thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-11-21 15:12:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into h3-integration

Browse Source
This commit is contained in:
alexey-milovidov 2019-06-30 16:19:11 +03:00 committed by GitHub
commit feafcb21bd
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
166 changed files with 2661 additions and 4713 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
.gitmodules vendored
View File

@ -88,3 +88,6 @@
[submodule "contrib/rapidjson"]
path = contrib/rapidjson
url = https://github.com/Tencent/rapidjson
[submodule "contrib/mimalloc"]
path = contrib/mimalloc
url = https://github.com/ClickHouse-Extras/mimalloc

121
CHANGELOG.md
View File

@ -1,3 +1,114 @@
## ClickHouse release 19.8.3.8, 2019-06-11
### New Features
* Added functions to work with JSON [#4686](https://github.com/yandex/ClickHouse/pull/4686) ([hcz](https://github.com/hczhcz)) [#5124](https://github.com/yandex/ClickHouse/pull/5124). ([Vitaly Baranov](https://github.com/vitlibar))
* Add a function basename, with a similar behaviour to a basename function, which exists in a lot of languages (`os.path.basename` in python, `basename` in PHP, etc...). Work with both an UNIX-like path or a Windows path. [#5136](https://github.com/yandex/ClickHouse/pull/5136) ([Guillaume Tassery](https://github.com/YiuRULE))
* Added `LIMIT n, m BY` or `LIMIT m OFFSET n BY` syntax to set offset of n for LIMIT BY clause. [#5138](https://github.com/yandex/ClickHouse/pull/5138) ([Anton Popov](https://github.com/CurtizJ))
* Added new data type `SimpleAggregateFunction`, which allows to have columns with light aggregation in an `AggregatingMergeTree`. This can only be used with simple functions like `any`, `anyLast`, `sum`, `min`, `max`. [#4629](https://github.com/yandex/ClickHouse/pull/4629) ([Boris Granveaud](https://github.com/bgranvea))
* Added support for non-constant arguments in function `ngramDistance` [#5198](https://github.com/yandex/ClickHouse/pull/5198) ([Danila Kutenin](https://github.com/danlark1))
* Added functions `skewPop`, `skewSamp`, `kurtPop` and `kurtSamp` to compute for sequence skewness, sample skewness, kurtosis and sample kurtosis respectively. [#5200](https://github.com/yandex/ClickHouse/pull/5200) ([hcz](https://github.com/hczhcz))
* Support rename operation for `MaterializeView` storage. [#5209](https://github.com/yandex/ClickHouse/pull/5209) ([Guillaume Tassery](https://github.com/YiuRULE))
* Added server which allows connecting to ClickHouse using MySQL client. [#4715](https://github.com/yandex/ClickHouse/pull/4715) ([Yuriy Baranov](https://github.com/yurriy))
* Add `toDecimal*OrZero` and `toDecimal*OrNull` functions. [#5291](https://github.com/yandex/ClickHouse/pull/5291) ([Artem Zuikov](https://github.com/4ertus2))
* Support Decimal types in functions: `quantile`, `quantiles`, `median`, `quantileExactWeighted`, `quantilesExactWeighted`, medianExactWeighted. [#5304](https://github.com/yandex/ClickHouse/pull/5304) ([Artem Zuikov](https://github.com/4ertus2))
* Added `toValidUTF8` function, which replaces all invalid UTF-8 characters by replacement character <20> (U+FFFD). [#5322](https://github.com/yandex/ClickHouse/pull/5322) ([Danila Kutenin](https://github.com/danlark1))
* Added `format` function. Formatting constant pattern (simplified Python format pattern) with the strings listed in the arguments. [#5330](https://github.com/yandex/ClickHouse/pull/5330) ([Danila Kutenin](https://github.com/danlark1))
* Added `system.detached_parts` table containing information about detached parts of `MergeTree` tables. [#5353](https://github.com/yandex/ClickHouse/pull/5353) ([akuzm](https://github.com/akuzm))
* Added `ngramSearch` function to calculate the non-symmetric difference between needle and haystack. [#5418](https://github.com/yandex/ClickHouse/pull/5418)[#5422](https://github.com/yandex/ClickHouse/pull/5422) ([Danila Kutenin](https://github.com/danlark1))
* Implementation of basic machine learning methods (stochastic linear regression and logistic regression) using aggregate functions interface. Has different strategies for updating model weights (simple gradient descent, momentum method, Nesterov method). Also supports mini-batches of custom size. [#4943](https://github.com/yandex/ClickHouse/pull/4943) ([Quid37](https://github.com/Quid37))
* Implementation of `geohashEncode` and `geohashDecode` functions. [#5003](https://github.com/yandex/ClickHouse/pull/5003) ([Vasily Nemkov](https://github.com/Enmk))
* Added aggregate function `timeSeriesGroupSum`, which can aggregate different time series that sample timestamp not alignment. It will use linear interpolation between two sample timestamp and then sum time-series together. Added aggregate function `timeSeriesGroupRateSum`, which calculates the rate of time-series and then sum rates together. [#4542](https://github.com/yandex/ClickHouse/pull/4542) ([Yangkuan Liu](https://github.com/LiuYangkuan))
* Added functions `IPv4CIDRtoIPv4Range` and `IPv6CIDRtoIPv6Range` to calculate the lower and higher bounds for an IP in the subnet using a CIDR. [#5095](https://github.com/yandex/ClickHouse/pull/5095) ([Guillaume Tassery](https://github.com/YiuRULE))
* Add a X-ClickHouse-Summary header when we send a query using HTTP with enabled setting `send_progress_in_http_headers`. Return the usual information of X-ClickHouse-Progress, with additional information like how many rows and bytes were inserted in the query. [#5116](https://github.com/yandex/ClickHouse/pull/5116) ([Guillaume Tassery](https://github.com/YiuRULE))
### Improvements
* Added `max_parts_in_total` setting for MergeTree family of tables (default: 100 000) that prevents unsafe specification of partition key #5166. [#5171](https://github.com/yandex/ClickHouse/pull/5171) ([alexey-milovidov](https://github.com/alexey-milovidov))
* `clickhouse-obfuscator`: derive seed for individual columns by combining initial seed with column name, not column position. This is intended to transform datasets with multiple related tables, so that tables will remain JOINable after transformation. [#5178](https://github.com/yandex/ClickHouse/pull/5178) ([alexey-milovidov](https://github.com/alexey-milovidov))
* Added functions `JSONExtractRaw`, `JSONExtractKeyAndValues`. Renamed functions `jsonExtract<type>` to `JSONExtract<type>`. When something goes wrong these functions return the correspondent values, not `NULL`. Modified function `JSONExtract`, now it gets the return type from its last parameter and doesn't inject nullables. Implemented fallback to RapidJSON in case AVX2 instructions are not available. Simdjson library updated to a new version. [#5235](https://github.com/yandex/ClickHouse/pull/5235) ([Vitaly Baranov](https://github.com/vitlibar))
* Now `if` and `multiIf` functions don't rely on the condition's `Nullable`, but rely on the branches for sql compatibility. [#5238](https://github.com/yandex/ClickHouse/pull/5238) ([Jian Wu](https://github.com/janplus))
* `In` predicate now generates `Null` result from `Null` input like the `Equal` function. [#5152](https://github.com/yandex/ClickHouse/pull/5152) ([Jian Wu](https://github.com/janplus))
* Check the time limit every (flush_interval / poll_timeout) number of rows from Kafka. This allows to break the reading from Kafka consumer more frequently and to check the time limits for the top-level streams [#5249](https://github.com/yandex/ClickHouse/pull/5249) ([Ivan](https://github.com/abyss7))
* Link rdkafka with bundled SASL. It should allow to use SASL SCRAM authentication [#5253](https://github.com/yandex/ClickHouse/pull/5253) ([Ivan](https://github.com/abyss7))
* Batched version of RowRefList for ALL JOINS. [#5267](https://github.com/yandex/ClickHouse/pull/5267) ([Artem Zuikov](https://github.com/4ertus2))
* clickhouse-server: more informative listen error messages. [#5268](https://github.com/yandex/ClickHouse/pull/5268) ([proller](https://github.com/proller))
* Support dictionaries in clickhouse-copier for functions in `<sharding_key>` [#5270](https://github.com/yandex/ClickHouse/pull/5270) ([proller](https://github.com/proller))
* Add new setting `kafka_commit_every_batch` to regulate Kafka committing policy.
It allows to set commit mode: after every batch of messages is handled, or after the whole block is written to the storage. It's a trade-off between losing some messages or reading them twice in some extreme situations. [#5308](https://github.com/yandex/ClickHouse/pull/5308) ([Ivan](https://github.com/abyss7))
* Make `windowFunnel` support other Unsigned Integer Types. [#5320](https://github.com/yandex/ClickHouse/pull/5320) ([sundyli](https://github.com/sundy-li))
* Allow to shadow virtual column `_table` in Merge engine. [#5325](https://github.com/yandex/ClickHouse/pull/5325) ([Ivan](https://github.com/abyss7))
* Make `sequenceMatch` aggregate functions support other unsigned Integer types [#5339](https://github.com/yandex/ClickHouse/pull/5339) ([sundyli](https://github.com/sundy-li))
* Better error messages if checksum mismatch is most likely caused by hardware failures. [#5355](https://github.com/yandex/ClickHouse/pull/5355) ([alexey-milovidov](https://github.com/alexey-milovidov))
* Check that underlying tables support sampling for `StorageMerge` [#5366](https://github.com/yandex/ClickHouse/pull/5366) ([Ivan](https://github.com/abyss7))
* Сlose MySQL connections after their usage in external dictionaries. It is related to issue #893. [#5395](https://github.com/yandex/ClickHouse/pull/5395) ([Clément Rodriguez](https://github.com/clemrodriguez))
* Improvements of MySQL Wire Protocol. Changed name of format to MySQLWire. Using RAII for calling RSA_free. Disabling SSL if context cannot be created. [#5419](https://github.com/yandex/ClickHouse/pull/5419) ([Yuriy Baranov](https://github.com/yurriy))
* clickhouse-client: allow to run with unaccessable history file (read-only, no disk space, file is directory, ...). [#5431](https://github.com/yandex/ClickHouse/pull/5431) ([proller](https://github.com/proller))
* Respect query settings in asynchronous INSERTs into Distributed tables. [#4936](https://github.com/yandex/ClickHouse/pull/4936) ([TCeason](https://github.com/TCeason))
* Renamed functions `leastSqr` to `simpleLinearRegression`, `LinearRegression` to `linearRegression`, `LogisticRegression` to `logisticRegression`. [#5391](https://github.com/yandex/ClickHouse/pull/5391) ([Nikolai Kochetov](https://github.com/KochetovNicolai))
### Performance Improvements
* Paralellize processing of parts in alter modify query. [#4639](https://github.com/yandex/ClickHouse/pull/4639) ([Ivan Kush](https://github.com/IvanKush))
* Optimizations in regular expressions extraction. [#5193](https://github.com/yandex/ClickHouse/pull/5193) [#5191](https://github.com/yandex/ClickHouse/pull/5191) ([Danila Kutenin](https://github.com/danlark1))
* Do not add right join key column to join result if it's used only in join on section. [#5260](https://github.com/yandex/ClickHouse/pull/5260) ([Artem Zuikov](https://github.com/4ertus2))
* Freeze the Kafka buffer after first empty response. It avoids multiple invokations of `ReadBuffer::next()` for empty result in some row-parsing streams. [#5283](https://github.com/yandex/ClickHouse/pull/5283) ([Ivan](https://github.com/abyss7))
* `concat` function optimization for multiple arguments. [#5357](https://github.com/yandex/ClickHouse/pull/5357) ([Danila Kutenin](https://github.com/danlark1))
* Query optimisation. Allow push down IN statement while rewriting commа/cross join into inner one. [#5396](https://github.com/yandex/ClickHouse/pull/5396) ([Artem Zuikov](https://github.com/4ertus2))
* Upgrade our LZ4 implementation with reference one to have faster decompression. [#5070](https://github.com/yandex/ClickHouse/pull/5070) ([Danila Kutenin](https://github.com/danlark1))
* Implemented MSD radix sort (based on kxsort), and partial sorting. [#5129](https://github.com/yandex/ClickHouse/pull/5129) ([Evgenii Pravda](https://github.com/kvinty))
### Bug Fixes
* Fix push require columns with join [#5192](https://github.com/yandex/ClickHouse/pull/5192) ([Winter Zhang](https://github.com/zhang2014))
* Fixed bug, when ClickHouse is run by systemd, the command `sudo service clickhouse-server forcerestart` was not working as expected. [#5204](https://github.com/yandex/ClickHouse/pull/5204) ([proller](https://github.com/proller))
* Fix http error codes in DataPartsExchange (interserver http server on 9009 port always returned code 200, even on errors). [#5216](https://github.com/yandex/ClickHouse/pull/5216) ([proller](https://github.com/proller))
* Fix SimpleAggregateFunction for String longer than MAX_SMALL_STRING_SIZE [#5311](https://github.com/yandex/ClickHouse/pull/5311) ([Azat Khuzhin](https://github.com/azat))
* Fix error for `Decimal` to `Nullable(Decimal)` conversion in IN. Support other Decimal to Decimal conversions (including different scales). [#5350](https://github.com/yandex/ClickHouse/pull/5350) ([Artem Zuikov](https://github.com/4ertus2))
* Fixed FPU clobbering in simdjson library that lead to wrong calculation of `uniqHLL` and `uniqCombined` aggregate function and math functions such as `log`. [#5354](https://github.com/yandex/ClickHouse/pull/5354) ([alexey-milovidov](https://github.com/alexey-milovidov))
* Fixed handling mixed const/nonconst cases in JSON functions. [#5435](https://github.com/yandex/ClickHouse/pull/5435) ([Vitaly Baranov](https://github.com/vitlibar))
* Fix `retention` function. Now all conditions that satisfy in a row of data are added to the data state. [#5119](https://github.com/yandex/ClickHouse/pull/5119) ([小路](https://github.com/nicelulu))
* Fix result type for `quantileExact` with Decimals. [#5304](https://github.com/yandex/ClickHouse/pull/5304) ([Artem Zuikov](https://github.com/4ertus2))
### Documentation
* Translate documentation for `CollapsingMergeTree` to chinese. [#5168](https://github.com/yandex/ClickHouse/pull/5168) ([张风啸](https://github.com/AlexZFX))
* Translate some documentation about table engines to chinese.
[#5134](https://github.com/yandex/ClickHouse/pull/5134)
[#5328](https://github.com/yandex/ClickHouse/pull/5328)
([never lee](https://github.com/neverlee))
### Build/Testing/Packaging Improvements
* Fix some sanitizer reports that show probable use-after-free.[#5139](https://github.com/yandex/ClickHouse/pull/5139) [#5143](https://github.com/yandex/ClickHouse/pull/5143) [#5393](https://github.com/yandex/ClickHouse/pull/5393) ([Ivan](https://github.com/abyss7))
* Move performance tests out of separate directories for convenience. [#5158](https://github.com/yandex/ClickHouse/pull/5158) ([alexey-milovidov](https://github.com/alexey-milovidov))
* Fix incorrect performance tests. [#5255](https://github.com/yandex/ClickHouse/pull/5255) ([alesapin](https://github.com/alesapin))
* Added a tool to calculate checksums caused by bit flips to debug hardware issues. [#5334](https://github.com/yandex/ClickHouse/pull/5334) ([alexey-milovidov](https://github.com/alexey-milovidov))
* Make runner script more usable. [#5340](https://github.com/yandex/ClickHouse/pull/5340)[#5360](https://github.com/yandex/ClickHouse/pull/5360) ([filimonov](https://github.com/filimonov))
* Add small instruction how to write performance tests. [#5408](https://github.com/yandex/ClickHouse/pull/5408) ([alesapin](https://github.com/alesapin))
* Add ability to make substitutions in create, fill and drop query in performance tests [#5367](https://github.com/yandex/ClickHouse/pull/5367) ([Olga Khvostikova](https://github.com/stavrolia))
## ClickHouse release 19.7.5.27, 2019-06-09
### New features
* Added bitmap related functions `bitmapHasAny` and `bitmapHasAll` analogous to `hasAny` and `hasAll` functions for arrays. [#5279](https://github.com/yandex/ClickHouse/pull/5279) ([Sergi Vladykin](https://github.com/svladykin))
### Bug Fixes
* Fix segfault on `minmax` INDEX with Null value. [#5246](https://github.com/yandex/ClickHouse/pull/5246) ([Nikita Vasilev](https://github.com/nikvas0))
* Mark all input columns in LIMIT BY as required output. It fixes 'Not found column' error in some distributed queries. [#5407](https://github.com/yandex/ClickHouse/pull/5407) ([Constantin S. Pan](https://github.com/kvap))
* Fix "Column '0' already exists" error in `SELECT .. PREWHERE` on column with DEFAULT [#5397](https://github.com/yandex/ClickHouse/pull/5397) ([proller](https://github.com/proller))
* Fix `ALTER MODIFY TTL` query on `ReplicatedMergeTree`. [#5539](https://github.com/yandex/ClickHouse/pull/5539/commits) ([Anton Popov](https://github.com/CurtizJ))
* Don't crash the server when Kafka consumers have failed to start. [#5285](https://github.com/yandex/ClickHouse/pull/5285) ([Ivan](https://github.com/abyss7))
* Fixed bitmap functions produce wrong result. [#5359](https://github.com/yandex/ClickHouse/pull/5359) ([Andy Yang](https://github.com/andyyzh))
* Fix element_count for hashed dictionary (do not include duplicates) [#5440](https://github.com/yandex/ClickHouse/pull/5440) ([Azat Khuzhin](https://github.com/azat))
* Use contents of environment variable TZ as the name for timezone. It helps to correctly detect default timezone in some cases.[#5443](https://github.com/yandex/ClickHouse/pull/5443) ([Ivan](https://github.com/abyss7))
* Do not try to convert integers in `dictGetT` functions, because it doesn't work correctly. Throw an exception instead. [#5446](https://github.com/yandex/ClickHouse/pull/5446) ([Artem Zuikov](https://github.com/4ertus2))
* Fix settings in ExternalData HTTP request. [#5455](https://github.com/yandex/ClickHouse/pull/5455) ([Danila
Kutenin](https://github.com/danlark1))
* Fix bug when parts were removed only from FS without dropping them from Zookeeper. [#5520](https://github.com/yandex/ClickHouse/pull/5520) ([alesapin](https://github.com/alesapin))
* Fix segmentation fault in `bitmapHasAny` function. [#5528](https://github.com/yandex/ClickHouse/pull/5528) ([Zhichang Yu](https://github.com/yuzhichang))
* Fixed error when replication connection pool doesn't retry to resolve host, even when DNS cache was dropped. [#5534](https://github.com/yandex/ClickHouse/pull/5534) ([alesapin](https://github.com/alesapin))
* Fixed `DROP INDEX IF EXISTS` query. Now `ALTER TABLE ... DROP INDEX IF EXISTS ...` query doesn't raise an exception if provided index does not exist. [#5524](https://github.com/yandex/ClickHouse/pull/5524) ([Gleb Novikov](https://github.com/NanoBjorn))
* Fix union all supertype column. There were cases with inconsistent data and column types of resulting columns. [#5503](https://github.com/yandex/ClickHouse/pull/5503) ([Artem Zuikov](https://github.com/4ertus2))
* Skip ZNONODE during DDL query processing. Before if another node removes the znode in task queue, the one that
did not process it, but already get list of children, will terminate the DDLWorker thread. [#5489](https://github.com/yandex/ClickHouse/pull/5489) ([Azat Khuzhin](https://github.com/azat))
* Fix INSERT into Distributed() table with MATERIALIZED column. [#5429](https://github.com/yandex/ClickHouse/pull/5429) ([Azat Khuzhin](https://github.com/azat))
## ClickHouse release 19.7.3.9, 2019-05-30
### New Features
@ -60,6 +171,16 @@ lee](https://github.com/neverlee))
[#5110](https://github.com/yandex/ClickHouse/pull/5110)
([proller](https://github.com/proller))
## ClickHouse release 19.6.3.18, 2019-06-13
### Bug Fixes
* Fixed IN condition pushdown for queries from table functions `mysql` and `odbc` and corresponding table engines. This fixes #3540 and #2384. [#5313](https://github.com/yandex/ClickHouse/pull/5313) ([alexey-milovidov](https://github.com/alexey-milovidov))
* Fix deadlock in Zookeeper. [#5297](https://github.com/yandex/ClickHouse/pull/5297) ([github1youlc](https://github.com/github1youlc))
* Allow quoted decimals in CSV. [#5284](https://github.com/yandex/ClickHouse/pull/5284) ([Artem Zuikov](https://github.com/4ertus2)
* Disallow conversion from float Inf/NaN into Decimals (throw exception). [#5282](https://github.com/yandex/ClickHouse/pull/5282) ([Artem Zuikov](https://github.com/4ertus2))
* Fix data race in rename query. [#5247](https://github.com/yandex/ClickHouse/pull/5247) ([Winter Zhang](https://github.com/zhang2014))
* Temporarily disable LFAlloc. Usage of LFAlloc might lead to a lot of MAP_FAILED in allocating UncompressedCache and in a result to crashes of queries at high loaded servers. [cfdba93](https://github.com/yandex/ClickHouse/commit/cfdba938ce22f16efeec504f7f90206a515b1280)([Danila Kutenin](https://github.com/danlark1))
## ClickHouse release 19.6.2.11, 2019-05-13
### New Features

32
CMakeLists.txt
View File

@ -16,6 +16,8 @@ set(CMAKE_LINK_DEPENDS_NO_SHARED 1) # Do not relink all depended targets on .so
set(CMAKE_CONFIGURATION_TYPES "RelWithDebInfo;Debug;Release;MinSizeRel" CACHE STRING "" FORCE)
set(CMAKE_DEBUG_POSTFIX "d" CACHE STRING "Generate debug library name with a postfix.") # To be consistent with CMakeLists from contrib libs.
include (cmake/arch.cmake)
option(ENABLE_IPO "Enable inter-procedural optimization (aka LTO)" OFF) # need cmake 3.9+
if(ENABLE_IPO)
cmake_policy(SET CMP0069 NEW)
@ -31,12 +33,12 @@ else()
message(STATUS "IPO/LTO not enabled.")
endif()
if (CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
if (COMPILER_GCC)
# Require at least gcc 7
if (CMAKE_CXX_COMPILER_VERSION VERSION_LESS 7 AND NOT CMAKE_VERSION VERSION_LESS 2.8.9)
message (FATAL_ERROR "GCC version must be at least 7. For example, if GCC 7 is available under gcc-7, g++-7 names, do the following: export CC=gcc-7 CXX=g++-7; rm -rf CMakeCache.txt CMakeFiles; and re run cmake or ./release.")
endif ()
elseif (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
elseif (COMPILER_CLANG)
# Require at least clang 5
if (CMAKE_CXX_COMPILER_VERSION VERSION_LESS 5)
message (FATAL_ERROR "Clang version must be at least 5.")
@ -81,7 +83,6 @@ endif ()
include (cmake/sanitize.cmake)
include (cmake/arch.cmake)
if (CMAKE_GENERATOR STREQUAL "Ninja")
# Turn on colored output. https://github.com/ninja-build/ninja/wiki/FAQ
@ -102,13 +103,12 @@ if (COMPILER_GCC AND CMAKE_CXX_COMPILER_VERSION VERSION_GREATER "8.3.0")
set (CXX_WARNING_FLAGS "${CXX_WARNING_FLAGS} -Wno-array-bounds")
endif ()
if (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
if (COMPILER_CLANG)
# clang: warning: argument unused during compilation: '-stdlib=libc++'
# clang: warning: argument unused during compilation: '-specs=/usr/share/dpkg/no-pie-compile.specs' [-Wunused-command-line-argument]
set (COMMON_WARNING_FLAGS "${COMMON_WARNING_FLAGS} -Wno-unused-command-line-argument")
endif ()
option (TEST_COVERAGE "Enables flags for test coverage" OFF)
option (ENABLE_TESTS "Enables tests" ON)
if (CMAKE_SYSTEM_PROCESSOR MATCHES "amd64|x86_64")
@ -128,7 +128,7 @@ string(REGEX MATCH "-?[0-9]+(.[0-9]+)?$" COMPILER_POSTFIX ${CMAKE_CXX_COMPILER})
find_program (LLD_PATH NAMES "lld${COMPILER_POSTFIX}" "lld")
find_program (GOLD_PATH NAMES "gold")
if (CMAKE_CXX_COMPILER_ID STREQUAL "Clang" AND LLD_PATH AND NOT LINKER_NAME)
if (COMPILER_CLANG AND LLD_PATH AND NOT LINKER_NAME)
set (LINKER_NAME "lld")
elseif (GOLD_PATH)
set (LINKER_NAME "gold")
@ -162,7 +162,7 @@ if (ARCH_NATIVE)
endif ()
# Special options for better optimized code with clang
#if (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
#if (COMPILER_CLANG)
# set (CMAKE_CXX_FLAGS_RELWITHDEBINFO "${CMAKE_CXX_FLAGS_RELWITHDEBINFO} -Wno-unused-command-line-argument -mllvm -inline-threshold=10000")
#endif ()
@ -177,6 +177,14 @@ else ()
set (CXX_FLAGS_INTERNAL_COMPILER "-std=c++1z")
endif ()
option(WITH_COVERAGE "Build with coverage." 0)
if(WITH_COVERAGE AND COMPILER_CLANG)
set(COMPILER_FLAGS "${COMPILER_FLAGS} -fprofile-instr-generate -fcoverage-mapping")
endif()
if(WITH_COVERAGE AND COMPILER_GCC)
set(COMPILER_FLAGS "${COMPILER_FLAGS} -fprofile-arcs -ftest-coverage")
endif()
set (CMAKE_BUILD_COLOR_MAKEFILE ON)
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${COMPILER_FLAGS} ${PLATFORM_EXTRA_CXX_FLAG} -fno-omit-frame-pointer ${COMMON_WARNING_FLAGS} ${CXX_WARNING_FLAGS}")
#set (CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} ${CMAKE_CXX_FLAGS_ADD}")
@ -188,10 +196,8 @@ set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${COMPILER_FLAGS} -fn
set (CMAKE_C_FLAGS_RELWITHDEBINFO "${CMAKE_C_FLAGS_RELWITHDEBINFO} -O3 ${CMAKE_C_FLAGS_ADD}")
set (CMAKE_C_FLAGS_DEBUG "${CMAKE_C_FLAGS_DEBUG} -O0 -g3 -ggdb3 -fno-inline ${CMAKE_C_FLAGS_ADD}")
include (cmake/use_libcxx.cmake)
# Set standard, system and compiler libraries explicitly.
# This is intended for more control of what we are linking.
@ -251,7 +257,6 @@ if (DEFAULT_LIBS)
set(CMAKE_CXX_STANDARD_LIBRARIES ${DEFAULT_LIBS})
endif ()
if (NOT MAKE_STATIC_LIBRARIES)
set(CMAKE_POSITION_INDEPENDENT_CODE ON)
endif ()
@ -268,11 +273,6 @@ if (USE_INCLUDE_WHAT_YOU_USE)
endif()
endif ()
# Flags for test coverage
if (TEST_COVERAGE)
set (CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fprofile-arcs -ftest-coverage -DIS_DEBUG")
endif (TEST_COVERAGE)
if (ENABLE_TESTS)
message (STATUS "Tests are enabled")
endif ()
@ -336,7 +336,7 @@ include (cmake/find_hdfs3.cmake) # uses protobuf
include (cmake/find_consistent-hashing.cmake)
include (cmake/find_base64.cmake)
include (cmake/find_hyperscan.cmake)
include (cmake/find_lfalloc.cmake)
include (cmake/find_mimalloc.cmake)
include (cmake/find_simdjson.cmake)
include (cmake/find_rapidjson.cmake)
find_contrib_lib(cityhash)

2
README.md
View File

@ -12,8 +12,6 @@ ClickHouse is an open-source column-oriented database management system that all
* You can also [fill this form](https://forms.yandex.com/surveys/meet-yandex-clickhouse-team/) to meet Yandex ClickHouse team in person.
## Upcoming Events
* [ClickHouse on HighLoad++ Siberia](https://www.highload.ru/siberia/2019/abstracts/5348) on June 24-25.
* [ClickHouse Meetup in Novosibirsk](https://events.yandex.ru/events/ClickHouse/26-June-2019/) on June 26.
* [ClickHouse Meetup in Minsk](https://yandex.ru/promo/metrica/clickhouse-minsk) on July 11.
* [ClickHouse Meetup in Shenzhen](https://www.huodongxing.com/event/3483759917300) on October 20.
* [ClickHouse Meetup in Shanghai](https://www.huodongxing.com/event/4483760336000) on October 27.

