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

Merge branch 'master' into nlp

Browse Source
This commit is contained in:
Nikolay Degterinsky 2021-05-28 10:26:42 +00:00
commit 63bb8abf89
2080 changed files with 75972 additions and 15696 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
.gitmodules vendored
View File

@ -17,6 +17,7 @@
[submodule "contrib/zlib-ng"]
path = contrib/zlib-ng
url = https://github.com/ClickHouse-Extras/zlib-ng.git
branch = clickhouse-new
[submodule "contrib/googletest"]
path = contrib/googletest
url = https://github.com/google/googletest.git
@ -227,3 +228,7 @@
[submodule "contrib/datasketches-cpp"]
path = contrib/datasketches-cpp
url = https://github.com/ClickHouse-Extras/datasketches-cpp.git
[submodule "contrib/yaml-cpp"]
path = contrib/yaml-cpp
url = https://github.com/ClickHouse-Extras/yaml-cpp.git

139
CHANGELOG.md
View File

@ -1,3 +1,142 @@
## ClickHouse release 21.5, 2021-05-20
#### Backward Incompatible Change
* Change comparison of integers and floating point numbers when integer is not exactly representable in the floating point data type. In new version comparison will return false as the rounding error will occur. Example: `9223372036854775808.0 != 9223372036854775808`, because the number `9223372036854775808` is not representable as floating point number exactly (and `9223372036854775808.0` is rounded to `9223372036854776000.0`). But in previous version the comparison will return as the numbers are equal, because if the floating point number `9223372036854776000.0` get converted back to UInt64, it will yield `9223372036854775808`. For the reference, the Python programming language also treats these numbers as equal. But this behaviour was dependend on CPU model (different results on AMD64 and AArch64 for some out-of-range numbers), so we make the comparison more precise. It will treat int and float numbers equal only if int is represented in floating point type exactly. [#22595](https://github.com/ClickHouse/ClickHouse/pull/22595) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Remove support for `argMin` and `argMax` for single `Tuple` argument. The code was not memory-safe. The feature was added by mistake and it is confusing for people. These functions can be reintroduced under different names later. This fixes [#22384](https://github.com/ClickHouse/ClickHouse/issues/22384) and reverts [#17359](https://github.com/ClickHouse/ClickHouse/issues/17359). [#23393](https://github.com/ClickHouse/ClickHouse/pull/23393) ([alexey-milovidov](https://github.com/alexey-milovidov)).
#### New Feature
* Added functions `dictGetChildren(dictionary, key)`, `dictGetDescendants(dictionary, key, level)`. Function `dictGetChildren` return all children as an array if indexes. It is a inverse transformation for `dictGetHierarchy`. Function `dictGetDescendants` return all descendants as if `dictGetChildren` was applied `level` times recursively. Zero `level` value is equivalent to infinity. Improved performance of `dictGetHierarchy`, `dictIsIn` functions. Closes [#14656](https://github.com/ClickHouse/ClickHouse/issues/14656). [#22096](https://github.com/ClickHouse/ClickHouse/pull/22096) ([Maksim Kita](https://github.com/kitaisreal)).
* Added function `dictGetOrNull`. It works like `dictGet`, but return `Null` in case key was not found in dictionary. Closes [#22375](https://github.com/ClickHouse/ClickHouse/issues/22375). [#22413](https://github.com/ClickHouse/ClickHouse/pull/22413) ([Maksim Kita](https://github.com/kitaisreal)).
* Added a table function `s3Cluster`, which allows to process files from `s3` in parallel on every node of a specified cluster. [#22012](https://github.com/ClickHouse/ClickHouse/pull/22012) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)).
* Added support for replicas and shards in MySQL/PostgreSQL table engine / table function. You can write `SELECT * FROM mysql('host{1,2}-{1|2}', ...)`. Closes [#20969](https://github.com/ClickHouse/ClickHouse/issues/20969). [#22217](https://github.com/ClickHouse/ClickHouse/pull/22217) ([Kseniia Sumarokova](https://github.com/kssenii)).
* Added `ALTER TABLE ... FETCH PART ...` query. It's similar to `FETCH PARTITION`, but fetches only one part. [#22706](https://github.com/ClickHouse/ClickHouse/pull/22706) ([turbo jason](https://github.com/songenjie)).
* Added a setting `max_distributed_depth` that limits the depth of recursive queries to `Distributed` tables. Closes [#20229](https://github.com/ClickHouse/ClickHouse/issues/20229). [#21942](https://github.com/ClickHouse/ClickHouse/pull/21942) ([flynn](https://github.com/ucasFL)).
#### Performance Improvement
* Improved performance of `intDiv` by dynamic dispatch for AVX2. This closes [#22314](https://github.com/ClickHouse/ClickHouse/issues/22314). [#23000](https://github.com/ClickHouse/ClickHouse/pull/23000) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Improved performance of reading from `ArrowStream` input format for sources other then local file (e.g. URL). [#22673](https://github.com/ClickHouse/ClickHouse/pull/22673) ([nvartolomei](https://github.com/nvartolomei)).
* Disabled compression by default when interacting with localhost (with clickhouse-client or server to server with distributed queries) via native protocol. It may improve performance of some import/export operations. This closes [#22234](https://github.com/ClickHouse/ClickHouse/issues/22234). [#22237](https://github.com/ClickHouse/ClickHouse/pull/22237) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Exclude values that does not belong to the shard from right part of IN section for distributed queries (under `optimize_skip_unused_shards_rewrite_in`, enabled by default, since it still requires `optimize_skip_unused_shards`). [#21511](https://github.com/ClickHouse/ClickHouse/pull/21511) ([Azat Khuzhin](https://github.com/azat)).
* Improved performance of reading a subset of columns with File-like table engine and column-oriented format like Parquet, Arrow or ORC. This closes [#issue:20129](https://github.com/ClickHouse/ClickHouse/issues/20129). [#21302](https://github.com/ClickHouse/ClickHouse/pull/21302) ([keenwolf](https://github.com/keen-wolf)).
* Allow to move more conditions to `PREWHERE` as it was before version 21.1 (adjustment of internal heuristics). Insufficient number of moved condtions could lead to worse performance. [#23397](https://github.com/ClickHouse/ClickHouse/pull/23397) ([Anton Popov](https://github.com/CurtizJ)).
* Improved performance of ODBC connections and fixed all the outstanding issues from the backlog. Using `nanodbc` library instead of `Poco::ODBC`. Closes [#9678](https://github.com/ClickHouse/ClickHouse/issues/9678). Add support for DateTime64 and Decimal* for ODBC table engine. Closes [#21961](https://github.com/ClickHouse/ClickHouse/issues/21961). Fixed issue with cyrillic text being truncated. Closes [#16246](https://github.com/ClickHouse/ClickHouse/issues/16246). Added connection pools for odbc bridge. [#21972](https://github.com/ClickHouse/ClickHouse/pull/21972) ([Kseniia Sumarokova](https://github.com/kssenii)).
#### Improvement
* Increase `max_uri_size` (the maximum size of URL in HTTP interface) to 1 MiB by default. This closes [#21197](https://github.com/ClickHouse/ClickHouse/issues/21197). [#22997](https://github.com/ClickHouse/ClickHouse/pull/22997) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Set `background_fetches_pool_size` to `8` that is better for production usage with frequent small insertions or slow ZooKeeper cluster. [#22945](https://github.com/ClickHouse/ClickHouse/pull/22945) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* FlatDictionary added `initial_array_size`, `max_array_size` options. [#22521](https://github.com/ClickHouse/ClickHouse/pull/22521) ([Maksim Kita](https://github.com/kitaisreal)).
* Add new setting `non_replicated_deduplication_window` for non-replicated MergeTree inserts deduplication. [#22514](https://github.com/ClickHouse/ClickHouse/pull/22514) ([alesapin](https://github.com/alesapin)).
* Update paths to the `CatBoost` model configs in config reloading. [#22434](https://github.com/ClickHouse/ClickHouse/pull/22434) ([Kruglov Pavel](https://github.com/Avogar)).
* Added `Decimal256` type support in dictionaries. `Decimal256` is experimental feature. Closes [#20979](https://github.com/ClickHouse/ClickHouse/issues/20979). [#22960](https://github.com/ClickHouse/ClickHouse/pull/22960) ([Maksim Kita](https://github.com/kitaisreal)).
* Enabled `async_socket_for_remote` by default (using less amount of OS threads for distributed queries). [#23683](https://github.com/ClickHouse/ClickHouse/pull/23683) ([Nikolai Kochetov](https://github.com/KochetovNicolai)).
* Fixed `quantile(s)TDigest`. Added special handling of singleton centroids according to tdunning/t-digest 3.2+. Also a bug with over-compression of centroids in implementation of earlier version of the algorithm was fixed. [#23314](https://github.com/ClickHouse/ClickHouse/pull/23314) ([Vladimir Chebotarev](https://github.com/excitoon)).
* Make function name `unhex` case insensitive for compatibility with MySQL. [#23229](https://github.com/ClickHouse/ClickHouse/pull/23229) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Implement functions `arrayHasAny`, `arrayHasAll`, `has`, `indexOf`, `countEqual` for generic case when types of array elements are different. In previous versions the functions `arrayHasAny`, `arrayHasAll` returned false and `has`, `indexOf`, `countEqual` thrown exception. Also add support for `Decimal` and big integer types in functions `has` and similar. This closes [#20272](https://github.com/ClickHouse/ClickHouse/issues/20272). [#23044](https://github.com/ClickHouse/ClickHouse/pull/23044) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Raised the threshold on max number of matches in result of the function `extractAllGroupsHorizontal`. [#23036](https://github.com/ClickHouse/ClickHouse/pull/23036) ([Vasily Nemkov](https://github.com/Enmk)).
* Do not perform `optimize_skip_unused_shards` for cluster with one node. [#22999](https://github.com/ClickHouse/ClickHouse/pull/22999) ([Azat Khuzhin](https://github.com/azat)).
* Added ability to run clickhouse-keeper (experimental drop-in replacement to ZooKeeper) with SSL. Config settings `keeper_server.tcp_port_secure` can be used for secure interaction between client and keeper-server. `keeper_server.raft_configuration.secure` can be used to enable internal secure communication between nodes. [#22992](https://github.com/ClickHouse/ClickHouse/pull/22992) ([alesapin](https://github.com/alesapin)).
* Added ability to flush buffer only in background for `Buffer` tables. [#22986](https://github.com/ClickHouse/ClickHouse/pull/22986) ([Azat Khuzhin](https://github.com/azat)).
* When selecting from MergeTree table with NULL in WHERE condition, in rare cases, exception was thrown. This closes [#20019](https://github.com/ClickHouse/ClickHouse/issues/20019). [#22978](https://github.com/ClickHouse/ClickHouse/pull/22978) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Fix error handling in Poco HTTP Client for AWS. [#22973](https://github.com/ClickHouse/ClickHouse/pull/22973) ([kreuzerkrieg](https://github.com/kreuzerkrieg)).
* Respect `max_part_removal_threads` for `ReplicatedMergeTree`. [#22971](https://github.com/ClickHouse/ClickHouse/pull/22971) ([Azat Khuzhin](https://github.com/azat)).
* Fix obscure corner case of MergeTree settings inactive_parts_to_throw_insert = 0 with inactive_parts_to_delay_insert > 0. [#22947](https://github.com/ClickHouse/ClickHouse/pull/22947) ([Azat Khuzhin](https://github.com/azat)).
* `dateDiff` now works with `DateTime64` arguments (even for values outside of `DateTime` range) [#22931](https://github.com/ClickHouse/ClickHouse/pull/22931) ([Vasily Nemkov](https://github.com/Enmk)).
* MaterializeMySQL (experimental feature): added an ability to replicate MySQL databases containing views without failing. This is accomplished by ignoring the views. [#22760](https://github.com/ClickHouse/ClickHouse/pull/22760) ([Christian](https://github.com/cfroystad)).
* Allow RBAC row policy via postgresql protocol. Closes [#22658](https://github.com/ClickHouse/ClickHouse/issues/22658). PostgreSQL protocol is enabled in configuration by default. [#22755](https://github.com/ClickHouse/ClickHouse/pull/22755) ([Kseniia Sumarokova](https://github.com/kssenii)).
* Add metric to track how much time is spend during waiting for Buffer layer lock. [#22725](https://github.com/ClickHouse/ClickHouse/pull/22725) ([Azat Khuzhin](https://github.com/azat)).
* Allow to use CTE in VIEW definition. This closes [#22491](https://github.com/ClickHouse/ClickHouse/issues/22491). [#22657](https://github.com/ClickHouse/ClickHouse/pull/22657) ([Amos Bird](https://github.com/amosbird)).
* Clear the rest of the screen and show cursor in `clickhouse-client` if previous program has left garbage in terminal. This closes [#16518](https://github.com/ClickHouse/ClickHouse/issues/16518). [#22634](https://github.com/ClickHouse/ClickHouse/pull/22634) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Make `round` function to behave consistently on non-x86_64 platforms. Rounding half to nearest even (Banker's rounding) is used. [#22582](https://github.com/ClickHouse/ClickHouse/pull/22582) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Correctly check structure of blocks of data that are sending by Distributed tables. [#22325](https://github.com/ClickHouse/ClickHouse/pull/22325) ([Azat Khuzhin](https://github.com/azat)).
* Allow publishing Kafka errors to a virtual column of Kafka engine, controlled by the `kafka_handle_error_mode` setting. [#21850](https://github.com/ClickHouse/ClickHouse/pull/21850) ([fastio](https://github.com/fastio)).
* Add aliases `simpleJSONExtract/simpleJSONHas` to `visitParam/visitParamExtract{UInt, Int, Bool, Float, Raw, String}`. Fixes [#21383](https://github.com/ClickHouse/ClickHouse/issues/21383). [#21519](https://github.com/ClickHouse/ClickHouse/pull/21519) ([fastio](https://github.com/fastio)).
* Add `clickhouse-library-bridge` for library dictionary source. Closes [#9502](https://github.com/ClickHouse/ClickHouse/issues/9502). [#21509](https://github.com/ClickHouse/ClickHouse/pull/21509) ([Kseniia Sumarokova](https://github.com/kssenii)).
* Forbid to drop a column if it's referenced by materialized view. Closes [#21164](https://github.com/ClickHouse/ClickHouse/issues/21164). [#21303](https://github.com/ClickHouse/ClickHouse/pull/21303) ([flynn](https://github.com/ucasFL)).
* Support dynamic interserver credentials (rotating credentials without downtime). [#14113](https://github.com/ClickHouse/ClickHouse/pull/14113) ([johnskopis](https://github.com/johnskopis)).
* Add support for Kafka storage with `Arrow` and `ArrowStream` format messages. [#23415](https://github.com/ClickHouse/ClickHouse/pull/23415) ([Chao Ma](https://github.com/godliness)).
* Fixed missing semicolon in exception message. The user may find this exception message unpleasant to read. [#23208](https://github.com/ClickHouse/ClickHouse/pull/23208) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Fixed missing whitespace in some exception messages about `LowCardinality` type. [#23207](https://github.com/ClickHouse/ClickHouse/pull/23207) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Some values were formatted with alignment in center in table cells in `Markdown` format. Not anymore. [#23096](https://github.com/ClickHouse/ClickHouse/pull/23096) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Remove non-essential details from suggestions in clickhouse-client. This closes [#22158](https://github.com/ClickHouse/ClickHouse/issues/22158). [#23040](https://github.com/ClickHouse/ClickHouse/pull/23040) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Correct calculation of `bytes_allocated` field in system.dictionaries for sparse_hashed dictionaries. [#22867](https://github.com/ClickHouse/ClickHouse/pull/22867) ([Azat Khuzhin](https://github.com/azat)).
* Fixed approximate total rows accounting for reverse reading from MergeTree. [#22726](https://github.com/ClickHouse/ClickHouse/pull/22726) ([Azat Khuzhin](https://github.com/azat)).
* Fix the case when it was possible to configure dictionary with clickhouse source that was looking to itself that leads to infinite loop. Closes [#14314](https://github.com/ClickHouse/ClickHouse/issues/14314). [#22479](https://github.com/ClickHouse/ClickHouse/pull/22479) ([Maksim Kita](https://github.com/kitaisreal)).
