exception->code()))
{
error_matches_hint = false;
fmt::print(stderr, "Expected server error code: {} but got: {} (query: {}).\n",
test_hint.serverError(), server_exception->code(), full_query);
test_hint.serverErrors(), server_exception->code(), full_query);
}
}
if (test_hint.clientError())
if (test_hint.hasClientErrors())
{
if (!client_exception)
{
error_matches_hint = false;
fmt::print(stderr, "Expected client error code '{}' but got no client error (query: {}).\n",
test_hint.clientError(), full_query);
test_hint.clientErrors(), full_query);
}
else if (client_exception->code() != test_hint.clientError())
else if (!test_hint.hasExpectedClientError(client_exception->code()))
{
error_matches_hint = false;
fmt::print(stderr, "Expected client error code '{}' but got '{}' (query: {}).\n",
test_hint.clientError(), client_exception->code(), full_query);
test_hint.clientErrors(), client_exception->code(), full_query);
}
}
if (!test_hint.clientError() && !test_hint.serverError())
if (!test_hint.hasClientErrors() && !test_hint.hasServerErrors())
{
// No error was expected but it still occurred. This is the
// default case without test hint, doesn't need additional
@ -1975,19 +1975,19 @@ bool ClientBase::executeMultiQuery(const String & all_queries_text)
}
else
{
if (test_hint.clientError())
if (test_hint.hasClientErrors())
{
error_matches_hint = false;
fmt::print(stderr,
"The query succeeded but the client error '{}' was expected (query: {}).\n",
test_hint.clientError(), full_query);
test_hint.clientErrors(), full_query);
}
if (test_hint.serverError())
if (test_hint.hasServerErrors())
{
error_matches_hint = false;
fmt::print(stderr,
"The query succeeded but the server error '{}' was expected (query: {}).\n",
test_hint.serverError(), full_query);
test_hint.serverErrors(), full_query);
}
}

2
src/Client/Connection.cpp
View File

@ -506,7 +506,7 @@ void Connection::sendQuery(
bool with_pending_data,
std::function<void(const Progress &)>)
{
OpenTelemetry::SpanHolder span("Connection::sendQuery()");
OpenTelemetry::SpanHolder span("Connection::sendQuery()", OpenTelemetry::CLIENT);
span.addAttribute("clickhouse.query_id", query_id_);
span.addAttribute("clickhouse.query", query);
span.addAttribute("target", [this] () { return this->getHost() + ":" + std::to_string(this->getPort()); });

10
src/Client/HedgedConnections.cpp
View File

@ -31,8 +31,6 @@ HedgedConnections::HedgedConnections(
: hedged_connections_factory(pool_, &context_->getSettingsRef(), timeouts_, table_to_check_)
, context(std::move(context_))
, settings(context->getSettingsRef())
, drain_timeout(settings.drain_timeout)
, allow_changing_replica_until_first_data_packet(settings.allow_changing_replica_until_first_data_packet)
, throttler(throttler_)
{
std::vector<Connection *> connections = hedged_connections_factory.getManyConnections(pool_mode);
@ -263,7 +261,7 @@ Packet HedgedConnections::drain()
while (!epoll.empty())
{
ReplicaLocation location = getReadyReplicaLocation(DrainCallback{drain_timeout});
ReplicaLocation location = getReadyReplicaLocation();
Packet packet = receivePacketFromReplica(location);
switch (packet.type)
{
@ -290,10 +288,10 @@ Packet HedgedConnections::drain()
Packet HedgedConnections::receivePacket()
{
std::lock_guard lock(cancel_mutex);
return receivePacketUnlocked({}, false /* is_draining */);
return receivePacketUnlocked({});
}
Packet HedgedConnections::receivePacketUnlocked(AsyncCallback async_callback, bool /* is_draining */)
Packet HedgedConnections::receivePacketUnlocked(AsyncCallback async_callback)
{
if (!sent_query)
throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot receive packets: no query sent.");
@ -413,7 +411,7 @@ Packet HedgedConnections::receivePacketFromReplica(const ReplicaLocation & repli
{
/// If we are allowed to change replica until the first data packet,
/// just restart timeout (if it hasn't expired yet). Otherwise disable changing replica with this offset.
if (allow_changing_replica_until_first_data_packet && !replica.is_change_replica_timeout_expired)
if (settings.allow_changing_replica_until_first_data_packet && !replica.is_change_replica_timeout_expired)
replica.change_replica_timeout.setRelative(hedged_connections_factory.getConnectionTimeouts().receive_data_timeout);
else
disableChangingReplica(replica_location);

8
src/Client/HedgedConnections.h
View File

@ -101,7 +101,7 @@ public:
Packet receivePacket() override;
Packet receivePacketUnlocked(AsyncCallback async_callback, bool is_draining) override;
Packet receivePacketUnlocked(AsyncCallback async_callback) override;
void disconnect() override;
@ -196,12 +196,6 @@ private:
Epoll epoll;
ContextPtr context;
const Settings & settings;
/// The following two fields are from settings but can be referenced outside the lifetime of
/// settings when connection is drained asynchronously.
Poco::Timespan drain_timeout;
bool allow_changing_replica_until_first_data_packet;
ThrottlerPtr throttler;
bool sent_query = false;
bool cancelled = false;