11
cmake/find_lfalloc.cmake
View File

@ -1,11 +0,0 @@
# TODO(danlark1). Disable LFAlloc for a while to fix mmap count problem
if (NOT OS_LINUX AND NOT SANITIZE AND NOT ARCH_ARM AND NOT ARCH_32 AND NOT ARCH_PPC64LE AND NOT OS_FREEBSD AND NOT APPLE)
option (ENABLE_LFALLOC "Set to FALSE to use system libgsasl library instead of bundled" ${NOT_UNBUNDLED})
endif ()
if (ENABLE_LFALLOC)
set (USE_LFALLOC 1)
set (USE_LFALLOC_RANDOM_HINT 1)
set (LFALLOC_INCLUDE_DIR ${ClickHouse_SOURCE_DIR}/contrib/lfalloc/src)
message (STATUS "Using lfalloc=${USE_LFALLOC}: ${LFALLOC_INCLUDE_DIR}")
endif ()

15
cmake/find_mimalloc.cmake Normal file
View File

@ -0,0 +1,15 @@
if (OS_LINUX AND NOT SANITIZE AND NOT ARCH_ARM AND NOT ARCH_32 AND NOT ARCH_PPC64LE)
option (ENABLE_MIMALLOC "Set to FALSE to disable usage of mimalloc for internal ClickHouse caches" ${NOT_UNBUNDLED})
endif ()
if (NOT EXISTS "${ClickHouse_SOURCE_DIR}/contrib/mimalloc/include/mimalloc.h")
message (WARNING "submodule contrib/mimalloc is missing. to fix try run: \n git submodule update --init --recursive")
return()
endif ()
if (ENABLE_MIMALLOC)
set (MIMALLOC_INCLUDE_DIR ${ClickHouse_SOURCE_DIR}/contrib/mimalloc/include)
set (USE_MIMALLOC 1)
set (MIMALLOC_LIBRARY mimalloc-static)
message (STATUS "Using mimalloc: ${MIMALLOC_INCLUDE_DIR} : ${MIMALLOC_LIBRARY}")
endif ()

12
contrib/CMakeLists.txt vendored
View File

@ -1,11 +1,11 @@
# Third-party libraries may have substandard code.
if (CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wno-unused-function -Wno-unused-variable -Wno-unused-but-set-variable -Wno-unused-result -Wno-deprecated-declarations -Wno-maybe-uninitialized -Wno-format -Wno-misleading-indentation -Wno-stringop-overflow -Wno-implicit-function-declaration -Wno-return-type -Wno-array-bounds -Wno-bool-compare -Wno-int-conversion -Wno-switch -Wno-stringop-truncation")
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-old-style-cast -Wno-unused-function -Wno-unused-variable -Wno-unused-but-set-variable -Wno-unused-result -Wno-deprecated-declarations -Wno-non-virtual-dtor -Wno-maybe-uninitialized -Wno-format -Wno-misleading-indentation -Wno-implicit-fallthrough -Wno-class-memaccess -Wno-sign-compare -Wno-array-bounds -Wno-missing-attributes -Wno-stringop-truncation -std=c++1z")
set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -w")
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -w -std=c++1z")
elseif (CMAKE_CXX_COMPILER_ID STREQUAL "Clang")
set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Wno-unused-function -Wno-unused-variable -Wno-unused-result -Wno-deprecated-declarations -Wno-format -Wno-parentheses-equality -Wno-tautological-constant-compare -Wno-tautological-constant-out-of-range-compare -Wno-implicit-function-declaration -Wno-return-type -Wno-pointer-bool-conversion -Wno-enum-conversion -Wno-int-conversion -Wno-switch -Wno-string-plus-int")
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wno-old-style-cast -Wno-unused-function -Wno-unused-variable -Wno-unused-result -Wno-deprecated-declarations -Wno-non-virtual-dtor -Wno-format -Wno-inconsistent-missing-override -std=c++1z")
set (CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -w")
set (CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -w -std=c++1z")
endif ()
set_property(DIRECTORY PROPERTY EXCLUDE_FROM_ALL 1)
@ -321,3 +321,7 @@ endif()
if (USE_SIMDJSON)
add_subdirectory (simdjson-cmake)
endif()
if (USE_MIMALLOC)
add_subdirectory (mimalloc)
endif()

1820
contrib/lfalloc/src/lf_allocX64.h
View File

File diff suppressed because it is too large Load Diff

23
contrib/lfalloc/src/lfmalloc.h
View File

@ -1,23 +0,0 @@
#pragma once
#include <string.h>
#include <stdlib.h>
#include "util/system/compiler.h"
namespace NMalloc {
volatile inline bool IsAllocatorCorrupted = false;
static inline void AbortFromCorruptedAllocator() {
IsAllocatorCorrupted = true;
abort();
}
struct TAllocHeader {
void* Block;
size_t AllocSize;
void Y_FORCE_INLINE Encode(void* block, size_t size, size_t signature) {
Block = block;
AllocSize = size | signature;
}
};
}

33
contrib/lfalloc/src/util/README.md
View File

@ -1,33 +0,0 @@
Style guide for the util folder is a stricter version of general style guide (mostly in terms of ambiguity resolution).
* all {} must be in K&R style
* &, * tied closer to a type, not to variable
* always use `using` not `typedef`
* even a single line block must be in braces {}:
```
if (A) {
B();
}
```
* _ at the end of private data member of a class - `First_`, `Second_`
* every .h file must be accompanied with corresponding .cpp to avoid a leakage and check that it is self contained
* prohibited to use `printf`-like functions
Things declared in the general style guide, which sometimes are missed:
* `template <`, not `template<`
* `noexcept`, not `throw ()` nor `throw()`, not required for destructors
* indents inside `namespace` same as inside `class`
Requirements for a new code (and for corrections in an old code which involves change of behaviour) in util:
* presence of UNIT-tests
* presence of comments in Doxygen style
* accessors without Get prefix (`Length()`, but not `GetLength()`)
This guide is not a mandatory as there is the general style guide.
Nevertheless if it is not followed, then a next `ya style .` run in the util folder will undeservedly update authors of some lines of code.
Thus before a commit it is recommended to run `ya style .` in the util folder.

51
contrib/lfalloc/src/util/system/atomic.h
View File

@ -1,51 +0,0 @@
#pragma once
#include "defaults.h"
using TAtomicBase = intptr_t;
using TAtomic = volatile TAtomicBase;
#if defined(__GNUC__)
#include "atomic_gcc.h"
#elif defined(_MSC_VER)
#include "atomic_win.h"
#else
#error unsupported platform
#endif
#if !defined(ATOMIC_COMPILER_BARRIER)
#define ATOMIC_COMPILER_BARRIER()
#endif
static inline TAtomicBase AtomicSub(TAtomic& a, TAtomicBase v) {
return AtomicAdd(a, -v);
}
static inline TAtomicBase AtomicGetAndSub(TAtomic& a, TAtomicBase v) {
return AtomicGetAndAdd(a, -v);
}
#if defined(USE_GENERIC_SETGET)
static inline TAtomicBase AtomicGet(const TAtomic& a) {
return a;
}
static inline void AtomicSet(TAtomic& a, TAtomicBase v) {
a = v;
}
#endif
static inline bool AtomicTryLock(TAtomic* a) {
return AtomicCas(a, 1, 0);
}
static inline bool AtomicTryAndTryLock(TAtomic* a) {
return (AtomicGet(*a) == 0) && AtomicTryLock(a);
}
static inline void AtomicUnlock(TAtomic* a) {
ATOMIC_COMPILER_BARRIER();
AtomicSet(*a, 0);
}
#include "atomic_ops.h"

90
contrib/lfalloc/src/util/system/atomic_gcc.h
View File

@ -1,90 +0,0 @@
#pragma once
#define ATOMIC_COMPILER_BARRIER() __asm__ __volatile__("" \
: \
: \
: "memory")
static inline TAtomicBase AtomicGet(const TAtomic& a) {
TAtomicBase tmp;
#if defined(_arm64_)
__asm__ __volatile__(
"ldar %x[value], %[ptr] \n\t"
: [value] "=r"(tmp)
: [ptr] "Q"(a)
: "memory");
#else
__atomic_load(&a, &tmp, __ATOMIC_ACQUIRE);
#endif
return tmp;
}
static inline void AtomicSet(TAtomic& a, TAtomicBase v) {
#if defined(_arm64_)
__asm__ __volatile__(
"stlr %x[value], %[ptr] \n\t"
: [ptr] "=Q"(a)
: [value] "r"(v)
: "memory");
#else
__atomic_store(&a, &v, __ATOMIC_RELEASE);
#endif
}
static inline intptr_t AtomicIncrement(TAtomic& p) {
return __atomic_add_fetch(&p, 1, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicGetAndIncrement(TAtomic& p) {
return __atomic_fetch_add(&p, 1, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicDecrement(TAtomic& p) {
return __atomic_sub_fetch(&p, 1, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicGetAndDecrement(TAtomic& p) {
return __atomic_fetch_sub(&p, 1, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicAdd(TAtomic& p, intptr_t v) {
return __atomic_add_fetch(&p, v, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicGetAndAdd(TAtomic& p, intptr_t v) {
return __atomic_fetch_add(&p, v, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicSwap(TAtomic* p, intptr_t v) {
(void)p; // disable strange 'parameter set but not used' warning on gcc
intptr_t ret;
__atomic_exchange(p, &v, &ret, __ATOMIC_SEQ_CST);
return ret;
}
static inline bool AtomicCas(TAtomic* a, intptr_t exchange, intptr_t compare) {
(void)a; // disable strange 'parameter set but not used' warning on gcc
return __atomic_compare_exchange(a, &compare, &exchange, false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicGetAndCas(TAtomic* a, intptr_t exchange, intptr_t compare) {
(void)a; // disable strange 'parameter set but not used' warning on gcc
__atomic_compare_exchange(a, &compare, &exchange, false, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST);
return compare;
}
static inline intptr_t AtomicOr(TAtomic& a, intptr_t b) {
return __atomic_or_fetch(&a, b, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicXor(TAtomic& a, intptr_t b) {
return __atomic_xor_fetch(&a, b, __ATOMIC_SEQ_CST);
}
static inline intptr_t AtomicAnd(TAtomic& a, intptr_t b) {
return __atomic_and_fetch(&a, b, __ATOMIC_SEQ_CST);
}
static inline void AtomicBarrier() {
__sync_synchronize();
}

189
contrib/lfalloc/src/util/system/atomic_ops.h
View File

@ -1,189 +0,0 @@
#pragma once
#include <type_traits>
template <typename T>
inline TAtomic* AsAtomicPtr(T volatile* target) {
return reinterpret_cast<TAtomic*>(target);
}
template <typename T>
inline const TAtomic* AsAtomicPtr(T const volatile* target) {
return reinterpret_cast<const TAtomic*>(target);
}
// integral types
template <typename T>
struct TAtomicTraits {
enum {
Castable = std::is_integral<T>::value && sizeof(T) == sizeof(TAtomicBase) && !std::is_const<T>::value,
};
};
template <typename T, typename TT>
using TEnableIfCastable = std::enable_if_t<TAtomicTraits<T>::Castable, TT>;
template <typename T>
inline TEnableIfCastable<T, T> AtomicGet(T const volatile& target) {
return static_cast<T>(AtomicGet(*AsAtomicPtr(&target)));
}
template <typename T>
inline TEnableIfCastable<T, void> AtomicSet(T volatile& target, TAtomicBase value) {
AtomicSet(*AsAtomicPtr(&target), value);
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicIncrement(T volatile& target) {
return static_cast<T>(AtomicIncrement(*AsAtomicPtr(&target)));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicGetAndIncrement(T volatile& target) {
return static_cast<T>(AtomicGetAndIncrement(*AsAtomicPtr(&target)));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicDecrement(T volatile& target) {
return static_cast<T>(AtomicDecrement(*AsAtomicPtr(&target)));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicGetAndDecrement(T volatile& target) {
return static_cast<T>(AtomicGetAndDecrement(*AsAtomicPtr(&target)));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicAdd(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicAdd(*AsAtomicPtr(&target), value));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicGetAndAdd(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicGetAndAdd(*AsAtomicPtr(&target), value));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicSub(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicSub(*AsAtomicPtr(&target), value));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicGetAndSub(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicGetAndSub(*AsAtomicPtr(&target), value));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicSwap(T volatile* target, TAtomicBase exchange) {
return static_cast<T>(AtomicSwap(AsAtomicPtr(target), exchange));
}
template <typename T>
inline TEnableIfCastable<T, bool> AtomicCas(T volatile* target, TAtomicBase exchange, TAtomicBase compare) {
return AtomicCas(AsAtomicPtr(target), exchange, compare);
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicGetAndCas(T volatile* target, TAtomicBase exchange, TAtomicBase compare) {
return static_cast<T>(AtomicGetAndCas(AsAtomicPtr(target), exchange, compare));
}
template <typename T>
inline TEnableIfCastable<T, bool> AtomicTryLock(T volatile* target) {
return AtomicTryLock(AsAtomicPtr(target));
}
template <typename T>
inline TEnableIfCastable<T, bool> AtomicTryAndTryLock(T volatile* target) {
return AtomicTryAndTryLock(AsAtomicPtr(target));
}
template <typename T>
inline TEnableIfCastable<T, void> AtomicUnlock(T volatile* target) {
AtomicUnlock(AsAtomicPtr(target));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicOr(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicOr(*AsAtomicPtr(&target), value));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicAnd(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicAnd(*AsAtomicPtr(&target), value));
}
template <typename T>
inline TEnableIfCastable<T, T> AtomicXor(T volatile& target, TAtomicBase value) {
return static_cast<T>(AtomicXor(*AsAtomicPtr(&target), value));
}
// pointer types
template <typename T>
inline T* AtomicGet(T* const volatile& target) {
return reinterpret_cast<T*>(AtomicGet(*AsAtomicPtr(&target)));
}
template <typename T>
inline void AtomicSet(T* volatile& target, T* value) {
AtomicSet(*AsAtomicPtr(&target), reinterpret_cast<TAtomicBase>(value));
}
using TNullPtr = decltype(nullptr);
template <typename T>
inline void AtomicSet(T* volatile& target, TNullPtr) {
AtomicSet(*AsAtomicPtr(&target), 0);
}
template <typename T>
inline T* AtomicSwap(T* volatile* target, T* exchange) {
return reinterpret_cast<T*>(AtomicSwap(AsAtomicPtr(target), reinterpret_cast<TAtomicBase>(exchange)));
}
template <typename T>
inline T* AtomicSwap(T* volatile* target, TNullPtr) {
return reinterpret_cast<T*>(AtomicSwap(AsAtomicPtr(target), 0));
}
template <typename T>
inline bool AtomicCas(T* volatile* target, T* exchange, T* compare) {
return AtomicCas(AsAtomicPtr(target), reinterpret_cast<TAtomicBase>(exchange), reinterpret_cast<TAtomicBase>(compare));
}
template <typename T>
inline T* AtomicGetAndCas(T* volatile* target, T* exchange, T* compare) {
return reinterpret_cast<T*>(AtomicGetAndCas(AsAtomicPtr(target), reinterpret_cast<TAtomicBase>(exchange), reinterpret_cast<TAtomicBase>(compare)));
}
template <typename T>
inline bool AtomicCas(T* volatile* target, T* exchange, TNullPtr) {
return AtomicCas(AsAtomicPtr(target), reinterpret_cast<TAtomicBase>(exchange), 0);
}
template <typename T>
inline T* AtomicGetAndCas(T* volatile* target, T* exchange, TNullPtr) {
return reinterpret_cast<T*>(AtomicGetAndCas(AsAtomicPtr(target), reinterpret_cast<TAtomicBase>(exchange), 0));
}
template <typename T>
inline bool AtomicCas(T* volatile* target, TNullPtr, T* compare) {
return AtomicCas(AsAtomicPtr(target), 0, reinterpret_cast<TAtomicBase>(compare));
}
template <typename T>
inline T* AtomicGetAndCas(T* volatile* target, TNullPtr, T* compare) {
return reinterpret_cast<T*>(AtomicGetAndCas(AsAtomicPtr(target), 0, reinterpret_cast<TAtomicBase>(compare)));
}
template <typename T>
inline bool AtomicCas(T* volatile* target, TNullPtr, TNullPtr) {
return AtomicCas(AsAtomicPtr(target), 0, 0);
}
template <typename T>
inline T* AtomicGetAndCas(T* volatile* target, TNullPtr, TNullPtr) {
return reinterpret_cast<T*>(AtomicGetAndCas(AsAtomicPtr(target), 0, 0));
}

114
contrib/lfalloc/src/util/system/atomic_win.h
View File

@ -1,114 +0,0 @@
#pragma once
#include <intrin.h>
#define USE_GENERIC_SETGET
#if defined(_i386_)
#pragma intrinsic(_InterlockedIncrement)
#pragma intrinsic(_InterlockedDecrement)
#pragma intrinsic(_InterlockedExchangeAdd)
#pragma intrinsic(_InterlockedExchange)
#pragma intrinsic(_InterlockedCompareExchange)
static inline intptr_t AtomicIncrement(TAtomic& a) {
return _InterlockedIncrement((volatile long*)&a);
}
static inline intptr_t AtomicGetAndIncrement(TAtomic& a) {
return _InterlockedIncrement((volatile long*)&a) - 1;
}
static inline intptr_t AtomicDecrement(TAtomic& a) {
return _InterlockedDecrement((volatile long*)&a);
}
static inline intptr_t AtomicGetAndDecrement(TAtomic& a) {
return _InterlockedDecrement((volatile long*)&a) + 1;
}
static inline intptr_t AtomicAdd(TAtomic& a, intptr_t b) {
return _InterlockedExchangeAdd((volatile long*)&a, b) + b;
}
static inline intptr_t AtomicGetAndAdd(TAtomic& a, intptr_t b) {
return _InterlockedExchangeAdd((volatile long*)&a, b);
}
static inline intptr_t AtomicSwap(TAtomic* a, intptr_t b) {
return _InterlockedExchange((volatile long*)a, b);
}
static inline bool AtomicCas(TAtomic* a, intptr_t exchange, intptr_t compare) {
return _InterlockedCompareExchange((volatile long*)a, exchange, compare) == compare;
}
static inline intptr_t AtomicGetAndCas(TAtomic* a, intptr_t exchange, intptr_t compare) {
return _InterlockedCompareExchange((volatile long*)a, exchange, compare);
}
#else // _x86_64_
#pragma intrinsic(_InterlockedIncrement64)
#pragma intrinsic(_InterlockedDecrement64)
#pragma intrinsic(_InterlockedExchangeAdd64)
#pragma intrinsic(_InterlockedExchange64)
#pragma intrinsic(_InterlockedCompareExchange64)
static inline intptr_t AtomicIncrement(TAtomic& a) {
return _InterlockedIncrement64((volatile __int64*)&a);
}
static inline intptr_t AtomicGetAndIncrement(TAtomic& a) {
return _InterlockedIncrement64((volatile __int64*)&a) - 1;
}
static inline intptr_t AtomicDecrement(TAtomic& a) {
return _InterlockedDecrement64((volatile __int64*)&a);
}
static inline intptr_t AtomicGetAndDecrement(TAtomic& a) {
return _InterlockedDecrement64((volatile __int64*)&a) + 1;
}
static inline intptr_t AtomicAdd(TAtomic& a, intptr_t b) {
return _InterlockedExchangeAdd64((volatile __int64*)&a, b) + b;
}
static inline intptr_t AtomicGetAndAdd(TAtomic& a, intptr_t b) {
return _InterlockedExchangeAdd64((volatile __int64*)&a, b);
}
static inline intptr_t AtomicSwap(TAtomic* a, intptr_t b) {
return _InterlockedExchange64((volatile __int64*)a, b);
}
static inline bool AtomicCas(TAtomic* a, intptr_t exchange, intptr_t compare) {
return _InterlockedCompareExchange64((volatile __int64*)a, exchange, compare) == compare;
}
static inline intptr_t AtomicGetAndCas(TAtomic* a, intptr_t exchange, intptr_t compare) {
return _InterlockedCompareExchange64((volatile __int64*)a, exchange, compare);
}
static inline intptr_t AtomicOr(TAtomic& a, intptr_t b) {
return _InterlockedOr64(&a, b) | b;
}
static inline intptr_t AtomicAnd(TAtomic& a, intptr_t b) {
return _InterlockedAnd64(&a, b) & b;
}
static inline intptr_t AtomicXor(TAtomic& a, intptr_t b) {
return _InterlockedXor64(&a, b) ^ b;
}
#endif // _x86_
//TODO
static inline void AtomicBarrier() {
TAtomic val = 0;
AtomicSwap(&val, 0);
}

617
contrib/lfalloc/src/util/system/compiler.h
View File

@ -1,617 +0,0 @@
#pragma once
// useful cross-platfrom definitions for compilers
/**
* @def Y_FUNC_SIGNATURE
*
* Use this macro to get pretty function name (see example).
*
* @code
* void Hi() {
* Cout << Y_FUNC_SIGNATURE << Endl;
* }
* template <typename T>
* void Do() {
* Cout << Y_FUNC_SIGNATURE << Endl;
* }
* int main() {
* Hi(); // void Hi()
* Do<int>(); // void Do() [T = int]
* Do<TString>(); // void Do() [T = TString]
* }
* @endcode
*/
#if defined(__GNUC__)
#define Y_FUNC_SIGNATURE __PRETTY_FUNCTION__
#elif defined(_MSC_VER)
#define Y_FUNC_SIGNATURE __FUNCSIG__
#else
#define Y_FUNC_SIGNATURE ""
#endif
#ifdef __GNUC__
#define Y_PRINTF_FORMAT(n, m) __attribute__((__format__(__printf__, n, m)))
#endif
#ifndef Y_PRINTF_FORMAT
#define Y_PRINTF_FORMAT(n, m)
#endif
#if defined(__clang__)
#define Y_NO_SANITIZE(...) __attribute__((no_sanitize(__VA_ARGS__)))
#endif
#if !defined(Y_NO_SANITIZE)
#define Y_NO_SANITIZE(...)
#endif
/**
* @def Y_DECLARE_UNUSED
*
* Macro is needed to silence compiler warning about unused entities (e.g. function or argument).
*
* @code
* Y_DECLARE_UNUSED int FunctionUsedSolelyForDebugPurposes();
* assert(FunctionUsedSolelyForDebugPurposes() == 42);
*
* void Foo(const int argumentUsedOnlyForDebugPurposes Y_DECLARE_UNUSED) {
* assert(argumentUsedOnlyForDebugPurposes == 42);
* // however you may as well omit `Y_DECLARE_UNUSED` and use `UNUSED` macro instead
* Y_UNUSED(argumentUsedOnlyForDebugPurposes);
* }
* @endcode
*/
#ifdef __GNUC__
#define Y_DECLARE_UNUSED __attribute__((unused))
#endif
#ifndef Y_DECLARE_UNUSED
#define Y_DECLARE_UNUSED
#endif
#if defined(__GNUC__)
#define Y_LIKELY(Cond) __builtin_expect(!!(Cond), 1)
#define Y_UNLIKELY(Cond) __builtin_expect(!!(Cond), 0)
#define Y_PREFETCH_READ(Pointer, Priority) __builtin_prefetch((const void*)(Pointer), 0, Priority)
#define Y_PREFETCH_WRITE(Pointer, Priority) __builtin_prefetch((const void*)(Pointer), 1, Priority)
#endif
/**
* @def Y_FORCE_INLINE
*
* Macro to use in place of 'inline' in function declaration/definition to force
* it to be inlined.
*/
#if !defined(Y_FORCE_INLINE)
#if defined(CLANG_COVERAGE)
#/* excessive __always_inline__ might significantly slow down compilation of an instrumented unit */
#define Y_FORCE_INLINE inline
#elif defined(_MSC_VER)
#define Y_FORCE_INLINE __forceinline
#elif defined(__GNUC__)
#/* Clang also defines __GNUC__ (as 4) */
#define Y_FORCE_INLINE inline __attribute__((__always_inline__))
#else
#define Y_FORCE_INLINE inline
#endif
#endif
/**
* @def Y_NO_INLINE
*
* Macro to use in place of 'inline' in function declaration/definition to
* prevent it from being inlined.
*/
#if !defined(Y_NO_INLINE)
#if defined(_MSC_VER)
#define Y_NO_INLINE __declspec(noinline)
#elif defined(__GNUC__) || defined(__INTEL_COMPILER)
#/* Clang also defines __GNUC__ (as 4) */
#define Y_NO_INLINE __attribute__((__noinline__))
#else
#define Y_NO_INLINE
#endif
#endif
//to cheat compiler about strict aliasing or similar problems
#if defined(__GNUC__)
#define Y_FAKE_READ(X) \
do { \
__asm__ __volatile__("" \
: \
: "m"(X)); \
} while (0)
#define Y_FAKE_WRITE(X) \
do { \
__asm__ __volatile__("" \
: "=m"(X)); \
} while (0)
#endif
#if !defined(Y_FAKE_READ)
#define Y_FAKE_READ(X)
#endif
#if !defined(Y_FAKE_WRITE)
#define Y_FAKE_WRITE(X)
#endif
#ifndef Y_PREFETCH_READ
#define Y_PREFETCH_READ(Pointer, Priority) (void)(const void*)(Pointer), (void)Priority
#endif
#ifndef Y_PREFETCH_WRITE
#define Y_PREFETCH_WRITE(Pointer, Priority) (void)(const void*)(Pointer), (void)Priority
#endif
#ifndef Y_LIKELY
#define Y_LIKELY(Cond) (Cond)
#define Y_UNLIKELY(Cond) (Cond)
#endif
#ifdef __GNUC__
#define _packed __attribute__((packed))
#else
#define _packed
#endif
#if defined(__GNUC__)
#define Y_WARN_UNUSED_RESULT __attribute__((warn_unused_result))
#endif
#ifndef Y_WARN_UNUSED_RESULT
#define Y_WARN_UNUSED_RESULT
#endif
#if defined(__GNUC__)
#define Y_HIDDEN __attribute__((visibility("hidden")))
#endif
#if !defined(Y_HIDDEN)
#define Y_HIDDEN
#endif
#if defined(__GNUC__)
#define Y_PUBLIC __attribute__((visibility("default")))
#endif
#if !defined(Y_PUBLIC)
#define Y_PUBLIC
#endif
#if !defined(Y_UNUSED) && !defined(__cplusplus)
#define Y_UNUSED(var) (void)(var)
#endif
#if !defined(Y_UNUSED) && defined(__cplusplus)
template <class... Types>
constexpr Y_FORCE_INLINE int Y_UNUSED(Types&&...) {
return 0;
};
#endif
/**
* @def Y_ASSUME
*
* Macro that tells the compiler that it can generate optimized code
* as if the given expression will always evaluate true.
* The behavior is undefined if it ever evaluates false.
*
* @code
* // factored into a function so that it's testable
* inline int Avg(int x, int y) {
* if (x >= 0 && y >= 0) {
* return (static_cast<unsigned>(x) + static_cast<unsigned>(y)) >> 1;
* } else {
* // a slower implementation
* }
* }
*
* // we know that xs and ys are non-negative from domain knowledge,
* // but we can't change the types of xs and ys because of API constrains
* int Foo(const TVector<int>& xs, const TVector<int>& ys) {
* TVector<int> avgs;
* avgs.resize(xs.size());
* for (size_t i = 0; i < xs.size(); ++i) {
* auto x = xs[i];
* auto y = ys[i];
* Y_ASSUME(x >= 0);
* Y_ASSUME(y >= 0);
* xs[i] = Avg(x, y);
* }
* }
* @endcode
*/
#if defined(__GNUC__)
#define Y_ASSUME(condition) ((condition) ? (void)0 : __builtin_unreachable())
#elif defined(_MSC_VER)
#define Y_ASSUME(condition) __assume(condition)
#else
#define Y_ASSUME(condition) Y_UNUSED(condition)
#endif
#ifdef __cplusplus
[[noreturn]]
#endif
Y_HIDDEN void _YandexAbort();
/**
* @def Y_UNREACHABLE
*
* Macro that marks the rest of the code branch unreachable.
* The behavior is undefined if it's ever reached.
*
* @code
* switch (i % 3) {
* case 0:
* return foo;
* case 1:
* return bar;
* case 2:
* return baz;
* default:
* Y_UNREACHABLE();
* }
* @endcode
*/
#if defined(__GNUC__) || defined(_MSC_VER)
#define Y_UNREACHABLE() Y_ASSUME(0)
#else
#define Y_UNREACHABLE() _YandexAbort()
#endif
#if defined(undefined_sanitizer_enabled)
#define _ubsan_enabled_
#endif
#ifdef __clang__
#if __has_feature(thread_sanitizer)
#define _tsan_enabled_
#endif
#if __has_feature(memory_sanitizer)
#define _msan_enabled_
#endif
#if __has_feature(address_sanitizer)
#define _asan_enabled_
#endif
#else
#if defined(thread_sanitizer_enabled) || defined(__SANITIZE_THREAD__)
#define _tsan_enabled_
#endif
#if defined(memory_sanitizer_enabled)
#define _msan_enabled_
#endif
#if defined(address_sanitizer_enabled) || defined(__SANITIZE_ADDRESS__)
#define _asan_enabled_
#endif
#endif
#if defined(_asan_enabled_) || defined(_msan_enabled_) || defined(_tsan_enabled_) || defined(_ubsan_enabled_)
#define _san_enabled_
#endif
#if defined(_MSC_VER)
#define __PRETTY_FUNCTION__ __FUNCSIG__
#endif
#if defined(__GNUC__)
#define Y_WEAK __attribute__((weak))
#else
#define Y_WEAK
#endif
#if defined(__CUDACC_VER_MAJOR__)
#define Y_CUDA_AT_LEAST(x, y) (__CUDACC_VER_MAJOR__ > x || (__CUDACC_VER_MAJOR__ == x && __CUDACC_VER_MINOR__ >= y))
#else
#define Y_CUDA_AT_LEAST(x, y) 0
#endif
// NVidia CUDA C++ Compiler did not know about noexcept keyword until version 9.0
#if !Y_CUDA_AT_LEAST(9, 0)
#if defined(__CUDACC__) && !defined(noexcept)
#define noexcept throw ()
#endif
#endif
#if defined(__GNUC__)
#define Y_COLD __attribute__((cold))
#define Y_LEAF __attribute__((leaf))
#define Y_WRAPPER __attribute__((artificial))
#else
#define Y_COLD
#define Y_LEAF
#define Y_WRAPPER
#endif
/**
* @def Y_PRAGMA
*
* Macro for use in other macros to define compiler pragma
* See below for other usage examples
*
* @code
* #if defined(__clang__) || defined(__GNUC__)
* #define Y_PRAGMA_NO_WSHADOW \
* Y_PRAGMA("GCC diagnostic ignored \"-Wshadow\"")
* #elif defined(_MSC_VER)
* #define Y_PRAGMA_NO_WSHADOW \
* Y_PRAGMA("warning(disable:4456 4457")
* #else
* #define Y_PRAGMA_NO_WSHADOW
* #endif
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA(x) _Pragma(x)
#elif defined(_MSC_VER)
#define Y_PRAGMA(x) __pragma(x)
#else
#define Y_PRAGMA(x)
#endif
/**
* @def Y_PRAGMA_DIAGNOSTIC_PUSH
*
* Cross-compiler pragma to save diagnostic settings
*
* @see
* GCC: https://gcc.gnu.org/onlinedocs/gcc/Diagnostic-Pragmas.html
* MSVC: https://msdn.microsoft.com/en-us/library/2c8f766e.aspx
* Clang: https://clang.llvm.org/docs/UsersManual.html#controlling-diagnostics-via-pragmas
*
* @code
* Y_PRAGMA_DIAGNOSTIC_PUSH
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA_DIAGNOSTIC_PUSH \
Y_PRAGMA("GCC diagnostic push")
#elif defined(_MSC_VER)
#define Y_PRAGMA_DIAGNOSTIC_PUSH \
Y_PRAGMA(warning(push))
#else
#define Y_PRAGMA_DIAGNOSTIC_PUSH
#endif
/**
* @def Y_PRAGMA_DIAGNOSTIC_POP
*
* Cross-compiler pragma to restore diagnostic settings
*
* @see
* GCC: https://gcc.gnu.org/onlinedocs/gcc/Diagnostic-Pragmas.html
* MSVC: https://msdn.microsoft.com/en-us/library/2c8f766e.aspx
* Clang: https://clang.llvm.org/docs/UsersManual.html#controlling-diagnostics-via-pragmas
*
* @code
* Y_PRAGMA_DIAGNOSTIC_POP
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA_DIAGNOSTIC_POP \
Y_PRAGMA("GCC diagnostic pop")
#elif defined(_MSC_VER)
#define Y_PRAGMA_DIAGNOSTIC_POP \
Y_PRAGMA(warning(pop))
#else
#define Y_PRAGMA_DIAGNOSTIC_POP
#endif
/**
* @def Y_PRAGMA_NO_WSHADOW
*
* Cross-compiler pragma to disable warnings about shadowing variables
*
* @code
* Y_PRAGMA_DIAGNOSTIC_PUSH
* Y_PRAGMA_NO_WSHADOW
*
* // some code which use variable shadowing, e.g.:
*
* for (int i = 0; i < 100; ++i) {
* Use(i);
*
* for (int i = 42; i < 100500; ++i) { // this i is shadowing previous i
* AnotherUse(i);
* }
* }
*
* Y_PRAGMA_DIAGNOSTIC_POP
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA_NO_WSHADOW \
Y_PRAGMA("GCC diagnostic ignored \"-Wshadow\"")
#elif defined(_MSC_VER)
#define Y_PRAGMA_NO_WSHADOW \
Y_PRAGMA(warning(disable : 4456 4457))
#else
#define Y_PRAGMA_NO_WSHADOW
#endif
/**
* @ def Y_PRAGMA_NO_UNUSED_FUNCTION
*
* Cross-compiler pragma to disable warnings about unused functions
*
* @see
* GCC: https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
* Clang: https://clang.llvm.org/docs/DiagnosticsReference.html#wunused-function
* MSVC: there is no such warning
*
* @code
* Y_PRAGMA_DIAGNOSTIC_PUSH
* Y_PRAGMA_NO_UNUSED_FUNCTION
*
* // some code which introduces a function which later will not be used, e.g.:
*
* void Foo() {
* }
*
* int main() {
* return 0; // Foo() never called
* }
*
* Y_PRAGMA_DIAGNOSTIC_POP
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA_NO_UNUSED_FUNCTION \
Y_PRAGMA("GCC diagnostic ignored \"-Wunused-function\"")
#else
#define Y_PRAGMA_NO_UNUSED_FUNCTION
#endif
/**
* @ def Y_PRAGMA_NO_UNUSED_PARAMETER
*
* Cross-compiler pragma to disable warnings about unused function parameters
*
* @see
* GCC: https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
* Clang: https://clang.llvm.org/docs/DiagnosticsReference.html#wunused-parameter
* MSVC: https://msdn.microsoft.com/en-us/library/26kb9fy0.aspx
*
* @code
* Y_PRAGMA_DIAGNOSTIC_PUSH
* Y_PRAGMA_NO_UNUSED_PARAMETER
*
* // some code which introduces a function with unused parameter, e.g.:
*
* void foo(int a) {
* // a is not referenced
* }
*
* int main() {
* foo(1);
* return 0;
* }
*
* Y_PRAGMA_DIAGNOSTIC_POP
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA_NO_UNUSED_PARAMETER \
Y_PRAGMA("GCC diagnostic ignored \"-Wunused-parameter\"")
#elif defined(_MSC_VER)
#define Y_PRAGMA_NO_UNUSED_PARAMETER \
Y_PRAGMA(warning(disable : 4100))
#else
#define Y_PRAGMA_NO_UNUSED_PARAMETER
#endif
/**
* @def Y_PRAGMA_NO_DEPRECATED
*
* Cross compiler pragma to disable warnings and errors about deprecated
*
* @see
* GCC: https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
* Clang: https://clang.llvm.org/docs/DiagnosticsReference.html#wdeprecated
* MSVC: https://docs.microsoft.com/en-us/cpp/error-messages/compiler-warnings/compiler-warning-level-3-c4996?view=vs-2017
*
* @code
* Y_PRAGMA_DIAGNOSTIC_PUSH
* Y_PRAGMA_NO_DEPRECATED
*
* [deprecated] void foo() {
* // ...
* }
*
* int main() {
* foo();
* return 0;
* }
*
* Y_PRAGMA_DIAGNOSTIC_POP
* @endcode
*/
#if defined(__clang__) || defined(__GNUC__)
#define Y_PRAGMA_NO_DEPRECATED \
Y_PRAGMA("GCC diagnostic ignored \"-Wdeprecated\"")
#elif defined(_MSC_VER)
#define Y_PRAGMA_NO_DEPRECATED \
Y_PRAGMA(warning(disable : 4996))
#else
#define Y_PRAGMA_NO_DEPRECATED
#endif
#if defined(__clang__) || defined(__GNUC__)
/**
* @def Y_CONST_FUNCTION
methods and functions, marked with this method are promised to:
1. do not have side effects
2. this method do not read global memory
NOTE: this attribute can't be set for methods that depend on data, pointed by this
this allow compilers to do hard optimization of that functions
NOTE: in common case this attribute can't be set if method have pointer-arguments
NOTE: as result there no any reason to discard result of such method
*/
#define Y_CONST_FUNCTION [[gnu::const]]
#endif
#if !defined(Y_CONST_FUNCTION)
#define Y_CONST_FUNCTION
#endif
#if defined(__clang__) || defined(__GNUC__)
/**
* @def Y_PURE_FUNCTION
methods and functions, marked with this method are promised to:
1. do not have side effects
2. result will be the same if no global memory changed
this allow compilers to do hard optimization of that functions
NOTE: as result there no any reason to discard result of such method
*/
#define Y_PURE_FUNCTION [[gnu::pure]]
#endif
#if !defined(Y_PURE_FUNCTION)
#define Y_PURE_FUNCTION
#endif
/**
* @ def Y_HAVE_INT128
*
* Defined when the compiler supports __int128 extension
*
* @code
*
* #if defined(Y_HAVE_INT128)
* __int128 myVeryBigInt = 12345678901234567890;
* #endif
*
* @endcode
*/
#if defined(__SIZEOF_INT128__)
#define Y_HAVE_INT128 1
#endif
/**
* XRAY macro must be passed to compiler if XRay is enabled.
*
* Define everything XRay-specific as a macro so that it doesn't cause errors
* for compilers that doesn't support XRay.
*/
#if defined(XRAY) && defined(__cplusplus)
#include <xray/xray_interface.h>
#define Y_XRAY_ALWAYS_INSTRUMENT [[clang::xray_always_instrument]]
#define Y_XRAY_NEVER_INSTRUMENT [[clang::xray_never_instrument]]
#define Y_XRAY_CUSTOM_EVENT(__string, __length) \
do { \
__xray_customevent(__string, __length); \
} while (0)
#else
#define Y_XRAY_ALWAYS_INSTRUMENT
#define Y_XRAY_NEVER_INSTRUMENT
#define Y_XRAY_CUSTOM_EVENT(__string, __length) \
do { \
} while (0)
#endif