#### Bug Fix
* Multiple fixes for hedged requests. Fixed an error `Can't initialize pipeline with empty pipe` for queries with `GLOBAL IN/JOIN` when the setting `use_hedged_requests` is enabled. Fixes [#23431](https://github.com/ClickHouse/ClickHouse/issues/23431). [#23805](https://github.com/ClickHouse/ClickHouse/pull/23805) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). Fixed a race condition in hedged connections which leads to crash. This fixes [#22161](https://github.com/ClickHouse/ClickHouse/issues/22161). [#22443](https://github.com/ClickHouse/ClickHouse/pull/22443) ([Kruglov Pavel](https://github.com/Avogar)). Fix possible crash in case if `unknown packet` was received from remote query (with `async_socket_for_remote` enabled). Fixes [#21167](https://github.com/ClickHouse/ClickHouse/issues/21167). [#23309](https://github.com/ClickHouse/ClickHouse/pull/23309) ([Nikolai Kochetov](https://github.com/KochetovNicolai)).
* Fixed the behavior when disabling `input_format_with_names_use_header ` setting discards all the input with CSVWithNames format. This fixes [#22406](https://github.com/ClickHouse/ClickHouse/issues/22406). [#23202](https://github.com/ClickHouse/ClickHouse/pull/23202) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)).
* Fixed remote JDBC bridge timeout connection issue. Closes [#9609](https://github.com/ClickHouse/ClickHouse/issues/9609). [#23771](https://github.com/ClickHouse/ClickHouse/pull/23771) ([Maksim Kita](https://github.com/kitaisreal), [alexey-milovidov](https://github.com/alexey-milovidov)).
* Fix the logic of initial load of `complex_key_hashed` if `update_field` is specified. Closes [#23800](https://github.com/ClickHouse/ClickHouse/issues/23800). [#23824](https://github.com/ClickHouse/ClickHouse/pull/23824) ([Maksim Kita](https://github.com/kitaisreal)).
* Fixed crash when `PREWHERE` and row policy filter are both in effect with empty result. [#23763](https://github.com/ClickHouse/ClickHouse/pull/23763) ([Amos Bird](https://github.com/amosbird)).
* Avoid possible "Cannot schedule a task" error (in case some exception had been occurred) on INSERT into Distributed. [#23744](https://github.com/ClickHouse/ClickHouse/pull/23744) ([Azat Khuzhin](https://github.com/azat)).
* Added an exception in case of completely the same values in both samples in aggregate function `mannWhitneyUTest`. This fixes [#23646](https://github.com/ClickHouse/ClickHouse/issues/23646). [#23654](https://github.com/ClickHouse/ClickHouse/pull/23654) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)).
* Fixed server fault when inserting data through HTTP caused an exception. This fixes [#23512](https://github.com/ClickHouse/ClickHouse/issues/23512). [#23643](https://github.com/ClickHouse/ClickHouse/pull/23643) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)).
* Fixed misinterpretation of some `LIKE` expressions with escape sequences. [#23610](https://github.com/ClickHouse/ClickHouse/pull/23610) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Fixed restart / stop command hanging. Closes [#20214](https://github.com/ClickHouse/ClickHouse/issues/20214). [#23552](https://github.com/ClickHouse/ClickHouse/pull/23552) ([filimonov](https://github.com/filimonov)).
* Fixed `COLUMNS` matcher in case of multiple JOINs in select query. Closes [#22736](https://github.com/ClickHouse/ClickHouse/issues/22736). [#23501](https://github.com/ClickHouse/ClickHouse/pull/23501) ([Maksim Kita](https://github.com/kitaisreal)).
* Fixed a crash when modifying column's default value when a column itself is used as `ReplacingMergeTree`'s parameter. [#23483](https://github.com/ClickHouse/ClickHouse/pull/23483) ([hexiaoting](https://github.com/hexiaoting)).
* Fixed corner cases in vertical merges with `ReplacingMergeTree`. In rare cases they could lead to fails of merges with exceptions like `Incomplete granules are not allowed while blocks are granules size`. [#23459](https://github.com/ClickHouse/ClickHouse/pull/23459) ([Anton Popov](https://github.com/CurtizJ)).
* Fixed bug that does not allow cast from empty array literal, to array with dimensions greater than 1, e.g. `CAST([] AS Array(Array(String)))`. Closes [#14476](https://github.com/ClickHouse/ClickHouse/issues/14476). [#23456](https://github.com/ClickHouse/ClickHouse/pull/23456) ([Maksim Kita](https://github.com/kitaisreal)).
* Fixed a bug when `deltaSum` aggregate function produced incorrect result after resetting the counter. [#23437](https://github.com/ClickHouse/ClickHouse/pull/23437) ([Russ Frank](https://github.com/rf)).
* Fixed `Cannot unlink file` error on unsuccessful creation of ReplicatedMergeTree table with multidisk configuration. This closes [#21755](https://github.com/ClickHouse/ClickHouse/issues/21755). [#23433](https://github.com/ClickHouse/ClickHouse/pull/23433) ([tavplubix](https://github.com/tavplubix)).
* Fixed incompatible constant expression generation during partition pruning based on virtual columns. This fixes https://github.com/ClickHouse/ClickHouse/pull/21401#discussion_r611888913. [#23366](https://github.com/ClickHouse/ClickHouse/pull/23366) ([Amos Bird](https://github.com/amosbird)).
* Fixed a crash when setting join_algorithm is set to 'auto' and Join is performed with a Dictionary. Close [#23002](https://github.com/ClickHouse/ClickHouse/issues/23002). [#23312](https://github.com/ClickHouse/ClickHouse/pull/23312) ([Vladimir](https://github.com/vdimir)).
* Don't relax NOT conditions during partition pruning. This fixes [#23305](https://github.com/ClickHouse/ClickHouse/issues/23305) and [#21539](https://github.com/ClickHouse/ClickHouse/issues/21539). [#23310](https://github.com/ClickHouse/ClickHouse/pull/23310) ([Amos Bird](https://github.com/amosbird)).
* Fixed very rare race condition on background cleanup of old blocks. It might cause a block not to be deduplicated if it's too close to the end of deduplication window. [#23301](https://github.com/ClickHouse/ClickHouse/pull/23301) ([tavplubix](https://github.com/tavplubix)).
* Fixed very rare (distributed) race condition between creation and removal of ReplicatedMergeTree tables. It might cause exceptions like `node doesn't exist` on attempt to create replicated table. Fixes [#21419](https://github.com/ClickHouse/ClickHouse/issues/21419). [#23294](https://github.com/ClickHouse/ClickHouse/pull/23294) ([tavplubix](https://github.com/tavplubix)).
* Fixed simple key dictionary from DDL creation if primary key is not first attribute. Fixes [#23236](https://github.com/ClickHouse/ClickHouse/issues/23236). [#23262](https://github.com/ClickHouse/ClickHouse/pull/23262) ([Maksim Kita](https://github.com/kitaisreal)).
* Fixed reading from ODBC when there are many long column names in a table. Closes [#8853](https://github.com/ClickHouse/ClickHouse/issues/8853). [#23215](https://github.com/ClickHouse/ClickHouse/pull/23215) ([Kseniia Sumarokova](https://github.com/kssenii)).
* MaterializeMySQL (experimental feature): fixed `Not found column` error when selecting from `MaterializeMySQL` with condition on key column. Fixes [#22432](https://github.com/ClickHouse/ClickHouse/issues/22432). [#23200](https://github.com/ClickHouse/ClickHouse/pull/23200) ([tavplubix](https://github.com/tavplubix)).
* Correct aliases handling if subquery was optimized to constant. Fixes [#22924](https://github.com/ClickHouse/ClickHouse/issues/22924). Fixes [#10401](https://github.com/ClickHouse/ClickHouse/issues/10401). [#23191](https://github.com/ClickHouse/ClickHouse/pull/23191) ([Maksim Kita](https://github.com/kitaisreal)).
* Server might fail to start if `data_type_default_nullable` setting is enabled in default profile, it's fixed. Fixes [#22573](https://github.com/ClickHouse/ClickHouse/issues/22573). [#23185](https://github.com/ClickHouse/ClickHouse/pull/23185) ([tavplubix](https://github.com/tavplubix)).
* Fixed a crash on shutdown which happened because of wrong accounting of current connections. [#23154](https://github.com/ClickHouse/ClickHouse/pull/23154) ([Vitaly Baranov](https://github.com/vitlibar)).
* Fixed `Table .inner_id... doesn't exist` error when selecting from Materialized View after detaching it from Atomic database and attaching back. [#23047](https://github.com/ClickHouse/ClickHouse/pull/23047) ([tavplubix](https://github.com/tavplubix)).
* Fix error `Cannot find column in ActionsDAG result` which may happen if subquery uses `untuple`. Fixes [#22290](https://github.com/ClickHouse/ClickHouse/issues/22290). [#22991](https://github.com/ClickHouse/ClickHouse/pull/22991) ([Nikolai Kochetov](https://github.com/KochetovNicolai)).
* Fix usage of constant columns of type `Map` with nullable values. [#22939](https://github.com/ClickHouse/ClickHouse/pull/22939) ([Anton Popov](https://github.com/CurtizJ)).
* fixed `formatDateTime()` on `DateTime64` and "%C" format specifier fixed `toDateTime64()` for large values and non-zero scale. [#22937](https://github.com/ClickHouse/ClickHouse/pull/22937) ([Vasily Nemkov](https://github.com/Enmk)).
* Fixed a crash when using `mannWhitneyUTest` and `rankCorr` with window functions. This fixes [#22728](https://github.com/ClickHouse/ClickHouse/issues/22728). [#22876](https://github.com/ClickHouse/ClickHouse/pull/22876) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)).
* LIVE VIEW (experimental feature): fixed possible hanging in concurrent DROP/CREATE of TEMPORARY LIVE VIEW in `TemporaryLiveViewCleaner`, [see](https://gist.github.com/vzakaznikov/0c03195960fc86b56bfe2bc73a90019e). [#22858](https://github.com/ClickHouse/ClickHouse/pull/22858) ([Vitaly Baranov](https://github.com/vitlibar)).
* Fixed pushdown of `HAVING` in case, when filter column is used in aggregation. [#22763](https://github.com/ClickHouse/ClickHouse/pull/22763) ([Anton Popov](https://github.com/CurtizJ)).
* Fixed possible hangs in Zookeeper requests in case of OOM exception. Fixes [#22438](https://github.com/ClickHouse/ClickHouse/issues/22438). [#22684](https://github.com/ClickHouse/ClickHouse/pull/22684) ([Nikolai Kochetov](https://github.com/KochetovNicolai)).
* Fixed wait for mutations on several replicas for ReplicatedMergeTree table engines. Previously, mutation/alter query may finish before mutation actually executed on other replicas. [#22669](https://github.com/ClickHouse/ClickHouse/pull/22669) ([alesapin](https://github.com/alesapin)).
* Fixed exception for Log with nested types without columns in the SELECT clause. [#22654](https://github.com/ClickHouse/ClickHouse/pull/22654) ([Azat Khuzhin](https://github.com/azat)).
* Fix unlimited wait for auxiliary AWS requests. [#22594](https://github.com/ClickHouse/ClickHouse/pull/22594) ([Vladimir Chebotarev](https://github.com/excitoon)).
* Fixed a crash when client closes connection very early [#22579](https://github.com/ClickHouse/ClickHouse/issues/22579). [#22591](https://github.com/ClickHouse/ClickHouse/pull/22591) ([nvartolomei](https://github.com/nvartolomei)).
* `Map` data type (experimental feature): fixed an incorrect formatting of function `map` in distributed queries. [#22588](https://github.com/ClickHouse/ClickHouse/pull/22588) ([foolchi](https://github.com/foolchi)).
* Fixed deserialization of empty string without newline at end of TSV format. This closes [#20244](https://github.com/ClickHouse/ClickHouse/issues/20244). Possible workaround without version update: set `input_format_null_as_default` to zero. It was zero in old versions. [#22527](https://github.com/ClickHouse/ClickHouse/pull/22527) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Fixed wrong cast of a column of `LowCardinality` type in Merge Join algorithm. Close [#22386](https://github.com/ClickHouse/ClickHouse/issues/22386), close [#22388](https://github.com/ClickHouse/ClickHouse/issues/22388). [#22510](https://github.com/ClickHouse/ClickHouse/pull/22510) ([Vladimir](https://github.com/vdimir)).
* Buffer overflow (on read) was possible in `tokenbf_v1` full text index. The excessive bytes are not used but the read operation may lead to crash in rare cases. This closes [#19233](https://github.com/ClickHouse/ClickHouse/issues/19233). [#22421](https://github.com/ClickHouse/ClickHouse/pull/22421) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Do not limit HTTP chunk size. Fixes [#21907](https://github.com/ClickHouse/ClickHouse/issues/21907). [#22322](https://github.com/ClickHouse/ClickHouse/pull/22322) ([Ivan](https://github.com/abyss7)).
* Fixed a bug, which leads to underaggregation of data in case of enabled `optimize_aggregation_in_order` and many parts in table. Slightly improve performance of aggregation with enabled `optimize_aggregation_in_order`. [#21889](https://github.com/ClickHouse/ClickHouse/pull/21889) ([Anton Popov](https://github.com/CurtizJ)).
* Check if table function view is used as a column. This complements #20350. [#21465](https://github.com/ClickHouse/ClickHouse/pull/21465) ([Amos Bird](https://github.com/amosbird)).
* Fix "unknown column" error for tables with `Merge` engine in queris with `JOIN` and aggregation. Closes [#18368](https://github.com/ClickHouse/ClickHouse/issues/18368), close [#22226](https://github.com/ClickHouse/ClickHouse/issues/22226). [#21370](https://github.com/ClickHouse/ClickHouse/pull/21370) ([Vladimir](https://github.com/vdimir)).
* Fixed name clashes in pushdown optimization. It caused incorrect `WHERE` filtration after FULL JOIN. Close [#20497](https://github.com/ClickHouse/ClickHouse/issues/20497). [#20622](https://github.com/ClickHouse/ClickHouse/pull/20622) ([Vladimir](https://github.com/vdimir)).
* Fixed very rare bug when quorum insert with `quorum_parallel=1` is not really "quorum" because of deduplication. [#18215](https://github.com/ClickHouse/ClickHouse/pull/18215) ([filimonov](https://github.com/filimonov) - reported, [alesapin](https://github.com/alesapin) - fixed).
#### Build/Testing/Packaging Improvement
* Run stateless tests in parallel in CI. [#22300](https://github.com/ClickHouse/ClickHouse/pull/22300) ([alesapin](https://github.com/alesapin)).
* Simplify debian packages. This fixes [#21698](https://github.com/ClickHouse/ClickHouse/issues/21698). [#22976](https://github.com/ClickHouse/ClickHouse/pull/22976) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Added support for ClickHouse build on Apple M1. [#21639](https://github.com/ClickHouse/ClickHouse/pull/21639) ([changvvb](https://github.com/changvvb)).
* Fixed ClickHouse Keeper build for MacOS. [#22860](https://github.com/ClickHouse/ClickHouse/pull/22860) ([alesapin](https://github.com/alesapin)).
* Fixed some tests on AArch64 platform. [#22596](https://github.com/ClickHouse/ClickHouse/pull/22596) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Added function alignment for possibly better performance. [#21431](https://github.com/ClickHouse/ClickHouse/pull/21431) ([Danila Kutenin](https://github.com/danlark1)).
* Adjust some tests to output identical results on amd64 and aarch64 (qemu). The result was depending on implementation specific CPU behaviour. [#22590](https://github.com/ClickHouse/ClickHouse/pull/22590) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Allow query profiling only on x86_64. See [#15174](https://github.com/ClickHouse/ClickHouse/issues/15174#issuecomment-812954965) and [#15638](https://github.com/ClickHouse/ClickHouse/issues/15638#issuecomment-703805337). This closes [#15638](https://github.com/ClickHouse/ClickHouse/issues/15638). [#22580](https://github.com/ClickHouse/ClickHouse/pull/22580) ([alexey-milovidov](https://github.com/alexey-milovidov)).
* Allow building with unbundled xz (lzma) using `USE_INTERNAL_XZ_LIBRARY=OFF` CMake option. [#22571](https://github.com/ClickHouse/ClickHouse/pull/22571) ([Kfir Itzhak](https://github.com/mastertheknife)).
* Enable bundled `openldap` on `ppc64le` [#22487](https://github.com/ClickHouse/ClickHouse/pull/22487) ([Kfir Itzhak](https://github.com/mastertheknife)).
* Disable incompatible libraries (platform specific typically) on `ppc64le` [#22475](https://github.com/ClickHouse/ClickHouse/pull/22475) ([Kfir Itzhak](https://github.com/mastertheknife)).
* Add Jepsen test in CI for clickhouse Keeper. [#22373](https://github.com/ClickHouse/ClickHouse/pull/22373) ([alesapin](https://github.com/alesapin)).
* Build `jemalloc` with support for [heap profiling](https://github.com/jemalloc/jemalloc/wiki/Use-Case%3A-Heap-Profiling). [#22834](https://github.com/ClickHouse/ClickHouse/pull/22834) ([nvartolomei](https://github.com/nvartolomei)).
* Avoid UB in `*Log` engines for rwlock unlock due to unlock from another thread. [#22583](https://github.com/ClickHouse/ClickHouse/pull/22583) ([Azat Khuzhin](https://github.com/azat)).
* Fixed UB by unlocking the rwlock of the TinyLog from the same thread. [#22560](https://github.com/ClickHouse/ClickHouse/pull/22560) ([Azat Khuzhin](https://github.com/azat)).
## ClickHouse release 21.4
### ClickHouse release 21.4.1 2021-04-12