36
src/Client/IConnections.cpp
View File

@ -1,36 +0,0 @@
#include <Client/IConnections.h>
#include <Poco/Net/SocketImpl.h>
namespace DB
{
namespace ErrorCodes
{
extern const int SOCKET_TIMEOUT;
}
/// This wrapper struct allows us to use Poco's socket polling code with a raw fd.
/// The only difference from Poco::Net::SocketImpl is that we don't close the fd in the destructor.
struct PocoSocketWrapper : public Poco::Net::SocketImpl
{
explicit PocoSocketWrapper(int fd)
{
reset(fd);
}
// Do not close fd.
~PocoSocketWrapper() override { reset(-1); }
};
void IConnections::DrainCallback::operator()(int fd, Poco::Timespan, const std::string & fd_description) const
{
if (!PocoSocketWrapper(fd).poll(drain_timeout, Poco::Net::Socket::SELECT_READ))
{
throw Exception(ErrorCodes::SOCKET_TIMEOUT,
"Read timeout ({} ms) while draining from {}",
drain_timeout.totalMilliseconds(),
fd_description);
}
}
}

8
src/Client/IConnections.h
View File

@ -13,12 +13,6 @@ namespace DB
class IConnections : boost::noncopyable
{
public:
struct DrainCallback
{
Poco::Timespan drain_timeout;
void operator()(int fd, Poco::Timespan, const std::string & fd_description = "") const;
};
/// Send all scalars to replicas.
virtual void sendScalarsData(Scalars & data) = 0;
/// Send all content of external tables to replicas.
@ -40,7 +34,7 @@ public:
virtual Packet receivePacket() = 0;
/// Version of `receivePacket` function without locking.
virtual Packet receivePacketUnlocked(AsyncCallback async_callback, bool is_draining) = 0;
virtual Packet receivePacketUnlocked(AsyncCallback async_callback) = 0;
/// Break all active connections.
virtual void disconnect() = 0;

32
src/Client/MultiplexedConnections.cpp
View File

@ -20,7 +20,7 @@ namespace ErrorCodes
MultiplexedConnections::MultiplexedConnections(Connection & connection, const Settings & settings_, const ThrottlerPtr & throttler)
: settings(settings_), drain_timeout(settings.drain_timeout), receive_timeout(settings.receive_timeout)
: settings(settings_)
{
connection.setThrottler(throttler);
@ -33,7 +33,7 @@ MultiplexedConnections::MultiplexedConnections(Connection & connection, const Se
MultiplexedConnections::MultiplexedConnections(std::shared_ptr<Connection> connection_ptr_, const Settings & settings_, const ThrottlerPtr & throttler)
: settings(settings_), drain_timeout(settings.drain_timeout), receive_timeout(settings.receive_timeout)
: settings(settings_)
, connection_ptr(connection_ptr_)
{
connection_ptr->setThrottler(throttler);
@ -46,8 +46,9 @@ MultiplexedConnections::MultiplexedConnections(std::shared_ptr<Connection> conne
}
MultiplexedConnections::MultiplexedConnections(
std::vector<IConnectionPool::Entry> && connections, const Settings & settings_, const ThrottlerPtr & throttler)
: settings(settings_), drain_timeout(settings.drain_timeout), receive_timeout(settings.receive_timeout)
std::vector<IConnectionPool::Entry> && connections,
const Settings & settings_, const ThrottlerPtr & throttler)
: settings(settings_)
{
/// If we didn't get any connections from pool and getMany() did not throw exceptions, this means that
/// `skip_unavailable_shards` was set. Then just return.
@ -206,7 +207,7 @@ void MultiplexedConnections::sendMergeTreeReadTaskResponse(const ParallelReadRes
Packet MultiplexedConnections::receivePacket()
{
std::lock_guard lock(cancel_mutex);
Packet packet = receivePacketUnlocked({}, false /* is_draining */);
Packet packet = receivePacketUnlocked({});
return packet;
}
@ -254,7 +255,7 @@ Packet MultiplexedConnections::drain()
while (hasActiveConnections())
{
Packet packet = receivePacketUnlocked(DrainCallback{drain_timeout}, true /* is_draining */);
Packet packet = receivePacketUnlocked({});
switch (packet.type)
{
@ -304,14 +305,14 @@ std::string MultiplexedConnections::dumpAddressesUnlocked() const
return buf.str();
}
Packet MultiplexedConnections::receivePacketUnlocked(AsyncCallback async_callback, bool is_draining)
Packet MultiplexedConnections::receivePacketUnlocked(AsyncCallback async_callback)
{
if (!sent_query)
throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot receive packets: no query sent.");
if (!hasActiveConnections())
throw Exception(ErrorCodes::LOGICAL_ERROR, "No more packets are available.");
ReplicaState & state = getReplicaForReading(is_draining);
ReplicaState & state = getReplicaForReading();
current_connection = state.connection;
if (current_connection == nullptr)
throw Exception(ErrorCodes::NO_AVAILABLE_REPLICA, "Logical error: no available replica");
@ -366,10 +367,9 @@ Packet MultiplexedConnections::receivePacketUnlocked(AsyncCallback async_callbac
return packet;
}
MultiplexedConnections::ReplicaState & MultiplexedConnections::getReplicaForReading(bool is_draining)
MultiplexedConnections::ReplicaState & MultiplexedConnections::getReplicaForReading()
{
/// Fast path when we only focus on one replica and are not draining the connection.
if (replica_states.size() == 1 && !is_draining)
if (replica_states.size() == 1)
return replica_states[0];
Poco::Net::Socket::SocketList read_list;
@ -390,7 +390,7 @@ MultiplexedConnections::ReplicaState & MultiplexedConnections::getReplicaForRead
Poco::Net::Socket::SocketList write_list;
Poco::Net::Socket::SocketList except_list;
auto timeout = is_draining ? drain_timeout : receive_timeout;
auto timeout = settings.receive_timeout;
int n = 0;
/// EINTR loop
@ -417,9 +417,7 @@ MultiplexedConnections::ReplicaState & MultiplexedConnections::getReplicaForRead
break;
}
/// We treat any error as timeout for simplicity.
/// And we also check if read_list is still empty just in case.
if (n <= 0 || read_list.empty())
if (n == 0)
{
const auto & addresses = dumpAddressesUnlocked();
for (ReplicaState & state : replica_states)
@ -438,7 +436,9 @@ MultiplexedConnections::ReplicaState & MultiplexedConnections::getReplicaForRead
}
}
/// TODO Motivation of rand is unclear.
/// TODO Absolutely wrong code: read_list could be empty; motivation of rand is unclear.
/// This code path is disabled by default.
auto & socket = read_list[thread_local_rng() % read_list.size()];
if (fd_to_replica_state_idx.empty())
{