168
contrib/lfalloc/src/util/system/defaults.h
View File

@ -1,168 +0,0 @@
#pragma once
#include "platform.h"
#if defined _unix_
#define LOCSLASH_C '/'
#define LOCSLASH_S "/"
#else
#define LOCSLASH_C '\\'
#define LOCSLASH_S "\\"
#endif // _unix_
#if defined(__INTEL_COMPILER) && defined(__cplusplus)
#include <new>
#endif
// low and high parts of integers
#if !defined(_win_)
#include <sys/param.h>
#endif
#if defined(BSD) || defined(_android_)
#if defined(BSD)
#include <machine/endian.h>
#endif
#if defined(_android_)
#include <endian.h>
#endif
#if (BYTE_ORDER == LITTLE_ENDIAN)
#define _little_endian_
#elif (BYTE_ORDER == BIG_ENDIAN)
#define _big_endian_
#else
#error unknown endian not supported
#endif
#elif (defined(_sun_) && !defined(__i386__)) || defined(_hpux_) || defined(WHATEVER_THAT_HAS_BIG_ENDIAN)
#define _big_endian_
#else
#define _little_endian_
#endif
// alignment
#if (defined(_sun_) && !defined(__i386__)) || defined(_hpux_) || defined(__alpha__) || defined(__ia64__) || defined(WHATEVER_THAT_NEEDS_ALIGNING_QUADS)
#define _must_align8_
#endif
#if (defined(_sun_) && !defined(__i386__)) || defined(_hpux_) || defined(__alpha__) || defined(__ia64__) || defined(WHATEVER_THAT_NEEDS_ALIGNING_LONGS)
#define _must_align4_
#endif
#if (defined(_sun_) && !defined(__i386__)) || defined(_hpux_) || defined(__alpha__) || defined(__ia64__) || defined(WHATEVER_THAT_NEEDS_ALIGNING_SHORTS)
#define _must_align2_
#endif
#if defined(__GNUC__)
#define alias_hack __attribute__((__may_alias__))
#endif
#ifndef alias_hack
#define alias_hack
#endif
#include "types.h"
#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L)
#define PRAGMA(x) _Pragma(#x)
#define RCSID(idstr) PRAGMA(comment(exestr, idstr))
#else
#define RCSID(idstr) static const char rcsid[] = idstr
#endif
#include "compiler.h"
#ifdef _win_
#include <malloc.h>
#elif defined(_sun_)
#include <alloca.h>
#endif
#ifdef NDEBUG
#define Y_IF_DEBUG(X)
#else
#define Y_IF_DEBUG(X) X
#endif
/**
* @def Y_ARRAY_SIZE
*
* This macro is needed to get number of elements in a statically allocated fixed size array. The
* expression is a compile-time constant and therefore can be used in compile time computations.
*
* @code
* enum ENumbers {
* EN_ONE,
* EN_TWO,
* EN_SIZE
* }
*
* const char* NAMES[] = {
* "one",
* "two"
* }
*
* static_assert(Y_ARRAY_SIZE(NAMES) == EN_SIZE, "you should define `NAME` for each enumeration");
* @endcode
*
* This macro also catches type errors. If you see a compiler error like "warning: division by zero
* is undefined" when using `Y_ARRAY_SIZE` then you are probably giving it a pointer.
*
* Since all of our code is expected to work on a 64 bit platform where pointers are 8 bytes we may
* falsefully accept pointers to types of sizes that are divisors of 8 (1, 2, 4 and 8).
*/
#if defined(__cplusplus)
namespace NArraySizePrivate {
template <class T>
struct TArraySize;
template <class T, size_t N>
struct TArraySize<T[N]> {
enum {
Result = N
};
};
template <class T, size_t N>
struct TArraySize<T (&)[N]> {
enum {
Result = N
};
};
}
#define Y_ARRAY_SIZE(arr) ((size_t)::NArraySizePrivate::TArraySize<decltype(arr)>::Result)
#else
#undef Y_ARRAY_SIZE
#define Y_ARRAY_SIZE(arr) \
((sizeof(arr) / sizeof((arr)[0])) / static_cast<size_t>(!(sizeof(arr) % sizeof((arr)[0]))))
#endif
#undef Y_ARRAY_BEGIN
#define Y_ARRAY_BEGIN(arr) (arr)
#undef Y_ARRAY_END
#define Y_ARRAY_END(arr) ((arr) + Y_ARRAY_SIZE(arr))
/**
* Concatenates two symbols, even if one of them is itself a macro.
*/
#define Y_CAT(X, Y) Y_CAT_I(X, Y)
#define Y_CAT_I(X, Y) Y_CAT_II(X, Y)
#define Y_CAT_II(X, Y) X##Y
#define Y_STRINGIZE(X) UTIL_PRIVATE_STRINGIZE_AUX(X)
#define UTIL_PRIVATE_STRINGIZE_AUX(X) #X
#if defined(__COUNTER__)
#define Y_GENERATE_UNIQUE_ID(N) Y_CAT(N, __COUNTER__)
#endif
#if !defined(Y_GENERATE_UNIQUE_ID)
#define Y_GENERATE_UNIQUE_ID(N) Y_CAT(N, __LINE__)
#endif
#define NPOS ((size_t)-1)

242
contrib/lfalloc/src/util/system/platform.h
View File

@ -1,242 +0,0 @@
#pragma once
// What OS ?
// our definition has the form _{osname}_
#if defined(_WIN64)
#define _win64_
#define _win32_
#elif defined(__WIN32__) || defined(_WIN32) // _WIN32 is also defined by the 64-bit compiler for backward compatibility
#define _win32_
#else
#define _unix_
#if defined(__sun__) || defined(sun) || defined(sparc) || defined(__sparc)
#define _sun_
#endif
#if defined(__hpux__)
#define _hpux_
#endif
#if defined(__linux__)
#define _linux_
#endif
#if defined(__FreeBSD__)
#define _freebsd_
#endif
#if defined(__CYGWIN__)
#define _cygwin_
#endif
#if defined(__APPLE__)
#define _darwin_
#endif
#if defined(__ANDROID__)
#define _android_
#endif
#endif
#if defined(__IOS__)
#define _ios_
#endif
#if defined(_linux_)
#if defined(_musl_)
//nothing to do
#elif defined(_android_)
#define _bionic_
#else
#define _glibc_
#endif
#endif
#if defined(_darwin_)
#define unix
#define __unix__
#endif
#if defined(_win32_) || defined(_win64_)
#define _win_
#endif
#if defined(__arm__) || defined(__ARM__) || defined(__ARM_NEON) || defined(__aarch64__) || defined(_M_ARM)
#if defined(__arm64) || defined(__arm64__) || defined(__aarch64__)
#define _arm64_
#else
#define _arm32_
#endif
#endif
#if defined(_arm64_) || defined(_arm32_)
#define _arm_
#endif
/* __ia64__ and __x86_64__ - defined by GNU C.
* _M_IA64, _M_X64, _M_AMD64 - defined by Visual Studio.
*
* Microsoft can define _M_IX86, _M_AMD64 (before Visual Studio 8)
* or _M_X64 (starting in Visual Studio 8).
*/
#if defined(__x86_64__) || defined(_M_X64) || defined(_M_AMD64)
#define _x86_64_
#endif
#if defined(__i386__) || defined(_M_IX86)
#define _i386_
#endif
#if defined(__ia64__) || defined(_M_IA64)
#define _ia64_
#endif
#if defined(__powerpc__)
#define _ppc_
#endif
#if defined(__powerpc64__)
#define _ppc64_
#endif
#if !defined(sparc) && !defined(__sparc) && !defined(__hpux__) && !defined(__alpha__) && !defined(_ia64_) && !defined(_x86_64_) && !defined(_arm_) && !defined(_i386_) && !defined(_ppc_) && !defined(_ppc64_)
#error "platform not defined, please, define one"
#endif
#if defined(_x86_64_) || defined(_i386_)
#define _x86_
#endif
#if defined(__MIC__)
#define _mic_
#define _k1om_
#endif
// stdio or MessageBox
#if defined(__CONSOLE__) || defined(_CONSOLE)
#define _console_
#endif
#if (defined(_win_) && !defined(_console_))
#define _windows_
#elif !defined(_console_)
#define _console_
#endif
#if defined(__SSE__) || defined(SSE_ENABLED)
#define _sse_
#endif
#if defined(__SSE2__) || defined(SSE2_ENABLED)
#define _sse2_
#endif
#if defined(__SSE3__) || defined(SSE3_ENABLED)
#define _sse3_
#endif
#if defined(__SSSE3__) || defined(SSSE3_ENABLED)
#define _ssse3_
#endif
#if defined(POPCNT_ENABLED)
#define _popcnt_
#endif
#if defined(__DLL__) || defined(_DLL)
#define _dll_
#endif
// 16, 32 or 64
#if defined(__sparc_v9__) || defined(_x86_64_) || defined(_ia64_) || defined(_arm64_) || defined(_ppc64_)
#define _64_
#else
#define _32_
#endif
/* All modern 64-bit Unix systems use scheme LP64 (long, pointers are 64-bit).
* Microsoft uses a different scheme: LLP64 (long long, pointers are 64-bit).
*
* Scheme LP64 LLP64
* char 8 8
* short 16 16
* int 32 32
* long 64 32
* long long 64 64
* pointer 64 64
*/
#if defined(_32_)
#define SIZEOF_PTR 4
#elif defined(_64_)
#define SIZEOF_PTR 8
#endif
#define PLATFORM_DATA_ALIGN SIZEOF_PTR
#if !defined(SIZEOF_PTR)
#error todo
#endif
#define SIZEOF_CHAR 1
#define SIZEOF_UNSIGNED_CHAR 1
#define SIZEOF_SHORT 2
#define SIZEOF_UNSIGNED_SHORT 2
#define SIZEOF_INT 4
#define SIZEOF_UNSIGNED_INT 4
#if defined(_32_)
#define SIZEOF_LONG 4
#define SIZEOF_UNSIGNED_LONG 4
#elif defined(_64_)
#if defined(_win_)
#define SIZEOF_LONG 4
#define SIZEOF_UNSIGNED_LONG 4
#else
#define SIZEOF_LONG 8
#define SIZEOF_UNSIGNED_LONG 8
#endif // _win_
#endif // _32_
#if !defined(SIZEOF_LONG)
#error todo
#endif
#define SIZEOF_LONG_LONG 8
#define SIZEOF_UNSIGNED_LONG_LONG 8
#undef SIZEOF_SIZE_T // in case we include <Python.h> which defines it, too
#define SIZEOF_SIZE_T SIZEOF_PTR
#if defined(__INTEL_COMPILER)
#pragma warning(disable 1292)
#pragma warning(disable 1469)
#pragma warning(disable 193)
#pragma warning(disable 271)
#pragma warning(disable 383)
#pragma warning(disable 424)
#pragma warning(disable 444)
#pragma warning(disable 584)
#pragma warning(disable 593)
#pragma warning(disable 981)
#pragma warning(disable 1418)
#pragma warning(disable 304)
#pragma warning(disable 810)
#pragma warning(disable 1029)
#pragma warning(disable 1419)
#pragma warning(disable 177)
#pragma warning(disable 522)
#pragma warning(disable 858)
#pragma warning(disable 111)
#pragma warning(disable 1599)
#pragma warning(disable 411)
#pragma warning(disable 304)
#pragma warning(disable 858)
#pragma warning(disable 444)
#pragma warning(disable 913)
#pragma warning(disable 310)
#pragma warning(disable 167)
#pragma warning(disable 180)
#pragma warning(disable 1572)
#endif
#if defined(_MSC_VER)
#undef _WINSOCKAPI_
#define _WINSOCKAPI_
#undef NOMINMAX
#define NOMINMAX
#endif

117
contrib/lfalloc/src/util/system/types.h
View File

@ -1,117 +0,0 @@
#pragma once
// DO_NOT_STYLE
#include "platform.h"
#include <inttypes.h>
typedef int8_t i8;
typedef int16_t i16;
typedef uint8_t ui8;
typedef uint16_t ui16;
typedef int yssize_t;
#define PRIYSZT "d"
#if defined(_darwin_) && defined(_32_)
typedef unsigned long ui32;
typedef long i32;
#else
typedef uint32_t ui32;
typedef int32_t i32;
#endif
#if defined(_darwin_) && defined(_64_)
typedef unsigned long ui64;
typedef long i64;
#else
typedef uint64_t ui64;
typedef int64_t i64;
#endif
#define LL(number) INT64_C(number)
#define ULL(number) UINT64_C(number)
// Macro for size_t and ptrdiff_t types
#if defined(_32_)
# if defined(_darwin_)
# define PRISZT "lu"
# undef PRIi32
# define PRIi32 "li"
# undef SCNi32
# define SCNi32 "li"
# undef PRId32
# define PRId32 "li"
# undef SCNd32
# define SCNd32 "li"
# undef PRIu32
# define PRIu32 "lu"
# undef SCNu32
# define SCNu32 "lu"
# undef PRIx32
# define PRIx32 "lx"
# undef SCNx32
# define SCNx32 "lx"
# elif !defined(_cygwin_)
# define PRISZT PRIu32
# else
# define PRISZT "u"
# endif
# define SCNSZT SCNu32
# define PRIPDT PRIi32
# define SCNPDT SCNi32
# define PRITMT PRIi32
# define SCNTMT SCNi32
#elif defined(_64_)
# if defined(_darwin_)
# define PRISZT "lu"
# undef PRIu64
# define PRIu64 PRISZT
# undef PRIx64
# define PRIx64 "lx"
# undef PRIX64
# define PRIX64 "lX"
# undef PRId64
# define PRId64 "ld"
# undef PRIi64
# define PRIi64 "li"
# undef SCNi64
# define SCNi64 "li"
# undef SCNu64
# define SCNu64 "lu"
# undef SCNx64
# define SCNx64 "lx"
# else
# define PRISZT PRIu64
# endif
# define SCNSZT SCNu64
# define PRIPDT PRIi64
# define SCNPDT SCNi64
# define PRITMT PRIi64
# define SCNTMT SCNi64
#else
# error "Unsupported platform"
#endif
// SUPERLONG
#if !defined(DONT_USE_SUPERLONG) && !defined(SUPERLONG_MAX)
#define SUPERLONG_MAX ~LL(0)
typedef i64 SUPERLONG;
#endif
// UNICODE
// UCS-2, native byteorder
typedef ui16 wchar16;
// internal symbol type: UTF-16LE
typedef wchar16 TChar;
typedef ui32 wchar32;
#if defined(_MSC_VER)
#include <basetsd.h>
typedef SSIZE_T ssize_t;
#define HAVE_SSIZE_T 1
#include <wchar.h>
#endif
#include <sys/types.h>

9
contrib/libhdfs3-cmake/CMake/Platform.cmake
View File

@ -15,9 +15,14 @@ IF(CMAKE_COMPILER_IS_GNUCXX)
STRING(REGEX MATCHALL "[0-9]+" GCC_COMPILER_VERSION ${GCC_COMPILER_VERSION})
LIST(LENGTH GCC_COMPILER_VERSION GCC_COMPILER_VERSION_LENGTH)
LIST(GET GCC_COMPILER_VERSION 0 GCC_COMPILER_VERSION_MAJOR)
LIST(GET GCC_COMPILER_VERSION 0 GCC_COMPILER_VERSION_MINOR)
if (GCC_COMPILER_VERSION_LENGTH GREATER 1)
LIST(GET GCC_COMPILER_VERSION 1 GCC_COMPILER_VERSION_MINOR)
else ()
set (GCC_COMPILER_VERSION_MINOR 0)
endif ()
SET(GCC_COMPILER_VERSION_MAJOR ${GCC_COMPILER_VERSION_MAJOR} CACHE INTERNAL "gcc major version")
SET(GCC_COMPILER_VERSION_MINOR ${GCC_COMPILER_VERSION_MINOR} CACHE INTERNAL "gcc minor version")

1
contrib/mimalloc vendored Submodule

@ -0,0 +1 @@
Subproject commit a787bdebce94bf3776dc0d1ad597917f479ab8d5

5
dbms/CMakeLists.txt
View File

@ -223,8 +223,9 @@ if(RE2_INCLUDE_DIR)
target_include_directories(clickhouse_common_io SYSTEM BEFORE PUBLIC ${RE2_INCLUDE_DIR})
endif()
if (USE_LFALLOC)
target_include_directories (clickhouse_common_io SYSTEM BEFORE PUBLIC ${LFALLOC_INCLUDE_DIR})
if (USE_MIMALLOC)
target_include_directories (clickhouse_common_io SYSTEM BEFORE PUBLIC ${MIMALLOC_INCLUDE_DIR})
target_link_libraries (clickhouse_common_io PRIVATE ${MIMALLOC_LIBRARY})
endif ()
if(CPUID_LIBRARY)

5
dbms/programs/client/readpassphrase/readpassphrase.h
View File

@ -29,6 +29,11 @@
//#include "includes.h"
#include "config_client.h"
// Should not be included on BSD systems, but if it happen...
#ifdef HAVE_READPASSPHRASE
# include_next <readpassphrase.h>
#endif
#ifndef HAVE_READPASSPHRASE
# ifdef __cplusplus

31
dbms/programs/copier/ClusterCopier.cpp
View File

@ -96,14 +96,19 @@ namespace
using DatabaseAndTableName = std::pair<String, String>;
String getDatabaseDotTable(const String & database, const String & table)
String getQuotedTable(const String & database, const String & table)
{
if (database.empty())
{
return backQuoteIfNeed(table);
}
return backQuoteIfNeed(database) + "." + backQuoteIfNeed(table);
}
String getDatabaseDotTable(const DatabaseAndTableName & db_and_table)
String getQuotedTable(const DatabaseAndTableName & db_and_table)
{
return getDatabaseDotTable(db_and_table.first, db_and_table.second);
return getQuotedTable(db_and_table.first, db_and_table.second);
}
@ -467,7 +472,7 @@ String DB::TaskShard::getDescription() const
std::stringstream ss;
ss << "N" << numberInCluster()
<< " (having a replica " << getHostNameExample()
<< ", pull table " + getDatabaseDotTable(task_table.table_pull)
<< ", pull table " + getQuotedTable(task_table.table_pull)
<< " of cluster " + task_table.cluster_pull_name << ")";
return ss.str();
}
@ -741,8 +746,10 @@ public:
{
auto zookeeper = context.getZooKeeper();
task_description_watch_callback = [this] (const Coordination::WatchResponse &)
task_description_watch_callback = [this] (const Coordination::WatchResponse & response)
{
if (response.error != Coordination::ZOK)
return;
UInt64 version = ++task_descprtion_version;
LOG_DEBUG(log, "Task description should be updated, local version " << version);
};
@ -1296,7 +1303,7 @@ protected:
/// Remove all status nodes
zookeeper->tryRemoveRecursive(current_shards_path);
String query = "ALTER TABLE " + getDatabaseDotTable(task_table.table_push);
String query = "ALTER TABLE " + getQuotedTable(task_table.table_push);
query += " DROP PARTITION " + task_partition.name + "";
/// TODO: use this statement after servers will be updated up to 1.1.54310
@ -1539,7 +1546,7 @@ protected:
auto get_select_query = [&] (const DatabaseAndTableName & from_table, const String & fields, String limit = "")
{
String query;
query += "SELECT " + fields + " FROM " + getDatabaseDotTable(from_table);
query += "SELECT " + fields + " FROM " + getQuotedTable(from_table);
/// TODO: Bad, it is better to rewrite with ASTLiteral(partition_key_field)
query += " WHERE (" + queryToString(task_table.engine_push_partition_key_ast) + " = (" + task_partition.name + " AS partition_key))";
if (!task_table.where_condition_str.empty())
@ -1677,7 +1684,7 @@ protected:
LOG_DEBUG(log, "Create destination tables. Query: " << query);
UInt64 shards = executeQueryOnCluster(task_table.cluster_push, query, create_query_push_ast, &task_cluster->settings_push,
PoolMode::GET_MANY);
LOG_DEBUG(log, "Destination tables " << getDatabaseDotTable(task_table.table_push) << " have been created on " << shards
LOG_DEBUG(log, "Destination tables " << getQuotedTable(task_table.table_push) << " have been created on " << shards
<< " shards of " << task_table.cluster_push->getShardCount());
}
@ -1699,7 +1706,7 @@ protected:
ASTPtr query_insert_ast;
{
String query;
query += "INSERT INTO " + getDatabaseDotTable(task_shard.table_split_shard) + " VALUES ";
query += "INSERT INTO " + getQuotedTable(task_shard.table_split_shard) + " VALUES ";
ParserQuery p_query(query.data() + query.size());
query_insert_ast = parseQuery(p_query, query, 0);
@ -1824,7 +1831,7 @@ protected:
String getRemoteCreateTable(const DatabaseAndTableName & table, Connection & connection, const Settings * settings = nullptr)
{
String query = "SHOW CREATE TABLE " + getDatabaseDotTable(table);
String query = "SHOW CREATE TABLE " + getQuotedTable(table);
Block block = getBlockWithAllStreamData(std::make_shared<RemoteBlockInputStream>(
connection, query, InterpreterShowCreateQuery::getSampleBlock(), context, settings));
@ -1887,7 +1894,7 @@ protected:
{
WriteBufferFromOwnString wb;
wb << "SELECT DISTINCT " << queryToString(task_table.engine_push_partition_key_ast) << " AS partition FROM"
<< " " << getDatabaseDotTable(task_shard.table_read_shard) << " ORDER BY partition DESC";
<< " " << getQuotedTable(task_shard.table_read_shard) << " ORDER BY partition DESC";
query = wb.str();
}
@ -1929,7 +1936,7 @@ protected:
{
WriteBufferFromOwnString wb;
wb << "SELECT 1"
<< " FROM "<< getDatabaseDotTable(task_shard.table_read_shard)
<< " FROM "<< getQuotedTable(task_shard.table_read_shard)
<< " WHERE " << queryToString(task_table.engine_push_partition_key_ast) << " = " << partition_quoted_name
<< " LIMIT 1";
query = wb.str();