11
CMakeLists.txt
View File

@ -36,7 +36,7 @@ option(FAIL_ON_UNSUPPORTED_OPTIONS_COMBINATION
if(FAIL_ON_UNSUPPORTED_OPTIONS_COMBINATION)
set(RECONFIGURE_MESSAGE_LEVEL FATAL_ERROR)
else()
set(RECONFIGURE_MESSAGE_LEVEL STATUS)
set(RECONFIGURE_MESSAGE_LEVEL WARNING)
endif()
enable_language(C CXX ASM)
@ -169,8 +169,8 @@ endif ()
option(ENABLE_TESTS "Provide unit_test_dbms target with Google.Test unit tests" ON)
option(ENABLE_EXAMPLES "Build all example programs in 'examples' subdirectories" OFF)
if (OS_LINUX AND NOT UNBUNDLED AND MAKE_STATIC_LIBRARIES AND NOT SPLIT_SHARED_LIBRARIES AND CMAKE_VERSION VERSION_GREATER "3.9.0")
# Only for Linux, x86_64.
if (OS_LINUX AND (ARCH_AMD64 OR ARCH_AARCH64) AND NOT UNBUNDLED AND MAKE_STATIC_LIBRARIES AND NOT SPLIT_SHARED_LIBRARIES AND CMAKE_VERSION VERSION_GREATER "3.9.0")
# Only for Linux, x86_64 or aarch64.
option(GLIBC_COMPATIBILITY "Enable compatibility with older glibc libraries." ON)
elseif(GLIBC_COMPATIBILITY)
message (${RECONFIGURE_MESSAGE_LEVEL} "Glibc compatibility cannot be enabled in current configuration")
@ -504,7 +504,6 @@ include (cmake/find/libuv.cmake) # for amqpcpp and cassandra
include (cmake/find/amqpcpp.cmake)
include (cmake/find/capnp.cmake)
include (cmake/find/llvm.cmake)
include (cmake/find/termcap.cmake) # for external static llvm
include (cmake/find/h3.cmake)
include (cmake/find/libxml2.cmake)
include (cmake/find/brotli.cmake)
@ -527,6 +526,7 @@ include (cmake/find/nanodbc.cmake)
include (cmake/find/rocksdb.cmake)
include (cmake/find/libpqxx.cmake)
include (cmake/find/nuraft.cmake)
include (cmake/find/yaml-cpp.cmake)
if(NOT USE_INTERNAL_PARQUET_LIBRARY)
@ -593,6 +593,9 @@ include_directories(${ConfigIncludePath})
# Add as many warnings as possible for our own code.
include (cmake/warnings.cmake)
# Check if needed compiler flags are supported
include (cmake/check_flags.cmake)
add_subdirectory (base)
add_subdirectory (src)
add_subdirectory (programs)

5
README.md
View File

@ -8,8 +8,11 @@ ClickHouse® is an open-source column-oriented database management system that a
* [Tutorial](https://clickhouse.tech/docs/en/getting_started/tutorial/) shows how to set up and query small ClickHouse cluster.
* [Documentation](https://clickhouse.tech/docs/en/) provides more in-depth information.
* [YouTube channel](https://www.youtube.com/c/ClickHouseDB) has a lot of content about ClickHouse in video format.
* [Slack](https://join.slack.com/t/clickhousedb/shared_invite/zt-nwwakmk4-xOJ6cdy0sJC3It8j348~IA) and [Telegram](https://telegram.me/clickhouse_en) allow to chat with ClickHouse users in real-time.
* [Slack](https://join.slack.com/t/clickhousedb/shared_invite/zt-qfort0u8-TWqK4wIP0YSdoDE0btKa1w) and [Telegram](https://telegram.me/clickhouse_en) allow to chat with ClickHouse users in real-time.
* [Blog](https://clickhouse.yandex/blog/en/) contains various ClickHouse-related articles, as well as announcements and reports about events.
* [Code Browser](https://clickhouse.tech/codebrowser/html_report/ClickHouse/index.html) with syntax highlight and navigation.
* [Contacts](https://clickhouse.tech/#contacts) can help to get your questions answered if there are any.
* You can also [fill this form](https://clickhouse.tech/#meet) to meet Yandex ClickHouse team in person.
## Upcoming Events
* [SF Bay Area ClickHouse Community Meetup (online)](https://www.meetup.com/San-Francisco-Bay-Area-ClickHouse-Meetup/events/278144089/) on 16 June 2021.

8
base/bridge/CMakeLists.txt
View File

@ -3,5 +3,11 @@ add_library (bridge
)
target_include_directories (daemon PUBLIC ..)
target_link_libraries (bridge PRIVATE daemon dbms Poco::Data Poco::Data::ODBC)
target_link_libraries (bridge
PRIVATE
daemon
dbms
Poco::Data
Poco::Data::ODBC
)

216
base/common/DecomposedFloat.h Normal file
View File

@ -0,0 +1,216 @@
#pragma once
#include <cstdint>
#include <cstddef>
#include <cstring>
#include <common/extended_types.h>
/// Allows to check the internals of IEEE-754 floating point number.
template <typename T> struct FloatTraits;
template <>
struct FloatTraits<float>
{
using UInt = uint32_t;
static constexpr size_t bits = 32;
static constexpr size_t exponent_bits = 8;
static constexpr size_t mantissa_bits = bits - exponent_bits - 1;
};
template <>
struct FloatTraits<double>
{
using UInt = uint64_t;
static constexpr size_t bits = 64;
static constexpr size_t exponent_bits = 11;
static constexpr size_t mantissa_bits = bits - exponent_bits - 1;
};
/// x = sign * (2 ^ normalized_exponent) * (1 + mantissa * 2 ^ -mantissa_bits)
/// x = sign * (2 ^ normalized_exponent + mantissa * 2 ^ (normalized_exponent - mantissa_bits))
template <typename T>
struct DecomposedFloat
{
using Traits = FloatTraits<T>;
DecomposedFloat(T x)
{
memcpy(&x_uint, &x, sizeof(x));
}
typename Traits::UInt x_uint;
bool is_negative() const
{
return x_uint >> (Traits::bits - 1);
}
/// Returns 0 for both +0. and -0.
int sign() const
{
return (exponent() == 0 && mantissa() == 0)
? 0
: (is_negative()
? -1
: 1);
}
uint16_t exponent() const
{
return (x_uint >> (Traits::mantissa_bits)) & (((1ull << (Traits::exponent_bits + 1)) - 1) >> 1);
}
int16_t normalized_exponent() const
{
return int16_t(exponent()) - ((1ull << (Traits::exponent_bits - 1)) - 1);
}
uint64_t mantissa() const
{
return x_uint & ((1ull << Traits::mantissa_bits) - 1);
}
int64_t mantissa_with_sign() const
{
return is_negative() ? -mantissa() : mantissa();
}
/// NOTE Probably floating point instructions can be better.
bool is_integer_in_representable_range() const
{
return x_uint == 0
|| (normalized_exponent() >= 0 /// The number is not less than one
/// The number is inside the range where every integer has exact representation in float
&& normalized_exponent() <= static_cast<int16_t>(Traits::mantissa_bits)
/// After multiplying by 2^exp, the fractional part becomes zero, means the number is integer
&& ((mantissa() & ((1ULL << (Traits::mantissa_bits - normalized_exponent())) - 1)) == 0));
}
/// Compare float with integer of arbitrary width (both signed and unsigned are supported). Assuming two's complement arithmetic.
/// Infinities are compared correctly. NaNs are treat similarly to infinities, so they can be less than all numbers.
/// (note that we need total order)
template <typename Int>
int compare(Int rhs)
{
if (rhs == 0)
return sign();
/// Different signs
if (is_negative() && rhs > 0)
return -1;
if (!is_negative() && rhs < 0)
return 1;
/// Fractional number with magnitude less than one
if (normalized_exponent() < 0)
{
if (!is_negative())
return rhs > 0 ? -1 : 1;
else
return rhs >= 0 ? -1 : 1;
}
/// The case of the most negative integer
if constexpr (is_signed_v<Int>)
{
if (rhs == std::numeric_limits<Int>::lowest())
{
assert(is_negative());
if (normalized_exponent() < static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>))
return 1;
if (normalized_exponent() > static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>))
return -1;
if (mantissa() == 0)
return 0;
else
return -1;
}
}
/// Too large number: abs(float) > abs(rhs). Also the case with infinities and NaN.
if (normalized_exponent() >= static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>))
return is_negative() ? -1 : 1;
using UInt = make_unsigned_t<Int>;
UInt uint_rhs = rhs < 0 ? -rhs : rhs;
/// Smaller octave: abs(rhs) < abs(float)
if (uint_rhs < (static_cast<UInt>(1) << normalized_exponent()))
return is_negative() ? -1 : 1;
/// Larger octave: abs(rhs) > abs(float)
if (normalized_exponent() + 1 < static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>)
&& uint_rhs >= (static_cast<UInt>(1) << (normalized_exponent() + 1)))
return is_negative() ? 1 : -1;
/// The same octave
/// uint_rhs == 2 ^ normalized_exponent + mantissa * 2 ^ (normalized_exponent - mantissa_bits)
bool large_and_always_integer = normalized_exponent() >= static_cast<int16_t>(Traits::mantissa_bits);
typename Traits::UInt a = large_and_always_integer
? mantissa() << (normalized_exponent() - Traits::mantissa_bits)
: mantissa() >> (Traits::mantissa_bits - normalized_exponent());
typename Traits::UInt b = uint_rhs - (static_cast<UInt>(1) << normalized_exponent());
if (a < b)
return is_negative() ? 1 : -1;
if (a > b)
return is_negative() ? -1 : 1;
/// Float has no fractional part means that the numbers are equal.
if (large_and_always_integer || (mantissa() & ((1ULL << (Traits::mantissa_bits - normalized_exponent())) - 1)) == 0)
return 0;
else
/// Float has fractional part means its abs value is larger.
return is_negative() ? -1 : 1;
}
template <typename Int>
bool equals(Int rhs)
{
return compare(rhs) == 0;
}
template <typename Int>
bool notEquals(Int rhs)
{
return compare(rhs) != 0;
}
template <typename Int>
bool less(Int rhs)
{
return compare(rhs) < 0;
}
template <typename Int>
bool greater(Int rhs)
{
return compare(rhs) > 0;
}
template <typename Int>
bool lessOrEquals(Int rhs)
{
return compare(rhs) <= 0;
}
template <typename Int>
bool greaterOrEquals(Int rhs)
{
return compare(rhs) >= 0;
}
};
using DecomposedFloat64 = DecomposedFloat<double>;
using DecomposedFloat32 = DecomposedFloat<float>;

79
base/common/arithmeticOverflow.h
View File

@ -56,27 +56,33 @@ namespace common
}
template <>
inline bool addOverflow(__int128 x, __int128 y, __int128 & res)
inline bool addOverflow(Int128 x, Int128 y, Int128 & res)
{
static constexpr __int128 min_int128 = minInt128();
static constexpr __int128 max_int128 = maxInt128();
res = addIgnoreOverflow(x, y);
return (y > 0 && x > max_int128 - y) || (y < 0 && x < min_int128 - y);
return (y > 0 && x > std::numeric_limits<Int128>::max() - y) ||
(y < 0 && x < std::numeric_limits<Int128>::min() - y);
}
template <>
inline bool addOverflow(wInt256 x, wInt256 y, wInt256 & res)
inline bool addOverflow(UInt128 x, UInt128 y, UInt128 & res)
{
res = addIgnoreOverflow(x, y);
return (y > 0 && x > std::numeric_limits<wInt256>::max() - y) ||
(y < 0 && x < std::numeric_limits<wInt256>::min() - y);
return x > std::numeric_limits<UInt128>::max() - y;
}
template <>
inline bool addOverflow(wUInt256 x, wUInt256 y, wUInt256 & res)
inline bool addOverflow(Int256 x, Int256 y, Int256 & res)
{
res = addIgnoreOverflow(x, y);
return x > std::numeric_limits<wUInt256>::max() - y;
return (y > 0 && x > std::numeric_limits<Int256>::max() - y) ||
(y < 0 && x < std::numeric_limits<Int256>::min() - y);
}
template <>
inline bool addOverflow(UInt256 x, UInt256 y, UInt256 & res)
{
res = addIgnoreOverflow(x, y);
return x > std::numeric_limits<UInt256>::max() - y;
}
template <typename T>
@ -104,24 +110,30 @@ namespace common
}
template <>
inline bool subOverflow(__int128 x, __int128 y, __int128 & res)
inline bool subOverflow(Int128 x, Int128 y, Int128 & res)
{
static constexpr __int128 min_int128 = minInt128();
static constexpr __int128 max_int128 = maxInt128();
res = subIgnoreOverflow(x, y);
return (y < 0 && x > max_int128 + y) || (y > 0 && x < min_int128 + y);
return (y < 0 && x > std::numeric_limits<Int128>::max() + y) ||
(y > 0 && x < std::numeric_limits<Int128>::min() + y);
}
template <>
inline bool subOverflow(wInt256 x, wInt256 y, wInt256 & res)
inline bool subOverflow(UInt128 x, UInt128 y, UInt128 & res)
{
res = subIgnoreOverflow(x, y);
return (y < 0 && x > std::numeric_limits<wInt256>::max() + y) ||
(y > 0 && x < std::numeric_limits<wInt256>::min() + y);
return x < y;
}
template <>
inline bool subOverflow(wUInt256 x, wUInt256 y, wUInt256 & res)
inline bool subOverflow(Int256 x, Int256 y, Int256 & res)
{
res = subIgnoreOverflow(x, y);
return (y < 0 && x > std::numeric_limits<Int256>::max() + y) ||
(y > 0 && x < std::numeric_limits<Int256>::min() + y);
}
template <>
inline bool subOverflow(UInt256 x, UInt256 y, UInt256 & res)
{
res = subIgnoreOverflow(x, y);
return x < y;
@ -151,36 +163,33 @@ namespace common
return __builtin_smulll_overflow(x, y, &res);
}
/// Overflow check is not implemented for big integers.
template <>
inline bool mulOverflow(__int128 x, __int128 y, __int128 & res)
inline bool mulOverflow(Int128 x, Int128 y, Int128 & res)
{
res = mulIgnoreOverflow(x, y);
if (!x || !y)
return false;
unsigned __int128 a = (x > 0) ? x : -x;
unsigned __int128 b = (y > 0) ? y : -y;
return mulIgnoreOverflow(a, b) / b != a;
return false;
}
template <>
inline bool mulOverflow(wInt256 x, wInt256 y, wInt256 & res)
inline bool mulOverflow(Int256 x, Int256 y, Int256 & res)
{
res = mulIgnoreOverflow(x, y);
if (!x || !y)
return false;
wInt256 a = (x > 0) ? x : -x;
wInt256 b = (y > 0) ? y : -y;
return mulIgnoreOverflow(a, b) / b != a;
return false;
}
template <>
inline bool mulOverflow(wUInt256 x, wUInt256 y, wUInt256 & res)
inline bool mulOverflow(UInt128 x, UInt128 y, UInt128 & res)
{
res = mulIgnoreOverflow(x, y);
if (!x || !y)
return false;
return res / y != x;
return false;
}
template <>
inline bool mulOverflow(UInt256 x, UInt256 y, UInt256 & res)
{
res = mulIgnoreOverflow(x, y);
return false;
}
}