9
src/Client/MultiplexedConnections.h
View File

@ -65,7 +65,7 @@ public:
void setReplicaInfo(ReplicaInfo value) override { replica_info = value; }
private:
Packet receivePacketUnlocked(AsyncCallback async_callback, bool is_draining) override;
Packet receivePacketUnlocked(AsyncCallback async_callback) override;
/// Internal version of `dumpAddresses` function without locking.
std::string dumpAddressesUnlocked() const;
@ -78,18 +78,13 @@ private:
};
/// Get a replica where you can read the data.
ReplicaState & getReplicaForReading(bool is_draining);
ReplicaState & getReplicaForReading();
/// Mark the replica as invalid.
void invalidateReplica(ReplicaState & replica_state);
const Settings & settings;
/// The following two fields are from settings but can be referenced outside the lifetime of
/// settings when connection is drained asynchronously.
Poco::Timespan drain_timeout;
Poco::Timespan receive_timeout;
/// The current number of valid connections to the replicas of this shard.
size_t active_connection_count = 0;

126
src/Client/TestHint.cpp
View File

@ -1,32 +1,15 @@
#include "TestHint.h"
#include <charconv>
#include <string_view>
#include <Client/TestHint.h>
#include <Common/Exception.h>
#include <Common/ErrorCodes.h>
#include <IO/ReadBufferFromString.h>
#include <IO/ReadHelpers.h>
#include <Parsers/Lexer.h>
#include <Common/ErrorCodes.h>
#include <Common/Exception.h>
namespace
namespace DB::ErrorCodes
{
/// Parse error as number or as a string (name of the error code const)
int parseErrorCode(DB::ReadBufferFromString & in)
{
int code = -1;
String code_name;
auto * pos = in.position();
tryReadText(code, in);
if (pos != in.position())
{
return code;
}
/// Try parse as string
readStringUntilWhitespace(code_name, in);
return DB::ErrorCodes::getErrorCodeByName(code_name);
}
extern const int CANNOT_PARSE_TEXT;
}
namespace DB
@ -60,8 +43,8 @@ TestHint::TestHint(const String & query_)
size_t pos_end = comment.find('}', pos_start);
if (pos_end != String::npos)
{
String hint(comment.begin() + pos_start + 1, comment.begin() + pos_end);
parse(hint, is_leading_hint);
Lexer comment_lexer(comment.c_str() + pos_start + 1, comment.c_str() + pos_end, 0);
parse(comment_lexer, is_leading_hint);
}
}
}
@ -69,27 +52,30 @@ TestHint::TestHint(const String & query_)
}
}
void TestHint::parse(const String & hint, bool is_leading_hint)
bool TestHint::hasExpectedClientError(int error)
{
ReadBufferFromString in(hint);
String item;
while (!in.eof())
{
readStringUntilWhitespace(item, in);
if (in.eof())
break;
skipWhitespaceIfAny(in);
if (!is_leading_hint)
{
if (item == "serverError")
server_error = parseErrorCode(in);
else if (item == "clientError")
client_error = parseErrorCode(in);
return std::find(client_errors.begin(), client_errors.end(), error) != client_errors.end();
}
bool TestHint::hasExpectedServerError(int error)
{
return std::find(server_errors.begin(), server_errors.end(), error) != server_errors.end();
}
void TestHint::parse(Lexer & comment_lexer, bool is_leading_hint)
{
std::unordered_set<std::string_view> commands{"echo", "echoOn", "echoOff"};
std::unordered_set<std::string_view> command_errors{
"serverError",
"clientError",
};
for (Token token = comment_lexer.nextToken(); !token.isEnd(); token = comment_lexer.nextToken())
{
String item = String(token.begin, token.end);
if (token.type == TokenType::BareWord && commands.contains(item))
{
if (item == "echo")
echo.emplace(true);
if (item == "echoOn")
@ -97,6 +83,56 @@ void TestHint::parse(const String & hint, bool is_leading_hint)
if (item == "echoOff")
echo.emplace(false);
}
else if (!is_leading_hint && token.type == TokenType::BareWord && command_errors.contains(item))
{
/// Everything after this must be a list of errors separated by comma
ErrorVector error_codes;
while (!token.isEnd())
{
token = comment_lexer.nextToken();
if (token.type == TokenType::Whitespace)
continue;
if (token.type == TokenType::Number)
{
int code;
auto [p, ec] = std::from_chars(token.begin, token.end, code);
if (p == token.begin)
throw DB::Exception(
DB::ErrorCodes::CANNOT_PARSE_TEXT,
"Could not parse integer number for errorcode: {}",
std::string_view(token.begin, token.end));
error_codes.push_back(code);
}
else if (token.type == TokenType::BareWord)
{
int code = DB::ErrorCodes::getErrorCodeByName(std::string_view(token.begin, token.end));
error_codes.push_back(code);
}
else
throw DB::Exception(
DB::ErrorCodes::CANNOT_PARSE_TEXT,
"Could not parse error code in {}: {}",
getTokenName(token.type),
std::string_view(token.begin, token.end));
do
{
token = comment_lexer.nextToken();
} while (!token.isEnd() && token.type == TokenType::Whitespace);
if (!token.isEnd() && token.type != TokenType::Comma)
throw DB::Exception(
DB::ErrorCodes::CANNOT_PARSE_TEXT,
"Could not parse error code. Expected ','. Got '{}'",
std::string_view(token.begin, token.end));
}
if (item == "serverError")
server_errors = error_codes;
else
client_errors = error_codes;
break;
}
}
}
}