6
dbms/src/AggregateFunctions/AggregateFunctionGroupUniqArray.cpp
View File

@ -43,11 +43,11 @@ static IAggregateFunction * createWithExtraTypes(const DataTypePtr & argument_ty
else if (which.idx == TypeIndex::DateTime) return new AggregateFunctionGroupUniqArrayDateTime<has_limit>(argument_type, std::forward<TArgs>(args)...);
else
{
/// Check that we can use plain version of AggreagteFunctionGroupUniqArrayGeneric
/// Check that we can use plain version of AggregateFunctionGroupUniqArrayGeneric
if (argument_type->isValueUnambiguouslyRepresentedInContiguousMemoryRegion())
return new AggreagteFunctionGroupUniqArrayGeneric<true, has_limit>(argument_type, std::forward<TArgs>(args)...);
return new AggregateFunctionGroupUniqArrayGeneric<true, has_limit>(argument_type, std::forward<TArgs>(args)...);
else
return new AggreagteFunctionGroupUniqArrayGeneric<false, has_limit>(argument_type, std::forward<TArgs>(args)...);
return new AggregateFunctionGroupUniqArrayGeneric<false, has_limit>(argument_type, std::forward<TArgs>(args)...);
}
}

14
dbms/src/AggregateFunctions/AggregateFunctionGroupUniqArray.h
View File

@ -122,7 +122,7 @@ public:
/// Generic implementation, it uses serialized representation as object descriptor.
struct AggreagteFunctionGroupUniqArrayGenericData
struct AggregateFunctionGroupUniqArrayGenericData
{
static constexpr size_t INIT_ELEMS = 2; /// adjustable
static constexpr size_t ELEM_SIZE = sizeof(HashSetCellWithSavedHash<StringRef, StringRefHash>);
@ -132,7 +132,7 @@ struct AggreagteFunctionGroupUniqArrayGenericData
};
/// Helper function for deserialize and insert for the class AggreagteFunctionGroupUniqArrayGeneric
/// Helper function for deserialize and insert for the class AggregateFunctionGroupUniqArrayGeneric
template <bool is_plain_column>
static StringRef getSerializationImpl(const IColumn & column, size_t row_num, Arena & arena);
@ -143,15 +143,15 @@ static void deserializeAndInsertImpl(StringRef str, IColumn & data_to);
* For such columns groupUniqArray() can be implemented more efficiently (especially for small numeric arrays).
*/
template <bool is_plain_column = false, typename Tlimit_num_elem = std::false_type>
class AggreagteFunctionGroupUniqArrayGeneric
: public IAggregateFunctionDataHelper<AggreagteFunctionGroupUniqArrayGenericData, AggreagteFunctionGroupUniqArrayGeneric<is_plain_column, Tlimit_num_elem>>
class AggregateFunctionGroupUniqArrayGeneric
: public IAggregateFunctionDataHelper<AggregateFunctionGroupUniqArrayGenericData, AggregateFunctionGroupUniqArrayGeneric<is_plain_column, Tlimit_num_elem>>
{
DataTypePtr & input_data_type;
static constexpr bool limit_num_elems = Tlimit_num_elem::value;
UInt64 max_elems;
using State = AggreagteFunctionGroupUniqArrayGenericData;
using State = AggregateFunctionGroupUniqArrayGenericData;
static StringRef getSerialization(const IColumn & column, size_t row_num, Arena & arena)
{
@ -164,8 +164,8 @@ class AggreagteFunctionGroupUniqArrayGeneric
}
public:
AggreagteFunctionGroupUniqArrayGeneric(const DataTypePtr & input_data_type, UInt64 max_elems_ = std::numeric_limits<UInt64>::max())
: IAggregateFunctionDataHelper<AggreagteFunctionGroupUniqArrayGenericData, AggreagteFunctionGroupUniqArrayGeneric<is_plain_column, Tlimit_num_elem>>({input_data_type}, {})
AggregateFunctionGroupUniqArrayGeneric(const DataTypePtr & input_data_type, UInt64 max_elems_ = std::numeric_limits<UInt64>::max())
: IAggregateFunctionDataHelper<AggregateFunctionGroupUniqArrayGenericData, AggregateFunctionGroupUniqArrayGeneric<is_plain_column, Tlimit_num_elem>>({input_data_type}, {})
, input_data_type(this->argument_types[0])
, max_elems(max_elems_) {}

6
dbms/src/AggregateFunctions/AggregateFunctionSequenceMatch.h
View File

@ -47,8 +47,7 @@ struct AggregateFunctionSequenceMatchData final
using Comparator = ComparePairFirst<std::less>;
bool sorted = true;
static constexpr size_t bytes_in_arena = 64;
PODArray<TimestampEvents, bytes_in_arena, AllocatorWithStackMemory<Allocator<false>, bytes_in_arena>> events_list;
PODArrayWithStackMemory<TimestampEvents, 64> events_list;
void add(const Timestamp timestamp, const Events & events)
{
@ -203,8 +202,7 @@ private:
PatternAction(const PatternActionType type, const std::uint64_t extra = 0) : type{type}, extra{extra} {}
};
static constexpr size_t bytes_on_stack = 64;
using PatternActions = PODArray<PatternAction, bytes_on_stack, AllocatorWithStackMemory<Allocator<false>, bytes_on_stack>>;
using PatternActions = PODArrayWithStackMemory<PatternAction, 64>;
Derived & derived() { return static_cast<Derived &>(*this); }

3
dbms/src/AggregateFunctions/AggregateFunctionTimeSeriesGroupSum.h
View File

@ -68,9 +68,8 @@ struct AggregateFunctionTimeSeriesGroupSumData
}
};
static constexpr size_t bytes_on_stack = 128;
typedef std::map<UInt64, Points> Series;
typedef PODArray<DataPoint, bytes_on_stack, AllocatorWithStackMemory<Allocator<false>, bytes_on_stack>> AggSeries;
typedef PODArrayWithStackMemory<DataPoint, 128> AggSeries;
Series ss;
AggSeries result;

5
dbms/src/AggregateFunctions/AggregateFunctionWindowFunnel.h
View File

@ -35,10 +35,7 @@ template <typename T>
struct AggregateFunctionWindowFunnelData
{
using TimestampEvent = std::pair<T, UInt8>;
static constexpr size_t bytes_on_stack = 64;
using TimestampEvents = PODArray<TimestampEvent, bytes_on_stack, AllocatorWithStackMemory<Allocator<false>, bytes_on_stack>>;
using TimestampEvents = PODArray<TimestampEvent, 64>;
using Comparator = ComparePairFirst;
bool sorted = true;

3
dbms/src/AggregateFunctions/QuantileExact.h
View File

@ -27,8 +27,7 @@ struct QuantileExact
{
/// The memory will be allocated to several elements at once, so that the state occupies 64 bytes.
static constexpr size_t bytes_in_arena = 64 - sizeof(PODArray<Value>);
using Array = PODArray<Value, bytes_in_arena, AllocatorWithStackMemory<Allocator<false>, bytes_in_arena>>;
using Array = PODArrayWithStackMemory<Value, bytes_in_arena>;
Array array;
void add(const Value & x)

3
dbms/src/AggregateFunctions/QuantileTDigest.h
View File

@ -86,8 +86,7 @@ class QuantileTDigest
/// The memory will be allocated to several elements at once, so that the state occupies 64 bytes.
static constexpr size_t bytes_in_arena = 128 - sizeof(PODArray<Centroid>) - sizeof(Count) - sizeof(UInt32);
using Summary = PODArray<Centroid, bytes_in_arena / sizeof(Centroid), AllocatorWithStackMemory<Allocator<false>, bytes_in_arena>>;
using Summary = PODArrayWithStackMemory<Centroid, bytes_in_arena>;
Summary summary;
Count count = 0;

3
dbms/src/AggregateFunctions/ReservoirSampler.h
View File

@ -194,8 +194,7 @@ private:
friend void rs_perf_test();
/// We allocate a little memory on the stack - to avoid allocations when there are many objects with a small number of elements.
static constexpr size_t bytes_on_stack = 64;
using Array = DB::PODArray<T, bytes_on_stack / sizeof(T), AllocatorWithStackMemory<Allocator<false>, bytes_on_stack>>;
using Array = DB::PODArrayWithStackMemory<T, 64>;
size_t sample_count;
size_t total_values = 0;

3
dbms/src/AggregateFunctions/ReservoirSamplerDeterministic.h
View File

@ -164,9 +164,8 @@ public:
private:
/// We allocate some memory on the stack to avoid allocations when there are many objects with a small number of elements.
static constexpr size_t bytes_on_stack = 64;
using Element = std::pair<T, UInt32>;
using Array = DB::PODArray<Element, bytes_on_stack / sizeof(Element), AllocatorWithStackMemory<Allocator<false>, bytes_on_stack>>;
using Array = DB::PODArray<Element, 64>;
size_t sample_count;
size_t total_values{};

6
dbms/src/AggregateFunctions/registerAggregateFunctions.cpp
View File

@ -28,6 +28,9 @@ void registerAggregateFunctionTopK(AggregateFunctionFactory &);
void registerAggregateFunctionsBitwise(AggregateFunctionFactory &);
void registerAggregateFunctionsBitmap(AggregateFunctionFactory &);
void registerAggregateFunctionsMaxIntersections(AggregateFunctionFactory &);
void registerAggregateFunctionHistogram(AggregateFunctionFactory &);
void registerAggregateFunctionRetention(AggregateFunctionFactory &);
void registerAggregateFunctionTimeSeriesGroupSum(AggregateFunctionFactory &);
void registerAggregateFunctionMLMethod(AggregateFunctionFactory &);
void registerAggregateFunctionEntropy(AggregateFunctionFactory &);
void registerAggregateFunctionSimpleLinearRegression(AggregateFunctionFactory &);
@ -41,9 +44,6 @@ void registerAggregateFunctionCombinatorMerge(AggregateFunctionCombinatorFactory
void registerAggregateFunctionCombinatorNull(AggregateFunctionCombinatorFactory &);
void registerAggregateFunctionCombinatorResample(AggregateFunctionCombinatorFactory &);
void registerAggregateFunctionHistogram(AggregateFunctionFactory & factory);
void registerAggregateFunctionRetention(AggregateFunctionFactory & factory);
void registerAggregateFunctionTimeSeriesGroupSum(AggregateFunctionFactory & factory);
void registerAggregateFunctions()
{
{

2
dbms/src/Columns/ColumnVector.cpp
View File

@ -33,7 +33,7 @@ template <typename T>
StringRef ColumnVector<T>::serializeValueIntoArena(size_t n, Arena & arena, char const *& begin) const
{
auto pos = arena.allocContinue(sizeof(T), begin);
unalignedStore(pos, data[n]);
unalignedStore<T>(pos, data[n]);
return StringRef(pos, sizeof(T));
}

53
dbms/src/Common/LFAllocator.cpp
View File

@ -1,53 +0,0 @@
#include <Common/config.h>
#if USE_LFALLOC
#include "LFAllocator.h"
#include <cstring>
#include <lf_allocX64.h>
namespace DB
{
void * LFAllocator::alloc(size_t size, size_t alignment)
{
if (alignment == 0)
return LFAlloc(size);
else
{
void * ptr;
int res = LFPosixMemalign(&ptr, alignment, size);
return res ? nullptr : ptr;
}
}
void LFAllocator::free(void * buf, size_t)
{
LFFree(buf);
}
void * LFAllocator::realloc(void * old_ptr, size_t, size_t new_size, size_t alignment)
{
if (old_ptr == nullptr)
{
void * result = LFAllocator::alloc(new_size, alignment);
return result;
}
if (new_size == 0)
{
LFFree(old_ptr);
return nullptr;
}
void * new_ptr = LFAllocator::alloc(new_size, alignment);
if (new_ptr == nullptr)
return nullptr;
size_t old_size = LFGetSize(old_ptr);
memcpy(new_ptr, old_ptr, ((old_size < new_size) ? old_size : new_size));
LFFree(old_ptr);
return new_ptr;
}
}
#endif

22
dbms/src/Common/LFAllocator.h
View File

@ -1,22 +0,0 @@
#pragma once
#include <Common/config.h>
#if !USE_LFALLOC
#error "do not include this file until USE_LFALLOC is set to 1"
#endif
#include <cstddef>
namespace DB
{
struct LFAllocator
{
static void * alloc(size_t size, size_t alignment = 0);
static void free(void * buf, size_t);
static void * realloc(void * buf, size_t, size_t new_size, size_t alignment = 0);
};
}

43
dbms/src/Common/MiAllocator.cpp Normal file
View File

@ -0,0 +1,43 @@
#include <Common/config.h>
#if USE_MIMALLOC
#include "MiAllocator.h"
#include <mimalloc.h>
namespace DB
{
void * MiAllocator::alloc(size_t size, size_t alignment)
{
if (alignment == 0)
return mi_malloc(size);
else
return mi_malloc_aligned(size, alignment);
}
void MiAllocator::free(void * buf, size_t)
{
mi_free(buf);
}
void * MiAllocator::realloc(void * old_ptr, size_t, size_t new_size, size_t alignment)
{
if (old_ptr == nullptr)
return alloc(new_size, alignment);
if (new_size == 0)
{
mi_free(old_ptr);
return nullptr;
}
if (alignment == 0)
return mi_realloc(old_ptr, alignment);
return mi_realloc_aligned(old_ptr, new_size, alignment);
}
}
#endif

28
dbms/src/Common/MiAllocator.h Normal file
View File

@ -0,0 +1,28 @@
#pragma once
#include <Common/config.h>
#if !USE_MIMALLOC
#error "do not include this file until USE_MIMALLOC is set to 1"
#endif
#include <cstddef>
namespace DB
{
/*
* This is a different allocator that is based on mimalloc (Microsoft malloc).
* It can be used separately from main allocator to catch heap corruptions and vulnerabilities (for example, for caches).
* We use MI_SECURE mode in mimalloc to achieve such behaviour.
*/
struct MiAllocator
{
static void * alloc(size_t size, size_t alignment = 0);
static void free(void * buf, size_t);
static void * realloc(void * old_ptr, size_t, size_t new_size, size_t alignment = 0);
};
}

31
dbms/src/Common/PODArray.h
View File

@ -45,7 +45,7 @@ inline constexpr size_t integerRoundUp(size_t value, size_t dividend)
* Only part of the std::vector interface is supported.
*
* The default constructor creates an empty object that does not allocate memory.
* Then the memory is allocated at least INITIAL_SIZE bytes.
* Then the memory is allocated at least initial_bytes bytes.
*
* If you insert elements with push_back, without making a `reserve`, then PODArray is about 2.5 times faster than std::vector.
*
@ -74,7 +74,7 @@ extern const char EmptyPODArray[EmptyPODArraySize];
/** Base class that depend only on size of element, not on element itself.
* You can static_cast to this class if you want to insert some data regardless to the actual type T.
*/
template <size_t ELEMENT_SIZE, size_t INITIAL_SIZE, typename TAllocator, size_t pad_right_, size_t pad_left_>
template <size_t ELEMENT_SIZE, size_t initial_bytes, typename TAllocator, size_t pad_right_, size_t pad_left_>
class PODArrayBase : private boost::noncopyable, private TAllocator /// empty base optimization
{
protected:
@ -161,7 +161,8 @@ protected:
{
// The allocated memory should be multiplication of ELEMENT_SIZE to hold the element, otherwise,
// memory issue such as corruption could appear in edge case.
realloc(std::max(((INITIAL_SIZE - 1) / ELEMENT_SIZE + 1) * ELEMENT_SIZE, minimum_memory_for_elements(1)),
realloc(std::max(integerRoundUp(initial_bytes, ELEMENT_SIZE),
minimum_memory_for_elements(1)),
std::forward<TAllocatorParams>(allocator_params)...);
}
else
@ -257,11 +258,11 @@ public:
}
};
template <typename T, size_t INITIAL_SIZE = 4096, typename TAllocator = Allocator<false>, size_t pad_right_ = 0, size_t pad_left_ = 0>
class PODArray : public PODArrayBase<sizeof(T), INITIAL_SIZE, TAllocator, pad_right_, pad_left_>
template <typename T, size_t initial_bytes = 4096, typename TAllocator = Allocator<false>, size_t pad_right_ = 0, size_t pad_left_ = 0>
class PODArray : public PODArrayBase<sizeof(T), initial_bytes, TAllocator, pad_right_, pad_left_>
{
protected:
using Base = PODArrayBase<sizeof(T), INITIAL_SIZE, TAllocator, pad_right_, pad_left_>;
using Base = PODArrayBase<sizeof(T), initial_bytes, TAllocator, pad_right_, pad_left_>;
T * t_start() { return reinterpret_cast<T *>(this->c_start); }
T * t_end() { return reinterpret_cast<T *>(this->c_end); }
@ -618,17 +619,23 @@ public:
}
};
template <typename T, size_t INITIAL_SIZE, typename TAllocator, size_t pad_right_>
void swap(PODArray<T, INITIAL_SIZE, TAllocator, pad_right_> & lhs, PODArray<T, INITIAL_SIZE, TAllocator, pad_right_> & rhs)
template <typename T, size_t initial_bytes, typename TAllocator, size_t pad_right_>
void swap(PODArray<T, initial_bytes, TAllocator, pad_right_> & lhs, PODArray<T, initial_bytes, TAllocator, pad_right_> & rhs)
{
lhs.swap(rhs);
}
/** For columns. Padding is enough to read and write xmm-register at the address of the last element. */
template <typename T, size_t INITIAL_SIZE = 4096, typename TAllocator = Allocator<false>>
using PaddedPODArray = PODArray<T, INITIAL_SIZE, TAllocator, 15, 16>;
template <typename T, size_t initial_bytes = 4096, typename TAllocator = Allocator<false>>
using PaddedPODArray = PODArray<T, initial_bytes, TAllocator, 15, 16>;
template <typename T, size_t stack_size_in_bytes>
using PODArrayWithStackMemory = PODArray<T, 0, AllocatorWithStackMemory<Allocator<false>, integerRoundUp(stack_size_in_bytes, sizeof(T))>>;
/** A helper for declaring PODArray that uses inline memory.
* The initial size is set to use all the inline bytes, since using less would
* only add some extra allocation calls.
*/
template <typename T, size_t inline_bytes,
size_t rounded_bytes = integerRoundUp(inline_bytes, sizeof(T))>
using PODArrayWithStackMemory = PODArray<T, rounded_bytes,
AllocatorWithStackMemory<Allocator<false>, rounded_bytes>>;
}

10
dbms/src/Common/ThreadPool.cpp
View File

@ -30,10 +30,18 @@ template <typename Thread>
template <typename ReturnType>
ReturnType ThreadPoolImpl<Thread>::scheduleImpl(Job job, int priority, std::optional<uint64_t> wait_microseconds)
{
auto on_error = []
auto on_error = [&]
{
if constexpr (std::is_same_v<ReturnType, void>)
{
if (first_exception)
{
std::exception_ptr exception;
std::swap(exception, first_exception);
std::rethrow_exception(exception);
}
throw DB::Exception("Cannot schedule a task", DB::ErrorCodes::CANNOT_SCHEDULE_TASK);
}
else
return false;
};

3
dbms/src/Common/config.h.in
View File

@ -8,7 +8,6 @@
#cmakedefine01 USE_CPUID
#cmakedefine01 USE_CPUINFO
#cmakedefine01 USE_BROTLI
#cmakedefine01 USE_LFALLOC
#cmakedefine01 USE_LFALLOC_RANDOM_HINT
#cmakedefine01 USE_MIMALLOC
#cmakedefine01 CLICKHOUSE_SPLIT_BINARY

5
dbms/src/Common/formatIPv6.cpp
View File

@ -10,7 +10,8 @@ namespace DB
{
// To be used in formatIPv4, maps a byte to it's string form prefixed with length (so save strlen call).
extern const char one_byte_to_string_lookup_table[256][4] = {
extern const char one_byte_to_string_lookup_table[256][4] =
{
{1, '0'}, {1, '1'}, {1, '2'}, {1, '3'}, {1, '4'}, {1, '5'}, {1, '6'}, {1, '7'}, {1, '8'}, {1, '9'},
{2, '1', '0'}, {2, '1', '1'}, {2, '1', '2'}, {2, '1', '3'}, {2, '1', '4'}, {2, '1', '5'}, {2, '1', '6'}, {2, '1', '7'}, {2, '1', '8'}, {2, '1', '9'},
{2, '2', '0'}, {2, '2', '1'}, {2, '2', '2'}, {2, '2', '3'}, {2, '2', '4'}, {2, '2', '5'}, {2, '2', '6'}, {2, '2', '7'}, {2, '2', '8'}, {2, '2', '9'},
@ -152,7 +153,7 @@ void formatIPv6(const unsigned char * src, char *& dst, UInt8 zeroed_tail_bytes_
}
/// Was it a trailing run of 0x00's?
if (best.base != -1 && (best.base + best.len) == words.size())
if (best.base != -1 && size_t(best.base + best.len) == words.size())
*dst++ = ':';
*dst++ = '\0';

12
dbms/src/Common/tests/CMakeLists.txt
View File

@ -41,9 +41,6 @@ target_link_libraries (compact_array PRIVATE clickhouse_common_io ${Boost_FILESY
add_executable (radix_sort radix_sort.cpp)
target_link_libraries (radix_sort PRIVATE clickhouse_common_io)
add_executable (shell_command_test shell_command_test.cpp)
target_link_libraries (shell_command_test PRIVATE clickhouse_common_io)
add_executable (arena_with_free_lists arena_with_free_lists.cpp)
target_link_libraries (arena_with_free_lists PRIVATE clickhouse_compression clickhouse_common_io)
@ -53,15 +50,6 @@ target_link_libraries (pod_array PRIVATE clickhouse_common_io)
add_executable (thread_creation_latency thread_creation_latency.cpp)
target_link_libraries (thread_creation_latency PRIVATE clickhouse_common_io)
add_executable (thread_pool thread_pool.cpp)
target_link_libraries (thread_pool PRIVATE clickhouse_common_io)
add_executable (thread_pool_2 thread_pool_2.cpp)
target_link_libraries (thread_pool_2 PRIVATE clickhouse_common_io)
add_executable (thread_pool_3 thread_pool_3.cpp)
target_link_libraries (thread_pool_3 PRIVATE clickhouse_common_io)
add_executable (multi_version multi_version.cpp)
target_link_libraries (multi_version PRIVATE clickhouse_common_io)
add_check(multi_version)

72
dbms/src/Common/tests/gtest_shell_command.cpp Normal file
View File

@ -0,0 +1,72 @@
#include <iostream>
#include <Core/Types.h>
#include <Common/ShellCommand.h>
#include <IO/copyData.h>
#include <IO/WriteBufferFromFileDescriptor.h>
#include <IO/ReadBufferFromString.h>
#include <IO/ReadHelpers.h>
#include <chrono>
#include <thread>
#pragma GCC diagnostic ignored "-Wsign-compare"
#ifdef __clang__
#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant"
#pragma clang diagnostic ignored "-Wundef"
#endif
#include <gtest/gtest.h>
using namespace DB;
TEST(ShellCommand, Execute)
{
auto command = ShellCommand::execute("echo 'Hello, world!'");
std::string res;
readStringUntilEOF(res, command->out);
command->wait();
EXPECT_EQ(res, "Hello, world!\n");
}
TEST(ShellCommand, ExecuteDirect)
{
auto command = ShellCommand::executeDirect("/bin/echo", {"Hello, world!"});
std::string res;
readStringUntilEOF(res, command->out);
command->wait();
EXPECT_EQ(res, "Hello, world!\n");
}
TEST(ShellCommand, ExecuteWithInput)
{
auto command = ShellCommand::execute("cat");
String in_str = "Hello, world!\n";
ReadBufferFromString in(in_str);
copyData(in, command->in);
command->in.close();
std::string res;
readStringUntilEOF(res, command->out);
command->wait();
EXPECT_EQ(res, "Hello, world!\n");
}
TEST(ShellCommand, AutoWait)
{
// <defunct> hunting:
for (int i = 0; i < 1000; ++i)
{
auto command = ShellCommand::execute("echo " + std::to_string(i));
//command->wait(); // now automatic
}
// std::cerr << "inspect me: ps auxwwf" << "\n";
// std::this_thread::sleep_for(std::chrono::seconds(100));
}

11
dbms/src/Common/tests/thread_pool.cpp → dbms/src/Common/tests/gtest_thread_pool_concurrent_wait.cpp
View File

@ -1,11 +1,18 @@
#include <Common/ThreadPool.h>
#pragma GCC diagnostic ignored "-Wsign-compare"
#ifdef __clang__
#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant"
#pragma clang diagnostic ignored "-Wundef"
#endif
#include <gtest/gtest.h>
/** Reproduces bug in ThreadPool.
* It get stuck if we call 'wait' many times from many other threads simultaneously.
*/
int main(int, char **)
TEST(ThreadPool, ConcurrentWait)
{
auto worker = []
{
@ -29,6 +36,4 @@ int main(int, char **)
waiting_pool.schedule([&pool]{ pool.wait(); });
waiting_pool.wait();
return 0;
}

32
dbms/src/Common/tests/gtest_thread_pool_limit.cpp Normal file
View File

@ -0,0 +1,32 @@
#include <atomic>
#include <iostream>
#include <Common/ThreadPool.h>
#pragma GCC diagnostic ignored "-Wsign-compare"
#ifdef __clang__
#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant"
#pragma clang diagnostic ignored "-Wundef"
#endif
#include <gtest/gtest.h>
/// Test for thread self-removal when number of free threads in pool is too large.
/// Just checks that nothing weird happens.
template <typename Pool>
int test()
{
Pool pool(10, 2, 10);
std::atomic<int> counter{0};
for (size_t i = 0; i < 10; ++i)
pool.schedule([&]{ ++counter; });
pool.wait();
return counter;
}
TEST(ThreadPool, ThreadRemoval)
{
EXPECT_EQ(test<FreeThreadPool>(), 10);
EXPECT_EQ(test<ThreadPool>(), 10);
}

14
dbms/src/Common/tests/thread_pool_2.cpp → dbms/src/Common/tests/gtest_thread_pool_loop.cpp
View File

@ -2,10 +2,17 @@
#include <iostream>
#include <Common/ThreadPool.h>
#pragma GCC diagnostic ignored "-Wsign-compare"
#ifdef __clang__
#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant"
#pragma clang diagnostic ignored "-Wundef"
#endif
#include <gtest/gtest.h>
int main(int, char **)
TEST(ThreadPool, Loop)
{
std::atomic<size_t> res{0};
std::atomic<int> res{0};
for (size_t i = 0; i < 1000; ++i)
{
@ -16,6 +23,5 @@ int main(int, char **)
pool.wait();
}
std::cerr << res << "\n";
return 0;
EXPECT_EQ(res, 16000);
}

38
dbms/src/Common/tests/gtest_thread_pool_schedule_exception.cpp Normal file
View File

@ -0,0 +1,38 @@
#include <iostream>
#include <stdexcept>
#include <Common/ThreadPool.h>
#pragma GCC diagnostic ignored "-Wsign-compare"
#ifdef __clang__
#pragma clang diagnostic ignored "-Wzero-as-null-pointer-constant"
#pragma clang diagnostic ignored "-Wundef"
#endif
#include <gtest/gtest.h>
bool check()
{
ThreadPool pool(10);
pool.schedule([]{ throw std::runtime_error("Hello, world!"); });
try
{
for (size_t i = 0; i < 100; ++i)
pool.schedule([]{}); /// An exception will be rethrown from this method.
}
catch (const std::runtime_error &)
{
return true;
}
pool.wait();
return false;
}
TEST(ThreadPool, ExceptionFromSchedule)
{
EXPECT_TRUE(check());
}

63
dbms/src/Common/tests/shell_command_test.cpp
View File

@ -1,63 +0,0 @@
#include <iostream>
#include <Core/Types.h>
#include <Common/ShellCommand.h>
#include <IO/copyData.h>
#include <IO/WriteBufferFromFileDescriptor.h>
#include <IO/ReadBufferFromString.h>
#include <chrono>
#include <thread>
using namespace DB;
int main(int, char **)
try
{
{
auto command = ShellCommand::execute("echo 'Hello, world!'");
WriteBufferFromFileDescriptor out(STDOUT_FILENO);
copyData(command->out, out);
command->wait();
}
{
auto command = ShellCommand::executeDirect("/bin/echo", {"Hello, world!"});
WriteBufferFromFileDescriptor out(STDOUT_FILENO);
copyData(command->out, out);
command->wait();
}
{
auto command = ShellCommand::execute("cat");
String in_str = "Hello, world!\n";
ReadBufferFromString in(in_str);
copyData(in, command->in);
command->in.close();
WriteBufferFromFileDescriptor out(STDOUT_FILENO);
copyData(command->out, out);
command->wait();
}
// <defunct> hunting:
for (int i = 0; i < 1000; ++i)
{
auto command = ShellCommand::execute("echo " + std::to_string(i));
//command->wait(); // now automatic
}
// std::cerr << "inspect me: ps auxwwf" << "\n";
// std::this_thread::sleep_for(std::chrono::seconds(100));
}
catch (...)
{
std::cerr << getCurrentExceptionMessage(false) << "\n";
return 1;
}

27
dbms/src/Common/tests/thread_pool_3.cpp
View File

@ -1,27 +0,0 @@
#include <mutex>
#include <iostream>
#include <Common/ThreadPool.h>
/// Test for thread self-removal when number of free threads in pool is too large.
/// Just checks that nothing weird happens.
template <typename Pool>
void test()
{
Pool pool(10, 2, 10);
std::mutex mutex;
for (size_t i = 0; i < 10; ++i)
pool.schedule([&]{ std::lock_guard lock(mutex); std::cerr << '.'; });
pool.wait();
}
int main(int, char **)
{
test<FreeThreadPool>();
std::cerr << '\n';
test<ThreadPool>();
std::cerr << '\n';
return 0;
}