49
base/common/extended_types.h
View File

@ -5,16 +5,14 @@
#include <common/types.h>
#include <common/wide_integer.h>
using Int128 = __int128;
using wInt256 = wide::integer<256, signed>;
using wUInt256 = wide::integer<256, unsigned>;
using Int128 = wide::integer<128, signed>;
using UInt128 = wide::integer<128, unsigned>;
using Int256 = wide::integer<256, signed>;
using UInt256 = wide::integer<256, unsigned>;
static_assert(sizeof(wInt256) == 32);
static_assert(sizeof(wUInt256) == 32);
static constexpr __int128 minInt128() { return static_cast<unsigned __int128>(1) << 127; }
static constexpr __int128 maxInt128() { return (static_cast<unsigned __int128>(1) << 127) - 1; }
static_assert(sizeof(Int256) == 32);
static_assert(sizeof(UInt256) == 32);
/// The standard library type traits, such as std::is_arithmetic, with one exception
/// (std::common_type), are "set in stone". Attempting to specialize them causes undefined behavior.
@ -26,7 +24,7 @@ struct is_signed
};
template <> struct is_signed<Int128> { static constexpr bool value = true; };
template <> struct is_signed<wInt256> { static constexpr bool value = true; };
template <> struct is_signed<Int256> { static constexpr bool value = true; };
template <typename T>
inline constexpr bool is_signed_v = is_signed<T>::value;
@ -37,7 +35,8 @@ struct is_unsigned
static constexpr bool value = std::is_unsigned_v<T>;
};
template <> struct is_unsigned<wUInt256> { static constexpr bool value = true; };
template <> struct is_unsigned<UInt128> { static constexpr bool value = true; };
template <> struct is_unsigned<UInt256> { static constexpr bool value = true; };
template <typename T>
inline constexpr bool is_unsigned_v = is_unsigned<T>::value;
@ -51,8 +50,9 @@ struct is_integer
};
template <> struct is_integer<Int128> { static constexpr bool value = true; };
template <> struct is_integer<wInt256> { static constexpr bool value = true; };
template <> struct is_integer<wUInt256> { static constexpr bool value = true; };
template <> struct is_integer<UInt128> { static constexpr bool value = true; };
template <> struct is_integer<Int256> { static constexpr bool value = true; };
template <> struct is_integer<UInt256> { static constexpr bool value = true; };
template <typename T>
inline constexpr bool is_integer_v = is_integer<T>::value;
@ -64,7 +64,11 @@ struct is_arithmetic
static constexpr bool value = std::is_arithmetic_v<T>;
};
template <> struct is_arithmetic<__int128> { static constexpr bool value = true; };
template <> struct is_arithmetic<Int128> { static constexpr bool value = true; };
template <> struct is_arithmetic<UInt128> { static constexpr bool value = true; };
template <> struct is_arithmetic<Int256> { static constexpr bool value = true; };
template <> struct is_arithmetic<UInt256> { static constexpr bool value = true; };
template <typename T>
inline constexpr bool is_arithmetic_v = is_arithmetic<T>::value;
@ -75,9 +79,10 @@ struct make_unsigned
typedef std::make_unsigned_t<T> type;
};
template <> struct make_unsigned<Int128> { using type = unsigned __int128; };
template <> struct make_unsigned<wInt256> { using type = wUInt256; };
template <> struct make_unsigned<wUInt256> { using type = wUInt256; };
template <> struct make_unsigned<Int128> { using type = UInt128; };
template <> struct make_unsigned<UInt128> { using type = UInt128; };
template <> struct make_unsigned<Int256> { using type = UInt256; };
template <> struct make_unsigned<UInt256> { using type = UInt256; };
template <typename T> using make_unsigned_t = typename make_unsigned<T>::type;
@ -87,8 +92,10 @@ struct make_signed
typedef std::make_signed_t<T> type;
};
template <> struct make_signed<wInt256> { using type = wInt256; };
template <> struct make_signed<wUInt256> { using type = wInt256; };
template <> struct make_signed<Int128> { using type = Int128; };
template <> struct make_signed<UInt128> { using type = Int128; };
template <> struct make_signed<Int256> { using type = Int256; };
template <> struct make_signed<UInt256> { using type = Int256; };
template <typename T> using make_signed_t = typename make_signed<T>::type;
@ -98,8 +105,10 @@ struct is_big_int
static constexpr bool value = false;
};
template <> struct is_big_int<wInt256> { static constexpr bool value = true; };
template <> struct is_big_int<wUInt256> { static constexpr bool value = true; };
template <> struct is_big_int<Int128> { static constexpr bool value = true; };
template <> struct is_big_int<UInt128> { static constexpr bool value = true; };
template <> struct is_big_int<Int256> { static constexpr bool value = true; };
template <> struct is_big_int<UInt256> { static constexpr bool value = true; };
template <typename T>
inline constexpr bool is_big_int_v = is_big_int<T>::value;

55
base/common/itoa.h
View File

@ -30,9 +30,8 @@
#include <cstddef>
#include <cstring>
#include <type_traits>
#include <common/extended_types.h>
using int128_t = __int128;
using uint128_t = unsigned __int128;
namespace impl
{
@ -106,7 +105,7 @@ using UnsignedOfSize = typename SelectType
uint16_t,
uint32_t,
uint64_t,
uint128_t
__uint128_t
>::Result;
/// Holds the result of dividing an unsigned N-byte variable by 10^N resulting in
@ -313,7 +312,8 @@ namespace convert
}
}
static inline int digits10(uint128_t x)
template <typename T>
static inline int digits10(T x)
{
if (x < 10ULL)
return 1;
@ -346,8 +346,11 @@ static inline int digits10(uint128_t x)
return 12 + digits10(x / 1000000000000ULL);
}
static inline char * writeUIntText(uint128_t x, char * p)
template <typename T>
static inline char * writeUIntText(T x, char * p)
{
static_assert(is_unsigned_v<T>);
int len = digits10(x);
auto pp = p + len;
while (x >= 100)
@ -370,14 +373,28 @@ static inline char * writeLeadingMinus(char * pos)
return pos + 1;
}
static inline char * writeSIntText(int128_t x, char * pos)
template <typename T>
static inline char * writeSIntText(T x, char * pos)
{
static constexpr int128_t min_int128 = uint128_t(1) << 127;
static_assert(std::is_same_v<T, Int128> || std::is_same_v<T, Int256>);
if (unlikely(x == min_int128))
using UnsignedT = make_unsigned_t<T>;
static constexpr T min_int = UnsignedT(1) << (sizeof(T) * 8 - 1);
if (unlikely(x == min_int))
{
memcpy(pos, "-170141183460469231731687303715884105728", 40);
return pos + 40;
if constexpr (std::is_same_v<T, Int128>)
{
const char * res = "-170141183460469231731687303715884105728";
memcpy(pos, res, strlen(res));
return pos + strlen(res);
}
else if constexpr (std::is_same_v<T, Int256>)
{
const char * res = "-57896044618658097711785492504343953926634992332820282019728792003956564819968";
memcpy(pos, res, strlen(res));
return pos + strlen(res);
}
}
if (x < 0)
@ -385,7 +402,7 @@ static inline char * writeSIntText(int128_t x, char * pos)
x = -x;
pos = writeLeadingMinus(pos);
}
return writeUIntText(static_cast<uint128_t>(x), pos);
return writeUIntText(UnsignedT(x), pos);
}
}
@ -403,13 +420,25 @@ inline char * itoa(char8_t i, char * p)
}
template <>
inline char * itoa<uint128_t>(uint128_t i, char * p)
inline char * itoa(UInt128 i, char * p)
{
return impl::writeUIntText(i, p);
}
template <>
inline char * itoa<int128_t>(int128_t i, char * p)
inline char * itoa(Int128 i, char * p)
{
return impl::writeSIntText(i, p);
}
template <>
inline char * itoa(UInt256 i, char * p)
{
return impl::writeUIntText(i, p);
}
template <>
inline char * itoa(Int256 i, char * p)
{
return impl::writeSIntText(i, p);
}

7
base/common/strong_typedef.h
View File

@ -4,7 +4,8 @@
#include <type_traits>
#include <utility>
template <class T, class Tag>
template <typename T, typename Tag>
struct StrongTypedef
{
private:
@ -38,14 +39,16 @@ public:
bool operator==(const Self & rhs) const { return t == rhs.t; }
bool operator<(const Self & rhs) const { return t < rhs.t; }
bool operator>(const Self & rhs) const { return t > rhs.t; }
T & toUnderType() { return t; }
const T & toUnderType() const { return t; }
};
namespace std
{
template <class T, class Tag>
template <typename T, typename Tag>
struct hash<StrongTypedef<T, Tag>>
{
size_t operator()(const StrongTypedef<T, Tag> & x) const

6
base/common/throwError.h
View File

@ -1,13 +1,15 @@
#pragma once
#include <stdexcept>
/// Throw DB::Exception-like exception before its definition.
/// DB::Exception derived from Poco::Exception derived from std::exception.
/// DB::Exception generally cought as Poco::Exception. std::exception generally has other catch blocks and could lead to other outcomes.
/// DB::Exception generally caught as Poco::Exception. std::exception generally has other catch blocks and could lead to other outcomes.
/// DB::Exception is not defined yet. It'd better to throw Poco::Exception but we do not want to include any big header here, even <string>.
/// So we throw some std::exception instead in the hope its catch block is the same as DB::Exception one.
template <typename T>
inline void throwError(const T & err)
[[noreturn]] inline void throwError(const T & err)
{
throw std::runtime_error(err);
}

47
base/common/wide_integer.h
View File

@ -58,9 +58,11 @@ public:
using signed_base_type = int64_t;
// ctors
constexpr integer() noexcept;
constexpr integer() noexcept = default;
template <typename T>
constexpr integer(T rhs) noexcept;
template <typename T>
constexpr integer(std::initializer_list<T> il) noexcept;
@ -108,9 +110,9 @@ public:
constexpr explicit operator bool() const noexcept;
template <class T>
using __integral_not_wide_integer_class = typename std::enable_if<std::is_arithmetic<T>::value, T>::type;
using _integral_not_wide_integer_class = typename std::enable_if<std::is_arithmetic<T>::value, T>::type;
template <class T, class = __integral_not_wide_integer_class<T>>
template <class T, class = _integral_not_wide_integer_class<T>>
constexpr operator T() const noexcept;
constexpr operator long double() const noexcept;
@ -119,25 +121,27 @@ public:
struct _impl;
base_type items[_impl::item_count];
private:
template <size_t Bits2, typename Signed2>
friend class integer;
friend class std::numeric_limits<integer<Bits, signed>>;
friend class std::numeric_limits<integer<Bits, unsigned>>;
base_type items[_impl::item_count];
};
template <typename T>
static constexpr bool ArithmeticConcept() noexcept;
template <class T1, class T2>
using __only_arithmetic = typename std::enable_if<ArithmeticConcept<T1>() && ArithmeticConcept<T2>()>::type;
using _only_arithmetic = typename std::enable_if<ArithmeticConcept<T1>() && ArithmeticConcept<T2>()>::type;
template <typename T>
static constexpr bool IntegralConcept() noexcept;
template <class T, class T2>
using __only_integer = typename std::enable_if<IntegralConcept<T>() && IntegralConcept<T2>()>::type;
using _only_integer = typename std::enable_if<IntegralConcept<T>() && IntegralConcept<T2>()>::type;
// Unary operators
template <size_t Bits, typename Signed>
@ -153,54 +157,55 @@ constexpr integer<Bits, Signed> operator+(const integer<Bits, Signed> & lhs) noe
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator*(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
std::common_type_t<Arithmetic, Arithmetic2> constexpr operator*(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator/(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
std::common_type_t<Arithmetic, Arithmetic2> constexpr operator/(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator+(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
std::common_type_t<Arithmetic, Arithmetic2> constexpr operator+(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator-(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
std::common_type_t<Arithmetic, Arithmetic2> constexpr operator-(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator%(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
std::common_type_t<Integral, Integral2> constexpr operator%(const Integral & rhs, const Integral2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator&(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
std::common_type_t<Integral, Integral2> constexpr operator&(const Integral & rhs, const Integral2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator|(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
std::common_type_t<Integral, Integral2> constexpr operator|(const Integral & rhs, const Integral2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
operator^(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
std::common_type_t<Integral, Integral2> constexpr operator^(const Integral & rhs, const Integral2 & lhs);
// TODO: Integral
template <size_t Bits, typename Signed>
constexpr integer<Bits, Signed> operator<<(const integer<Bits, Signed> & lhs, int n) noexcept;
template <size_t Bits, typename Signed>
constexpr integer<Bits, Signed> operator>>(const integer<Bits, Signed> & lhs, int n) noexcept;
@ -217,32 +222,32 @@ constexpr integer<Bits, Signed> operator>>(const integer<Bits, Signed> & lhs, In
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator<(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
constexpr bool operator<(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator>(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
constexpr bool operator>(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator<=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
constexpr bool operator<=(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator>=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
constexpr bool operator>=(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator==(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
constexpr bool operator==(const Arithmetic & rhs, const Arithmetic2 & lhs);
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator!=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
constexpr bool operator!=(const Arithmetic & rhs, const Arithmetic2 & lhs);
}