67
src/Client/TestHint.h
View File

@ -1,21 +1,30 @@
#pragma once
#include <optional>
#include <vector>
#include <fmt/format.h>
#include <Core/Types.h>
namespace DB
{
class Lexer;
/// Checks expected server and client error codes.
///
/// The following comment hints are supported:
///
/// - "-- { serverError 60 }" -- in case of you are expecting server error.
/// - "-- { serverError 16, 36 }" -- in case of you are expecting one of the 2 errors.
///
/// - "-- { clientError 20 }" -- in case of you are expecting client error.
/// - "-- { clientError 20, 60, 92 }" -- It's expected that the client will return one of the 3 errors.
///
/// - "-- { serverError FUNCTION_THROW_IF_VALUE_IS_NON_ZERO }" -- by error name.
/// - "-- { serverError NO_SUCH_COLUMN_IN_TABLE, BAD_ARGUMENTS }" -- by error name.
///
/// - "-- { clientError FUNCTION_THROW_IF_VALUE_IS_NON_ZERO }" -- by error name.
///
@ -43,29 +52,73 @@ namespace DB
class TestHint
{
public:
using ErrorVector = std::vector<int>;
TestHint(const String & query_);
int serverError() const { return server_error; }
int clientError() const { return client_error; }
const auto & serverErrors() const { return server_errors; }
const auto & clientErrors() const { return client_errors; }
std::optional<bool> echoQueries() const { return echo; }
bool hasClientErrors() { return !client_errors.empty(); }
bool hasServerErrors() { return !server_errors.empty(); }
bool hasExpectedClientError(int error);
bool hasExpectedServerError(int error);
private:
const String & query;
int server_error = 0;
int client_error = 0;
ErrorVector server_errors{};
ErrorVector client_errors{};
std::optional<bool> echo;
void parse(const String & hint, bool is_leading_hint);
void parse(Lexer & comment_lexer, bool is_leading_hint);
bool allErrorsExpected(int actual_server_error, int actual_client_error) const
{
return (server_error || client_error) && (server_error == actual_server_error) && (client_error == actual_client_error);
if (actual_server_error && std::find(server_errors.begin(), server_errors.end(), actual_server_error) == server_errors.end())
return false;
if (!actual_server_error && server_errors.size())
return false;
if (actual_client_error && std::find(client_errors.begin(), client_errors.end(), actual_client_error) == client_errors.end())
return false;
if (!actual_client_error && client_errors.size())
return false;
return true;
}
bool lostExpectedError(int actual_server_error, int actual_client_error) const
{
return (server_error && !actual_server_error) || (client_error && !actual_client_error);
return (server_errors.size() && !actual_server_error) || (client_errors.size() && !actual_client_error);
}
};
}
template <>
struct fmt::formatter<DB::TestHint::ErrorVector>
{
static constexpr auto parse(format_parse_context & ctx)
{
const auto * it = ctx.begin();
const auto * end = ctx.end();
/// Only support {}.
if (it != end && *it != '}')
throw format_error("Invalid format");
return it;
}
template <typename FormatContext>
auto format(const DB::TestHint::ErrorVector & ErrorVector, FormatContext & ctx)
{
if (ErrorVector.empty())
return format_to(ctx.out(), "{}", 0);
else if (ErrorVector.size() == 1)
return format_to(ctx.out(), "{}", ErrorVector[0]);
else
return format_to(ctx.out(), "[{}]", fmt::join(ErrorVector, ", "));
}
};

2
src/Columns/tests/gtest_weak_hash_32.cpp
View File

@ -14,7 +14,7 @@
#include <DataTypes/DataTypesNumber.h>
#include <Common/WeakHash.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <unordered_map>
#include <iostream>

2
src/Common/BinStringDecodeHelper.h
View File

@ -1,6 +1,6 @@
#pragma once
#include <Common/hex.h>
#include <base/hex.h>
namespace DB
{