4
dbms/src/Compression/CompressionCodecDelta.cpp
View File

@ -48,7 +48,7 @@ void compressDataForType(const char * source, UInt32 source_size, char * dest)
while (source < source_end)
{
T curr_src = unalignedLoad<T>(source);
unalignedStore(dest, curr_src - prev_src);
unalignedStore<T>(dest, curr_src - prev_src);
prev_src = curr_src;
source += sizeof(T);
@ -67,7 +67,7 @@ void decompressDataForType(const char * source, UInt32 source_size, char * dest)
while (source < source_end)
{
accumulator += unalignedLoad<T>(source);
unalignedStore(dest, accumulator);
unalignedStore<T>(dest, accumulator);
source += sizeof(T);
dest += sizeof(T);

12
dbms/src/Compression/CompressionCodecDoubleDelta.cpp
View File

@ -90,7 +90,7 @@ UInt32 compressDataForType(const char * source, UInt32 source_size, char * dest)
const char * source_end = source + source_size;
const UInt32 items_count = source_size / sizeof(T);
unalignedStore(dest, items_count);
unalignedStore<UInt32>(dest, items_count);
dest += sizeof(items_count);
T prev_value{};
@ -99,7 +99,7 @@ UInt32 compressDataForType(const char * source, UInt32 source_size, char * dest)
if (source < source_end)
{
prev_value = unalignedLoad<T>(source);
unalignedStore(dest, prev_value);
unalignedStore<T>(dest, prev_value);
source += sizeof(prev_value);
dest += sizeof(prev_value);
@ -109,7 +109,7 @@ UInt32 compressDataForType(const char * source, UInt32 source_size, char * dest)
{
const T curr_value = unalignedLoad<T>(source);
prev_delta = static_cast<DeltaType>(curr_value - prev_value);
unalignedStore(dest, prev_delta);
unalignedStore<T>(dest, prev_delta);
source += sizeof(curr_value);
dest += sizeof(prev_delta);
@ -164,7 +164,7 @@ void decompressDataForType(const char * source, UInt32 source_size, char * dest)
if (source < source_end)
{
prev_value = unalignedLoad<T>(source);
unalignedStore(dest, prev_value);
unalignedStore<T>(dest, prev_value);
source += sizeof(prev_value);
dest += sizeof(prev_value);
@ -174,7 +174,7 @@ void decompressDataForType(const char * source, UInt32 source_size, char * dest)
{
prev_delta = unalignedLoad<DeltaType>(source);
prev_value = static_cast<T>(prev_value + prev_delta);
unalignedStore(dest, prev_value);
unalignedStore<T>(dest, prev_value);
source += sizeof(prev_delta);
dest += sizeof(prev_value);
@ -209,7 +209,7 @@ void decompressDataForType(const char * source, UInt32 source_size, char * dest)
// else if first bit is zero, no need to read more data.
const T curr_value = static_cast<T>(prev_value + prev_delta + double_delta);
unalignedStore(dest, curr_value);
unalignedStore<T>(dest, curr_value);
dest += sizeof(curr_value);
prev_delta = curr_value - prev_value;

8
dbms/src/Compression/CompressionCodecGorilla.cpp
View File

@ -94,7 +94,7 @@ UInt32 compressDataForType(const char * source, UInt32 source_size, char * dest)
const UInt32 items_count = source_size / sizeof(T);
unalignedStore(dest, items_count);
unalignedStore<UInt32>(dest, items_count);
dest += sizeof(items_count);
T prev_value{};
@ -104,7 +104,7 @@ UInt32 compressDataForType(const char * source, UInt32 source_size, char * dest)
if (source < source_end)
{
prev_value = unalignedLoad<T>(source);
unalignedStore(dest, prev_value);
unalignedStore<T>(dest, prev_value);
source += sizeof(prev_value);
dest += sizeof(prev_value);
@ -166,7 +166,7 @@ void decompressDataForType(const char * source, UInt32 source_size, char * dest)
if (source < source_end)
{
prev_value = unalignedLoad<T>(source);
unalignedStore(dest, prev_value);
unalignedStore<T>(dest, prev_value);
source += sizeof(prev_value);
dest += sizeof(prev_value);
@ -210,7 +210,7 @@ void decompressDataForType(const char * source, UInt32 source_size, char * dest)
}
// else: 0b0 prefix - use prev_value
unalignedStore(dest, curr_value);
unalignedStore<T>(dest, curr_value);
dest += sizeof(curr_value);
prev_xored_info = curr_xored_info;

2
dbms/src/Compression/CompressionCodecT64.cpp
View File

@ -390,7 +390,7 @@ void decompressData(const char * src, UInt32 bytes_size, char * dst, UInt32 unco
{
_T min_value = min;
for (UInt32 i = 0; i < num_elements; ++i, dst += sizeof(_T))
unalignedStore(dst, min_value);
unalignedStore<_T>(dst, min_value);
return;
}

6
dbms/src/Compression/LZ4_decompress_faster.cpp
View File

@ -200,7 +200,7 @@ inline void copyOverlap8Shuffle(UInt8 * op, const UInt8 *& match, const size_t o
0, 1, 2, 3, 4, 5, 6, 0,
};
unalignedStore(op, vtbl1_u8(unalignedLoad<uint8x8_t>(match), unalignedLoad<uint8x8_t>(masks + 8 * offset)));
unalignedStore<uint8x8_t>(op, vtbl1_u8(unalignedLoad<uint8x8_t>(match), unalignedLoad<uint8x8_t>(masks + 8 * offset)));
match += masks[offset];
}
@ -328,10 +328,10 @@ inline void copyOverlap16Shuffle(UInt8 * op, const UInt8 *& match, const size_t
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 0,
};
unalignedStore(op,
unalignedStore<uint8x8_t>(op,
vtbl2_u8(unalignedLoad<uint8x8x2_t>(match), unalignedLoad<uint8x8_t>(masks + 16 * offset)));
unalignedStore(op + 8,
unalignedStore<uint8x8_t>(op + 8,
vtbl2_u8(unalignedLoad<uint8x8x2_t>(match), unalignedLoad<uint8x8_t>(masks + 16 * offset + 8)));
match += masks[offset];

2
dbms/src/Core/Settings.h
View File

@ -323,7 +323,7 @@ struct Settings : public SettingsCollection<Settings>
M(SettingBool, external_table_functions_use_nulls, true, "If it is set to true, external table functions will implicitly use Nullable type if needed. Otherwise NULLs will be substituted with default values. Currently supported only for 'mysql' table function.") \
M(SettingBool, allow_experimental_data_skipping_indices, false, "If it is set to true, data skipping indices can be used in CREATE TABLE/ALTER TABLE queries.") \
\
M(SettingBool, allow_hyperscan, true, "Allow functions that use Hyperscan library. Disable to avoid potentially long compilation times and excessive resource usage.") \
M(SettingBool, allow_hyperscan, 1, "Allow functions that use Hyperscan library. Disable to avoid potentially long compilation times and excessive resource usage.") \
M(SettingBool, allow_simdjson, 1, "Allow using simdjson library in 'JSON*' functions if AVX2 instructions are available. If disabled rapidjson will be used.") \
\
M(SettingUInt64, max_partitions_per_insert_block, 100, "Limit maximum number of partitions in single INSERTed block. Zero means unlimited. Throw exception if the block contains too many partitions. This setting is a safety threshold, because using large number of partitions is a common misconception.")

8
dbms/src/DataStreams/MarkInCompressedFile.h
View File

@ -7,8 +7,8 @@
#include <Common/PODArray.h>
#include <Common/config.h>
#if USE_LFALLOC
#include <Common/LFAllocator.h>
#if USE_MIMALLOC
#include <Common/MiAllocator.h>
#endif
namespace DB
@ -43,8 +43,8 @@ struct MarkInCompressedFile
}
};
#if USE_LFALLOC
using MarksInCompressedFile = PODArray<MarkInCompressedFile, 4096, LFAllocator>;
#if USE_MIMALLOC
using MarksInCompressedFile = PODArray<MarkInCompressedFile, 4096, MiAllocator>;
#else
using MarksInCompressedFile = PODArray<MarkInCompressedFile>;
#endif

60
dbms/src/Dictionaries/Embedded/TechDataHierarchy.cpp
View File

@ -1,60 +0,0 @@
#include "config_core.h"
#if USE_MYSQL
# include "TechDataHierarchy.h"
# include <common/logger_useful.h>
# include <mysqlxx/PoolWithFailover.h>
static constexpr auto config_key = "mysql_metrica";
void TechDataHierarchy::reload()
{
Logger * log = &Logger::get("TechDataHierarchy");
LOG_DEBUG(log, "Loading tech data hierarchy.");
mysqlxx::PoolWithFailover pool(config_key);
mysqlxx::Pool::Entry conn = pool.Get();
{
mysqlxx::Query q = conn->query("SELECT Id, COALESCE(Parent_Id, 0) FROM OS2");
LOG_TRACE(log, q.str());
mysqlxx::UseQueryResult res = q.use();
while (mysqlxx::Row row = res.fetch())
{
UInt64 child = row[0].getUInt();
UInt64 parent = row[1].getUInt();
if (child > 255 || parent > 255)
throw Poco::Exception("Too large OS id (> 255).");
os_parent[child] = parent;
}
}
{
mysqlxx::Query q = conn->query("SELECT Id, COALESCE(ParentId, 0) FROM SearchEngines");
LOG_TRACE(log, q.str());
mysqlxx::UseQueryResult res = q.use();
while (mysqlxx::Row row = res.fetch())
{
UInt64 child = row[0].getUInt();
UInt64 parent = row[1].getUInt();
if (child > 255 || parent > 255)
throw Poco::Exception("Too large search engine id (> 255).");
se_parent[child] = parent;
}
}
}
bool TechDataHierarchy::isConfigured(const Poco::Util::AbstractConfiguration & config)
{
return config.has(config_key);
}
#endif

76
dbms/src/Dictionaries/Embedded/TechDataHierarchy.h
View File

@ -1,76 +0,0 @@
#pragma once
#include <common/Types.h>
#include <ext/singleton.h>
namespace Poco
{
namespace Util
{
class AbstractConfiguration;
}
class Logger;
}
/** @brief Class that lets you know if a search engine or operating system belongs
* another search engine or operating system, respectively.
* Information about the hierarchy of regions is downloaded from the database.
*/
class TechDataHierarchy
{
private:
UInt8 os_parent[256]{};
UInt8 se_parent[256]{};
public:
void reload();
/// Has corresponding section in configuration file.
static bool isConfigured(const Poco::Util::AbstractConfiguration & config);
/// The "belongs" relation.
bool isOSIn(UInt8 lhs, UInt8 rhs) const
{
while (lhs != rhs && os_parent[lhs])
lhs = os_parent[lhs];
return lhs == rhs;
}
bool isSEIn(UInt8 lhs, UInt8 rhs) const
{
while (lhs != rhs && se_parent[lhs])
lhs = se_parent[lhs];
return lhs == rhs;
}
UInt8 OSToParent(UInt8 x) const { return os_parent[x]; }
UInt8 SEToParent(UInt8 x) const { return se_parent[x]; }
/// To the topmost ancestor.
UInt8 OSToMostAncestor(UInt8 x) const
{
while (os_parent[x])
x = os_parent[x];
return x;
}
UInt8 SEToMostAncestor(UInt8 x) const
{
while (se_parent[x])
x = se_parent[x];
return x;
}
};
class TechDataHierarchySingleton : public ext::singleton<TechDataHierarchySingleton>, public TechDataHierarchy
{
};

2
dbms/src/Formats/JSONEachRowRowInputStream.h
View File

@ -13,7 +13,7 @@ class ReadBuffer;
/** A stream for reading data in JSON format, where each row is represented by a separate JSON object.
* Objects can be separated by feed return, other whitespace characters in any number and possibly a comma.
* Objects can be separated by line feed, other whitespace characters in any number and possibly a comma.
* Fields can be listed in any order (including, in different lines there may be different order),
* and some fields may be missing.
*/

9
dbms/src/Functions/FunctionsEmbeddedDictionaries.cpp
View File

@ -16,15 +16,6 @@ void registerFunctionsEmbeddedDictionaries(FunctionFactory & factory)
factory.registerFunction<FunctionRegionIn>();
factory.registerFunction<FunctionRegionHierarchy>();
factory.registerFunction<FunctionRegionToName>();
#if USE_MYSQL
factory.registerFunction<FunctionOSToRoot>();
factory.registerFunction<FunctionSEToRoot>();
factory.registerFunction<FunctionOSIn>();
factory.registerFunction<FunctionSEIn>();
factory.registerFunction<FunctionOSHierarchy>();
factory.registerFunction<FunctionSEHierarchy>();
#endif
}
}

111
dbms/src/Functions/FunctionsEmbeddedDictionaries.h
View File

@ -18,11 +18,6 @@
#include <Common/config.h>
#include <Common/typeid_cast.h>
#include "config_core.h"
#if USE_MYSQL
#include <Dictionaries/Embedded/TechDataHierarchy.h>
#endif
namespace DB
{
@ -98,41 +93,6 @@ struct RegionHierarchyImpl
};
#if USE_MYSQL
struct OSToRootImpl
{
static UInt8 apply(UInt8 x, const TechDataHierarchy & hierarchy) { return hierarchy.OSToMostAncestor(x); }
};
struct SEToRootImpl
{
static UInt8 apply(UInt8 x, const TechDataHierarchy & hierarchy) { return hierarchy.SEToMostAncestor(x); }
};
struct OSInImpl
{
static bool apply(UInt32 x, UInt32 y, const TechDataHierarchy & hierarchy) { return hierarchy.isOSIn(x, y); }
};
struct SEInImpl
{
static bool apply(UInt32 x, UInt32 y, const TechDataHierarchy & hierarchy) { return hierarchy.isSEIn(x, y); }
};
struct OSHierarchyImpl
{
static UInt8 toParent(UInt8 x, const TechDataHierarchy & hierarchy) { return hierarchy.OSToParent(x); }
};
struct SEHierarchyImpl
{
static UInt8 toParent(UInt8 x, const TechDataHierarchy & hierarchy) { return hierarchy.SEToParent(x); }
};
#endif
/** Auxiliary thing, allowing to get from the dictionary a specific dictionary, corresponding to the point of view
* (the dictionary key passed as function argument).
* Example: when calling regionToCountry(x, 'ua'), a dictionary can be used, in which Crimea refers to Ukraine.
@ -515,18 +475,6 @@ struct NameRegionHierarchy { static constexpr auto name = "regionHie
struct NameRegionIn { static constexpr auto name = "regionIn"; };
#if USE_MYSQL
struct NameOSToRoot { static constexpr auto name = "OSToRoot"; };
struct NameSEToRoot { static constexpr auto name = "SEToRoot"; };
struct NameOSIn { static constexpr auto name = "OSIn"; };
struct NameSEIn { static constexpr auto name = "SEIn"; };
struct NameOSHierarchy { static constexpr auto name = "OSHierarchy"; };
struct NameSEHierarchy { static constexpr auto name = "SEHierarchy"; };
#endif
struct FunctionRegionToCity :
public FunctionTransformWithDictionary<UInt32, RegionToCityImpl, RegionsHierarchyGetter, NameRegionToCity>
{
@ -609,65 +557,6 @@ struct FunctionRegionHierarchy :
};
#if USE_MYSQL
struct FunctionOSToRoot :
public FunctionTransformWithDictionary<UInt8, OSToRootImpl, IdentityDictionaryGetter<TechDataHierarchy>, NameOSToRoot>
{
static FunctionPtr create(const Context & context)
{
return std::make_shared<base_type>(context.getEmbeddedDictionaries().getTechDataHierarchy());
}
};
struct FunctionSEToRoot :
public FunctionTransformWithDictionary<UInt8, SEToRootImpl, IdentityDictionaryGetter<TechDataHierarchy>, NameSEToRoot>
{
static FunctionPtr create(const Context & context)
{
return std::make_shared<base_type>(context.getEmbeddedDictionaries().getTechDataHierarchy());
}
};
struct FunctionOSIn :
public FunctionIsInWithDictionary<UInt8, OSInImpl, IdentityDictionaryGetter<TechDataHierarchy>, NameOSIn>
{
static FunctionPtr create(const Context & context)
{
return std::make_shared<base_type>(context.getEmbeddedDictionaries().getTechDataHierarchy());
}
};
struct FunctionSEIn :
public FunctionIsInWithDictionary<UInt8, SEInImpl, IdentityDictionaryGetter<TechDataHierarchy>, NameSEIn>
{
static FunctionPtr create(const Context & context)
{
return std::make_shared<base_type>(context.getEmbeddedDictionaries().getTechDataHierarchy());
}
};
struct FunctionOSHierarchy :
public FunctionHierarchyWithDictionary<UInt8, OSHierarchyImpl, IdentityDictionaryGetter<TechDataHierarchy>, NameOSHierarchy>
{
static FunctionPtr create(const Context & context)
{
return std::make_shared<base_type>(context.getEmbeddedDictionaries().getTechDataHierarchy());
}
};
struct FunctionSEHierarchy :
public FunctionHierarchyWithDictionary<UInt8, SEHierarchyImpl, IdentityDictionaryGetter<TechDataHierarchy>, NameSEHierarchy>
{
static FunctionPtr create(const Context & context)
{
return std::make_shared<base_type>(context.getEmbeddedDictionaries().getTechDataHierarchy());
}
};
#endif
/// Converts a region's numeric identifier to a name in the specified language using a dictionary.
class FunctionRegionToName : public IFunction
{

8
dbms/src/Functions/FunctionsRandom.cpp
View File

@ -57,10 +57,10 @@ void RandImpl::execute(char * output, size_t size)
for (const char * end = output + size; output < end; output += 16)
{
unalignedStore(output, generator0.next());
unalignedStore(output + 4, generator1.next());
unalignedStore(output + 8, generator2.next());
unalignedStore(output + 12, generator3.next());
unalignedStore<UInt32>(output, generator0.next());
unalignedStore<UInt32>(output + 4, generator1.next());
unalignedStore<UInt32>(output + 8, generator2.next());
unalignedStore<UInt32>(output + 12, generator3.next());
}
/// It is guaranteed (by PaddedPODArray) that we can overwrite up to 15 bytes after end.

3
dbms/src/Functions/FunctionsVisitParam.h
View File

@ -91,8 +91,7 @@ struct ExtractBool
struct ExtractRaw
{
static constexpr size_t bytes_on_stack = 64;
using ExpectChars = PODArray<char, bytes_on_stack, AllocatorWithStackMemory<Allocator<false>, bytes_on_stack>>;
using ExpectChars = PODArrayWithStackMemory<char, 64>;
static void extract(const UInt8 * pos, const UInt8 * end, ColumnString::Chars & res_data)
{

113
dbms/src/Functions/URL/domain.h
View File

@ -3,44 +3,117 @@
#include "protocol.h"
#include <common/find_symbols.h>
#include <cstring>
#include <Common/StringUtils/StringUtils.h>
namespace DB
{
namespace
{
inline StringRef checkAndReturnHost(const Pos & pos, const Pos & dot_pos, const Pos & start_of_host)
{
if (!dot_pos || start_of_host >= pos || pos - dot_pos == 1)
return StringRef{};
auto after_dot = *(dot_pos + 1);
if (after_dot == ':' || after_dot == '/' || after_dot == '?' || after_dot == '#')
return StringRef{};
return StringRef(start_of_host, pos - start_of_host);
}
}
/// Extracts host from given url.
inline StringRef getURLHost(const char * data, size_t size)
{
Pos pos = data;
Pos end = data + size;
if (end == (pos = find_first_symbols<'/'>(pos, end)))
return {};
if (pos != data)
if (*pos == '/' && *(pos + 1) == '/')
{
StringRef scheme = getURLScheme(data, size);
Pos scheme_end = data + scheme.size;
// Colon must follows after scheme.
if (pos - scheme_end != 1 || *scheme_end != ':')
return {};
pos += 2;
}
else
{
Pos scheme_end = data + std::min(size, 16UL);
for (++pos; pos < scheme_end; ++pos)
{
if (!isAlphaNumericASCII(*pos))
{
switch (*pos)
{
case '.':
case '-':
case '+':
break;
case ' ': /// restricted symbols
case '\t':
case '<':
case '>':
case '%':
case '{':
case '}':
case '|':
case '\\':
case '^':
case '~':
case '[':
case ']':
case ';':
case '=':
case '&':
return StringRef{};
default:
goto exloop;
}
}
}
exloop: if ((scheme_end - pos) > 2 && *pos == ':' && *(pos + 1) == '/' && *(pos + 2) == '/')
pos += 3;
else
pos = data;
}
if (end - pos < 2 || *(pos) != '/' || *(pos + 1) != '/')
return {};
pos += 2;
const char * start_of_host = pos;
Pos dot_pos = nullptr;
auto start_of_host = pos;
for (; pos < end; ++pos)
{
if (*pos == '@')
start_of_host = pos + 1;
else if (*pos == ':' || *pos == '/' || *pos == '?' || *pos == '#')
switch (*pos)
{
case '.':
dot_pos = pos;
break;
case ':': /// end symbols
case '/':
case '?':
case '#':
return checkAndReturnHost(pos, dot_pos, start_of_host);
case '@': /// myemail@gmail.com
start_of_host = pos + 1;
break;
case ' ': /// restricted symbols in whole URL
case '\t':
case '<':
case '>':
case '%':
case '{':
case '}':
case '|':
case '\\':
case '^':
case '~':
case '[':
case ']':
case ';':
case '=':
case '&':
return StringRef{};
}
}
return (pos == start_of_host) ? StringRef{} : StringRef(start_of_host, pos - start_of_host);
return checkAndReturnHost(pos, dot_pos, start_of_host);
}
template <bool without_www>

15
dbms/src/Functions/array/flatten.cpp → dbms/src/Functions/array/arrayFlatten.cpp
View File

@ -12,13 +12,13 @@ namespace ErrorCodes
extern const int ILLEGAL_COLUMN;
}
/// flatten([[1, 2, 3], [4, 5]]) = [1, 2, 3, 4, 5] - flatten array.
class FunctionFlatten : public IFunction
/// arrayFlatten([[1, 2, 3], [4, 5]]) = [1, 2, 3, 4, 5] - flatten array.
class ArrayFlatten : public IFunction
{
public:
static constexpr auto name = "flatten";
static constexpr auto name = "arrayFlatten";
static FunctionPtr create(const Context &) { return std::make_shared<FunctionFlatten>(); }
static FunctionPtr create(const Context &) { return std::make_shared<ArrayFlatten>(); }
size_t getNumberOfArguments() const override { return 1; }
bool useDefaultImplementationForConstants() const override { return true; }
@ -80,7 +80,7 @@ result: Row 1: [1, 2, 3], Row2: [4]
const ColumnArray * src_col = checkAndGetColumn<ColumnArray>(block.getByPosition(arguments[0]).column.get());
if (!src_col)
throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName() + " in argument of function 'flatten'",
throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName() + " in argument of function 'arrayFlatten'",
ErrorCodes::ILLEGAL_COLUMN);
const IColumn::Offsets & src_offsets = src_col->getOffsets();
@ -118,9 +118,10 @@ private:
};
void registerFunctionFlatten(FunctionFactory & factory)
void registerFunctionArrayFlatten(FunctionFactory & factory)
{
factory.registerFunction<FunctionFlatten>();
factory.registerFunction<ArrayFlatten>();
factory.registerAlias("flatten", "arrayFlatten", FunctionFactory::CaseInsensitive);
}
}

4
dbms/src/Functions/array/registerFunctionsArray.cpp
View File

@ -30,7 +30,7 @@ void registerFunctionArrayEnumerateUniqRanked(FunctionFactory &);
void registerFunctionArrayEnumerateDenseRanked(FunctionFactory &);
void registerFunctionArrayUniq(FunctionFactory &);
void registerFunctionArrayDistinct(FunctionFactory &);
void registerFunctionFlatten(FunctionFactory &);
void registerFunctionArrayFlatten(FunctionFactory &);
void registerFunctionArrayWithConstant(FunctionFactory &);
void registerFunctionsArray(FunctionFactory & factory)
@ -62,7 +62,7 @@ void registerFunctionsArray(FunctionFactory & factory)
registerFunctionArrayEnumerateDenseRanked(factory);
registerFunctionArrayUniq(factory);
registerFunctionArrayDistinct(factory);
registerFunctionFlatten(factory);
registerFunctionArrayFlatten(factory);
registerFunctionArrayWithConstant(factory);
}

0
dbms/src/Functions/array/empty.cpp → dbms/src/Functions/empty.cpp
View File

0
dbms/src/Functions/array/notEmpty.cpp → dbms/src/Functions/notEmpty.cpp
View File

11
dbms/src/Functions/registerFunctions.cpp
View File

@ -1,5 +1,6 @@
#include <Functions/FunctionFactory.h>
#include <Functions/registerFunctions.h>
#include "config_core.h"
#include "config_functions.h"
@ -41,16 +42,11 @@ void registerFunctionsGeo(FunctionFactory &);
void registerFunctionsNull(FunctionFactory &);
void registerFunctionsFindCluster(FunctionFactory &);
void registerFunctionsJSON(FunctionFactory &);
void registerFunctionTransform(FunctionFactory &);
#if USE_H3
void registerFunctionGeoToH3(FunctionFactory &);
#endif
#if USE_ICU
void registerFunctionConvertCharset(FunctionFactory &);
#endif
void registerFunctions()
{
auto & factory = FunctionFactory::instance();
@ -88,15 +84,10 @@ void registerFunctions()
registerFunctionsNull(factory);
registerFunctionsFindCluster(factory);
registerFunctionsJSON(factory);
registerFunctionTransform(factory);
#if USE_H3
registerFunctionGeoToH3(factory);
#endif
#if USE_ICU
registerFunctionConvertCharset(factory);
#endif
}
}

12
dbms/src/Functions/registerFunctionsMiscellaneous.cpp
View File

@ -1,3 +1,5 @@
#include "config_core.h"
namespace DB
{
@ -45,6 +47,11 @@ void registerFunctionJoinGet(FunctionFactory &);
void registerFunctionFilesystem(FunctionFactory &);
void registerFunctionEvalMLMethod(FunctionFactory &);
void registerFunctionBasename(FunctionFactory &);
void registerFunctionTransform(FunctionFactory &);
#if USE_ICU
void registerFunctionConvertCharset(FunctionFactory &);
#endif
void registerFunctionsMiscellaneous(FunctionFactory & factory)
{
@ -90,6 +97,11 @@ void registerFunctionsMiscellaneous(FunctionFactory & factory)
registerFunctionFilesystem(factory);
registerFunctionEvalMLMethod(factory);
registerFunctionBasename(factory);
registerFunctionTransform(factory);
#if USE_ICU
registerFunctionConvertCharset(factory);
#endif
}
}

9
dbms/src/IO/BitHelpers.h
View File

@ -5,6 +5,15 @@
#include <Core/Types.h>
#include <Common/BitHelpers.h>
#if defined(__OpenBSD__) || defined(__FreeBSD__)
# include <sys/endian.h>
#elif defined(__APPLE__)
# include <libkern/OSByteOrder.h>
# define htobe64(x) OSSwapHostToBigInt64(x)
# define be64toh(x) OSSwapBigToHostInt64(x)
#endif
namespace DB
{

8
dbms/src/IO/UncompressedCache.h
View File

@ -7,8 +7,8 @@
#include <IO/BufferWithOwnMemory.h>
#include <Common/config.h>
#if USE_LFALLOC
#include <Common/LFAllocator.h>
#if USE_MIMALLOC
#include <Common/MiAllocator.h>
#endif
@ -25,8 +25,8 @@ namespace DB
struct UncompressedCacheCell
{
#if USE_LFALLOC
Memory<LFAllocator> data;
#if USE_MIMALLOC
Memory<MiAllocator> data;
#else
Memory<> data;
#endif

30
dbms/src/Interpreters/BloomFilter.cpp
View File

@ -1,5 +1,4 @@
#include <Interpreters/BloomFilter.h>
#include <city.h>
@ -9,14 +8,13 @@ namespace DB
static constexpr UInt64 SEED_GEN_A = 845897321;
static constexpr UInt64 SEED_GEN_B = 217728422;
StringBloomFilter::StringBloomFilter(size_t size_, size_t hashes_, size_t seed_)
BloomFilter::BloomFilter(size_t size_, size_t hashes_, size_t seed_)
: size(size_), hashes(hashes_), seed(seed_), words((size + sizeof(UnderType) - 1) / sizeof(UnderType)), filter(words, 0) {}
StringBloomFilter::StringBloomFilter(const StringBloomFilter & bloom_filter)
BloomFilter::BloomFilter(const BloomFilter & bloom_filter)
: size(bloom_filter.size), hashes(bloom_filter.hashes), seed(bloom_filter.seed), words(bloom_filter.words), filter(bloom_filter.filter) {}
bool StringBloomFilter::find(const char * data, size_t len)
bool BloomFilter::find(const char * data, size_t len)
{
size_t hash1 = CityHash_v1_0_2::CityHash64WithSeed(data, len, seed);
size_t hash2 = CityHash_v1_0_2::CityHash64WithSeed(data, len, SEED_GEN_A * seed + SEED_GEN_B);
@ -30,7 +28,7 @@ bool StringBloomFilter::find(const char * data, size_t len)
return true;
}
void StringBloomFilter::add(const char * data, size_t len)
void BloomFilter::add(const char * data, size_t len)
{
size_t hash1 = CityHash_v1_0_2::CityHash64WithSeed(data, len, seed);
size_t hash2 = CityHash_v1_0_2::CityHash64WithSeed(data, len, SEED_GEN_A * seed + SEED_GEN_B);
@ -42,12 +40,12 @@ void StringBloomFilter::add(const char * data, size_t len)
}
}
void StringBloomFilter::clear()
void BloomFilter::clear()
{
filter.assign(words, 0);
}
bool StringBloomFilter::contains(const StringBloomFilter & bf)
bool BloomFilter::contains(const BloomFilter & bf)
{
for (size_t i = 0; i < words; ++i)
{
@ -57,7 +55,7 @@ bool StringBloomFilter::contains(const StringBloomFilter & bf)
return true;
}
UInt64 StringBloomFilter::isEmpty() const
UInt64 BloomFilter::isEmpty() const
{
for (size_t i = 0; i < words; ++i)
if (filter[i] != 0)
@ -65,7 +63,7 @@ UInt64 StringBloomFilter::isEmpty() const
return true;
}
bool operator== (const StringBloomFilter & a, const StringBloomFilter & b)
bool operator== (const BloomFilter & a, const BloomFilter & b)
{
for (size_t i = 0; i < a.words; ++i)
if (a.filter[i] != b.filter[i])
@ -73,4 +71,16 @@ bool operator== (const StringBloomFilter & a, const StringBloomFilter & b)
return true;
}
void BloomFilter::addHashWithSeed(const UInt64 & hash, const UInt64 & hash_seed)
{
size_t pos = CityHash_v1_0_2::Hash128to64(CityHash_v1_0_2::uint128(hash, hash_seed)) % (8 * size);
filter[pos / (8 * sizeof(UnderType))] |= (1ULL << (pos % (8 * sizeof(UnderType))));
}
bool BloomFilter::findHashWithSeed(const UInt64 & hash, const UInt64 & hash_seed)
{
size_t pos = CityHash_v1_0_2::Hash128to64(CityHash_v1_0_2::uint128(hash, hash_seed)) % (8 * size);
return bool(filter[pos / (8 * sizeof(UnderType))] & (1ULL << (pos % (8 * sizeof(UnderType)))));
}
}