188
base/common/wide_integer_impl.h
View File

@ -5,6 +5,7 @@
/// (See at http://www.boost.org/LICENSE_1_0.txt)
#include "throwError.h"
#include <cmath>
#include <cfloat>
#include <cassert>
@ -81,7 +82,7 @@ public:
res.items[T::_impl::big(0)] = std::numeric_limits<typename wide::integer<Bits, Signed>::signed_base_type>::min();
return res;
}
return 0;
return wide::integer<Bits, Signed>(0);
}
static constexpr wide::integer<Bits, Signed> max() noexcept
@ -176,7 +177,7 @@ struct integer<Bits, Signed>::_impl
constexpr static bool is_negative(const integer<B, T> & n) noexcept
{
if constexpr (std::is_same_v<T, signed>)
return static_cast<signed_base_type>(n.items[big(0)]) < 0;
return static_cast<signed_base_type>(n.items[integer<B, T>::_impl::big(0)]) < 0;
else
return false;
}
@ -193,40 +194,36 @@ struct integer<Bits, Signed>::_impl
template <size_t B, class S>
constexpr static integer<B, S> make_positive(const integer<B, S> & n) noexcept
{
return is_negative(n) ? operator_unary_minus(n) : n;
return is_negative(n) ? integer<B, S>(operator_unary_minus(n)) : n;
}
template <typename T>
__attribute__((no_sanitize("undefined"))) constexpr static auto to_Integral(T f) noexcept
{
if constexpr (std::is_same_v<T, __int128>)
return f;
else if constexpr (std::is_signed_v<T>)
if constexpr (std::is_signed_v<T>)
return static_cast<int64_t>(f);
else
return static_cast<uint64_t>(f);
}
template <typename Integral>
constexpr static void wide_integer_from_bultin(integer<Bits, Signed> & self, Integral rhs) noexcept
constexpr static void wide_integer_from_builtin(integer<Bits, Signed> & self, Integral rhs) noexcept
{
self.items[0] = _impl::to_Integral(rhs);
if constexpr (std::is_same_v<Integral, __int128>)
self.items[1] = rhs >> base_bits;
static_assert(sizeof(Integral) <= sizeof(base_type));
constexpr const unsigned start = (sizeof(Integral) == 16) ? 2 : 1;
self.items[0] = _impl::to_Integral(rhs);
if constexpr (std::is_signed_v<Integral>)
{
if (rhs < 0)
{
for (unsigned i = start; i < item_count; ++i)
for (size_t i = 1; i < item_count; ++i)
self.items[i] = -1;
return;
}
}
for (unsigned i = start; i < item_count; ++i)
for (size_t i = 1; i < item_count; ++i)
self.items[i] = 0;
}
@ -239,7 +236,8 @@ struct integer<Bits, Signed>::_impl
* a_(n - 1) = a_n * max_int + b2, a_n <= max_int <- base case.
*/
template <class T>
constexpr static void set_multiplier(integer<Bits, Signed> & self, T t) noexcept {
constexpr static void set_multiplier(integer<Bits, Signed> & self, T t) noexcept
{
constexpr uint64_t max_int = std::numeric_limits<uint64_t>::max();
/// Implementation specific behaviour on overflow (if we don't check here, stack overflow will triggered in bigint_cast).
@ -260,7 +258,8 @@ struct integer<Bits, Signed>::_impl
self += static_cast<uint64_t>(t - alpha * static_cast<T>(max_int)); // += b_i
}
constexpr static void wide_integer_from_bultin(integer<Bits, Signed>& self, double rhs) noexcept {
constexpr static void wide_integer_from_builtin(integer<Bits, Signed>& self, double rhs) noexcept
{
constexpr int64_t max_int = std::numeric_limits<int64_t>::max();
constexpr int64_t min_int = std::numeric_limits<int64_t>::min();
@ -383,13 +382,13 @@ struct integer<Bits, Signed>::_impl
if (bit_shift)
lhs.items[big(items_shift)] |= std::numeric_limits<base_type>::max() << (base_bits - bit_shift);
for (unsigned i = item_count - items_shift; i < items_shift; ++i)
lhs.items[little(i)] = std::numeric_limits<base_type>::max();
for (unsigned i = 0; i < items_shift; ++i)
lhs.items[big(i)] = std::numeric_limits<base_type>::max();
}
else
{
for (unsigned i = item_count - items_shift; i < items_shift; ++i)
lhs.items[little(i)] = 0;
for (unsigned i = 0; i < items_shift; ++i)
lhs.items[big(i)] = 0;
}
return lhs;
@ -397,23 +396,23 @@ struct integer<Bits, Signed>::_impl
private:
template <typename T>
constexpr static base_type get_item(const T & x, unsigned number)
constexpr static base_type get_item(const T & x, unsigned idx)
{
if constexpr (IsWideInteger<T>::value)
{
if (number < T::_impl::item_count)
return x.items[number];
if (idx < T::_impl::item_count)
return x.items[idx];
return 0;
}
else
{
if constexpr (sizeof(T) <= sizeof(base_type))
{
if (!number)
if (0 == idx)
return x;
}
else if (number * sizeof(base_type) < sizeof(T))
return x >> (number * base_bits); // & std::numeric_limits<base_type>::max()
else if (idx * sizeof(base_type) < sizeof(T))
return x >> (idx * base_bits); // & std::numeric_limits<base_type>::max()
return 0;
}
}
@ -439,7 +438,7 @@ private:
for (unsigned i = 1; i < item_count; ++i)
{
if (underflows[i-1])
if (underflows[i - 1])
{
base_type & res_item = res.items[little(i)];
if (res_item == 0)
@ -472,7 +471,7 @@ private:
for (unsigned i = 1; i < item_count; ++i)
{
if (overflows[i-1])
if (overflows[i - 1])
{
base_type & res_item = res.items[little(i)];
++res_item;
@ -532,6 +531,17 @@ private:
res.items[little(2)] = r12 >> 64;
return res;
}
else if constexpr (Bits == 128 && sizeof(base_type) == 8)
{
using CompilerUInt128 = unsigned __int128;
CompilerUInt128 a = (CompilerUInt128(lhs.items[1]) << 64) + lhs.items[0];
CompilerUInt128 b = (CompilerUInt128(rhs.items[1]) << 64) + rhs.items[0];
CompilerUInt128 c = a * b;
integer<Bits, Signed> res;
res.items[0] = c;
res.items[1] = c >> 64;
return res;
}
else
{
integer<Bits, Signed> res{};
@ -657,7 +667,7 @@ public:
}
template <typename T>
constexpr static bool operator_more(const integer<Bits, Signed> & lhs, const T & rhs) noexcept
constexpr static bool operator_greater(const integer<Bits, Signed> & lhs, const T & rhs) noexcept
{
if constexpr (should_keep_size<T>())
{
@ -677,7 +687,7 @@ public:
else
{
static_assert(IsWideInteger<T>::value);
return std::common_type_t<integer<Bits, Signed>, T>::_impl::operator_more(T(lhs), rhs);
return std::common_type_t<integer<Bits, Signed>, T>::_impl::operator_greater(T(lhs), rhs);
}
}
@ -764,7 +774,6 @@ public:
}
}
private:
template <typename T>
constexpr static bool is_zero(const T & x)
{
@ -781,46 +790,65 @@ private:
}
/// returns quotient as result and remainder in numerator.
template <typename T>
constexpr static T divide(T & numerator, T && denominator)
template <size_t Bits2>
constexpr static integer<Bits2, unsigned> divide(integer<Bits2, unsigned> & numerator, integer<Bits2, unsigned> denominator)
{
if (is_zero(denominator))
throwError("divide by zero");
static_assert(std::is_unsigned_v<Signed>);
T & n = numerator;
T & d = denominator;
T x = 1;
T quotient = 0;
while (!operator_more(d, n) && operator_eq(operator_amp(shift_right(d, base_bits * item_count - 1), 1), 0))
if constexpr (Bits == 128 && sizeof(base_type) == 8)
{
x = shift_left(x, 1);
d = shift_left(d, 1);
using CompilerUInt128 = unsigned __int128;
CompilerUInt128 a = (CompilerUInt128(numerator.items[1]) << 64) + numerator.items[0];
CompilerUInt128 b = (CompilerUInt128(denominator.items[1]) << 64) + denominator.items[0];
CompilerUInt128 c = a / b;
integer<Bits, Signed> res;
res.items[0] = c;
res.items[1] = c >> 64;
CompilerUInt128 remainder = a - b * c;
numerator.items[0] = remainder;
numerator.items[1] = remainder >> 64;
return res;
}
while (!operator_eq(x, 0))
if (is_zero(denominator))
throwError("Division by zero");
integer<Bits2, unsigned> x = 1;
integer<Bits2, unsigned> quotient = 0;
while (!operator_greater(denominator, numerator) && is_zero(operator_amp(shift_right(denominator, Bits2 - 1), 1)))
{
if (!operator_more(d, n))
x = shift_left(x, 1);
denominator = shift_left(denominator, 1);
}
while (!is_zero(x))
{
if (!operator_greater(denominator, numerator))
{
n = operator_minus(n, d);
numerator = operator_minus(numerator, denominator);
quotient = operator_pipe(quotient, x);
}
x = shift_right(x, 1);
d = shift_right(d, 1);
denominator = shift_right(denominator, 1);
}
return quotient;
}
public:
template <typename T>
constexpr static auto operator_slash(const integer<Bits, Signed> & lhs, const T & rhs)
{
if constexpr (should_keep_size<T>())
{
integer<Bits, Signed> numerator = make_positive(lhs);
integer<Bits, Signed> quotient = divide(numerator, make_positive(integer<Bits, Signed>(rhs)));
integer<Bits, unsigned> numerator = make_positive(lhs);
integer<Bits, unsigned> denominator = make_positive(integer<Bits, Signed>(rhs));
integer<Bits, unsigned> quotient = integer<Bits, unsigned>::_impl::divide(numerator, std::move(denominator));
if (std::is_same_v<Signed, signed> && is_negative(rhs) != is_negative(lhs))
quotient = operator_unary_minus(quotient);
@ -838,8 +866,9 @@ public:
{
if constexpr (should_keep_size<T>())
{
integer<Bits, Signed> remainder = make_positive(lhs);
divide(remainder, make_positive(integer<Bits, Signed>(rhs)));
integer<Bits, unsigned> remainder = make_positive(lhs);
integer<Bits, unsigned> denominator = make_positive(integer<Bits, Signed>(rhs));
integer<Bits, unsigned>::_impl::divide(remainder, std::move(denominator));
if (std::is_same_v<Signed, signed> && is_negative(lhs))
remainder = operator_unary_minus(remainder);
@ -905,7 +934,7 @@ public:
++c;
}
else
throwError("invalid char from");
throwError("Invalid char from");
}
}
else
@ -913,7 +942,7 @@ public:
while (*c)
{
if (*c < '0' || *c > '9')
throwError("invalid char from");
throwError("Invalid char from");
res = multiply(res, 10U);
res = plus(res, *c - '0');
@ -930,11 +959,6 @@ public:
// Members
template <size_t Bits, typename Signed>
constexpr integer<Bits, Signed>::integer() noexcept
: items{}
{}
template <size_t Bits, typename Signed>
template <typename T>
constexpr integer<Bits, Signed>::integer(T rhs) noexcept
@ -943,7 +967,7 @@ constexpr integer<Bits, Signed>::integer(T rhs) noexcept
if constexpr (IsWideInteger<T>::value)
_impl::wide_integer_from_wide_integer(*this, rhs);
else
_impl::wide_integer_from_bultin(*this, rhs);
_impl::wide_integer_from_builtin(*this, rhs);
}
template <size_t Bits, typename Signed>
@ -956,10 +980,19 @@ constexpr integer<Bits, Signed>::integer(std::initializer_list<T> il) noexcept
if constexpr (IsWideInteger<T>::value)
_impl::wide_integer_from_wide_integer(*this, *il.begin());
else
_impl::wide_integer_from_bultin(*this, *il.begin());
_impl::wide_integer_from_builtin(*this, *il.begin());
}
else if (il.size() == 0)
{
_impl::wide_integer_from_builtin(*this, 0);
}
else
_impl::wide_integer_from_bultin(*this, 0);
{
auto it = il.begin();
for (size_t i = 0; i < _impl::item_count; ++i)
if (it < il.end())
items[i] = *it;
}
}
template <size_t Bits, typename Signed>
@ -974,7 +1007,7 @@ template <size_t Bits, typename Signed>
template <typename T>
constexpr integer<Bits, Signed> & integer<Bits, Signed>::operator=(T rhs) noexcept
{
_impl::wide_integer_from_bultin(*this, rhs);
_impl::wide_integer_from_builtin(*this, rhs);
return *this;
}
@ -1057,7 +1090,7 @@ constexpr integer<Bits, Signed> & integer<Bits, Signed>::operator>>=(int n) noex
{
if (static_cast<size_t>(n) >= Bits)
{
if (is_negative(*this))
if (_impl::is_negative(*this))
*this = -1;
else
*this = 0;
@ -1107,16 +1140,17 @@ template <size_t Bits, typename Signed>
template <class T, class>
constexpr integer<Bits, Signed>::operator T() const noexcept
{
if constexpr (std::is_same_v<T, __int128>)
{
static_assert(Bits >= 128);
return (__int128(items[1]) << 64) | items[0];
}
else
{
static_assert(std::numeric_limits<T>::is_integer);
return items[0];
}
static_assert(std::numeric_limits<T>::is_integer);
/// NOTE: memcpy will suffice, but unfortunately, this function is constexpr.
using UnsignedT = std::make_unsigned_t<T>;
UnsignedT res{};
for (unsigned i = 0; i < _impl::item_count && i < (sizeof(T) + sizeof(base_type) - 1) / sizeof(base_type); ++i)
res += UnsignedT(items[i]) << (sizeof(base_type) * 8 * i);
return res;
}
template <size_t Bits, typename Signed>
@ -1280,7 +1314,7 @@ template <size_t Bits, typename Signed>
constexpr integer<Bits, Signed> operator<<(const integer<Bits, Signed> & lhs, int n) noexcept
{
if (static_cast<size_t>(n) >= Bits)
return 0;
return integer<Bits, Signed>(0);
if (n <= 0)
return lhs;
return integer<Bits, Signed>::_impl::shift_left(lhs, n);
@ -1289,7 +1323,7 @@ template <size_t Bits, typename Signed>
constexpr integer<Bits, Signed> operator>>(const integer<Bits, Signed> & lhs, int n) noexcept
{
if (static_cast<size_t>(n) >= Bits)
return 0;
return integer<Bits, Signed>(0);
if (n <= 0)
return lhs;
return integer<Bits, Signed>::_impl::shift_right(lhs, n);
@ -1309,7 +1343,7 @@ constexpr bool operator<(const Arithmetic & lhs, const Arithmetic2 & rhs)
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator>(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs)
{
return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_more(lhs, rhs);
return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_greater(lhs, rhs);
}
template <typename Arithmetic, typename Arithmetic2, class>
constexpr bool operator>(const Arithmetic & lhs, const Arithmetic2 & rhs)
@ -1332,7 +1366,7 @@ constexpr bool operator<=(const Arithmetic & lhs, const Arithmetic2 & rhs)
template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
constexpr bool operator>=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs)
{
return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_more(lhs, rhs)
return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_greater(lhs, rhs)
|| std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_eq(lhs, rhs);
}
template <typename Arithmetic, typename Arithmetic2, class>

34
base/common/wide_integer_to_string.h
View File

@ -1,9 +1,12 @@
#pragma once
#include <string>
#include <ostream>
#include <fmt/format.h>
#include "wide_integer.h"
namespace wide
{
@ -33,3 +36,34 @@ inline std::string to_string(const integer<Bits, Signed> & n)
}
}
template <size_t Bits, typename Signed>
std::ostream & operator<<(std::ostream & out, const wide::integer<Bits, Signed> & value)
{
return out << to_string(value);
}
/// See https://fmt.dev/latest/api.html#formatting-user-defined-types
template <size_t Bits, typename Signed>
struct fmt::formatter<wide::integer<Bits, Signed>>
{
constexpr auto parse(format_parse_context & ctx)
{
auto it = ctx.begin();
auto end = ctx.end();
/// Only support {}.
if (it != end && *it != '}')
throw format_error("invalid format");
return it;
}
template <typename FormatContext>
auto format(const wide::integer<Bits, Signed> & value, FormatContext & ctx)
{
return format_to(ctx.out(), "{}", to_string(value));
}
};

7
base/daemon/BaseDaemon.cpp
View File

@ -468,7 +468,7 @@ void BaseDaemon::reloadConfiguration()
* instead of using files specified in config.xml.
* (It's convenient to log in console when you start server without any command line parameters.)
*/
config_path = config().getString("config-file", "config.xml");
config_path = config().getString("config-file", getDefaultConfigFileName());
DB::ConfigProcessor config_processor(config_path, false, true);
config_processor.setConfigPath(Poco::Path(config_path).makeParent().toString());
loaded_config = config_processor.loadConfig(/* allow_zk_includes = */ true);
@ -516,6 +516,11 @@ std::string BaseDaemon::getDefaultCorePath() const
return "/opt/cores/";
}
std::string BaseDaemon::getDefaultConfigFileName() const
{
return "config.xml";
}
void BaseDaemon::closeFDs()
{
#if defined(OS_FREEBSD) || defined(OS_DARWIN)

2
base/daemon/BaseDaemon.h
View File

@ -149,6 +149,8 @@ protected:
virtual std::string getDefaultCorePath() const;
virtual std::string getDefaultConfigFileName() const;
std::optional<DB::StatusFile> pid_file;
std::atomic_bool is_cancelled{false};

2
base/daemon/SentryWriter.cpp
View File

@ -102,7 +102,7 @@ void SentryWriter::initialize(Poco::Util::LayeredConfiguration & config)
auto * logger = &Poco::Logger::get("SentryWriter");
if (config.getBool("send_crash_reports.enabled", false))
{
if (debug || (strlen(VERSION_OFFICIAL) > 0))
if (debug || (strlen(VERSION_OFFICIAL) > 0)) //-V560
{
enabled = true;
}

2
base/glibc-compatibility/CMakeLists.txt
View File

@ -15,7 +15,7 @@ if (GLIBC_COMPATIBILITY)
add_headers_and_sources(glibc_compatibility .)
add_headers_and_sources(glibc_compatibility musl)
if (ARCH_ARM)
if (ARCH_AARCH64)
list (APPEND glibc_compatibility_sources musl/aarch64/syscall.s musl/aarch64/longjmp.s)
set (musl_arch_include_dir musl/aarch64)
elseif (ARCH_AMD64)

3
base/glibc-compatibility/musl/lgamma.c
View File

@ -78,6 +78,9 @@
*
*/
// Disable warnings by PVS-Studio
//-V::GA
static const double
pi = 3.14159265358979311600e+00, /* 0x400921FB, 0x54442D18 */
a0 = 7.72156649015328655494e-02, /* 0x3FB3C467, 0xE37DB0C8 */

3
base/glibc-compatibility/musl/lgammal.c
View File

@ -85,6 +85,9 @@
*
*/
// Disable warnings by PVS-Studio
//-V::GA
#include <stdint.h>
#include <math.h>
#include "libm.h"

6
base/glibc-compatibility/musl/libm.h
View File

@ -155,7 +155,7 @@ static inline long double fp_barrierl(long double x)
static inline void fp_force_evalf(float x)
{
volatile float y;
y = x;
y = x; //-V1001
}
#endif
@ -164,7 +164,7 @@ static inline void fp_force_evalf(float x)
static inline void fp_force_eval(double x)
{
volatile double y;
y = x;
y = x; //-V1001
}
#endif
@ -173,7 +173,7 @@ static inline void fp_force_eval(double x)
static inline void fp_force_evall(long double x)
{
volatile long double y;
y = x;
y = x; //-V1001
}
#endif

3
base/glibc-compatibility/musl/powf.c
View File

@ -3,6 +3,9 @@
* SPDX-License-Identifier: MIT
*/
// Disable warnings by PVS-Studio
//-V::GA
#include <math.h>
#include <stdint.h>
#include "libm.h"

64
base/loggers/Loggers.cpp
View File

@ -40,7 +40,7 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
split->addTextLog(log, text_log_max_priority);
auto current_logger = config.getString("logger", "");
if (config_logger == current_logger)
if (config_logger == current_logger) //-V1051
return;
config_logger = current_logger;
@ -51,12 +51,22 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
/// Use extended interface of Channel for more comprehensive logging.
split = new DB::OwnSplitChannel();
auto log_level = config.getString("logger.level", "trace");
auto log_level_string = config.getString("logger.level", "trace");
/// different channels (log, console, syslog) may have different loglevels configured
/// The maximum (the most verbose) of those will be used as default for Poco loggers
int max_log_level = 0;
const auto log_path = config.getString("logger.log", "");
if (!log_path.empty())
{
createDirectory(log_path);
std::cerr << "Logging " << log_level << " to " << log_path << std::endl;
std::cerr << "Logging " << log_level_string << " to " << log_path << std::endl;
auto log_level = Poco::Logger::parseLevel(log_level_string);
if (log_level > max_log_level)
{
max_log_level = log_level;
}
// Set up two channel chains.
log_file = new Poco::FileChannel;
@ -69,9 +79,10 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
log_file->setProperty(Poco::FileChannel::PROP_ROTATEONOPEN, config.getRawString("logger.rotateOnOpen", "false"));
log_file->open();
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter(this);
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter;
Poco::AutoPtr<DB::OwnFormattingChannel> log = new DB::OwnFormattingChannel(pf, log_file);
log->setLevel(log_level);
split->addChannel(log);
}
@ -79,6 +90,15 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
if (!errorlog_path.empty())
{
createDirectory(errorlog_path);
// NOTE: we don't use notice & critical in the code, so in practice error log collects fatal & error & warning.
// (!) Warnings are important, they require attention and should never be silenced / ignored.
auto errorlog_level = Poco::Logger::parseLevel(config.getString("logger.errorlog_level", "notice"));
if (errorlog_level > max_log_level)
{
max_log_level = errorlog_level;
}
std::cerr << "Logging errors to " << errorlog_path << std::endl;
error_log_file = new Poco::FileChannel;
@ -90,20 +110,22 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
error_log_file->setProperty(Poco::FileChannel::PROP_FLUSH, config.getRawString("logger.flush", "true"));
error_log_file->setProperty(Poco::FileChannel::PROP_ROTATEONOPEN, config.getRawString("logger.rotateOnOpen", "false"));
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter(this);
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter;
Poco::AutoPtr<DB::OwnFormattingChannel> errorlog = new DB::OwnFormattingChannel(pf, error_log_file);
errorlog->setLevel(Poco::Message::PRIO_NOTICE);
errorlog->setLevel(errorlog_level);
errorlog->open();
split->addChannel(errorlog);
}
/// "dynamic_layer_selection" is needed only for Yandex.Metrika, that share part of ClickHouse code.
/// We don't need this configuration parameter.
if (config.getBool("logger.use_syslog", false) || config.getBool("dynamic_layer_selection", false))
if (config.getBool("logger.use_syslog", false))
{
//const std::string & cmd_name = commandName();
auto syslog_level = Poco::Logger::parseLevel(config.getString("logger.syslog_level", log_level_string));
if (syslog_level > max_log_level)
{
max_log_level = syslog_level;
}
if (config.has("logger.syslog.address"))
{
@ -127,9 +149,11 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
}
syslog_channel->open();
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter(this, OwnPatternFormatter::ADD_LAYER_TAG);
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter;
Poco::AutoPtr<DB::OwnFormattingChannel> log = new DB::OwnFormattingChannel(pf, syslog_channel);
log->setLevel(syslog_level);
split->addChannel(log);
}
@ -141,9 +165,17 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
{
bool color_enabled = config.getBool("logger.color_terminal", color_logs_by_default);
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter(this, OwnPatternFormatter::ADD_NOTHING, color_enabled);
auto console_log_level_string = config.getString("logger.console_log_level", log_level_string);
auto console_log_level = Poco::Logger::parseLevel(console_log_level_string);
if (console_log_level > max_log_level)
{
max_log_level = console_log_level;
}
Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter(color_enabled);
Poco::AutoPtr<DB::OwnFormattingChannel> log = new DB::OwnFormattingChannel(pf, new Poco::ConsoleChannel);
logger.warning("Logging " + log_level + " to console");
logger.warning("Logging " + console_log_level_string + " to console");
log->setLevel(console_log_level);
split->addChannel(log);
}
@ -152,17 +184,17 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
logger.setChannel(split);
// Global logging level (it can be overridden for specific loggers).
logger.setLevel(log_level);
logger.setLevel(max_log_level);
// Set level to all already created loggers
std::vector<std::string> names;
//logger_root = Logger::root();
logger.root().names(names);
for (const auto & name : names)
logger.root().get(name).setLevel(log_level);
logger.root().get(name).setLevel(max_log_level);
// Attach to the root logger.
logger.root().setLevel(log_level);
logger.root().setLevel(max_log_level);
logger.root().setChannel(logger.getChannel());
// Explicitly specified log levels for specific loggers.