4
src/Common/CurrentMetrics.cpp
View File

@ -84,10 +84,6 @@
M(MMappedFileBytes, "Sum size of mmapped file regions.") \
M(MMappedAllocs, "Total number of mmapped allocations") \
M(MMappedAllocBytes, "Sum bytes of mmapped allocations") \
M(AsyncDrainedConnections, "Number of connections drained asynchronously.") \
M(ActiveAsyncDrainedConnections, "Number of active connections drained asynchronously.") \
M(SyncDrainedConnections, "Number of connections drained synchronously.") \
M(ActiveSyncDrainedConnections, "Number of active connections drained synchronously.") \
M(AsynchronousReadWait, "Number of threads waiting for asynchronous read.") \
M(PendingAsyncInsert, "Number of asynchronous inserts that are waiting for flush.") \
M(KafkaConsumers, "Number of active Kafka consumers") \

19
src/Common/CurrentThread.cpp
View File

@ -110,23 +110,4 @@ ThreadGroupStatusPtr CurrentThread::getGroup()
return current_thread->getThreadGroup();
}
MemoryTracker * CurrentThread::getUserMemoryTracker()
{
if (unlikely(!current_thread))
return nullptr;
auto * tracker = current_thread->memory_tracker.getParent();
while (tracker && tracker->level != VariableContext::User)
tracker = tracker->getParent();
return tracker;
}
void CurrentThread::flushUntrackedMemory()
{
if (unlikely(!current_thread))
return;
current_thread->flushUntrackedMemory();
}
}

6
src/Common/CurrentThread.h
View File

@ -40,12 +40,6 @@ public:
/// Group to which belongs current thread
static ThreadGroupStatusPtr getGroup();
/// MemoryTracker for user that owns current thread if any
static MemoryTracker * getUserMemoryTracker();
/// Adjust counters in MemoryTracker hierarchy if untracked_memory is not 0.
static void flushUntrackedMemory();
/// A logs queue used by TCPHandler to pass logs to a client
static void attachInternalTextLogsQueue(const std::shared_ptr<InternalTextLogsQueue> & logs_queue,
LogsLevel client_logs_level);

69
src/Common/LoggingFormatStringHelpers.cpp
View File

@ -1,7 +1,76 @@
#include <Common/LoggingFormatStringHelpers.h>
#include <Common/SipHash.h>
#include <Common/thread_local_rng.h>
[[noreturn]] void functionThatFailsCompilationOfConstevalFunctions(const char * error)
{
throw std::runtime_error(error);
}
std::unordered_map<UInt64, std::pair<time_t, size_t>> LogFrequencyLimiterIml::logged_messages;
time_t LogFrequencyLimiterIml::last_cleanup = 0;
std::mutex LogFrequencyLimiterIml::mutex;
void LogFrequencyLimiterIml::log(Poco::Message & message)
{
std::string_view pattern = message.getFormatString();
if (pattern.empty())
{
/// Do not filter messages without a format string
if (auto * channel = logger->getChannel())
channel->log(message);
return;
}
SipHash hash;
hash.update(logger->name());
/// Format strings are compile-time constants, so they are uniquely identified by pointer and size
hash.update(pattern.data());
hash.update(pattern.size());
time_t now = time(nullptr);
size_t skipped_similar_messages = 0;
bool need_cleanup;
bool need_log;
{
std::lock_guard lock(mutex);
need_cleanup = last_cleanup + 300 <= now;
auto & info = logged_messages[hash.get64()];
need_log = info.first + min_interval_s <= now;
if (need_log)
{
skipped_similar_messages = info.second;
info.first = now;
info.second = 0;
}
else
{
++info.second;
}
}
/// We don't need all threads to do cleanup, just randomize
if (need_cleanup && thread_local_rng() % 100 == 0)
cleanup();
/// The message it too frequent, skip it for now
/// NOTE It's not optimal because we format the message first and only then check if we need to actually write it, see LOG_IMPL macro
if (!need_log)
return;
if (skipped_similar_messages)
message.appendText(fmt::format(" (skipped {} similar messages)", skipped_similar_messages));
if (auto * channel = logger->getChannel())
channel->log(message);
}
void LogFrequencyLimiterIml::cleanup(time_t too_old_threshold_s)
{
time_t now = time(nullptr);
time_t old = now - too_old_threshold_s;
std::lock_guard lock(mutex);
std::erase_if(logged_messages, [old](const auto & elem) { return elem.second.first < old; });
last_cleanup = now;
}