24
dbms/src/Interpreters/BloomFilter.h
View File

@ -1,15 +1,17 @@
#pragma once
#include <Core/Types.h>
#include <vector>
#include <Core/Types.h>
#include <Common/PODArray.h>
#include <Common/Allocator.h>
#include <Columns/ColumnVector.h>
namespace DB
{
/// Bloom filter for strings.
class StringBloomFilter
class BloomFilter
{
public:
using UnderType = UInt64;
using Container = std::vector<UnderType>;
@ -17,16 +19,19 @@ public:
/// size -- size of filter in bytes.
/// hashes -- number of used hash functions.
/// seed -- random seed for hash functions generation.
StringBloomFilter(size_t size_, size_t hashes_, size_t seed_);
StringBloomFilter(const StringBloomFilter & bloom_filter);
BloomFilter(size_t size_, size_t hashes_, size_t seed_);
BloomFilter(const BloomFilter & bloom_filter);
bool find(const char * data, size_t len);
void add(const char * data, size_t len);
void clear();
void addHashWithSeed(const UInt64 & hash, const UInt64 & hash_seed);
bool findHashWithSeed(const UInt64 & hash, const UInt64 & hash_seed);
/// Checks if this contains everything from another bloom filter.
/// Bloom filters must have equal size and seed.
bool contains(const StringBloomFilter & bf);
bool contains(const BloomFilter & bf);
const Container & getFilter() const { return filter; }
Container & getFilter() { return filter; }
@ -34,7 +39,7 @@ public:
/// For debug.
UInt64 isEmpty() const;
friend bool operator== (const StringBloomFilter & a, const StringBloomFilter & b);
friend bool operator== (const BloomFilter & a, const BloomFilter & b);
private:
size_t size;
@ -44,7 +49,8 @@ private:
Container filter;
};
using BloomFilterPtr = std::shared_ptr<BloomFilter>;
bool operator== (const StringBloomFilter & a, const StringBloomFilter & b);
bool operator== (const BloomFilter & a, const BloomFilter & b);
}

207
dbms/src/Interpreters/BloomFilterHash.h Normal file
View File

@ -0,0 +1,207 @@
#pragma once
#include <Columns/IColumn.h>
#include <Columns/ColumnConst.h>
#include <Columns/ColumnsNumber.h>
#include <Columns/ColumnString.h>
#include <Columns/ColumnFixedString.h>
#include <DataTypes/IDataType.h>
#include <DataTypes/DataTypesNumber.h>
#include <ext/bit_cast.h>
#include <Common/HashTable/Hash.h>
namespace DB
{
namespace ErrorCodes
{
extern const int ILLEGAL_COLUMN;
}
struct BloomFilterHash
{
static constexpr UInt64 bf_hash_seed[15] = {
13635471485423070496ULL, 10336109063487487899ULL, 17779957404565211594ULL, 8988612159822229247ULL, 4954614162757618085ULL,
12980113590177089081ULL, 9263883436177860930ULL, 3656772712723269762ULL, 10362091744962961274ULL, 7582936617938287249ULL,
15033938188484401405ULL, 18286745649494826751ULL, 6852245486148412312ULL, 8886056245089344681ULL, 10151472371158292780ULL
};
static ColumnPtr hashWithField(const IDataType * data_type, const Field & field)
{
WhichDataType which(data_type);
if (which.isUInt() || which.isDateOrDateTime())
return ColumnConst::create(ColumnUInt64::create(1, intHash64(field.safeGet<UInt64>())), 1);
else if (which.isInt() || which.isEnum())
return ColumnConst::create(ColumnUInt64::create(1, intHash64(ext::bit_cast<UInt64>(field.safeGet<Int64>()))), 1);
else if (which.isFloat32() || which.isFloat64())
return ColumnConst::create(ColumnUInt64::create(1, intHash64(ext::bit_cast<UInt64>(field.safeGet<Float64>()))), 1);
else if (which.isString() || which.isFixedString())
{
const auto & value = field.safeGet<String>();
return ColumnConst::create(ColumnUInt64::create(1, CityHash_v1_0_2::CityHash64(value.data(), value.size())), 1);
}
else
throw Exception("Unexpected type " + data_type->getName() + " of bloom filter index.", ErrorCodes::LOGICAL_ERROR);
}
static ColumnPtr hashWithColumn(const DataTypePtr & data_type, const ColumnPtr & column, size_t pos, size_t limit)
{
auto index_column = ColumnUInt64::create(limit);
ColumnUInt64::Container & index_column_vec = index_column->getData();
getAnyTypeHash<true>(&*data_type, &*column, index_column_vec, pos);
return index_column;
}
template <bool is_first>
static void getAnyTypeHash(const IDataType * data_type, const IColumn * column, ColumnUInt64::Container & vec, size_t pos)
{
WhichDataType which(data_type);
if (which.isUInt8()) getNumberTypeHash<UInt8, is_first>(column, vec, pos);
else if (which.isUInt16()) getNumberTypeHash<UInt16, is_first>(column, vec, pos);
else if (which.isUInt32()) getNumberTypeHash<UInt32, is_first>(column, vec, pos);
else if (which.isUInt64()) getNumberTypeHash<UInt64, is_first>(column, vec, pos);
else if (which.isInt8()) getNumberTypeHash<Int8, is_first>(column, vec, pos);
else if (which.isInt16()) getNumberTypeHash<Int16, is_first>(column, vec, pos);
else if (which.isInt32()) getNumberTypeHash<Int32, is_first>(column, vec, pos);
else if (which.isInt64()) getNumberTypeHash<Int64, is_first>(column, vec, pos);
else if (which.isEnum8()) getNumberTypeHash<Int8, is_first>(column, vec, pos);
else if (which.isEnum16()) getNumberTypeHash<Int16, is_first>(column, vec, pos);
else if (which.isDate()) getNumberTypeHash<UInt16, is_first>(column, vec, pos);
else if (which.isDateTime()) getNumberTypeHash<UInt32, is_first>(column, vec, pos);
else if (which.isFloat32()) getNumberTypeHash<Float32, is_first>(column, vec, pos);
else if (which.isFloat64()) getNumberTypeHash<Float64, is_first>(column, vec, pos);
else if (which.isString()) getStringTypeHash<is_first>(column, vec, pos);
else if (which.isFixedString()) getStringTypeHash<is_first>(column, vec, pos);
else throw Exception("Unexpected type " + data_type->getName() + " of bloom filter index.", ErrorCodes::LOGICAL_ERROR);
}
template <typename Type, bool is_first>
static void getNumberTypeHash(const IColumn * column, ColumnUInt64::Container & vec, size_t pos)
{
const auto * index_column = typeid_cast<const ColumnVector<Type> *>(column);
if (unlikely(!index_column))
throw Exception("Illegal column type was passed to the bloom filter index.", ErrorCodes::ILLEGAL_COLUMN);
const typename ColumnVector<Type>::Container & vec_from = index_column->getData();
/// Because we're missing the precision of float in the Field.h
/// to be consistent, we need to convert Float32 to Float64 processing, also see: BloomFilterHash::hashWithField
if constexpr (std::is_same_v<ColumnVector<Type>, ColumnFloat32>)
{
for (size_t index = 0, size = vec.size(); index < size; ++index)
{
UInt64 hash = intHash64(ext::bit_cast<UInt64>(Float64(vec_from[index + pos])));
if constexpr (is_first)
vec[index] = hash;
else
vec[index] = CityHash_v1_0_2::Hash128to64(CityHash_v1_0_2::uint128(vec[index], hash));
}
}
else
{
for (size_t index = 0, size = vec.size(); index < size; ++index)
{
UInt64 hash = intHash64(ext::bit_cast<UInt64>(vec_from[index + pos]));
if constexpr (is_first)
vec[index] = hash;
else
vec[index] = CityHash_v1_0_2::Hash128to64(CityHash_v1_0_2::uint128(vec[index], hash));
}
}
}
template <bool is_first>
static void getStringTypeHash(const IColumn * column, ColumnUInt64::Container & vec, size_t pos)
{
if (const auto * index_column = typeid_cast<const ColumnString *>(column))
{
const ColumnString::Chars & data = index_column->getChars();
const ColumnString::Offsets & offsets = index_column->getOffsets();
ColumnString::Offset current_offset = pos;
for (size_t index = 0, size = vec.size(); index < size; ++index)
{
UInt64 city_hash = CityHash_v1_0_2::CityHash64(
reinterpret_cast<const char *>(&data[current_offset]), offsets[index + pos] - current_offset - 1);
if constexpr (is_first)
vec[index] = city_hash;
else
vec[index] = CityHash_v1_0_2::Hash128to64(CityHash_v1_0_2::uint128(vec[index], city_hash));
current_offset = offsets[index + pos];
}
}
else if (const auto * fixed_string_index_column = typeid_cast<const ColumnFixedString *>(column))
{
size_t fixed_len = fixed_string_index_column->getN();
const auto & data = fixed_string_index_column->getChars();
for (size_t index = 0, size = vec.size(); index < size; ++index)
{
UInt64 city_hash = CityHash_v1_0_2::CityHash64(reinterpret_cast<const char *>(&data[(index + pos) * fixed_len]), fixed_len);
if constexpr (is_first)
vec[index] = city_hash;
else
vec[index] = CityHash_v1_0_2::Hash128to64(CityHash_v1_0_2::uint128(vec[index], city_hash));
}
}
else
throw Exception("Illegal column type was passed to the bloom filter index.", ErrorCodes::ILLEGAL_COLUMN);
}
static std::pair<size_t, size_t> calculationBestPractices(double max_conflict_probability)
{
static const size_t MAX_BITS_PER_ROW = 20;
static const size_t MAX_HASH_FUNCTION_COUNT = 15;
/// For the smallest index per level in probability_lookup_table
static const size_t min_probability_index_each_bits[] = {0, 0, 1, 2, 3, 3, 4, 5, 6, 6, 7, 8, 8, 9, 10, 10, 11, 12, 12, 13, 14};
static const long double probability_lookup_table[MAX_BITS_PER_ROW + 1][MAX_HASH_FUNCTION_COUNT] =
{
{1.0}, /// dummy, 0 bits per row
{1.0, 1.0},
{1.0, 0.393, 0.400},
{1.0, 0.283, 0.237, 0.253},
{1.0, 0.221, 0.155, 0.147, 0.160},
{1.0, 0.181, 0.109, 0.092, 0.092, 0.101}, // 5
{1.0, 0.154, 0.0804, 0.0609, 0.0561, 0.0578, 0.0638},
{1.0, 0.133, 0.0618, 0.0423, 0.0359, 0.0347, 0.0364},
{1.0, 0.118, 0.0489, 0.0306, 0.024, 0.0217, 0.0216, 0.0229},
{1.0, 0.105, 0.0397, 0.0228, 0.0166, 0.0141, 0.0133, 0.0135, 0.0145},
{1.0, 0.0952, 0.0329, 0.0174, 0.0118, 0.00943, 0.00844, 0.00819, 0.00846}, // 10
{1.0, 0.0869, 0.0276, 0.0136, 0.00864, 0.0065, 0.00552, 0.00513, 0.00509},
{1.0, 0.08, 0.0236, 0.0108, 0.00646, 0.00459, 0.00371, 0.00329, 0.00314},
{1.0, 0.074, 0.0203, 0.00875, 0.00492, 0.00332, 0.00255, 0.00217, 0.00199, 0.00194},
{1.0, 0.0689, 0.0177, 0.00718, 0.00381, 0.00244, 0.00179, 0.00146, 0.00129, 0.00121, 0.0012},
{1.0, 0.0645, 0.0156, 0.00596, 0.003, 0.00183, 0.00128, 0.001, 0.000852, 0.000775, 0.000744}, // 15
{1.0, 0.0606, 0.0138, 0.005, 0.00239, 0.00139, 0.000935, 0.000702, 0.000574, 0.000505, 0.00047, 0.000459},
{1.0, 0.0571, 0.0123, 0.00423, 0.00193, 0.00107, 0.000692, 0.000499, 0.000394, 0.000335, 0.000302, 0.000287, 0.000284},
{1.0, 0.054, 0.0111, 0.00362, 0.00158, 0.000839, 0.000519, 0.00036, 0.000275, 0.000226, 0.000198, 0.000183, 0.000176},
{1.0, 0.0513, 0.00998, 0.00312, 0.0013, 0.000663, 0.000394, 0.000264, 0.000194, 0.000155, 0.000132, 0.000118, 0.000111, 0.000109},
{1.0, 0.0488, 0.00906, 0.0027, 0.00108, 0.00053, 0.000303, 0.000196, 0.00014, 0.000108, 8.89e-05, 7.77e-05, 7.12e-05, 6.79e-05, 6.71e-05} // 20
};
for (size_t bits_per_row = 1; bits_per_row < MAX_BITS_PER_ROW; ++bits_per_row)
{
if (probability_lookup_table[bits_per_row][min_probability_index_each_bits[bits_per_row]] <= max_conflict_probability)
{
size_t max_size_of_hash_functions = min_probability_index_each_bits[bits_per_row];
for (size_t size_of_hash_functions = max_size_of_hash_functions; size_of_hash_functions > 0; --size_of_hash_functions)
if (probability_lookup_table[bits_per_row][size_of_hash_functions] > max_conflict_probability)
return std::pair<size_t, size_t>(bits_per_row, size_of_hash_functions + 1);
}
}
return std::pair<size_t, size_t>(MAX_BITS_PER_ROW - 1, min_probability_index_each_bits[MAX_BITS_PER_ROW - 1]);
}
};
}

4
dbms/src/Interpreters/Compiler.cpp
View File

@ -262,8 +262,8 @@ void Compiler::compile(
" -I " << compiler_headers << "/dbms/src/"
" -isystem " << compiler_headers << "/contrib/cityhash102/include/"
" -isystem " << compiler_headers << "/contrib/libpcg-random/include/"
#if USE_LFALLOC
" -isystem " << compiler_headers << "/contrib/lfalloc/src/"
#if USE_MIMALLOC
" -isystem " << compiler_headers << "/contrib/mimalloc/include/"
#endif
" -isystem " << compiler_headers << INTERNAL_DOUBLE_CONVERSION_INCLUDE_DIR
" -isystem " << compiler_headers << INTERNAL_Poco_Foundation_INCLUDE_DIR

14
dbms/src/Interpreters/Context.cpp
View File

@ -244,15 +244,12 @@ struct ContextShared
return;
shutdown_called = true;
{
std::lock_guard lock(mutex);
/** After system_logs have been shut down it is guaranteed that no system table gets created or written to.
* Note that part changes at shutdown won't be logged to part log.
*/
/** After this point, system logs will shutdown their threads and no longer write any data.
* It will prevent recreation of system tables at shutdown.
* Note that part changes at shutdown won't be logged to part log.
*/
system_logs.reset();
}
if (system_logs)
system_logs->shutdown();
/** At this point, some tables may have threads that block our mutex.
* To shutdown them correctly, we will copy the current list of tables,
@ -280,6 +277,7 @@ struct ContextShared
/// Preemptive destruction is important, because these objects may have a refcount to ContextShared (cyclic reference).
/// TODO: Get rid of this.
system_logs.reset();
embedded_dictionaries.reset();
external_dictionaries.reset();
external_models.reset();

184
dbms/src/Interpreters/DDLWorker.cpp
View File

@ -1,6 +1,9 @@
#include <Interpreters/DDLWorker.h>
#include <Parsers/ASTAlterQuery.h>
#include <Parsers/ASTDropQuery.h>
#include <Parsers/ASTOptimizeQuery.h>
#include <Parsers/ASTQueryWithOnCluster.h>
#include <Parsers/ASTQueryWithTableAndOutput.h>
#include <Parsers/ParserQuery.h>
#include <Parsers/parseQuery.h>
#include <Parsers/queryToString.h>
@ -29,6 +32,7 @@
#include <Common/ZooKeeper/KeeperException.h>
#include <Common/ZooKeeper/Lock.h>
#include <Common/isLocalAddress.h>
#include <Storages/StorageReplicatedMergeTree.h>
#include <Poco/Timestamp.h>
#include <random>
#include <pcg_random.hpp>
@ -614,14 +618,24 @@ void DDLWorker::processTask(DDLTask & task, const ZooKeeperPtr & zookeeper)
String rewritten_query = queryToString(rewritten_ast);
LOG_DEBUG(log, "Executing query: " << rewritten_query);
if (const auto * ast_alter = rewritten_ast->as<ASTAlterQuery>())
if (auto query_with_table = dynamic_cast<ASTQueryWithTableAndOutput *>(rewritten_ast.get()); query_with_table)
{
processTaskAlter(task, ast_alter, rewritten_query, task.entry_path, zookeeper);
String database = query_with_table->database.empty() ? context.getCurrentDatabase() : query_with_table->database;
StoragePtr storage = context.tryGetTable(database, query_with_table->table);
/// For some reason we check consistency of cluster definition only
/// in case of ALTER query, but not in case of CREATE/DROP etc.
/// It's strange, but this behaviour exits for a long and we cannot change it.
if (storage && query_with_table->as<ASTAlterQuery>())
checkShardConfig(query_with_table->table, task, storage);
if (storage && taskShouldBeExecutedOnLeader(rewritten_ast, storage))
tryExecuteQueryOnLeaderReplica(task, storage, rewritten_query, task.entry_path, zookeeper);
else
tryExecuteQuery(rewritten_query, task, task.execution_status);
}
else
{
tryExecuteQuery(rewritten_query, task, task.execution_status);
}
}
catch (const Coordination::Exception &)
{
@ -646,43 +660,52 @@ void DDLWorker::processTask(DDLTask & task, const ZooKeeperPtr & zookeeper)
}
void DDLWorker::processTaskAlter(
DDLTask & task,
const ASTAlterQuery * ast_alter,
const String & rewritten_query,
const String & node_path,
const ZooKeeperPtr & zookeeper)
bool DDLWorker::taskShouldBeExecutedOnLeader(const ASTPtr ast_ddl, const StoragePtr storage) const
{
String database = ast_alter->database.empty() ? context.getCurrentDatabase() : ast_alter->database;
StoragePtr storage = context.getTable(database, ast_alter->table);
/// Pure DROP queries have to be executed on each node separately
if (auto query = ast_ddl->as<ASTDropQuery>(); query && query->kind != ASTDropQuery::Kind::Truncate)
return false;
bool execute_once_on_replica = storage->supportsReplication();
bool execute_on_leader_replica = false;
if (!ast_ddl->as<ASTAlterQuery>() && !ast_ddl->as<ASTOptimizeQuery>() && !ast_ddl->as<ASTDropQuery>())
return false;
for (const auto & command : ast_alter->command_list->commands)
{
if (!isSupportedAlterType(command->type))
throw Exception("Unsupported type of ALTER query", ErrorCodes::NOT_IMPLEMENTED);
return storage->supportsReplication();
}
if (execute_once_on_replica)
execute_on_leader_replica |= command->type == ASTAlterCommand::DROP_PARTITION;
}
void DDLWorker::checkShardConfig(const String & table, const DDLTask & task, StoragePtr storage) const
{
const auto & shard_info = task.cluster->getShardsInfo().at(task.host_shard_num);
bool config_is_replicated_shard = shard_info.hasInternalReplication();
if (execute_once_on_replica && !config_is_replicated_shard)
if (storage->supportsReplication() && !config_is_replicated_shard)
{
throw Exception("Table " + ast_alter->table + " is replicated, but shard #" + toString(task.host_shard_num + 1) +
throw Exception("Table '" + table + "' is replicated, but shard #" + toString(task.host_shard_num + 1) +
" isn't replicated according to its cluster definition."
" Possibly <internal_replication>true</internal_replication> is forgotten in the cluster config.",
ErrorCodes::INCONSISTENT_CLUSTER_DEFINITION);
}
if (!execute_once_on_replica && config_is_replicated_shard)
if (!storage->supportsReplication() && config_is_replicated_shard)
{
throw Exception("Table " + ast_alter->table + " isn't replicated, but shard #" + toString(task.host_shard_num + 1) +
throw Exception("Table '" + table + "' isn't replicated, but shard #" + toString(task.host_shard_num + 1) +
" is replicated according to its cluster definition", ErrorCodes::INCONSISTENT_CLUSTER_DEFINITION);
}
}
bool DDLWorker::tryExecuteQueryOnLeaderReplica(
DDLTask & task,
StoragePtr storage,
const String & rewritten_query,
const String & node_path,
const ZooKeeperPtr & zookeeper)
{
StorageReplicatedMergeTree * replicated_storage = dynamic_cast<StorageReplicatedMergeTree *>(storage.get());
/// If we will develop new replicated storage
if (!replicated_storage)
throw Exception("Storage type '" + storage->getName() + "' is not supported by distributed DDL", ErrorCodes::NOT_IMPLEMENTED);
/// Generate unique name for shard node, it will be used to execute the query by only single host
/// Shard node name has format 'replica_name1,replica_name2,...,replica_nameN'
@ -701,70 +724,73 @@ void DDLWorker::processTaskAlter(
return res;
};
if (execute_once_on_replica)
String shard_node_name = get_shard_name(task.cluster->getShardsAddresses().at(task.host_shard_num));
String shard_path = node_path + "/shards/" + shard_node_name;
String is_executed_path = shard_path + "/executed";
zookeeper->createAncestors(shard_path + "/");
auto is_already_executed = [&]() -> bool
{
String shard_node_name = get_shard_name(task.cluster->getShardsAddresses().at(task.host_shard_num));
String shard_path = node_path + "/shards/" + shard_node_name;
String is_executed_path = shard_path + "/executed";
zookeeper->createAncestors(shard_path + "/");
bool is_executed_by_any_replica = false;
String executed_by;
if (zookeeper->tryGet(is_executed_path, executed_by))
{
auto lock = createSimpleZooKeeperLock(zookeeper, shard_path, "lock", task.host_id_str);
pcg64 rng(randomSeed());
LOG_DEBUG(log, "Task " << task.entry_name << " has already been executed by leader replica ("
<< executed_by << ") of the same shard.");
return true;
}
auto is_already_executed = [&]() -> bool
return false;
};
pcg64 rng(randomSeed());
auto lock = createSimpleZooKeeperLock(zookeeper, shard_path, "lock", task.host_id_str);
static const size_t max_tries = 20;
bool executed_by_leader = false;
for (size_t num_tries = 0; num_tries < max_tries; ++num_tries)
{
if (is_already_executed())
{
executed_by_leader = true;
break;
}
StorageReplicatedMergeTree::Status status;
replicated_storage->getStatus(status);
/// Leader replica take lock
if (status.is_leader && lock->tryLock())
{
if (is_already_executed())
{
String executed_by;
if (zookeeper->tryGet(is_executed_path, executed_by))
{
is_executed_by_any_replica = true;
LOG_DEBUG(log, "Task " << task.entry_name << " has already been executed by another replica ("
<< executed_by << ") of the same shard.");
return true;
}
return false;
};
static const size_t max_tries = 20;
for (size_t num_tries = 0; num_tries < max_tries; ++num_tries)
{
if (is_already_executed())
break;
if (lock->tryLock())
{
if (is_already_executed())
break;
tryExecuteQuery(rewritten_query, task, task.execution_status);
if (execute_on_leader_replica && task.execution_status.code == ErrorCodes::NOT_IMPLEMENTED)
{
/// TODO: it is ok to receive exception "host is not leader"
}
zookeeper->create(is_executed_path, task.host_id_str, zkutil::CreateMode::Persistent);
lock->unlock();
is_executed_by_any_replica = true;
break;
}
std::this_thread::sleep_for(std::chrono::milliseconds(std::uniform_int_distribution<long>(0, 1000)(rng)));
executed_by_leader = true;
break;
}
/// If the leader will unexpectedly changed this method will return false
/// and on the next iteration new leader will take lock
if (tryExecuteQuery(rewritten_query, task, task.execution_status))
{
zookeeper->create(is_executed_path, task.host_id_str, zkutil::CreateMode::Persistent);
executed_by_leader = true;
break;
}
}
if (!is_executed_by_any_replica)
{
task.execution_status = ExecutionStatus(ErrorCodes::NOT_IMPLEMENTED,
"Cannot enqueue replicated DDL query for a replicated shard");
}
/// Does nothing if wasn't previously locked
lock->unlock();
std::this_thread::sleep_for(std::chrono::milliseconds(std::uniform_int_distribution<long>(0, 1000)(rng)));
}
else
/// Not executed by leader so was not executed at all
if (!executed_by_leader)
{
tryExecuteQuery(rewritten_query, task, task.execution_status);
task.execution_status = ExecutionStatus(ErrorCodes::NOT_IMPLEMENTED,
"Cannot execute replicated DDL query on leader");
return false;
}
return true;
}

17
dbms/src/Interpreters/DDLWorker.h
View File

@ -5,6 +5,7 @@
#include <Common/CurrentThread.h>
#include <Common/ThreadPool.h>
#include <common/logger_useful.h>
#include <Storages/IStorage.h>
#include <atomic>
#include <chrono>
@ -54,12 +55,22 @@ private:
/// Returns true and sets current_task if entry parsed and the check is passed
bool initAndCheckTask(const String & entry_name, String & out_reason, const ZooKeeperPtr & zookeeper);
void processTask(DDLTask & task, const ZooKeeperPtr & zookeeper);
void processTaskAlter(
/// Check that query should be executed on leader replica only
bool taskShouldBeExecutedOnLeader(const ASTPtr ast_ddl, StoragePtr storage) const;
/// Check that shard has consistent config with table
void checkShardConfig(const String & table, const DDLTask & taks, StoragePtr storage) const;
/// Executes query only on leader replica in case of replicated table.
/// Queries like TRUNCATE/ALTER .../OPTIMIZE have to be executed only on one node of shard.
/// Most of these queries can be executed on non-leader replica, but actually they still send
/// query via RemoteBlockOutputStream to leader, so to avoid such "2-phase" query execution we
/// execute query directly on leader.
bool tryExecuteQueryOnLeaderReplica(
DDLTask & task,
const ASTAlterQuery * ast_alter,
StoragePtr storage,
const String & rewritten_query,
const String & node_path,
const ZooKeeperPtr & zookeeper);

18
dbms/src/Interpreters/EmbeddedDictionaries.cpp
View File

@ -1,12 +1,10 @@
#include <Dictionaries/Embedded/RegionsHierarchies.h>
#include <Dictionaries/Embedded/RegionsNames.h>
#include <Dictionaries/Embedded/TechDataHierarchy.h>
#include <Dictionaries/Embedded/IGeoDictionariesLoader.h>
#include <Interpreters/Context.h>
#include <Interpreters/EmbeddedDictionaries.h>
#include <Common/setThreadName.h>
#include <Common/Exception.h>
#include "config_core.h"
#include <common/logger_useful.h>
#include <Poco/Util/Application.h>
@ -74,22 +72,6 @@ bool EmbeddedDictionaries::reloadImpl(const bool throw_on_error, const bool forc
bool was_exception = false;
#if USE_MYSQL
DictionaryReloader<TechDataHierarchy> reload_tech_data = [=] (const Poco::Util::AbstractConfiguration & config)
-> std::unique_ptr<TechDataHierarchy>
{
if (!TechDataHierarchy::isConfigured(config))
return {};
auto dictionary = std::make_unique<TechDataHierarchy>();
dictionary->reload();
return dictionary;
};
if (!reloadDictionary<TechDataHierarchy>(tech_data_hierarchy, reload_tech_data, throw_on_error, force_reload))
was_exception = true;
#endif
DictionaryReloader<RegionsHierarchies> reload_regions_hierarchies = [=] (const Poco::Util::AbstractConfiguration & config)
{
return geo_dictionaries_loader->reloadRegionsHierarchies(config);

7
dbms/src/Interpreters/EmbeddedDictionaries.h
View File

@ -10,7 +10,6 @@
namespace Poco { class Logger; namespace Util { class AbstractConfiguration; } }
class RegionsHierarchies;
class TechDataHierarchy;
class RegionsNames;
class IGeoDictionariesLoader;
@ -30,7 +29,6 @@ private:
Context & context;
MultiVersion<RegionsHierarchies> regions_hierarchies;
MultiVersion<TechDataHierarchy> tech_data_hierarchy;
MultiVersion<RegionsNames> regions_names;
std::unique_ptr<IGeoDictionariesLoader> geo_dictionaries_loader;
@ -85,11 +83,6 @@ public:
return regions_hierarchies.get();
}
MultiVersion<TechDataHierarchy>::Version getTechDataHierarchy() const
{
return tech_data_hierarchy.get();
}
MultiVersion<RegionsNames>::Version getRegionsNames() const
{
return regions_names.get();