9
base/loggers/Loggers.h
View File

@ -8,6 +8,7 @@
#include <Interpreters/TextLog.h>
#include "OwnSplitChannel.h"
namespace Poco::Util
{
class AbstractConfiguration;
@ -21,16 +22,8 @@ public:
/// Close log files. On next log write files will be reopened.
void closeLogs(Poco::Logger & logger);
std::optional<size_t> getLayer() const
{
return layer; /// layer set in inheritor class BaseDaemonApplication.
}
void setTextLog(std::shared_ptr<DB::TextLog> log, int max_priority);
protected:
std::optional<size_t> layer;
private:
Poco::AutoPtr<Poco::FileChannel> log_file;
Poco::AutoPtr<Poco::FileChannel> error_log_file;

3
base/loggers/OwnFormattingChannel.h
View File

@ -22,6 +22,9 @@ public:
void setLevel(Poco::Message::Priority priority_) { priority = priority_; }
// Poco::Logger::parseLevel returns ints
void setLevel(int level) { priority = static_cast<Poco::Message::Priority>(level); }
void open() override
{
if (pChannel)

19
base/loggers/OwnPatternFormatter.cpp
View File

@ -13,31 +13,18 @@
#include "Loggers.h"
OwnPatternFormatter::OwnPatternFormatter(const Loggers * loggers_, OwnPatternFormatter::Options options_, bool color_)
: Poco::PatternFormatter(""), loggers(loggers_), options(options_), color(color_)
OwnPatternFormatter::OwnPatternFormatter(bool color_)
: Poco::PatternFormatter(""), color(color_)
{
}
void OwnPatternFormatter::formatExtended(const DB::ExtendedLogMessage & msg_ext, std::string & text)
void OwnPatternFormatter::formatExtended(const DB::ExtendedLogMessage & msg_ext, std::string & text) const
{
DB::WriteBufferFromString wb(text);
const Poco::Message & msg = msg_ext.base;
/// For syslog: tag must be before message and first whitespace.
/// This code is only used in Yandex.Metrika and unneeded in ClickHouse.
if ((options & ADD_LAYER_TAG) && loggers)
{
auto layer = loggers->getLayer();
if (layer)
{
writeCString("layer[", wb);
DB::writeIntText(*layer, wb);
writeCString("]: ", wb);
}
}
/// Change delimiters in date for compatibility with old logs.
DB::writeDateTimeText<'.', ':'>(msg_ext.time_seconds, wb);

13
base/loggers/OwnPatternFormatter.h
View File

@ -24,20 +24,11 @@ class Loggers;
class OwnPatternFormatter : public Poco::PatternFormatter
{
public:
/// ADD_LAYER_TAG is needed only for Yandex.Metrika, that share part of ClickHouse code.
enum Options
{
ADD_NOTHING = 0,
ADD_LAYER_TAG = 1 << 0
};
OwnPatternFormatter(const Loggers * loggers_, Options options_ = ADD_NOTHING, bool color_ = false);
OwnPatternFormatter(bool color_ = false);
void format(const Poco::Message & msg, std::string & text) override;
void formatExtended(const DB::ExtendedLogMessage & msg_ext, std::string & text);
void formatExtended(const DB::ExtendedLogMessage & msg_ext, std::string & text) const;
private:
const Loggers * loggers;
Options options;
bool color;
};

10
base/mysqlxx/PoolWithFailover.cpp
View File

@ -78,6 +78,8 @@ PoolWithFailover::PoolWithFailover(
const RemoteDescription & addresses,
const std::string & user,
const std::string & password,
unsigned default_connections_,
unsigned max_connections_,
size_t max_tries_)
: max_tries(max_tries_)
, shareable(false)
@ -85,7 +87,13 @@ PoolWithFailover::PoolWithFailover(
/// Replicas have the same priority, but traversed replicas are moved to the end of the queue.
for (const auto & [host, port] : addresses)
{
replicas_by_priority[0].emplace_back(std::make_shared<Pool>(database, host, user, password, port));
replicas_by_priority[0].emplace_back(std::make_shared<Pool>(database,
host, user, password, port,
/* socket_ = */ "",
MYSQLXX_DEFAULT_TIMEOUT,
MYSQLXX_DEFAULT_RW_TIMEOUT,
default_connections_,
max_connections_));
}
}

2
base/mysqlxx/PoolWithFailover.h
View File

@ -115,6 +115,8 @@ namespace mysqlxx
const RemoteDescription & addresses,
const std::string & user,
const std::string & password,
unsigned default_connections_ = MYSQLXX_POOL_WITH_FAILOVER_DEFAULT_START_CONNECTIONS,
unsigned max_connections_ = MYSQLXX_POOL_WITH_FAILOVER_DEFAULT_MAX_CONNECTIONS,
size_t max_tries_ = MYSQLXX_POOL_WITH_FAILOVER_DEFAULT_MAX_TRIES);
PoolWithFailover(const PoolWithFailover & other);

65
base/pcg-random/pcg_extras.hpp
View File

@ -447,69 +447,6 @@ inline SrcIter uneven_copy(SrcIter src_first,
std::integral_constant<bool, DEST_IS_SMALLER>{});
}
/* generate_to, fill in a fixed-size array of integral type using a SeedSeq
* (actually works for any random-access iterator)
*/
template <size_t size, typename SeedSeq, typename DestIter>
inline void generate_to_impl(SeedSeq&& generator, DestIter dest,
std::true_type)
{
generator.generate(dest, dest+size);
}
template <size_t size, typename SeedSeq, typename DestIter>
void generate_to_impl(SeedSeq&& generator, DestIter dest,
std::false_type)
{
typedef typename std::iterator_traits<DestIter>::value_type dest_t;
constexpr auto DEST_SIZE = sizeof(dest_t);
constexpr auto GEN_SIZE = sizeof(uint32_t);
constexpr bool GEN_IS_SMALLER = GEN_SIZE < DEST_SIZE;
constexpr size_t FROM_ELEMS =
GEN_IS_SMALLER
? size * ((DEST_SIZE+GEN_SIZE-1) / GEN_SIZE)
: (size + (GEN_SIZE / DEST_SIZE) - 1)
/ ((GEN_SIZE / DEST_SIZE) + GEN_IS_SMALLER);
// this odd code ^^^^^^^^^^^^^^^^^ is work-around for
// a bug: http://llvm.org/bugs/show_bug.cgi?id=21287
if (FROM_ELEMS <= 1024) {
uint32_t buffer[FROM_ELEMS];
generator.generate(buffer, buffer+FROM_ELEMS);
uneven_copy(buffer, dest, dest+size);
} else {
uint32_t* buffer = static_cast<uint32_t*>(malloc(GEN_SIZE * FROM_ELEMS));
generator.generate(buffer, buffer+FROM_ELEMS);
uneven_copy(buffer, dest, dest+size);
free(static_cast<void*>(buffer));
}
}
template <size_t size, typename SeedSeq, typename DestIter>
inline void generate_to(SeedSeq&& generator, DestIter dest)
{
typedef typename std::iterator_traits<DestIter>::value_type dest_t;
constexpr bool IS_32BIT = sizeof(dest_t) == sizeof(uint32_t);
generate_to_impl<size>(std::forward<SeedSeq>(generator), dest,
std::integral_constant<bool, IS_32BIT>{});
}
/* generate_one, produce a value of integral type using a SeedSeq
* (optionally, we can have it produce more than one and pick which one
* we want)
*/
template <typename UInt, size_t i = 0UL, size_t N = i+1UL, typename SeedSeq>
inline UInt generate_one(SeedSeq&& generator)
{
UInt result[N];
generate_to<N>(std::forward<SeedSeq>(generator), result);
return result[i];
}
template <typename RngType>
auto bounded_rand(RngType& rng, typename RngType::result_type upper_bound)
-> typename RngType::result_type
@ -517,7 +454,7 @@ auto bounded_rand(RngType& rng, typename RngType::result_type upper_bound)
typedef typename RngType::result_type rtype;
rtype threshold = (RngType::max() - RngType::min() + rtype(1) - upper_bound)
% upper_bound;
for (;;) {
for (;;) { //-V1044
rtype r = rng() - RngType::min();
if (r >= threshold)
return r % upper_bound;

8
base/pcg-random/pcg_random.hpp
View File

@ -928,7 +928,7 @@ struct rxs_m_xs_mixin {
constexpr bitcount_t shift = bits - xtypebits;
constexpr bitcount_t mask = (1 << opbits) - 1;
bitcount_t rshift =
opbits ? bitcount_t(internal >> (bits - opbits)) & mask : 0;
opbits ? bitcount_t(internal >> (bits - opbits)) & mask : 0; //-V547
internal ^= internal >> (opbits + rshift);
internal *= mcg_multiplier<itype>::multiplier();
xtype result = internal >> shift;
@ -950,7 +950,7 @@ struct rxs_m_xs_mixin {
internal *= mcg_unmultiplier<itype>::unmultiplier();
bitcount_t rshift = opbits ? (internal >> (bits - opbits)) & mask : 0;
bitcount_t rshift = opbits ? (internal >> (bits - opbits)) & mask : 0; //-V547
internal = unxorshift(internal, bits, opbits + rshift);
return internal;
@ -975,7 +975,7 @@ struct rxs_m_mixin {
: 2;
constexpr bitcount_t shift = bits - xtypebits;
constexpr bitcount_t mask = (1 << opbits) - 1;
bitcount_t rshift = opbits ? (internal >> (bits - opbits)) & mask : 0;
bitcount_t rshift = opbits ? (internal >> (bits - opbits)) & mask : 0; //-V547
internal ^= internal >> (opbits + rshift);
internal *= mcg_multiplier<itype>::multiplier();
xtype result = internal >> shift;
@ -1366,7 +1366,7 @@ void extended<table_pow2,advance_pow2,baseclass,extvalclass,kdd>::selfinit()
// - any strange correlations would only be apparent if we
// were to backstep the generator so that the base generator
// was generating the same values again
result_type xdiff = baseclass::operator()() - baseclass::operator()();
result_type xdiff = baseclass::operator()() - baseclass::operator()(); //-V501
for (size_t i = 0; i < table_size; ++i) {
data_[i] = baseclass::operator()() ^ xdiff;
}

10
cmake/autogenerated_versions.txt
View File

@ -1,9 +1,9 @@
# This strings autochanged from release_lib.sh:
SET(VERSION_REVISION 54451)
SET(VERSION_REVISION 54452)
SET(VERSION_MAJOR 21)
SET(VERSION_MINOR 6)
SET(VERSION_MINOR 7)
SET(VERSION_PATCH 1)
SET(VERSION_GITHASH 96fced4c3cf432fb0b401d2ab01f0c56e5f74a96)
SET(VERSION_DESCRIBE v21.6.1.1-prestable)
SET(VERSION_STRING 21.6.1.1)
SET(VERSION_GITHASH 976ccc2e908ac3bc28f763bfea8134ea0a121b40)
SET(VERSION_DESCRIBE v21.7.1.1-prestable)
SET(VERSION_STRING 21.7.1.1)
# end of autochange

6
cmake/check_flags.cmake Normal file
View File

@ -0,0 +1,6 @@
include (CheckCXXCompilerFlag)
include (CheckCCompilerFlag)
check_cxx_compiler_flag("-Wsuggest-destructor-override" HAS_SUGGEST_DESTRUCTOR_OVERRIDE)
check_cxx_compiler_flag("-Wshadow" HAS_SHADOW)
check_cxx_compiler_flag("-Wsuggest-override" HAS_SUGGEST_OVERRIDE)