61
src/Common/LoggingFormatStringHelpers.h
View File

@ -1,6 +1,11 @@
#pragma once
#include <base/defines.h>
#include <base/types.h>
#include <fmt/format.h>
#include <mutex>
#include <unordered_map>
#include <Poco/Logger.h>
#include <Poco/Message.h>
struct PreformattedMessage;
consteval void formatStringCheckArgsNumImpl(std::string_view str, size_t nargs);
@ -156,3 +161,59 @@ struct CheckArgsNumHelperImpl
template <typename... Args> using CheckArgsNumHelper = CheckArgsNumHelperImpl<std::type_identity_t<Args>...>;
template <typename... Args> void formatStringCheckArgsNum(CheckArgsNumHelper<Args...>, Args &&...) {}
/// This wrapper helps to avoid too frequent and noisy log messages.
/// For each pair (logger_name, format_string) it remembers when such a message was logged the last time.
/// The message will not be logged again if less than min_interval_s seconds passed since the previously logged message.
class LogFrequencyLimiterIml
{
/// Hash(logger_name, format_string) -> (last_logged_time_s, skipped_messages_count)
static std::unordered_map<UInt64, std::pair<time_t, size_t>> logged_messages;
static time_t last_cleanup;
static std::mutex mutex;
Poco::Logger * logger;
time_t min_interval_s;
public:
LogFrequencyLimiterIml(Poco::Logger * logger_, time_t min_interval_s_) : logger(logger_), min_interval_s(min_interval_s_) {}
LogFrequencyLimiterIml & operator -> () { return *this; }
bool is(Poco::Message::Priority priority) { return logger->is(priority); }
LogFrequencyLimiterIml * getChannel() {return this; }
const String & name() const { return logger->name(); }
void log(Poco::Message & message);
/// Clears messages that were logged last time more than too_old_threshold_s seconds ago
static void cleanup(time_t too_old_threshold_s = 600);
Poco::Logger * getLogger() { return logger; }
};
/// This wrapper is useful to save formatted message into a String before sending it to a logger
class LogToStrImpl
{
String & out_str;
Poco::Logger * logger;
std::unique_ptr<LogFrequencyLimiterIml> maybe_nested;
bool propagate_to_actual_log = true;
public:
LogToStrImpl(String & out_str_, Poco::Logger * logger_) : out_str(out_str_), logger(logger_) {}
LogToStrImpl(String & out_str_, std::unique_ptr<LogFrequencyLimiterIml> && maybe_nested_)
: out_str(out_str_), logger(maybe_nested_->getLogger()), maybe_nested(std::move(maybe_nested_)) {}
LogToStrImpl & operator -> () { return *this; }
bool is(Poco::Message::Priority priority) { propagate_to_actual_log &= logger->is(priority); return true; }
LogToStrImpl * getChannel() {return this; }
const String & name() const { return logger->name(); }
void log(Poco::Message & message)
{
out_str = message.getText();
if (!propagate_to_actual_log)
return;
if (maybe_nested)
maybe_nested->log(message);
else if (auto * channel = logger->getChannel())
channel->log(message);
}
};

5
src/Common/OpenTelemetryTraceContext.cpp
View File

@ -3,7 +3,7 @@
#include <random>
#include <base/getThreadId.h>
#include <Common/Exception.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Core/Settings.h>
#include <IO/Operators.h>
@ -92,7 +92,7 @@ bool Span::addAttributeImpl(std::string_view name, std::string_view value) noexc
return true;
}
SpanHolder::SpanHolder(std::string_view _operation_name)
SpanHolder::SpanHolder(std::string_view _operation_name, SpanKind _kind)
{
if (!current_thread_trace_context.isTraceEnabled())
{
@ -106,6 +106,7 @@ SpanHolder::SpanHolder(std::string_view _operation_name)
this->parent_span_id = current_thread_trace_context.span_id;
this->span_id = thread_local_rng(); // create a new id for this span
this->operation_name = _operation_name;
this->kind = _kind;
this->start_time_us
= std::chrono::duration_cast<std::chrono::microseconds>(std::chrono::system_clock::now().time_since_epoch()).count();

26
src/Common/OpenTelemetryTraceContext.h
View File

@ -13,6 +13,29 @@ class ReadBuffer;
namespace OpenTelemetry
{
/// See https://opentelemetry.io/docs/reference/specification/trace/api/#spankind
enum SpanKind
{
/// Default value. Indicates that the span represents an internal operation within an application,
/// as opposed to an operations with remote parents or children.
INTERNAL = 0,
/// Indicates that the span covers server-side handling of a synchronous RPC or other remote request.
/// This span is often the child of a remote CLIENT span that was expected to wait for a response.
SERVER = 1,
/// Indicates that the span describes a request to some remote service.
/// This span is usually the parent of a remote SERVER span and does not end until the response is received.
CLIENT = 2,
/// 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.
/// In messaging scenarios with batching, tracing individual messages requires a new PRODUCER span per message to be created.
PRODUCER = 3,
/// Indicates that the span describes a child of an asynchronous PRODUCER request
CONSUMER = 4
};
struct Span
{
UUID trace_id{};
@ -21,6 +44,7 @@ struct Span
String operation_name;
UInt64 start_time_us = 0;
UInt64 finish_time_us = 0;
SpanKind kind = INTERNAL;
Map attributes;
/// Following methods are declared as noexcept to make sure they're exception safe.
@ -155,7 +179,7 @@ using TracingContextHolderPtr = std::unique_ptr<TracingContextHolder>;
/// Once it's created or destructed, it automatically maitains the tracing context on the thread that it lives.
struct SpanHolder : public Span
{
SpanHolder(std::string_view);
SpanHolder(std::string_view, SpanKind _kind = INTERNAL);
~SpanHolder();
/// Finish a span explicitly if needed.