50
dbms/src/Interpreters/ProcessList.cpp
View File

@ -87,10 +87,9 @@ ProcessList::EntryPtr ProcessList::insert(const String & query_, const IAST * as
{
std::unique_lock lock(mutex);
const auto max_wait_ms = settings.queue_max_wait_ms.totalMilliseconds();
if (!is_unlimited_query && max_size && processes.size() >= max_size)
{
auto max_wait_ms = settings.queue_max_wait_ms.totalMilliseconds();
if (!max_wait_ms || !have_space.wait_for(lock, std::chrono::milliseconds(max_wait_ms), [&]{ return processes.size() < max_size; }))
throw Exception("Too many simultaneous queries. Maximum: " + toString(max_size), ErrorCodes::TOO_MANY_SIMULTANEOUS_QUERIES);
}
@ -117,20 +116,41 @@ ProcessList::EntryPtr ProcessList::insert(const String & query_, const IAST * as
+ ", maximum: " + settings.max_concurrent_queries_for_user.toString(),
ErrorCodes::TOO_MANY_SIMULTANEOUS_QUERIES);
auto range = user_process_list->second.queries.equal_range(client_info.current_query_id);
if (range.first != range.second)
auto running_query = user_process_list->second.queries.find(client_info.current_query_id);
if (running_query != user_process_list->second.queries.end())
{
if (!settings.replace_running_query)
throw Exception("Query with id = " + client_info.current_query_id + " is already running.",
ErrorCodes::QUERY_WITH_SAME_ID_IS_ALREADY_RUNNING);
/// Ask queries to cancel. They will check this flag.
for (auto it = range.first; it != range.second; ++it)
it->second->is_killed.store(true, std::memory_order_relaxed);
}
running_query->second->is_killed.store(true, std::memory_order_relaxed);
if (!max_wait_ms || !have_space.wait_for(lock, std::chrono::milliseconds(max_wait_ms), [&]
{
running_query = user_process_list->second.queries.find(client_info.current_query_id);
if (running_query == user_process_list->second.queries.end())
return true;
running_query->second->is_killed.store(true, std::memory_order_relaxed);
return false;
}))
throw Exception("Query with id = " + client_info.current_query_id + " is already running and can't be stopped",
ErrorCodes::QUERY_WITH_SAME_ID_IS_ALREADY_RUNNING);
}
}
}
/// Check other users running query with our query_id
for (const auto & user_process_list : user_to_queries)
{
if (user_process_list.first == client_info.current_user)
continue;
if (auto running_query = user_process_list.second.queries.find(client_info.current_query_id); running_query != user_process_list.second.queries.end())
throw Exception("Query with id = " + client_info.current_query_id + " is already running by user " + user_process_list.first,
ErrorCodes::QUERY_WITH_SAME_ID_IS_ALREADY_RUNNING);
}
auto process_it = processes.emplace(processes.end(),
query_, client_info, settings.max_memory_usage, settings.memory_tracker_fault_probability, priorities.insert(settings.priority));
@ -226,17 +246,12 @@ ProcessListEntry::~ProcessListEntry()
bool found = false;
auto range = user_process_list.queries.equal_range(query_id);
if (range.first != range.second)
if (auto running_query = user_process_list.queries.find(query_id); running_query != user_process_list.queries.end())
{
for (auto jt = range.first; jt != range.second; ++jt)
if (running_query->second == process_list_element_ptr)
{
if (jt->second == process_list_element_ptr)
{
user_process_list.queries.erase(jt);
found = true;
break;
}
user_process_list.queries.erase(running_query->first);
found = true;
}
}
@ -245,8 +260,7 @@ ProcessListEntry::~ProcessListEntry()
LOG_ERROR(&Logger::get("ProcessList"), "Logical error: cannot find query by query_id and pointer to ProcessListElement in ProcessListForUser");
std::terminate();
}
parent.have_space.notify_one();
parent.have_space.notify_all();
/// If there are no more queries for the user, then we will reset memory tracker and network throttler.
if (user_process_list.queries.empty())

2
dbms/src/Interpreters/ProcessList.h
View File

@ -203,7 +203,7 @@ struct ProcessListForUser
ProcessListForUser();
/// query_id -> ProcessListElement(s). There can be multiple queries with the same query_id as long as all queries except one are cancelled.
using QueryToElement = std::unordered_multimap<String, QueryStatus *>;
using QueryToElement = std::unordered_map<String, QueryStatus *>;
QueryToElement queries;
ProfileEvents::Counters user_performance_counters{VariableContext::User, &ProfileEvents::global_counters};

2
dbms/src/Interpreters/Set.cpp
View File

@ -465,7 +465,7 @@ MergeTreeSetIndex::MergeTreeSetIndex(const Columns & set_elements, std::vector<K
* 1: the intersection of the set and the range is non-empty
* 2: the range contains elements not in the set
*/
BoolMask MergeTreeSetIndex::mayBeTrueInRange(const std::vector<Range> & key_ranges, const DataTypes & data_types)
BoolMask MergeTreeSetIndex::checkInRange(const std::vector<Range> & key_ranges, const DataTypes & data_types)
{
size_t tuple_size = indexes_mapping.size();

4
dbms/src/Interpreters/Set.h
View File

@ -167,7 +167,7 @@ using Sets = std::vector<SetPtr>;
class IFunction;
using FunctionPtr = std::shared_ptr<IFunction>;
/// Class for mayBeTrueInRange function.
/// Class for checkInRange function.
class MergeTreeSetIndex
{
public:
@ -185,7 +185,7 @@ public:
size_t size() const { return ordered_set.at(0)->size(); }
BoolMask mayBeTrueInRange(const std::vector<Range> & key_ranges, const DataTypes & data_types);
BoolMask checkInRange(const std::vector<Range> & key_ranges, const DataTypes & data_types);
private:
Columns ordered_set;

6
dbms/src/Interpreters/SystemLog.cpp
View File

@ -50,6 +50,12 @@ SystemLogs::SystemLogs(Context & global_context, const Poco::Util::AbstractConfi
SystemLogs::~SystemLogs()
{
shutdown();
}
void SystemLogs::shutdown()
{
if (query_log)
query_log->shutdown();

107
dbms/src/Interpreters/SystemLog.h
View File

@ -2,6 +2,7 @@
#include <thread>
#include <atomic>
#include <condition_variable>
#include <boost/noncopyable.hpp>
#include <common/logger_useful.h>
#include <Core/Types.h>
@ -67,6 +68,8 @@ struct SystemLogs
SystemLogs(Context & global_context, const Poco::Util::AbstractConfiguration & config);
~SystemLogs();
void shutdown();
std::shared_ptr<QueryLog> query_log; /// Used to log queries.
std::shared_ptr<QueryThreadLog> query_thread_log; /// Used to log query threads.
std::shared_ptr<PartLog> part_log; /// Used to log operations with parts
@ -101,22 +104,10 @@ public:
/** Append a record into log.
* Writing to table will be done asynchronously and in case of failure, record could be lost.
*/
void add(const LogElement & element)
{
if (is_shutdown)
return;
/// Without try we could block here in case of queue overflow.
if (!queue.tryPush({false, element}))
LOG_ERROR(log, "SystemLog queue is full");
}
void add(const LogElement & element);
/// Flush data in the buffer to disk
void flush()
{
if (!is_shutdown)
flushImpl(false);
}
void flush();
/// Stop the background flush thread before destructor. No more data will be written.
void shutdown();
@ -130,7 +121,15 @@ protected:
const size_t flush_interval_milliseconds;
std::atomic<bool> is_shutdown{false};
using QueueItem = std::pair<bool, LogElement>; /// First element is shutdown flag for thread.
enum class EntryType
{
LOG_ELEMENT = 0,
AUTO_FLUSH,
FORCE_FLUSH,
SHUTDOWN,
};
using QueueItem = std::pair<EntryType, LogElement>;
/// Queue is bounded. But its size is quite large to not block in all normal cases.
ConcurrentBoundedQueue<QueueItem> queue {DBMS_SYSTEM_LOG_QUEUE_SIZE};
@ -140,7 +139,6 @@ protected:
* than accumulation of large amount of log records (for example, for query log - processing of large amount of queries).
*/
std::vector<LogElement> data;
std::mutex data_mutex;
Logger * log;
@ -157,7 +155,13 @@ protected:
bool is_prepared = false;
void prepareTable();
void flushImpl(bool quiet);
std::mutex flush_mutex;
std::mutex condvar_mutex;
std::condition_variable flush_condvar;
bool force_flushing = false;
/// flushImpl can be executed only in saving_thread.
void flushImpl(EntryType reason);
};
@ -178,6 +182,37 @@ SystemLog<LogElement>::SystemLog(Context & context_,
}
template <typename LogElement>
void SystemLog<LogElement>::add(const LogElement & element)
{
if (is_shutdown)
return;
/// Without try we could block here in case of queue overflow.
if (!queue.tryPush({EntryType::LOG_ELEMENT, element}))
LOG_ERROR(log, "SystemLog queue is full");
}
template <typename LogElement>
void SystemLog<LogElement>::flush()
{
if (is_shutdown)
return;
std::lock_guard flush_lock(flush_mutex);
force_flushing = true;
/// Tell thread to execute extra flush.
queue.push({EntryType::FORCE_FLUSH, {}});
/// Wait for flush being finished.
std::unique_lock lock(condvar_mutex);
while (force_flushing)
flush_condvar.wait(lock);
}
template <typename LogElement>
void SystemLog<LogElement>::shutdown()
{
@ -186,7 +221,7 @@ void SystemLog<LogElement>::shutdown()
return;
/// Tell thread to shutdown.
queue.push({true, {}});
queue.push({EntryType::SHUTDOWN, {}});
saving_thread.join();
}
@ -219,16 +254,10 @@ void SystemLog<LogElement>::threadFunction()
QueueItem element;
bool has_element = false;
bool is_empty;
{
std::unique_lock lock(data_mutex);
is_empty = data.empty();
}
/// data.size() is increased only in this function
/// TODO: get rid of data and queue duality
if (is_empty)
if (data.empty())
{
queue.pop(element);
has_element = true;
@ -242,25 +271,27 @@ void SystemLog<LogElement>::threadFunction()
if (has_element)
{
if (element.first)
if (element.first == EntryType::SHUTDOWN)
{
/// Shutdown.
/// NOTE: MergeTree engine can write data even it is already in shutdown state.
flush();
flushImpl(element.first);
break;
}
else
else if (element.first == EntryType::FORCE_FLUSH)
{
std::unique_lock lock(data_mutex);
data.push_back(element.second);
flushImpl(element.first);
time_after_last_write.restart();
continue;
}
else
data.push_back(element.second);
}
size_t milliseconds_elapsed = time_after_last_write.elapsed() / 1000000;
if (milliseconds_elapsed >= flush_interval_milliseconds)
{
/// Write data to a table.
flushImpl(true);
flushImpl(EntryType::AUTO_FLUSH);
time_after_last_write.restart();
}
}
@ -275,13 +306,11 @@ void SystemLog<LogElement>::threadFunction()
template <typename LogElement>
void SystemLog<LogElement>::flushImpl(bool quiet)
void SystemLog<LogElement>::flushImpl(EntryType reason)
{
std::unique_lock lock(data_mutex);
try
{
if (quiet && data.empty())
if ((reason == EntryType::AUTO_FLUSH || reason == EntryType::SHUTDOWN) && data.empty())
return;
LOG_TRACE(log, "Flushing system log");
@ -320,6 +349,12 @@ void SystemLog<LogElement>::flushImpl(bool quiet)
/// In case of exception, also clean accumulated data - to avoid locking.
data.clear();
}
if (reason == EntryType::FORCE_FLUSH)
{
std::lock_guard lock(condvar_mutex);
force_flushing = false;
flush_condvar.notify_one();
}
}

2
dbms/src/Storages/IStorage.cpp
View File

@ -157,7 +157,7 @@ void IStorage::check(const Names & column_names) const
{
if (columns_map.end() == columns_map.find(name))
throw Exception(
"There is no column with name " + name + " in table. There are columns: " + list_of_columns,
"There is no column with name " + backQuote(name) + " in table " + getTableName() + ". There are columns: " + list_of_columns,
ErrorCodes::NO_SUCH_COLUMN_IN_TABLE);
if (unique_names.end() != unique_names.find(name))

39
dbms/src/Storages/MergeTree/KeyCondition.cpp
View File

@ -810,7 +810,7 @@ String KeyCondition::toString() const
*/
template <typename F>
static bool forAnyParallelogram(
static BoolMask forAnyParallelogram(
size_t key_size,
const Field * key_left,
const Field * key_right,
@ -866,16 +866,15 @@ static bool forAnyParallelogram(
for (size_t i = prefix_size + 1; i < key_size; ++i)
parallelogram[i] = Range();
if (callback(parallelogram))
return true;
BoolMask result(false, false);
result = result | callback(parallelogram);
/// [x1] x [y1 .. +inf)
if (left_bounded)
{
parallelogram[prefix_size] = Range(key_left[prefix_size]);
if (forAnyParallelogram(key_size, key_left, key_right, true, false, parallelogram, prefix_size + 1, callback))
return true;
result = result | forAnyParallelogram(key_size, key_left, key_right, true, false, parallelogram, prefix_size + 1, callback);
}
/// [x2] x (-inf .. y2]
@ -883,15 +882,14 @@ static bool forAnyParallelogram(
if (right_bounded)
{
parallelogram[prefix_size] = Range(key_right[prefix_size]);
if (forAnyParallelogram(key_size, key_left, key_right, false, true, parallelogram, prefix_size + 1, callback))
return true;
result = result | forAnyParallelogram(key_size, key_left, key_right, false, true, parallelogram, prefix_size + 1, callback);
}
return false;
return result;
}
bool KeyCondition::mayBeTrueInRange(
BoolMask KeyCondition::checkInRange(
size_t used_key_size,
const Field * left_key,
const Field * right_key,
@ -917,7 +915,7 @@ bool KeyCondition::mayBeTrueInRange(
return forAnyParallelogram(used_key_size, left_key, right_key, true, right_bounded, key_ranges, 0,
[&] (const std::vector<Range> & key_ranges_parallelogram)
{
auto res = mayBeTrueInParallelogram(key_ranges_parallelogram, data_types);
auto res = checkInParallelogram(key_ranges_parallelogram, data_types);
/* std::cerr << "Parallelogram: ";
for (size_t i = 0, size = key_ranges.size(); i != size; ++i)
@ -928,11 +926,11 @@ bool KeyCondition::mayBeTrueInRange(
});
}
std::optional<Range> KeyCondition::applyMonotonicFunctionsChainToRange(
Range key_range,
MonotonicFunctionsChain & functions,
DataTypePtr current_type
)
DataTypePtr current_type)
{
for (auto & func : functions)
{
@ -965,7 +963,7 @@ std::optional<Range> KeyCondition::applyMonotonicFunctionsChainToRange(
return key_range;
}
bool KeyCondition::mayBeTrueInParallelogram(const std::vector<Range> & parallelogram, const DataTypes & data_types) const
BoolMask KeyCondition::checkInParallelogram(const std::vector<Range> & parallelogram, const DataTypes & data_types) const
{
std::vector<BoolMask> rpn_stack;
for (size_t i = 0; i < rpn.size(); ++i)
@ -1013,7 +1011,7 @@ bool KeyCondition::mayBeTrueInParallelogram(const std::vector<Range> & parallelo
if (!element.set_index)
throw Exception("Set for IN is not created yet", ErrorCodes::LOGICAL_ERROR);
rpn_stack.emplace_back(element.set_index->mayBeTrueInRange(parallelogram, data_types));
rpn_stack.emplace_back(element.set_index->checkInRange(parallelogram, data_types));
if (element.function == RPNElement::FUNCTION_NOT_IN_SET)
rpn_stack.back() = !rpn_stack.back();
}
@ -1048,22 +1046,23 @@ bool KeyCondition::mayBeTrueInParallelogram(const std::vector<Range> & parallelo
}
if (rpn_stack.size() != 1)
throw Exception("Unexpected stack size in KeyCondition::mayBeTrueInRange", ErrorCodes::LOGICAL_ERROR);
throw Exception("Unexpected stack size in KeyCondition::checkInRange", ErrorCodes::LOGICAL_ERROR);
return rpn_stack[0].can_be_true;
return rpn_stack[0];
}
bool KeyCondition::mayBeTrueInRange(
BoolMask KeyCondition::checkInRange(
size_t used_key_size, const Field * left_key, const Field * right_key, const DataTypes & data_types) const
{
return mayBeTrueInRange(used_key_size, left_key, right_key, data_types, true);
return checkInRange(used_key_size, left_key, right_key, data_types, true);
}
bool KeyCondition::mayBeTrueAfter(
BoolMask KeyCondition::getMaskAfter(
size_t used_key_size, const Field * left_key, const DataTypes & data_types) const
{
return mayBeTrueInRange(used_key_size, left_key, nullptr, data_types, false);
return checkInRange(used_key_size, left_key, nullptr, data_types, false);
}

14
dbms/src/Storages/MergeTree/KeyCondition.h
View File

@ -235,17 +235,17 @@ public:
const Names & key_column_names,
const ExpressionActionsPtr & key_expr);
/// Whether the condition is feasible in the key range.
/// Whether the condition and its negation are (independently) feasible in the key range.
/// left_key and right_key must contain all fields in the sort_descr in the appropriate order.
/// data_types - the types of the key columns.
bool mayBeTrueInRange(size_t used_key_size, const Field * left_key, const Field * right_key, const DataTypes & data_types) const;
BoolMask checkInRange(size_t used_key_size, const Field * left_key, const Field * right_key, const DataTypes & data_types) const;
/// Whether the condition is feasible in the direct product of single column ranges specified by `parallelogram`.
bool mayBeTrueInParallelogram(const std::vector<Range> & parallelogram, const DataTypes & data_types) const;
/// Whether the condition and its negation are feasible in the direct product of single column ranges specified by `parallelogram`.
BoolMask checkInParallelogram(const std::vector<Range> & parallelogram, const DataTypes & data_types) const;
/// Is the condition valid in a semi-infinite (not limited to the right) key range.
/// Are the condition and its negation valid in a semi-infinite (not limited to the right) key range.
/// left_key must contain all the fields in the sort_descr in the appropriate order.
bool mayBeTrueAfter(size_t used_key_size, const Field * left_key, const DataTypes & data_types) const;
BoolMask getMaskAfter(size_t used_key_size, const Field * left_key, const DataTypes & data_types) const;
/// Checks that the index can not be used.
bool alwaysUnknownOrTrue() const;
@ -330,7 +330,7 @@ public:
static const AtomMap atom_map;
private:
bool mayBeTrueInRange(
BoolMask checkInRange(
size_t used_key_size,
const Field * left_key,
const Field * right_key,

16
dbms/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp
View File

@ -265,8 +265,8 @@ BlockInputStreams MergeTreeDataSelectExecutor::readFromParts(
if (part->isEmpty())
continue;
if (minmax_idx_condition && !minmax_idx_condition->mayBeTrueInParallelogram(
part->minmax_idx.parallelogram, data.minmax_idx_column_types))
if (minmax_idx_condition && !minmax_idx_condition->checkInParallelogram(
part->minmax_idx.parallelogram, data.minmax_idx_column_types).can_be_true)
continue;
if (max_block_numbers_to_read)
@ -518,7 +518,7 @@ BlockInputStreams MergeTreeDataSelectExecutor::readFromParts(
RangesInDataParts parts_with_ranges;
std::vector<std::pair<MergeTreeIndexPtr, IndexConditionPtr>> useful_indices;
std::vector<std::pair<MergeTreeIndexPtr, MergeTreeIndexConditionPtr>> useful_indices;
for (const auto & index : data.skip_indices)
{
auto condition = index->createIndexCondition(query_info, context);
@ -950,8 +950,8 @@ MarkRanges MergeTreeDataSelectExecutor::markRangesFromPKRange(
for (size_t i = 0; i < used_key_size; ++i)
index[i]->get(range.begin, index_left[i]);
may_be_true = key_condition.mayBeTrueAfter(
used_key_size, index_left.data(), data.primary_key_data_types);
may_be_true = key_condition.getMaskAfter(
used_key_size, index_left.data(), data.primary_key_data_types).can_be_true;
}
else
{
@ -964,8 +964,8 @@ MarkRanges MergeTreeDataSelectExecutor::markRangesFromPKRange(
index[i]->get(range.end, index_right[i]);
}
may_be_true = key_condition.mayBeTrueInRange(
used_key_size, index_left.data(), index_right.data(), data.primary_key_data_types);
may_be_true = key_condition.checkInRange(
used_key_size, index_left.data(), index_right.data(), data.primary_key_data_types).can_be_true;
}
if (!may_be_true)
@ -998,7 +998,7 @@ MarkRanges MergeTreeDataSelectExecutor::markRangesFromPKRange(
MarkRanges MergeTreeDataSelectExecutor::filterMarksUsingIndex(
MergeTreeIndexPtr index,
IndexConditionPtr condition,
MergeTreeIndexConditionPtr condition,
MergeTreeData::DataPartPtr part,
const MarkRanges & ranges,
const Settings & settings) const

2
dbms/src/Storages/MergeTree/MergeTreeDataSelectExecutor.h
View File

@ -84,7 +84,7 @@ private:
MarkRanges filterMarksUsingIndex(
MergeTreeIndexPtr index,
IndexConditionPtr condition,
MergeTreeIndexConditionPtr condition,
MergeTreeData::DataPartPtr part,
const MarkRanges & ranges,
const Settings & settings) const;

62
dbms/src/Storages/MergeTree/MergeTreeIndexAggregatorBloomFilter.cpp Normal file
View File

@ -0,0 +1,62 @@
#include <Storages/MergeTree/MergeTreeIndexAggregatorBloomFilter.h>
#include <ext/bit_cast.h>
#include <Columns/ColumnString.h>
#include <Columns/ColumnsNumber.h>
#include <Columns/ColumnFixedString.h>
#include <Common/HashTable/Hash.h>
#include <DataTypes/DataTypesNumber.h>
#include <Interpreters/BloomFilterHash.h>
namespace DB
{
namespace ErrorCodes
{
extern const int LOGICAL_ERROR;
extern const int ILLEGAL_COLUMN;
}
MergeTreeIndexAggregatorBloomFilter::MergeTreeIndexAggregatorBloomFilter(
size_t bits_per_row_, size_t hash_functions_, const Names & columns_name_)
: bits_per_row(bits_per_row_), hash_functions(hash_functions_), index_columns_name(columns_name_)
{
}
bool MergeTreeIndexAggregatorBloomFilter::empty() const
{
return !total_rows;
}
MergeTreeIndexGranulePtr MergeTreeIndexAggregatorBloomFilter::getGranuleAndReset()
{
const auto granule = std::make_shared<MergeTreeIndexGranuleBloomFilter>(bits_per_row, hash_functions, total_rows, granule_index_blocks);
total_rows = 0;
granule_index_blocks.clear();
return granule;
}
void MergeTreeIndexAggregatorBloomFilter::update(const Block & block, size_t * pos, size_t limit)
{
if (*pos >= block.rows())
throw Exception("The provided position is not less than the number of block rows. Position: " + toString(*pos) + ", Block rows: " +
toString(block.rows()) + ".", ErrorCodes::LOGICAL_ERROR);
Block granule_index_block;
size_t max_read_rows = std::min(block.rows() - *pos, limit);
for (size_t index = 0; index < index_columns_name.size(); ++index)
{
const auto & column_and_type = block.getByName(index_columns_name[index]);
const auto & index_column = BloomFilterHash::hashWithColumn(column_and_type.type, column_and_type.column, *pos, max_read_rows);
granule_index_block.insert({std::move(index_column), std::make_shared<DataTypeUInt64>(), column_and_type.name});
}
*pos += max_read_rows;
total_rows += max_read_rows;
granule_index_blocks.push_back(granule_index_block);
}
}

29
dbms/src/Storages/MergeTree/MergeTreeIndexAggregatorBloomFilter.h Normal file
View File

@ -0,0 +1,29 @@
#pragma once
#include <Storages/MergeTree/MergeTreeIndices.h>
#include <Storages/MergeTree/MergeTreeIndexGranuleBloomFilter.h>
namespace DB
{
class MergeTreeIndexAggregatorBloomFilter : public IMergeTreeIndexAggregator
{
public:
MergeTreeIndexAggregatorBloomFilter(size_t bits_per_row_, size_t hash_functions_, const Names & columns_name_);
bool empty() const override;
MergeTreeIndexGranulePtr getGranuleAndReset() override;
void update(const Block & block, size_t * pos, size_t limit) override;
private:
size_t bits_per_row;
size_t hash_functions;
const Names index_columns_name;
size_t total_rows = 0;
Blocks granule_index_blocks;
};
}

110
dbms/src/Storages/MergeTree/MergeTreeIndexBloomFilter.cpp Normal file
View File

@ -0,0 +1,110 @@
#include <Storages/MergeTree/MergeTreeIndexBloomFilter.h>
#include <Storages/MergeTree/MergeTreeData.h>
#include <Interpreters/SyntaxAnalyzer.h>
#include <Interpreters/ExpressionAnalyzer.h>
#include <Core/Types.h>
#include <ext/bit_cast.h>
#include <Parsers/ASTLiteral.h>
#include <IO/ReadHelpers.h>
#include <IO/WriteHelpers.h>
#include <DataTypes/DataTypeNullable.h>
#include <Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h>
#include <Parsers/queryToString.h>
#include <Columns/ColumnConst.h>
#include <Interpreters/BloomFilterHash.h>
namespace DB
{
namespace ErrorCodes
{
extern const int LOGICAL_ERROR;
extern const int INCORRECT_QUERY;
}
MergeTreeIndexBloomFilter::MergeTreeIndexBloomFilter(
const String & name_, const ExpressionActionsPtr & expr_, const Names & columns_, const DataTypes & data_types_, const Block & header_,
size_t granularity_, size_t bits_per_row_, size_t hash_functions_)
: IMergeTreeIndex(name_, expr_, columns_, data_types_, header_, granularity_), bits_per_row(bits_per_row_),
hash_functions(hash_functions_)
{
}
MergeTreeIndexGranulePtr MergeTreeIndexBloomFilter::createIndexGranule() const
{
return std::make_shared<MergeTreeIndexGranuleBloomFilter>(bits_per_row, hash_functions, columns.size());
}
bool MergeTreeIndexBloomFilter::mayBenefitFromIndexForIn(const ASTPtr & node) const
{
const String & column_name = node->getColumnName();
for (const auto & name : columns)
if (column_name == name)
return true;
if (const auto * func = typeid_cast<const ASTFunction *>(node.get()))
{
for (const auto & children : func->arguments->children)
if (mayBenefitFromIndexForIn(children))
return true;
}
return false;
}
MergeTreeIndexAggregatorPtr MergeTreeIndexBloomFilter::createIndexAggregator() const
{
return std::make_shared<MergeTreeIndexAggregatorBloomFilter>(bits_per_row, hash_functions, columns);
}
MergeTreeIndexConditionPtr MergeTreeIndexBloomFilter::createIndexCondition(const SelectQueryInfo & query_info, const Context & context) const
{
return std::make_shared<MergeTreeIndexConditionBloomFilter>(query_info, context, header, hash_functions);
}
static void assertIndexColumnsType(const Block & header)
{
if (!header || !header.columns())
throw Exception("Index must have columns.", ErrorCodes::INCORRECT_QUERY);
const DataTypes & columns_data_types = header.getDataTypes();
for (size_t index = 0; index < columns_data_types.size(); ++index)
{
WhichDataType which(columns_data_types[index]);
if (!which.isUInt() && !which.isInt() && !which.isString() && !which.isFixedString() && !which.isFloat() &&
!which.isDateOrDateTime() && !which.isEnum())
throw Exception("Unexpected type " + columns_data_types[index]->getName() + " of bloom filter index.",
ErrorCodes::ILLEGAL_COLUMN);
}
}
std::unique_ptr<IMergeTreeIndex> bloomFilterIndexCreatorNew(
const NamesAndTypesList & columns, std::shared_ptr<ASTIndexDeclaration> node, const Context & context)
{
if (node->name.empty())
throw Exception("Index must have unique name.", ErrorCodes::INCORRECT_QUERY);
ASTPtr expr_list = MergeTreeData::extractKeyExpressionList(node->expr->clone());
auto syntax = SyntaxAnalyzer(context, {}).analyze(expr_list, columns);
auto index_expr = ExpressionAnalyzer(expr_list, syntax, context).getActions(false);
auto index_sample = ExpressionAnalyzer(expr_list, syntax, context).getActions(true)->getSampleBlock();
assertIndexColumnsType(index_sample);
double max_conflict_probability = 0.025;
if (node->type->arguments && !node->type->arguments->children.empty())
max_conflict_probability = typeid_cast<const ASTLiteral &>(*node->type->arguments->children[0]).value.get<Float64>();
const auto & bits_per_row_and_size_of_hash_functions = BloomFilterHash::calculationBestPractices(max_conflict_probability);
return std::make_unique<MergeTreeIndexBloomFilter>(
node->name, std::move(index_expr), index_sample.getNames(), index_sample.getDataTypes(), index_sample, node->granularity,
bits_per_row_and_size_of_hash_functions.first, bits_per_row_and_size_of_hash_functions.second);
}
}