98
cmake/find/llvm.cmake
View File

@ -1,102 +1,34 @@
if (APPLE OR SPLIT_SHARED_LIBRARIES OR NOT ARCH_AMD64)
if (APPLE OR SPLIT_SHARED_LIBRARIES OR NOT ARCH_AMD64 OR SANITIZE STREQUAL "undefined")
set (ENABLE_EMBEDDED_COMPILER OFF CACHE INTERNAL "")
endif()
option (ENABLE_EMBEDDED_COMPILER "Set to TRUE to enable support for 'compile_expressions' option for query execution" ${ENABLE_LIBRARIES})
# Broken in macos. TODO: update clang, re-test, enable on Apple
if (ENABLE_EMBEDDED_COMPILER AND NOT SPLIT_SHARED_LIBRARIES AND ARCH_AMD64 AND NOT (SANITIZE STREQUAL "undefined"))
option (USE_INTERNAL_LLVM_LIBRARY "Use bundled or system LLVM library." ${NOT_UNBUNDLED})
endif()
option (ENABLE_EMBEDDED_COMPILER "Enable support for 'compile_expressions' option for query execution" ON)
if (NOT ENABLE_EMBEDDED_COMPILER)
if(USE_INTERNAL_LLVM_LIBRARY)
message (${RECONFIGURE_MESSAGE_LEVEL} "Cannot use internal LLVM library with ENABLE_EMBEDDED_COMPILER=OFF")
endif()
set (USE_EMBEDDED_COMPILER 0)
return()
endif()
if (NOT EXISTS "${ClickHouse_SOURCE_DIR}/contrib/llvm/llvm/CMakeLists.txt")
if (USE_INTERNAL_LLVM_LIBRARY)
message (WARNING "submodule contrib/llvm is missing. to fix try run: \n git submodule update --init --recursive")
message (${RECONFIGURE_MESSAGE_LEVEL} "Can't fidd internal LLVM library")
endif()
set (MISSING_INTERNAL_LLVM_LIBRARY 1)
message (${RECONFIGURE_MESSAGE_LEVEL} "submodule /contrib/llvm is missing. to fix try run: \n git submodule update --init --recursive")
endif ()
if (NOT USE_INTERNAL_LLVM_LIBRARY)
set (LLVM_PATHS "/usr/local/lib/llvm" "/usr/lib/llvm")
set (USE_EMBEDDED_COMPILER 1)
foreach(llvm_v 11.1 11)
if (NOT LLVM_FOUND)
find_package (LLVM ${llvm_v} CONFIG PATHS ${LLVM_PATHS})
endif ()
endforeach ()
set (LLVM_FOUND 1)
set (LLVM_VERSION "12.0.0bundled")
set (LLVM_INCLUDE_DIRS
"${ClickHouse_SOURCE_DIR}/contrib/llvm/llvm/include"
"${ClickHouse_BINARY_DIR}/contrib/llvm/llvm/include"
)
set (LLVM_LIBRARY_DIRS "${ClickHouse_BINARY_DIR}/contrib/llvm/llvm")
if (LLVM_FOUND)
# Remove dynamically-linked zlib and libedit from LLVM's dependencies:
set_target_properties(LLVMSupport PROPERTIES INTERFACE_LINK_LIBRARIES "-lpthread;LLVMDemangle;${ZLIB_LIBRARIES}")
set_target_properties(LLVMLineEditor PROPERTIES INTERFACE_LINK_LIBRARIES "LLVMSupport")
option(LLVM_HAS_RTTI "Enable if LLVM was build with RTTI enabled" ON)
set (USE_EMBEDDED_COMPILER 1)
else()
message (${RECONFIGURE_MESSAGE_LEVEL} "Can't find system LLVM")
set (USE_EMBEDDED_COMPILER 0)
endif()
if (LLVM_FOUND AND OS_LINUX AND USE_LIBCXX AND NOT FORCE_LLVM_WITH_LIBCXX)
message(WARNING "Option USE_INTERNAL_LLVM_LIBRARY is not set but the LLVM library from OS packages "
"in Linux is incompatible with libc++ ABI. LLVM Will be disabled. Force: -DFORCE_LLVM_WITH_LIBCXX=ON")
message (${RECONFIGURE_MESSAGE_LEVEL} "Unsupported LLVM configuration, cannot enable LLVM")
set (LLVM_FOUND 0)
set (USE_EMBEDDED_COMPILER 0)
endif ()
endif()
if(NOT LLVM_FOUND AND NOT MISSING_INTERNAL_LLVM_LIBRARY)
if (CMAKE_CURRENT_SOURCE_DIR STREQUAL CMAKE_CURRENT_BINARY_DIR)
message(WARNING "Option ENABLE_EMBEDDED_COMPILER is set but internal LLVM library cannot build if build directory is the same as source directory.")
set (LLVM_FOUND 0)
set (USE_EMBEDDED_COMPILER 0)
elseif (SPLIT_SHARED_LIBRARIES)
# llvm-tablegen cannot find shared libraries that we build. Probably can be easily fixed.
message(WARNING "Option USE_INTERNAL_LLVM_LIBRARY is not compatible with SPLIT_SHARED_LIBRARIES. Build of LLVM will be disabled.")
set (LLVM_FOUND 0)
set (USE_EMBEDDED_COMPILER 0)
elseif (NOT ARCH_AMD64)
# It's not supported yet, but you can help.
message(WARNING "Option USE_INTERNAL_LLVM_LIBRARY is only available for x86_64. Build of LLVM will be disabled.")
set (LLVM_FOUND 0)
set (USE_EMBEDDED_COMPILER 0)
elseif (SANITIZE STREQUAL "undefined")
# llvm-tblgen, that is used during LLVM build, doesn't work with UBSan.
message(WARNING "Option USE_INTERNAL_LLVM_LIBRARY does not work with UBSan, because 'llvm-tblgen' tool from LLVM has undefined behaviour. Build of LLVM will be disabled.")
set (LLVM_FOUND 0)
set (USE_EMBEDDED_COMPILER 0)
else ()
set (USE_INTERNAL_LLVM_LIBRARY ON)
set (LLVM_FOUND 1)
set (USE_EMBEDDED_COMPILER 1)
set (LLVM_VERSION "9.0.0bundled")
set (LLVM_INCLUDE_DIRS
"${ClickHouse_SOURCE_DIR}/contrib/llvm/llvm/include"
"${ClickHouse_BINARY_DIR}/contrib/llvm/llvm/include"
)
set (LLVM_LIBRARY_DIRS "${ClickHouse_BINARY_DIR}/contrib/llvm/llvm")
endif()
endif()
if (LLVM_FOUND)
message(STATUS "LLVM include Directory: ${LLVM_INCLUDE_DIRS}")
message(STATUS "LLVM library Directory: ${LLVM_LIBRARY_DIRS}")
message(STATUS "LLVM C++ compiler flags: ${LLVM_CXXFLAGS}")
else()
message (${RECONFIGURE_MESSAGE_LEVEL} "Can't enable LLVM")
endif()
message(STATUS "LLVM include Directory: ${LLVM_INCLUDE_DIRS}")
message(STATUS "LLVM library Directory: ${LLVM_LIBRARY_DIRS}")
message(STATUS "LLVM C++ compiler flags: ${LLVM_CXXFLAGS}")
# This list was generated by listing all LLVM libraries, compiling the binary and removing all libraries while it still compiles.
set (REQUIRED_LLVM_LIBRARIES
LLVMOrcJIT
LLVMExecutionEngine
LLVMRuntimeDyld
LLVMX86CodeGen

17
cmake/find/termcap.cmake
View File

@ -1,17 +0,0 @@
if (ENABLE_EMBEDDED_COMPILER AND NOT USE_INTERNAL_LLVM_LIBRARY AND USE_STATIC_LIBRARIES)
find_library (TERMCAP_LIBRARY tinfo)
if (NOT TERMCAP_LIBRARY)
find_library (TERMCAP_LIBRARY ncurses)
endif()
if (NOT TERMCAP_LIBRARY)
find_library (TERMCAP_LIBRARY termcap)
endif()
if (NOT TERMCAP_LIBRARY)
message (FATAL_ERROR "Statically Linking external LLVM requires termcap")
endif()
target_link_libraries(LLVMSupport INTERFACE ${TERMCAP_LIBRARY})
message (STATUS "Using termcap: ${TERMCAP_LIBRARY}")
endif()

9
cmake/find/yaml-cpp.cmake Normal file
View File

@ -0,0 +1,9 @@
option(USE_YAML_CPP "Enable yaml-cpp" ${ENABLE_LIBRARIES})
if (NOT USE_YAML_CPP)
return()
endif()
if (NOT EXISTS "${ClickHouse_SOURCE_DIR}/contrib/yaml-cpp")
message (ERROR "submodule contrib/yaml-cpp is missing. to fix try run: \n git submodule update --init --recursive")
endif()

9
contrib/CMakeLists.txt vendored
View File

@ -50,6 +50,10 @@ add_subdirectory (replxx-cmake)
add_subdirectory (unixodbc-cmake)
add_subdirectory (nanodbc-cmake)
if (USE_YAML_CPP)
add_subdirectory (yaml-cpp-cmake)
endif()
if (USE_INTERNAL_XZ_LIBRARY)
add_subdirectory (xz)
endif()
@ -205,11 +209,12 @@ elseif(GTEST_SRC_DIR)
target_compile_definitions(gtest INTERFACE GTEST_HAS_POSIX_RE=0)
endif()
if (USE_EMBEDDED_COMPILER AND USE_INTERNAL_LLVM_LIBRARY)
if (USE_EMBEDDED_COMPILER)
# ld: unknown option: --color-diagnostics
if (APPLE)
set (LINKER_SUPPORTS_COLOR_DIAGNOSTICS 0 CACHE INTERNAL "")
endif ()
set (LLVM_ENABLE_EH 1 CACHE INTERNAL "")
set (LLVM_ENABLE_RTTI 1 CACHE INTERNAL "")
set (LLVM_ENABLE_PIC 0 CACHE INTERNAL "")
@ -224,8 +229,6 @@ if (USE_EMBEDDED_COMPILER AND USE_INTERNAL_LLVM_LIBRARY)
set (CMAKE_CXX_STANDARD ${CMAKE_CXX_STANDARD_bak})
unset (CMAKE_CXX_STANDARD_bak)
target_include_directories(LLVMSupport SYSTEM BEFORE PRIVATE ${ZLIB_INCLUDE_DIR})
endif ()
if (USE_INTERNAL_LIBGSASL_LIBRARY)

2
contrib/NuRaft vendored

@ -1 +1 @@
Subproject commit 377f8e77491d9f66ce8e32e88aae19dffe8dc4d7
Subproject commit 95d6bbba579b3a4e4c2dede954f541ff6f3dba51

2
contrib/avro vendored

@ -1 +1 @@
Subproject commit 92caca2d42fc9a97e34e95f963593539d32ed331
Subproject commit 1ee16d8c5a7808acff5cf0475f771195d9aa3faa

2
contrib/boringssl vendored

@ -1 +1 @@
Subproject commit 83c1cda8a0224dc817cbad2966c7ed4acc35f02a
Subproject commit a6a2e2ab3e44d97ce98e51c558e989f211de7eb3

2
contrib/cppkafka vendored

@ -1 +1 @@
Subproject commit b06e64ef5bffd636d918a742c689f69130c1dbab
Subproject commit 57a599d99c540e647bcd0eb9ea77c523cca011b3

2
contrib/datasketches-cpp vendored

@ -1 +1 @@
Subproject commit f915d35b2de676683493c86c585141a1e1c83334
Subproject commit 7d73d7610db31d4e1ecde0fb3a7ee90ef371207f

2
contrib/grpc vendored

@ -1 +1 @@
Subproject commit 1085a941238e66b13e3fb89c310533745380acbc
Subproject commit 60c986e15cae70aade721d26badabab1f822fdd6

2
contrib/librdkafka vendored

@ -1 +1 @@
Subproject commit cf11d0aa36d4738f2c9bf4377807661660f1be76
Subproject commit 43491d33ca2826531d1e3cae70d4bf1e5249e3c9

2
contrib/libunwind vendored

@ -1 +1 @@
Subproject commit 8fe25d7dc70f2a4ea38c3e5a33fa9d4199b67a5a
Subproject commit a491c27b33109a842d577c0f7ac5f5f218859181

2
contrib/llvm vendored

@ -1 +1 @@
Subproject commit cfaf365cf96918999d09d976ec736b4518cf5d02
Subproject commit e5751459412bce1391fb7a2e9bbc01e131bf72f1

2
contrib/poco vendored

@ -1 +1 @@
Subproject commit b7d9ec16ee33ca76643d5fcd907ea9a33285640a
Subproject commit 5994506908028612869fee627d68d8212dfe7c1e

2
contrib/re2 vendored

@ -1 +1 @@
Subproject commit 7cf8b88e8f70f97fd4926b56aa87e7f53b2717e0
Subproject commit 13ebb377c6ad763ca61d12dd6f88b1126bd0b911

2
contrib/re2_st/re2_transform.cmake
View File

@ -1,7 +1,7 @@
file (READ ${SOURCE_FILENAME} CONTENT)
string (REGEX REPLACE "using re2::RE2;" "" CONTENT "${CONTENT}")
string (REGEX REPLACE "using re2::LazyRE2;" "" CONTENT "${CONTENT}")
string (REGEX REPLACE "namespace re2" "namespace re2_st" CONTENT "${CONTENT}")
string (REGEX REPLACE "namespace re2 {" "namespace re2_st {" CONTENT "${CONTENT}")
string (REGEX REPLACE "re2::" "re2_st::" CONTENT "${CONTENT}")
string (REGEX REPLACE "\"re2/" "\"re2_st/" CONTENT "${CONTENT}")
string (REGEX REPLACE "(.\\*?_H)" "\\1_ST" CONTENT "${CONTENT}")

2
contrib/rocksdb vendored

@ -1 +1 @@
Subproject commit 54a0decabbcf4c0bb5cf7befa9c597f28289bff5
Subproject commit 07c77549a20b63ff6981b400085eba36bb5c80c4

2
contrib/simdjson vendored

@ -1 +1 @@
Subproject commit 95b4870e20be5f97d9dcf63b23b1c6f520c366c1
Subproject commit 8df32cea3359cb30120795da6020b3b73da01d38

1
contrib/yaml-cpp vendored Submodule

@ -0,0 +1 @@
Subproject commit 0c86adac6d117ee2b4afcedb8ade19036ca0327d

39
contrib/yaml-cpp-cmake/CMakeLists.txt Normal file
View File

@ -0,0 +1,39 @@
set (LIBRARY_DIR ${ClickHouse_SOURCE_DIR}/contrib/yaml-cpp)
set (SRCS
${LIBRARY_DIR}/src/binary.cpp
${LIBRARY_DIR}/src/emitterutils.cpp
${LIBRARY_DIR}/src/null.cpp
${LIBRARY_DIR}/src/scantoken.cpp
${LIBRARY_DIR}/src/convert.cpp
${LIBRARY_DIR}/src/exceptions.cpp
${LIBRARY_DIR}/src/ostream_wrapper.cpp
${LIBRARY_DIR}/src/simplekey.cpp
${LIBRARY_DIR}/src/depthguard.cpp
${LIBRARY_DIR}/src/exp.cpp
${LIBRARY_DIR}/src/parse.cpp
${LIBRARY_DIR}/src/singledocparser.cpp
${LIBRARY_DIR}/src/directives.cpp
${LIBRARY_DIR}/src/memory.cpp
${LIBRARY_DIR}/src/parser.cpp
${LIBRARY_DIR}/src/stream.cpp
${LIBRARY_DIR}/src/emit.cpp
${LIBRARY_DIR}/src/nodebuilder.cpp
${LIBRARY_DIR}/src/regex_yaml.cpp
${LIBRARY_DIR}/src/tag.cpp
${LIBRARY_DIR}/src/emitfromevents.cpp
${LIBRARY_DIR}/src/node.cpp
${LIBRARY_DIR}/src/scanner.cpp
${LIBRARY_DIR}/src/emitter.cpp
${LIBRARY_DIR}/src/node_data.cpp
${LIBRARY_DIR}/src/scanscalar.cpp
${LIBRARY_DIR}/src/emitterstate.cpp
${LIBRARY_DIR}/src/nodeevents.cpp
${LIBRARY_DIR}/src/scantag.cpp
)
add_library (yaml-cpp ${SRCS})
target_include_directories(yaml-cpp PRIVATE ${LIBRARY_DIR}/include/yaml-cpp)
target_include_directories(yaml-cpp SYSTEM BEFORE PUBLIC ${LIBRARY_DIR}/include)

2
contrib/zlib-ng vendored

@ -1 +1 @@
Subproject commit 5cc4d232020dc66d1d6c5438834457e2a2f6127b
Subproject commit db232d30b4c72fd58e6d7eae2d12cebf9c3d90db

2
contrib/zstd vendored

@ -1 +1 @@
Subproject commit 10f0e6993f9d2f682da6d04aa2385b7d53cbb4ee
Subproject commit a488ba114ec17ea1054b9057c26a046fc122b3b6

9
contrib/zstd-cmake/CMakeLists.txt
View File

@ -66,6 +66,7 @@ SET(Sources
"${LIBRARY_DIR}/compress/zstd_compress.c"
"${LIBRARY_DIR}/compress/zstd_compress_literals.c"
"${LIBRARY_DIR}/compress/zstd_compress_sequences.c"
"${LIBRARY_DIR}/compress/zstd_compress_superblock.c"
"${LIBRARY_DIR}/compress/zstd_double_fast.c"
"${LIBRARY_DIR}/compress/zstd_fast.c"
"${LIBRARY_DIR}/compress/zstd_lazy.c"
@ -95,16 +96,19 @@ SET(Headers
"${LIBRARY_DIR}/common/pool.h"
"${LIBRARY_DIR}/common/threading.h"
"${LIBRARY_DIR}/common/xxhash.h"
"${LIBRARY_DIR}/common/zstd_errors.h"
"${LIBRARY_DIR}/common/zstd_deps.h"
"${LIBRARY_DIR}/common/zstd_internal.h"
"${LIBRARY_DIR}/common/zstd_trace.h"
"${LIBRARY_DIR}/compress/hist.h"
"${LIBRARY_DIR}/compress/zstd_compress_internal.h"
"${LIBRARY_DIR}/compress/zstd_compress_literals.h"
"${LIBRARY_DIR}/compress/zstd_compress_sequences.h"
"${LIBRARY_DIR}/compress/zstd_compress_superblock.h"
"${LIBRARY_DIR}/compress/zstd_cwksp.h"
"${LIBRARY_DIR}/compress/zstd_double_fast.h"
"${LIBRARY_DIR}/compress/zstd_fast.h"
"${LIBRARY_DIR}/compress/zstd_lazy.h"
"${LIBRARY_DIR}/compress/zstd_ldm_geartab.h"
"${LIBRARY_DIR}/compress/zstd_ldm.h"
"${LIBRARY_DIR}/compress/zstdmt_compress.h"
"${LIBRARY_DIR}/compress/zstd_opt.h"
@ -113,7 +117,8 @@ SET(Headers
"${LIBRARY_DIR}/decompress/zstd_decompress_internal.h"
"${LIBRARY_DIR}/dictBuilder/cover.h"
"${LIBRARY_DIR}/dictBuilder/divsufsort.h"
"${LIBRARY_DIR}/dictBuilder/zdict.h"
"${LIBRARY_DIR}/zdict.h"
"${LIBRARY_DIR}/zstd_errors.h"
"${LIBRARY_DIR}/zstd.h")
SET(ZSTD_LEGACY_SUPPORT true)

4
debian/changelog vendored
View File

@ -1,5 +1,5 @@
clickhouse (21.6.1.1) unstable; urgency=low
clickhouse (21.7.1.1) unstable; urgency=low
* Modified source code
-- clickhouse-release <clickhouse-release@yandex-team.ru> Tue, 20 Apr 2021 01:48:16 +0300
-- clickhouse-release <clickhouse-release@yandex-team.ru> Thu, 20 May 2021 22:23:29 +0300

2
docker/client/Dockerfile
View File

@ -1,7 +1,7 @@
FROM ubuntu:18.04
ARG repository="deb https://repo.clickhouse.tech/deb/stable/ main/"
ARG version=21.6.1.*
ARG version=21.7.1.*
RUN apt-get update \
&& apt-get install --yes --no-install-recommends \