2
src/Common/ProfileEvents.cpp
View File

@ -166,6 +166,8 @@
\
M(WaitMarksLoadMicroseconds, "Time spent loading marks") \
M(BackgroundLoadingMarksTasks, "Number of background tasks for loading marks") \
M(LoadedMarksCount, "Number of marks loaded (total across columns).") \
M(LoadedMarksMemoryBytes, "Size of in-memory representations of loaded marks.") \
\
M(Merge, "Number of launched background merges.") \
M(MergedRows, "Rows read for background merges. This is the number of rows before merge.") \

2
src/Common/SymbolIndex.cpp
View File

@ -1,7 +1,7 @@
#if defined(__ELF__) && !defined(OS_FREEBSD)
#include <Common/SymbolIndex.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <algorithm>
#include <optional>

2
src/Common/escapeForFileName.cpp
View File

@ -1,4 +1,4 @@
#include <Common/hex.h>
#include <base/hex.h>
#include <Common/StringUtils/StringUtils.h>
#include <Common/escapeForFileName.h>

2
src/Common/formatIPv6.cpp
View File

@ -1,5 +1,5 @@
#include <Common/formatIPv6.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Common/StringUtils/StringUtils.h>
#include <base/range.h>

2
src/Common/formatIPv6.h
View File

@ -7,7 +7,7 @@
#include <utility>
#include <base/range.h>
#include <base/unaligned.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <Common/StringUtils/StringUtils.h>
constexpr size_t IPV4_BINARY_LENGTH = 4;

2
src/Common/getHashOfLoadedBinary.cpp
View File

@ -4,7 +4,7 @@
#include <link.h>
#include <array>
#include <Common/hex.h>
#include <base/hex.h>
static int callback(dl_phdr_info * info, size_t, void * data)

2
src/Common/getMappedArea.cpp
View File

@ -4,7 +4,7 @@
#if defined(OS_LINUX)
#include <Common/StringUtils/StringUtils.h>
#include <Common/hex.h>
#include <base/hex.h>
#include <IO/ReadBufferFromFile.h>
#include <IO/ReadHelpers.h>

92
src/Common/hex.cpp
View File

@ -1,92 +0,0 @@
#include <Common/hex.h>
const char * const hex_digit_to_char_uppercase_table = "0123456789ABCDEF";
const char * const hex_digit_to_char_lowercase_table = "0123456789abcdef";
const char * const hex_byte_to_char_uppercase_table =
"000102030405060708090A0B0C0D0E0F"
"101112131415161718191A1B1C1D1E1F"
"202122232425262728292A2B2C2D2E2F"
"303132333435363738393A3B3C3D3E3F"
"404142434445464748494A4B4C4D4E4F"
"505152535455565758595A5B5C5D5E5F"
"606162636465666768696A6B6C6D6E6F"
"707172737475767778797A7B7C7D7E7F"
"808182838485868788898A8B8C8D8E8F"
"909192939495969798999A9B9C9D9E9F"
"A0A1A2A3A4A5A6A7A8A9AAABACADAEAF"
"B0B1B2B3B4B5B6B7B8B9BABBBCBDBEBF"
"C0C1C2C3C4C5C6C7C8C9CACBCCCDCECF"
"D0D1D2D3D4D5D6D7D8D9DADBDCDDDEDF"
"E0E1E2E3E4E5E6E7E8E9EAEBECEDEEEF"
"F0F1F2F3F4F5F6F7F8F9FAFBFCFDFEFF";
const char * const hex_byte_to_char_lowercase_table =
"000102030405060708090a0b0c0d0e0f"
"101112131415161718191a1b1c1d1e1f"
"202122232425262728292a2b2c2d2e2f"
"303132333435363738393a3b3c3d3e3f"
"404142434445464748494a4b4c4d4e4f"
"505152535455565758595a5b5c5d5e5f"
"606162636465666768696a6b6c6d6e6f"
"707172737475767778797a7b7c7d7e7f"
"808182838485868788898a8b8c8d8e8f"
"909192939495969798999a9b9c9d9e9f"
"a0a1a2a3a4a5a6a7a8a9aaabacadaeaf"
"b0b1b2b3b4b5b6b7b8b9babbbcbdbebf"
"c0c1c2c3c4c5c6c7c8c9cacbcccdcecf"
"d0d1d2d3d4d5d6d7d8d9dadbdcdddedf"
"e0e1e2e3e4e5e6e7e8e9eaebecedeeef"
"f0f1f2f3f4f5f6f7f8f9fafbfcfdfeff";
const char * const hex_char_to_digit_table =
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\xff\xff\xff\xff\xff\xff" //0-9
"\xff\x0a\x0b\x0c\x0d\x0e\x0f\xff\xff\xff\xff\xff\xff\xff\xff\xff" //A-Z
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\x0a\x0b\x0c\x0d\x0e\x0f\xff\xff\xff\xff\xff\xff\xff\xff\xff" //a-z
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff\xff";
const char * const bin_byte_to_char_table =
"0000000000000001000000100000001100000100000001010000011000000111"
"0000100000001001000010100000101100001100000011010000111000001111"
"0001000000010001000100100001001100010100000101010001011000010111"
"0001100000011001000110100001101100011100000111010001111000011111"
"0010000000100001001000100010001100100100001001010010011000100111"
"0010100000101001001010100010101100101100001011010010111000101111"
"0011000000110001001100100011001100110100001101010011011000110111"
"0011100000111001001110100011101100111100001111010011111000111111"
"0100000001000001010000100100001101000100010001010100011001000111"
"0100100001001001010010100100101101001100010011010100111001001111"
"0101000001010001010100100101001101010100010101010101011001010111"
"0101100001011001010110100101101101011100010111010101111001011111"
"0110000001100001011000100110001101100100011001010110011001100111"
"0110100001101001011010100110101101101100011011010110111001101111"
"0111000001110001011100100111001101110100011101010111011001110111"
"0111100001111001011110100111101101111100011111010111111001111111"
"1000000010000001100000101000001110000100100001011000011010000111"
"1000100010001001100010101000101110001100100011011000111010001111"
"1001000010010001100100101001001110010100100101011001011010010111"
"1001100010011001100110101001101110011100100111011001111010011111"
"1010000010100001101000101010001110100100101001011010011010100111"
"1010100010101001101010101010101110101100101011011010111010101111"
"1011000010110001101100101011001110110100101101011011011010110111"
"1011100010111001101110101011101110111100101111011011111010111111"
"1100000011000001110000101100001111000100110001011100011011000111"
"1100100011001001110010101100101111001100110011011100111011001111"
"1101000011010001110100101101001111010100110101011101011011010111"
"1101100011011001110110101101101111011100110111011101111011011111"
"1110000011100001111000101110001111100100111001011110011011100111"
"1110100011101001111010101110101111101100111011011110111011101111"
"1111000011110001111100101111001111110100111101011111011011110111"
"1111100011111001111110101111101111111100111111011111111011111111";