31
dbms/src/Storages/MergeTree/MergeTreeIndexBloomFilter.h Normal file
View File

@ -0,0 +1,31 @@
#pragma once
#include <Interpreters/BloomFilter.h>
#include <Storages/MergeTree/MergeTreeIndices.h>
#include <Storages/MergeTree/MergeTreeIndexGranuleBloomFilter.h>
#include <Storages/MergeTree/MergeTreeIndexAggregatorBloomFilter.h>
namespace DB
{
class MergeTreeIndexBloomFilter : public IMergeTreeIndex
{
public:
MergeTreeIndexBloomFilter(
const String & name_, const ExpressionActionsPtr & expr_, const Names & columns_, const DataTypes & data_types_,
const Block & header_, size_t granularity_, size_t bits_per_row_, size_t hash_functions_);
MergeTreeIndexGranulePtr createIndexGranule() const override;
MergeTreeIndexAggregatorPtr createIndexAggregator() const override;
MergeTreeIndexConditionPtr createIndexCondition(const SelectQueryInfo & query_info, const Context & context) const override;
bool mayBenefitFromIndexForIn(const ASTPtr & node) const override;
private:
size_t bits_per_row;
size_t hash_functions;
};
}

352
dbms/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp Normal file
View File

@ -0,0 +1,352 @@
#include <Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h>
#include <Interpreters/QueryNormalizer.h>
#include <Interpreters/BloomFilterHash.h>
#include <Common/HashTable/ClearableHashMap.h>
#include <Storages/MergeTree/RPNBuilder.h>
#include <Storages/MergeTree/MergeTreeIndexGranuleBloomFilter.h>
#include <DataTypes/DataTypeTuple.h>
#include <Columns/ColumnConst.h>
#include <ext/bit_cast.h>
#include <Parsers/ASTSubquery.h>
#include <Parsers/ASTIdentifier.h>
#include <Columns/ColumnTuple.h>
#include <Interpreters/castColumn.h>
#include <Interpreters/convertFieldToType.h>
namespace DB
{
namespace
{
PreparedSetKey getPreparedSetKey(const ASTPtr & node, const DataTypePtr & data_type)
{
/// If the data type is tuple, let's try unbox once
if (node->as<ASTSubquery>() || node->as<ASTIdentifier>())
return PreparedSetKey::forSubquery(*node);
if (const auto * date_type_tuple = typeid_cast<const DataTypeTuple *>(&*data_type))
return PreparedSetKey::forLiteral(*node, date_type_tuple->getElements());
return PreparedSetKey::forLiteral(*node, DataTypes(1, data_type));
}
ColumnWithTypeAndName getPreparedSetInfo(const SetPtr & prepared_set)
{
if (prepared_set->getDataTypes().size() == 1)
return {prepared_set->getSetElements()[0], prepared_set->getDataTypes()[0], "dummy"};
return {ColumnTuple::create(prepared_set->getSetElements()), std::make_shared<DataTypeTuple>(prepared_set->getDataTypes()), "dummy"};
}
bool maybeTrueOnBloomFilter(const IColumn * hash_column, const BloomFilterPtr & bloom_filter, size_t hash_functions)
{
const auto const_column = typeid_cast<const ColumnConst *>(hash_column);
const auto non_const_column = typeid_cast<const ColumnUInt64 *>(hash_column);
if (!const_column && !non_const_column)
throw Exception("LOGICAL ERROR: hash column must be Const Column or UInt64 Column.", ErrorCodes::LOGICAL_ERROR);
if (const_column)
{
for (size_t index = 0; index < hash_functions; ++index)
if (!bloom_filter->findHashWithSeed(const_column->getValue<UInt64>(), BloomFilterHash::bf_hash_seed[index]))
return false;
return true;
}
else
{
bool missing_rows = true;
const ColumnUInt64::Container & data = non_const_column->getData();
for (size_t index = 0, size = data.size(); missing_rows && index < size; ++index)
{
bool match_row = true;
for (size_t hash_index = 0; match_row && hash_index < hash_functions; ++hash_index)
match_row = bloom_filter->findHashWithSeed(data[index], BloomFilterHash::bf_hash_seed[hash_index]);
missing_rows = !match_row;
}
return !missing_rows;
}
}
}
MergeTreeIndexConditionBloomFilter::MergeTreeIndexConditionBloomFilter(
const SelectQueryInfo & info, const Context & context, const Block & header, size_t hash_functions)
: header(header), context(context), query_info(info), hash_functions(hash_functions)
{
auto atomFromAST = [this](auto & node, auto &, auto & constants, auto & out) { return traverseAtomAST(node, constants, out); };
rpn = std::move(RPNBuilder<RPNElement>(info, context, atomFromAST).extractRPN());
}
bool MergeTreeIndexConditionBloomFilter::alwaysUnknownOrTrue() const
{
std::vector<bool> rpn_stack;
for (const auto & element : rpn)
{
if (element.function == RPNElement::FUNCTION_UNKNOWN
|| element.function == RPNElement::ALWAYS_TRUE)
{
rpn_stack.push_back(true);
}
else if (element.function == RPNElement::FUNCTION_EQUALS
|| element.function == RPNElement::FUNCTION_NOT_EQUALS
|| element.function == RPNElement::FUNCTION_IN
|| element.function == RPNElement::FUNCTION_NOT_IN
|| element.function == RPNElement::ALWAYS_FALSE)
{
rpn_stack.push_back(false);
}
else if (element.function == RPNElement::FUNCTION_NOT)
{
// do nothing
}
else if (element.function == RPNElement::FUNCTION_AND)
{
auto arg1 = rpn_stack.back();
rpn_stack.pop_back();
auto arg2 = rpn_stack.back();
rpn_stack.back() = arg1 && arg2;
}
else if (element.function == RPNElement::FUNCTION_OR)
{
auto arg1 = rpn_stack.back();
rpn_stack.pop_back();
auto arg2 = rpn_stack.back();
rpn_stack.back() = arg1 || arg2;
}
else
throw Exception("Unexpected function type in KeyCondition::RPNElement", ErrorCodes::LOGICAL_ERROR);
}
return rpn_stack[0];
}
bool MergeTreeIndexConditionBloomFilter::mayBeTrueOnGranule(const MergeTreeIndexGranuleBloomFilter * granule) const
{
std::vector<BoolMask> rpn_stack;
const auto & filters = granule->getFilters();
for (const auto & element : rpn)
{
if (element.function == RPNElement::FUNCTION_UNKNOWN)
{
rpn_stack.emplace_back(true, true);
}
else if (element.function == RPNElement::FUNCTION_IN
|| element.function == RPNElement::FUNCTION_NOT_IN
|| element.function == RPNElement::FUNCTION_EQUALS
|| element.function == RPNElement::FUNCTION_NOT_EQUALS)
{
bool match_rows = true;
const auto & predicate = element.predicate;
for (size_t index = 0; match_rows && index < predicate.size(); ++index)
{
const auto & query_index_hash = predicate[index];
const auto & filter = filters[query_index_hash.first];
const ColumnPtr & hash_column = query_index_hash.second;
match_rows = maybeTrueOnBloomFilter(&*hash_column, filter, hash_functions);
}
rpn_stack.emplace_back(match_rows, !match_rows);
if (element.function == RPNElement::FUNCTION_NOT_EQUALS || element.function == RPNElement::FUNCTION_NOT_IN)
rpn_stack.back() = !rpn_stack.back();
}
else if (element.function == RPNElement::FUNCTION_NOT)
{
rpn_stack.back() = !rpn_stack.back();
}
else if (element.function == RPNElement::FUNCTION_OR)
{
auto arg1 = rpn_stack.back();
rpn_stack.pop_back();
auto arg2 = rpn_stack.back();
rpn_stack.back() = arg1 | arg2;
}
else if (element.function == RPNElement::FUNCTION_AND)
{
auto arg1 = rpn_stack.back();
rpn_stack.pop_back();
auto arg2 = rpn_stack.back();
rpn_stack.back() = arg1 & arg2;
}
else if (element.function == RPNElement::ALWAYS_TRUE)
{
rpn_stack.emplace_back(true, false);
}
else if (element.function == RPNElement::ALWAYS_FALSE)
{
rpn_stack.emplace_back(false, true);
}
else
throw Exception("Unexpected function type in KeyCondition::RPNElement", ErrorCodes::LOGICAL_ERROR);
}
if (rpn_stack.size() != 1)
throw Exception("Unexpected stack size in KeyCondition::mayBeTrueInRange", ErrorCodes::LOGICAL_ERROR);
return rpn_stack[0].can_be_true;
}
bool MergeTreeIndexConditionBloomFilter::traverseAtomAST(const ASTPtr & node, Block & block_with_constants, RPNElement & out)
{
{
Field const_value;
DataTypePtr const_type;
if (KeyCondition::getConstant(node, block_with_constants, const_value, const_type))
{
if (const_value.getType() == Field::Types::UInt64 || const_value.getType() == Field::Types::Int64 ||
const_value.getType() == Field::Types::Float64)
{
/// Zero in all types is represented in memory the same way as in UInt64.
out.function = const_value.get<UInt64>() ? RPNElement::ALWAYS_TRUE : RPNElement::ALWAYS_FALSE;
return true;
}
}
}
if (const auto * function = node->as<ASTFunction>())
{
const ASTs & arguments = function->arguments->children;
if (arguments.size() != 2)
return false;
if (functionIsInOrGlobalInOperator(function->name))
{
if (const auto & prepared_set = getPreparedSet(arguments[1]))
return traverseASTIn(function->name, arguments[0], prepared_set, out);
}
else if (function->name == "equals" || function->name == "notEquals")
{
Field const_value;
DataTypePtr const_type;
if (KeyCondition::getConstant(arguments[1], block_with_constants, const_value, const_type))
return traverseASTEquals(function->name, arguments[0], const_type, const_value, out);
else if (KeyCondition::getConstant(arguments[0], block_with_constants, const_value, const_type))
return traverseASTEquals(function->name, arguments[1], const_type, const_value, out);
}
}
return false;
}
bool MergeTreeIndexConditionBloomFilter::traverseASTIn(
const String & function_name, const ASTPtr & key_ast, const SetPtr & prepared_set, RPNElement & out)
{
const auto & prepared_info = getPreparedSetInfo(prepared_set);
return traverseASTIn(function_name, key_ast, prepared_info.type, prepared_info.column, out);
}
bool MergeTreeIndexConditionBloomFilter::traverseASTIn(
const String & function_name, const ASTPtr & key_ast, const DataTypePtr & type, const ColumnPtr & column, RPNElement & out)
{
if (header.has(key_ast->getColumnName()))
{
size_t row_size = column->size();
size_t position = header.getPositionByName(key_ast->getColumnName());
const DataTypePtr & index_type = header.getByPosition(position).type;
const auto & converted_column = castColumn(ColumnWithTypeAndName{column, type, ""}, index_type, context);
out.predicate.emplace_back(std::make_pair(position, BloomFilterHash::hashWithColumn(index_type, converted_column, 0, row_size)));
if (function_name == "in" || function_name == "globalIn")
out.function = RPNElement::FUNCTION_IN;
if (function_name == "notIn" || function_name == "globalNotIn")
out.function = RPNElement::FUNCTION_NOT_IN;
return true;
}
if (const auto * function = key_ast->as<ASTFunction>())
{
WhichDataType which(type);
if (which.isTuple() && function->name == "tuple")
{
const auto & tuple_column = typeid_cast<const ColumnTuple *>(column.get());
const auto & tuple_data_type = typeid_cast<const DataTypeTuple *>(type.get());
const ASTs & arguments = typeid_cast<const ASTExpressionList &>(*function->arguments).children;
if (tuple_data_type->getElements().size() != arguments.size() || tuple_column->getColumns().size() != arguments.size())
throw Exception("Illegal types of arguments of function " + function_name, ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
bool match_with_subtype = false;
const auto & sub_columns = tuple_column->getColumns();
const auto & sub_data_types = tuple_data_type->getElements();
for (size_t index = 0; index < arguments.size(); ++index)
match_with_subtype |= traverseASTIn(function_name, arguments[index], sub_data_types[index], sub_columns[index], out);
return match_with_subtype;
}
}
return false;
}
bool MergeTreeIndexConditionBloomFilter::traverseASTEquals(
const String & function_name, const ASTPtr & key_ast, const DataTypePtr & value_type, const Field & value_field, RPNElement & out)
{
if (header.has(key_ast->getColumnName()))
{
size_t position = header.getPositionByName(key_ast->getColumnName());
const DataTypePtr & index_type = header.getByPosition(position).type;
Field converted_field = convertFieldToType(value_field, *index_type, &*value_type);
out.predicate.emplace_back(std::make_pair(position, BloomFilterHash::hashWithField(&*index_type, converted_field)));
out.function = function_name == "equals" ? RPNElement::FUNCTION_EQUALS : RPNElement::FUNCTION_NOT_EQUALS;
return true;
}
if (const auto * function = key_ast->as<ASTFunction>())
{
WhichDataType which(value_type);
if (which.isTuple() && function->name == "tuple")
{
const TupleBackend & tuple = get<const Tuple &>(value_field).toUnderType();
const auto value_tuple_data_type = typeid_cast<const DataTypeTuple *>(value_type.get());
const ASTs & arguments = typeid_cast<const ASTExpressionList &>(*function->arguments).children;
if (tuple.size() != arguments.size())
throw Exception("Illegal types of arguments of function " + function_name, ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
bool match_with_subtype = false;
const DataTypes & subtypes = value_tuple_data_type->getElements();
for (size_t index = 0; index < tuple.size(); ++index)
match_with_subtype |= traverseASTEquals(function_name, arguments[index], subtypes[index], tuple[index], out);
return match_with_subtype;
}
}
return false;
}
SetPtr MergeTreeIndexConditionBloomFilter::getPreparedSet(const ASTPtr & node)
{
if (header.has(node->getColumnName()))
{
const auto & column_and_type = header.getByName(node->getColumnName());
const auto & prepared_set_it = query_info.sets.find(getPreparedSetKey(node, column_and_type.type));
if (prepared_set_it != query_info.sets.end() && prepared_set_it->second->hasExplicitSetElements())
return prepared_set_it->second;
}
else
{
for (const auto & prepared_set_it : query_info.sets)
if (prepared_set_it.first.ast_hash == node->getTreeHash() && prepared_set_it.second->hasExplicitSetElements())
return prepared_set_it.second;
}
return DB::SetPtr();
}
}

74
dbms/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h Normal file
View File

@ -0,0 +1,74 @@
#pragma once
#include <Columns/IColumn.h>
#include <Interpreters/BloomFilter.h>
#include <Storages/MergeTree/KeyCondition.h>
#include <Storages/MergeTree/MergeTreeIndices.h>
#include <Storages/MergeTree/MergeTreeIndexGranuleBloomFilter.h>
namespace DB
{
class MergeTreeIndexConditionBloomFilter : public IMergeTreeIndexCondition
{
public:
struct RPNElement
{
enum Function
{
/// Atoms of a Boolean expression.
FUNCTION_EQUALS,
FUNCTION_NOT_EQUALS,
FUNCTION_IN,
FUNCTION_NOT_IN,
FUNCTION_UNKNOWN, /// Can take any value.
/// Operators of the logical expression.
FUNCTION_NOT,
FUNCTION_AND,
FUNCTION_OR,
/// Constants
ALWAYS_FALSE,
ALWAYS_TRUE,
};
RPNElement(Function function_ = FUNCTION_UNKNOWN) : function(function_) {}
Function function = FUNCTION_UNKNOWN;
std::vector<std::pair<size_t, ColumnPtr>> predicate;
};
MergeTreeIndexConditionBloomFilter(const SelectQueryInfo & info, const Context & context, const Block & header, size_t hash_functions);
bool alwaysUnknownOrTrue() const override;
bool mayBeTrueOnGranule(MergeTreeIndexGranulePtr granule) const override
{
if (const auto & bf_granule = typeid_cast<const MergeTreeIndexGranuleBloomFilter *>(granule.get()))
return mayBeTrueOnGranule(bf_granule);
throw Exception("LOGICAL ERROR: require bloom filter index granule.", ErrorCodes::LOGICAL_ERROR);
}
private:
const Block & header;
const Context & context;
const SelectQueryInfo & query_info;
const size_t hash_functions;
std::vector<RPNElement> rpn;
SetPtr getPreparedSet(const ASTPtr & node);
bool mayBeTrueOnGranule(const MergeTreeIndexGranuleBloomFilter * granule) const;
bool traverseAtomAST(const ASTPtr & node, Block & block_with_constants, RPNElement & out);
bool traverseASTIn(const String & function_name, const ASTPtr & key_ast, const SetPtr & prepared_set, RPNElement & out);
bool traverseASTIn(
const String & function_name, const ASTPtr & key_ast, const DataTypePtr & type, const ColumnPtr & column, RPNElement & out);
bool traverseASTEquals(
const String & function_name, const ASTPtr & key_ast, const DataTypePtr & value_type, const Field & value_field, RPNElement & out);
};
}

92
dbms/src/Storages/MergeTree/MergeTreeBloomFilterIndex.cpp → dbms/src/Storages/MergeTree/MergeTreeIndexFullText.cpp
View File

@ -1,4 +1,4 @@
#include <Storages/MergeTree/MergeTreeBloomFilterIndex.h>
#include <Storages/MergeTree/MergeTreeIndexFullText.h>
#include <Common/StringUtils/StringUtils.h>
#include <Common/UTF8Helpers.h>
@ -31,7 +31,7 @@ namespace ErrorCodes
/// Adds all tokens from string to bloom filter.
static void stringToBloomFilter(
const char * data, size_t size, const std::unique_ptr<ITokenExtractor> & token_extractor, StringBloomFilter & bloom_filter)
const char * data, size_t size, const std::unique_ptr<ITokenExtractor> & token_extractor, BloomFilter & bloom_filter)
{
size_t cur = 0;
size_t token_start = 0;
@ -42,7 +42,7 @@ static void stringToBloomFilter(
/// Adds all tokens from like pattern string to bloom filter. (Because like pattern can contain `\%` and `\_`.)
static void likeStringToBloomFilter(
const String & data, const std::unique_ptr<ITokenExtractor> & token_extractor, StringBloomFilter & bloom_filter)
const String & data, const std::unique_ptr<ITokenExtractor> & token_extractor, BloomFilter & bloom_filter)
{
size_t cur = 0;
String token;
@ -51,24 +51,23 @@ static void likeStringToBloomFilter(
}
MergeTreeBloomFilterIndexGranule::MergeTreeBloomFilterIndexGranule(const MergeTreeBloomFilterIndex & index)
MergeTreeIndexGranuleFullText::MergeTreeIndexGranuleFullText(const MergeTreeIndexFullText & index)
: IMergeTreeIndexGranule()
, index(index)
, bloom_filters(
index.columns.size(), StringBloomFilter(index.bloom_filter_size, index.bloom_filter_hashes, index.seed))
index.columns.size(), BloomFilter(index.bloom_filter_size, index.bloom_filter_hashes, index.seed))
, has_elems(false) {}
void MergeTreeBloomFilterIndexGranule::serializeBinary(WriteBuffer & ostr) const
void MergeTreeIndexGranuleFullText::serializeBinary(WriteBuffer & ostr) const
{
if (empty())
throw Exception(
"Attempt to write empty minmax index " + backQuote(index.name), ErrorCodes::LOGICAL_ERROR);
throw Exception("Attempt to write empty minmax index " + backQuote(index.name), ErrorCodes::LOGICAL_ERROR);
for (const auto & bloom_filter : bloom_filters)
ostr.write(reinterpret_cast<const char *>(bloom_filter.getFilter().data()), index.bloom_filter_size);
}
void MergeTreeBloomFilterIndexGranule::deserializeBinary(ReadBuffer & istr)
void MergeTreeIndexGranuleFullText::deserializeBinary(ReadBuffer & istr)
{
for (auto & bloom_filter : bloom_filters)
{
@ -78,17 +77,17 @@ void MergeTreeBloomFilterIndexGranule::deserializeBinary(ReadBuffer & istr)
}
MergeTreeBloomFilterIndexAggregator::MergeTreeBloomFilterIndexAggregator(const MergeTreeBloomFilterIndex & index)
: index(index), granule(std::make_shared<MergeTreeBloomFilterIndexGranule>(index)) {}
MergeTreeIndexAggregatorFullText::MergeTreeIndexAggregatorFullText(const MergeTreeIndexFullText & index)
: index(index), granule(std::make_shared<MergeTreeIndexGranuleFullText>(index)) {}
MergeTreeIndexGranulePtr MergeTreeBloomFilterIndexAggregator::getGranuleAndReset()
MergeTreeIndexGranulePtr MergeTreeIndexAggregatorFullText::getGranuleAndReset()
{
auto new_granule = std::make_shared<MergeTreeBloomFilterIndexGranule>(index);
auto new_granule = std::make_shared<MergeTreeIndexGranuleFullText>(index);
new_granule.swap(granule);
return new_granule;
}
void MergeTreeBloomFilterIndexAggregator::update(const Block & block, size_t * pos, size_t limit)
void MergeTreeIndexAggregatorFullText::update(const Block & block, size_t * pos, size_t limit)
{
if (*pos >= block.rows())
throw Exception(
@ -111,14 +110,14 @@ void MergeTreeBloomFilterIndexAggregator::update(const Block & block, size_t * p
}
const BloomFilterCondition::AtomMap BloomFilterCondition::atom_map
const MergeTreeConditionFullText::AtomMap MergeTreeConditionFullText::atom_map
{
{
"notEquals",
[] (RPNElement & out, const Field & value, const MergeTreeBloomFilterIndex & idx)
[] (RPNElement & out, const Field & value, const MergeTreeIndexFullText & idx)
{
out.function = RPNElement::FUNCTION_NOT_EQUALS;
out.bloom_filter = std::make_unique<StringBloomFilter>(
out.bloom_filter = std::make_unique<BloomFilter>(
idx.bloom_filter_size, idx.bloom_filter_hashes, idx.seed);
const auto & str = value.get<String>();
@ -128,10 +127,10 @@ const BloomFilterCondition::AtomMap BloomFilterCondition::atom_map
},
{
"equals",
[] (RPNElement & out, const Field & value, const MergeTreeBloomFilterIndex & idx)
[] (RPNElement & out, const Field & value, const MergeTreeIndexFullText & idx)
{
out.function = RPNElement::FUNCTION_EQUALS;
out.bloom_filter = std::make_unique<StringBloomFilter>(
out.bloom_filter = std::make_unique<BloomFilter>(
idx.bloom_filter_size, idx.bloom_filter_hashes, idx.seed);
const auto & str = value.get<String>();
@ -141,10 +140,10 @@ const BloomFilterCondition::AtomMap BloomFilterCondition::atom_map
},
{
"like",
[] (RPNElement & out, const Field & value, const MergeTreeBloomFilterIndex & idx)
[] (RPNElement & out, const Field & value, const MergeTreeIndexFullText & idx)
{
out.function = RPNElement::FUNCTION_LIKE;
out.bloom_filter = std::make_unique<StringBloomFilter>(
out.bloom_filter = std::make_unique<BloomFilter>(
idx.bloom_filter_size, idx.bloom_filter_hashes, idx.seed);
const auto & str = value.get<String>();
@ -154,7 +153,7 @@ const BloomFilterCondition::AtomMap BloomFilterCondition::atom_map
},
{
"notIn",
[] (RPNElement & out, const Field &, const MergeTreeBloomFilterIndex &)
[] (RPNElement & out, const Field &, const MergeTreeIndexFullText &)
{
out.function = RPNElement::FUNCTION_NOT_IN;
return true;
@ -162,7 +161,7 @@ const BloomFilterCondition::AtomMap BloomFilterCondition::atom_map
},
{
"in",
[] (RPNElement & out, const Field &, const MergeTreeBloomFilterIndex &)
[] (RPNElement & out, const Field &, const MergeTreeIndexFullText &)
{
out.function = RPNElement::FUNCTION_IN;
return true;
@ -170,24 +169,21 @@ const BloomFilterCondition::AtomMap BloomFilterCondition::atom_map
},
};
BloomFilterCondition::BloomFilterCondition(
MergeTreeConditionFullText::MergeTreeConditionFullText(
const SelectQueryInfo & query_info,
const Context & context,
const MergeTreeBloomFilterIndex & index_) : index(index_), prepared_sets(query_info.sets)
const MergeTreeIndexFullText & index_) : index(index_), prepared_sets(query_info.sets)
{
rpn = std::move(
RPNBuilder<RPNElement>(
query_info, context,
[this] (const ASTPtr & node,
const Context & /* context */,
Block & block_with_constants,
RPNElement & out) -> bool
[this] (const ASTPtr & node, const Context & /* context */, Block & block_with_constants, RPNElement & out) -> bool
{
return this->atomFromAST(node, block_with_constants, out);
}).extractRPN());
}
bool BloomFilterCondition::alwaysUnknownOrTrue() const
bool MergeTreeConditionFullText::alwaysUnknownOrTrue() const
{
/// Check like in KeyCondition.
std::vector<bool> rpn_stack;
@ -234,10 +230,10 @@ bool BloomFilterCondition::alwaysUnknownOrTrue() const
return rpn_stack[0];
}
bool BloomFilterCondition::mayBeTrueOnGranule(MergeTreeIndexGranulePtr idx_granule) const
bool MergeTreeConditionFullText::mayBeTrueOnGranule(MergeTreeIndexGranulePtr idx_granule) const
{
std::shared_ptr<MergeTreeBloomFilterIndexGranule> granule
= std::dynamic_pointer_cast<MergeTreeBloomFilterIndexGranule>(idx_granule);
std::shared_ptr<MergeTreeIndexGranuleFullText> granule
= std::dynamic_pointer_cast<MergeTreeIndexGranuleFullText>(idx_granule);
if (!granule)
throw Exception(
"BloomFilter index condition got a granule with the wrong type.", ErrorCodes::LOGICAL_ERROR);
@ -314,16 +310,16 @@ bool BloomFilterCondition::mayBeTrueOnGranule(MergeTreeIndexGranulePtr idx_granu
rpn_stack.emplace_back(true, false);
}
else
throw Exception("Unexpected function type in KeyCondition::RPNElement", ErrorCodes::LOGICAL_ERROR);
throw Exception("Unexpected function type in BloomFilterCondition::RPNElement", ErrorCodes::LOGICAL_ERROR);
}
if (rpn_stack.size() != 1)
throw Exception("Unexpected stack size in KeyCondition::mayBeTrueInRange", ErrorCodes::LOGICAL_ERROR);
throw Exception("Unexpected stack size in BloomFilterCondition::mayBeTrueOnGranule", ErrorCodes::LOGICAL_ERROR);
return rpn_stack[0].can_be_true;
}
bool BloomFilterCondition::getKey(const ASTPtr & node, size_t & key_column_num)
bool MergeTreeConditionFullText::getKey(const ASTPtr & node, size_t & key_column_num)
{
auto it = std::find(index.columns.begin(), index.columns.end(), node->getColumnName());
if (it == index.columns.end())
@ -333,7 +329,7 @@ bool BloomFilterCondition::getKey(const ASTPtr & node, size_t & key_column_num)
return true;
}
bool BloomFilterCondition::atomFromAST(
bool MergeTreeConditionFullText::atomFromAST(
const ASTPtr & node, Block & block_with_constants, RPNElement & out)
{
Field const_value;
@ -399,7 +395,7 @@ bool BloomFilterCondition::atomFromAST(
return false;
}
bool BloomFilterCondition::tryPrepareSetBloomFilter(
bool MergeTreeConditionFullText::tryPrepareSetBloomFilter(
const ASTs & args,
RPNElement & out)
{
@ -454,7 +450,7 @@ bool BloomFilterCondition::tryPrepareSetBloomFilter(
if (data_type->getTypeId() != TypeIndex::String && data_type->getTypeId() != TypeIndex::FixedString)
return false;
std::vector<std::vector<StringBloomFilter>> bloom_filters;
std::vector<std::vector<BloomFilter>> bloom_filters;
std::vector<size_t> key_position;
Columns columns = prepared_set->getSetElements();
@ -480,23 +476,23 @@ bool BloomFilterCondition::tryPrepareSetBloomFilter(
}
MergeTreeIndexGranulePtr MergeTreeBloomFilterIndex::createIndexGranule() const
MergeTreeIndexGranulePtr MergeTreeIndexFullText::createIndexGranule() const
{
return std::make_shared<MergeTreeBloomFilterIndexGranule>(*this);
return std::make_shared<MergeTreeIndexGranuleFullText>(*this);
}
MergeTreeIndexAggregatorPtr MergeTreeBloomFilterIndex::createIndexAggregator() const
MergeTreeIndexAggregatorPtr MergeTreeIndexFullText::createIndexAggregator() const
{
return std::make_shared<MergeTreeBloomFilterIndexAggregator>(*this);
return std::make_shared<MergeTreeIndexAggregatorFullText>(*this);
}
IndexConditionPtr MergeTreeBloomFilterIndex::createIndexCondition(
MergeTreeIndexConditionPtr MergeTreeIndexFullText::createIndexCondition(
const SelectQueryInfo & query, const Context & context) const
{
return std::make_shared<BloomFilterCondition>(query, context, *this);
return std::make_shared<MergeTreeConditionFullText>(query, context, *this);
};
bool MergeTreeBloomFilterIndex::mayBenefitFromIndexForIn(const ASTPtr & node) const
bool MergeTreeIndexFullText::mayBenefitFromIndexForIn(const ASTPtr & node) const
{
return std::find(std::cbegin(columns), std::cend(columns), node->getColumnName()) != std::cend(columns);
}
@ -679,7 +675,7 @@ std::unique_ptr<IMergeTreeIndex> bloomFilterIndexCreator(
auto tokenizer = std::make_unique<NgramTokenExtractor>(n);
return std::make_unique<MergeTreeBloomFilterIndex>(
return std::make_unique<MergeTreeIndexFullText>(
node->name, std::move(index_expr), columns, data_types, sample, node->granularity,
bloom_filter_size, bloom_filter_hashes, seed, std::move(tokenizer));
}
@ -697,7 +693,7 @@ std::unique_ptr<IMergeTreeIndex> bloomFilterIndexCreator(
auto tokenizer = std::make_unique<SplitTokenExtractor>();
return std::make_unique<MergeTreeBloomFilterIndex>(
return std::make_unique<MergeTreeIndexFullText>(
node->name, std::move(index_expr), columns, data_types, sample, node->granularity,
bloom_filter_size, bloom_filter_hashes, seed, std::move(tokenizer));
}

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