145
src/Common/hex.h
View File

@ -1,145 +0,0 @@
#pragma once
#include <string>
/// Maps 0..15 to 0..9A..F or 0..9a..f correspondingly.
extern const char * const hex_digit_to_char_uppercase_table;
extern const char * const hex_digit_to_char_lowercase_table;
inline char hexDigitUppercase(unsigned char c)
{
return hex_digit_to_char_uppercase_table[c];
}
inline char hexDigitLowercase(unsigned char c)
{
return hex_digit_to_char_lowercase_table[c];
}
#include <cstring>
#include <cstddef>
#include <base/types.h>
/// Maps 0..255 to 00..FF or 00..ff correspondingly
extern const char * const hex_byte_to_char_uppercase_table;
extern const char * const hex_byte_to_char_lowercase_table;
inline void writeHexByteUppercase(UInt8 byte, void * out)
{
memcpy(out, &hex_byte_to_char_uppercase_table[static_cast<size_t>(byte) * 2], 2);
}
inline void writeHexByteLowercase(UInt8 byte, void * out)
{
memcpy(out, &hex_byte_to_char_lowercase_table[static_cast<size_t>(byte) * 2], 2);
}
extern const char * const bin_byte_to_char_table;
inline void writeBinByte(UInt8 byte, void * out)
{
memcpy(out, &bin_byte_to_char_table[static_cast<size_t>(byte) * 8], 8);
}
/// Produces hex representation of an unsigned int with leading zeros (for checksums)
template <typename TUInt>
inline void writeHexUIntImpl(TUInt uint_, char * out, const char * const table)
{
union
{
TUInt value;
UInt8 uint8[sizeof(TUInt)];
};
value = uint_;
for (size_t i = 0; i < sizeof(TUInt); ++i)
{
if constexpr (std::endian::native == std::endian::little)
memcpy(out + i * 2, &table[static_cast<size_t>(uint8[sizeof(TUInt) - 1 - i]) * 2], 2);
else
memcpy(out + i * 2, &table[static_cast<size_t>(uint8[i]) * 2], 2);
}
}
template <typename TUInt>
inline void writeHexUIntUppercase(TUInt uint_, char * out)
{
writeHexUIntImpl(uint_, out, hex_byte_to_char_uppercase_table);
}
template <typename TUInt>
inline void writeHexUIntLowercase(TUInt uint_, char * out)
{
writeHexUIntImpl(uint_, out, hex_byte_to_char_lowercase_table);
}
template <typename TUInt>
std::string getHexUIntUppercase(TUInt uint_)
{
std::string res(sizeof(TUInt) * 2, '\0');
writeHexUIntUppercase(uint_, res.data());
return res;
}
template <typename TUInt>
std::string getHexUIntLowercase(TUInt uint_)
{
std::string res(sizeof(TUInt) * 2, '\0');
writeHexUIntLowercase(uint_, res.data());
return res;
}
/// Maps 0..9, A..F, a..f to 0..15. Other chars are mapped to implementation specific value.
extern const char * const hex_char_to_digit_table;
inline UInt8 unhex(char c)
{
return hex_char_to_digit_table[static_cast<UInt8>(c)];
}
inline UInt8 unhex2(const char * data)
{
return
static_cast<UInt8>(unhex(data[0])) * 0x10
+ static_cast<UInt8>(unhex(data[1]));
}
inline UInt16 unhex4(const char * data)
{
return
static_cast<UInt16>(unhex(data[0])) * 0x1000
+ static_cast<UInt16>(unhex(data[1])) * 0x100
+ static_cast<UInt16>(unhex(data[2])) * 0x10
+ static_cast<UInt16>(unhex(data[3]));
}
template <typename TUInt>
TUInt unhexUInt(const char * data)
{
TUInt res = 0;
if constexpr ((sizeof(TUInt) <= 8) || ((sizeof(TUInt) % 8) != 0))
{
for (size_t i = 0; i < sizeof(TUInt) * 2; ++i, ++data)
{
res <<= 4;
res += unhex(*data);
}
}
else
{
for (size_t i = 0; i < sizeof(TUInt) / 8; ++i, data += 16)
{
res <<= 64;
res += unhexUInt<UInt64>(data);
}
}
return res;
}

Some files were not shown because too many files have changed in this diff Show More