4
docker/packager/packager
View File

@ -154,6 +154,10 @@ def parse_env_variables(build_type, compiler, sanitizer, package_type, image_typ
if clang_tidy:
cmake_flags.append('-DENABLE_CLANG_TIDY=1')
cmake_flags.append('-DENABLE_UTILS=1')
cmake_flags.append('-DUSE_GTEST=1')
cmake_flags.append('-DENABLE_TESTS=1')
cmake_flags.append('-DENABLE_EXAMPLES=1')
# Don't stop on first error to find more clang-tidy errors in one run.
result.append('NINJA_FLAGS=-k0')

2
docker/server/Dockerfile
View File

@ -1,7 +1,7 @@
FROM ubuntu:20.04
ARG repository="deb https://repo.clickhouse.tech/deb/stable/ main/"
ARG version=21.6.1.*
ARG version=21.7.1.*
ARG gosu_ver=1.10
# set non-empty deb_location_url url to create a docker image

2
docker/test/Dockerfile
View File

@ -1,7 +1,7 @@
FROM ubuntu:18.04
ARG repository="deb https://repo.clickhouse.tech/deb/stable/ main/"
ARG version=21.6.1.*
ARG version=21.7.1.*
RUN apt-get update && \
apt-get install -y apt-transport-https dirmngr && \

42
docker/test/fasttest/run.sh
View File

@ -73,7 +73,7 @@ function start_server
--path "$FASTTEST_DATA"
--user_files_path "$FASTTEST_DATA/user_files"
--top_level_domains_path "$FASTTEST_DATA/top_level_domains"
--keeper_server.log_storage_path "$FASTTEST_DATA/coordination"
--keeper_server.storage_path "$FASTTEST_DATA/coordination"
)
clickhouse-server "${opts[@]}" &>> "$FASTTEST_OUTPUT/server.log" &
server_pid=$!
@ -308,6 +308,8 @@ function run_tests
01354_order_by_tuple_collate_const
01355_ilike
01411_bayesian_ab_testing
01798_uniq_theta_sketch
01799_long_uniq_theta_sketch
collate
collation
_orc_
@ -370,35 +372,19 @@ function run_tests
# Depends on AWS
01801_s3_cluster
# Depends on LLVM JIT
01072_nullable_jit
01852_jit_if
01865_jit_comparison_constant_result
01871_merge_tree_compile_expressions
)
(time clickhouse-test --hung-check -j 8 --order=random --use-skip-list --no-long --testname --shard --zookeeper --skip "${TESTS_TO_SKIP[@]}" -- "$FASTTEST_FOCUS" 2>&1 ||:) | ts '%Y-%m-%d %H:%M:%S' | tee "$FASTTEST_OUTPUT/test_log.txt"
# substr is to remove semicolon after test name
readarray -t FAILED_TESTS < <(awk '/\[ FAIL|TIMEOUT|ERROR \]/ { print substr($3, 1, length($3)-1) }' "$FASTTEST_OUTPUT/test_log.txt" | tee "$FASTTEST_OUTPUT/failed-parallel-tests.txt")
# We will rerun sequentially any tests that have failed during parallel run.
# They might have failed because there was some interference from other tests
# running concurrently. If they fail even in seqential mode, we will report them.
# FIXME All tests that require exclusive access to the server must be
# explicitly marked as `sequential`, and `clickhouse-test` must detect them and
# run them in a separate group after all other tests. This is faster and also
# explicit instead of guessing.
if [[ -n "${FAILED_TESTS[*]}" ]]
then
stop_server ||:
# Clean the data so that there is no interference from the previous test run.
rm -rf "$FASTTEST_DATA"/{{meta,}data,user_files,coordination} ||:
start_server
echo "Going to run again: ${FAILED_TESTS[*]}"
clickhouse-test --hung-check --order=random --no-long --testname --shard --zookeeper "${FAILED_TESTS[@]}" 2>&1 | ts '%Y-%m-%d %H:%M:%S' | tee -a "$FASTTEST_OUTPUT/test_log.txt"
else
echo "No failed tests"
fi
time clickhouse-test --hung-check -j 8 --order=random --use-skip-list \
--no-long --testname --shard --zookeeper --skip "${TESTS_TO_SKIP[@]}" \
-- "$FASTTEST_FOCUS" 2>&1 \
| ts '%Y-%m-%d %H:%M:%S' \
| tee "$FASTTEST_OUTPUT/test_log.txt"
}
case "$stage" in

5
docker/test/fuzzer/query-fuzzer-tweaks-users.xml
View File

@ -14,11 +14,6 @@
<max_memory_usage>
<max>10G</max>
</max_memory_usage>
<!-- Not ready for production -->
<compile_expressions>
<readonly />
</compile_expressions>
</constraints>
</default>
</profiles>

42
docker/test/fuzzer/run-fuzzer.sh
View File

@ -126,6 +126,20 @@ continue
echo "Fuzzer exit code is $fuzzer_exit_code"
# If the server dies, most often the fuzzer returns code 210: connetion
# refused, and sometimes also code 32: attempt to read after eof. For
# simplicity, check again whether the server is accepting connections, using
# clickhouse-client. We don't check for existence of server process, because
# the process is still present while the server is terminating and not
# accepting the connections anymore.
if clickhouse-client --query "select 1 format Null"
then
server_died=0
else
echo "Server live check returns $?"
server_died=1
fi
clickhouse-client --query "select elapsed, query from system.processes" ||:
killall clickhouse-server ||:
for _ in {1..10}
@ -187,25 +201,29 @@ case "$stage" in
# Make files with status and description we'll show for this check on Github
task_exit_code=$fuzzer_exit_code
if [ "$fuzzer_exit_code" == 143 ]
if [ "$server_died" == 1 ]
then
# SIGTERM -- the fuzzer was killed by timeout, which means a normal run.
echo "success" > status.txt
echo "OK" > description.txt
task_exit_code=0
elif [ "$fuzzer_exit_code" == 210 ]
then
# Lost connection to the server. This probably means that the server died
# with abort.
# The server has died.
task_exit_code=210
echo "failure" > status.txt
if ! grep -ao "Received signal.*\|Logical error.*\|Assertion.*failed\|Failed assertion.*\|.*runtime error: .*\|.*is located.*\|SUMMARY: AddressSanitizer:.*\|SUMMARY: MemorySanitizer:.*\|SUMMARY: ThreadSanitizer:.*\|.*_LIBCPP_ASSERT.*" server.log > description.txt
then
echo "Lost connection to server. See the logs." > description.txt
fi
elif [ "$fuzzer_exit_code" == "143" ] || [ "$fuzzer_exit_code" == "0" ]
then
# Variants of a normal run:
# 0 -- fuzzing ended earlier than timeout.
# 143 -- SIGTERM -- the fuzzer was killed by timeout.
task_exit_code=0
echo "success" > status.txt
echo "OK" > description.txt
else
# Something different -- maybe the fuzzer itself died? Don't grep the
# server log in this case, because we will find a message about normal
# server termination (Received signal 15), which is confusing.
# The server was alive, but the fuzzer returned some error. Probably this
# is a problem in the fuzzer itself. Don't grep the server log in this
# case, because we will find a message about normal server termination
# (Received signal 15), which is confusing.
task_exit_code=$fuzzer_exit_code
echo "failure" > status.txt
echo "Fuzzer failed ($fuzzer_exit_code). See the logs." > description.txt
fi

3
docker/test/integration/runner/Dockerfile
View File

@ -80,7 +80,8 @@ RUN python3 -m pip install \
redis \
tzlocal \
urllib3 \
requests-kerberos
requests-kerberos \
pyhdfs
COPY modprobe.sh /usr/local/bin/modprobe
COPY dockerd-entrypoint.sh /usr/local/bin/

92
docker/test/integration/runner/compose/docker_compose_keeper.yml Normal file
View File

@ -0,0 +1,92 @@
version: '2.3'
services:
zoo1:
image: ${image:-yandex/clickhouse-integration-test}
restart: always
user: ${user:-}
volumes:
- type: bind
source: ${keeper_binary:-}
target: /usr/bin/clickhouse
- type: bind
source: ${keeper_config_dir1:-}
target: /etc/clickhouse-keeper
- type: bind
source: ${keeper_logs_dir1:-}
target: /var/log/clickhouse-keeper
- type: ${keeper_fs:-tmpfs}
source: ${keeper_db_dir1:-}
target: /var/lib/clickhouse-keeper
entrypoint: "clickhouse keeper --config=/etc/clickhouse-keeper/keeper_config1.xml --log-file=/var/log/clickhouse-keeper/clickhouse-keeper.log --errorlog-file=/var/log/clickhouse-keeper/clickhouse-keeper.err.log"
cap_add:
- SYS_PTRACE
- NET_ADMIN
- IPC_LOCK
- SYS_NICE
security_opt:
- label:disable
dns_opt:
- attempts:2
- timeout:1
- inet6
- rotate
zoo2:
image: ${image:-yandex/clickhouse-integration-test}
restart: always
user: ${user:-}
volumes:
- type: bind
source: ${keeper_binary:-}
target: /usr/bin/clickhouse
- type: bind
source: ${keeper_config_dir2:-}
target: /etc/clickhouse-keeper
- type: bind
source: ${keeper_logs_dir2:-}
target: /var/log/clickhouse-keeper
- type: ${keeper_fs:-tmpfs}
source: ${keeper_db_dir2:-}
target: /var/lib/clickhouse-keeper
entrypoint: "clickhouse keeper --config=/etc/clickhouse-keeper/keeper_config2.xml --log-file=/var/log/clickhouse-keeper/clickhouse-keeper.log --errorlog-file=/var/log/clickhouse-keeper/clickhouse-keeper.err.log"
cap_add:
- SYS_PTRACE
- NET_ADMIN
- IPC_LOCK
- SYS_NICE
security_opt:
- label:disable
dns_opt:
- attempts:2
- timeout:1
- inet6
- rotate
zoo3:
image: ${image:-yandex/clickhouse-integration-test}
restart: always
user: ${user:-}
volumes:
- type: bind
source: ${keeper_binary:-}
target: /usr/bin/clickhouse
- type: bind
source: ${keeper_config_dir3:-}
target: /etc/clickhouse-keeper
- type: bind
source: ${keeper_logs_dir3:-}
target: /var/log/clickhouse-keeper
- type: ${keeper_fs:-tmpfs}
source: ${keeper_db_dir3:-}
target: /var/lib/clickhouse-keeper
entrypoint: "clickhouse keeper --config=/etc/clickhouse-keeper/keeper_config3.xml --log-file=/var/log/clickhouse-keeper/clickhouse-keeper.log --errorlog-file=/var/log/clickhouse-keeper/clickhouse-keeper.err.log"
cap_add:
- SYS_PTRACE
- NET_ADMIN
- IPC_LOCK
- SYS_NICE
security_opt:
- label:disable
dns_opt:
- attempts:2
- timeout:1
- inet6
- rotate

4
docker/test/integration/runner/dockerd-entrypoint.sh
View File

@ -27,6 +27,10 @@ while true; do
done
set -e
# cleanup for retry run if volume is not recreated
docker kill "$(docker ps -aq)" || true
docker rm "$(docker ps -aq)" || true
echo "Start tests"
export CLICKHOUSE_TESTS_SERVER_BIN_PATH=/clickhouse
export CLICKHOUSE_TESTS_CLIENT_BIN_PATH=/clickhouse

96
docker/test/performance-comparison/compare.sh
View File

@ -552,6 +552,63 @@ create table query_metric_stats_denorm engine File(TSVWithNamesAndTypes,
order by test, query_index, metric_name
;
" 2> >(tee -a analyze/errors.log 1>&2)
# Fetch historical query variability thresholds from the CI database
clickhouse-local --query "
left join file('analyze/report-thresholds.tsv', TSV,
'test text, report_threshold float') thresholds
on query_metric_stats.test = thresholds.test
"
if [ -v CHPC_DATABASE_URL ]
then
set +x # Don't show password in the log
client=(clickhouse-client
# Surprisingly, clickhouse-client doesn't understand --host 127.0.0.1:9000
# so I have to extract host and port with clickhouse-local. I tried to use
# Poco URI parser to support this in the client, but it's broken and can't
# parse host:port.
$(clickhouse-local --query "with '${CHPC_DATABASE_URL}' as url select '--host ' || domain(url) || ' --port ' || toString(port(url)) format TSV")
--secure
--user "${CHPC_DATABASE_USER}"
--password "${CHPC_DATABASE_PASSWORD}"
--config "right/config/client_config.xml"
--database perftest
--date_time_input_format=best_effort)
# Precision is going to be 1.5 times worse for PRs. How do I know it? I ran this:
# SELECT quantilesExact(0., 0.1, 0.5, 0.75, 0.95, 1.)(p / m)
# FROM
# (
# SELECT
# quantileIf(0.95)(stat_threshold, pr_number = 0) AS m,
# quantileIf(0.95)(stat_threshold, (pr_number != 0) AND (abs(diff) < stat_threshold)) AS p
# FROM query_metrics_v2
# WHERE (event_date > (today() - toIntervalMonth(1))) AND (metric = 'client_time')
# GROUP BY
# test,
# query_index,
# query_display_name
# HAVING count(*) > 100
# )
# The file can be empty if the server is inaccessible, so we can't use TSVWithNamesAndTypes.
"${client[@]}" --query "
select test, query_index,
quantileExact(0.99)(abs(diff)) max_diff,
quantileExactIf(0.99)(stat_threshold, abs(diff) < stat_threshold) * 1.5 max_stat_threshold,
query_display_name
from query_metrics_v2
where event_date > now() - interval 1 month
and metric = 'client_time'
and pr_number = 0
group by test, query_index, query_display_name
having count(*) > 100
" > analyze/historical-thresholds.tsv
else
touch analyze/historical-thresholds.tsv
fi
}
# Analyze results
@ -596,6 +653,26 @@ create view query_metric_stats as
diff float, stat_threshold float')
;
create table report_thresholds engine File(TSVWithNamesAndTypes, 'report/thresholds.tsv')
as select
query_display_names.test test, query_display_names.query_index query_index,
ceil(greatest(0.1, historical_thresholds.max_diff,
test_thresholds.report_threshold), 2) changed_threshold,
ceil(greatest(0.2, historical_thresholds.max_stat_threshold,
test_thresholds.report_threshold + 0.1), 2) unstable_threshold,
query_display_names.query_display_name query_display_name
from query_display_names
left join file('analyze/historical-thresholds.tsv', TSV,
'test text, query_index int, max_diff float, max_stat_threshold float,
query_display_name text') historical_thresholds
on query_display_names.test = historical_thresholds.test
and query_display_names.query_index = historical_thresholds.query_index
and query_display_names.query_display_name = historical_thresholds.query_display_name
left join file('analyze/report-thresholds.tsv', TSV,
'test text, report_threshold float') test_thresholds
on query_display_names.test = test_thresholds.test
;
-- Main statistics for queries -- query time as reported in query log.
create table queries engine File(TSVWithNamesAndTypes, 'report/queries.tsv')
as select
@ -610,23 +687,23 @@ create table queries engine File(TSVWithNamesAndTypes, 'report/queries.tsv')
-- uncaught regressions, because for the default 7 runs we do for PRs,
-- the randomization distribution has only 16 values, so the max quantile
-- is actually 0.9375.
abs(diff) > report_threshold and abs(diff) >= stat_threshold as changed_fail,
abs(diff) > report_threshold - 0.05 and abs(diff) >= stat_threshold as changed_show,
abs(diff) > changed_threshold and abs(diff) >= stat_threshold as changed_fail,
abs(diff) > changed_threshold - 0.05 and abs(diff) >= stat_threshold as changed_show,
not changed_fail and stat_threshold > report_threshold + 0.10 as unstable_fail,
not changed_show and stat_threshold > report_threshold - 0.05 as unstable_show,
not changed_fail and stat_threshold > unstable_threshold as unstable_fail,
not changed_show and stat_threshold > unstable_threshold - 0.05 as unstable_show,
left, right, diff, stat_threshold,
if(report_threshold > 0, report_threshold, 0.10) as report_threshold,
query_metric_stats.test test, query_metric_stats.query_index query_index,
query_display_name
query_display_names.query_display_name query_display_name
from query_metric_stats
left join file('analyze/report-thresholds.tsv', TSV,
'test text, report_threshold float') thresholds
on query_metric_stats.test = thresholds.test
left join query_display_names
on query_metric_stats.test = query_display_names.test
and query_metric_stats.query_index = query_display_names.query_index
left join report_thresholds
on query_display_names.test = report_thresholds.test
and query_display_names.query_index = report_thresholds.query_index
and query_display_names.query_display_name = report_thresholds.query_display_name
-- 'server_time' is rounded down to ms, which might be bad for very short queries.
-- Use 'client_time' instead.
where metric_name = 'client_time'
@ -889,7 +966,6 @@ create table all_query_metrics_tsv engine File(TSV, 'report/all-query-metrics.ts
order by test, query_index;
" 2> >(tee -a report/errors.log 1>&2)
# Prepare source data for metrics and flamegraphs for queries that were profiled
# by perf.py.
for version in {right,left}

3
docker/test/performance-comparison/config/users.d/perf-comparison-tweaks-users.xml
View File

@ -20,6 +20,9 @@
<!-- mmap shows some improvements in perf tests -->
<min_bytes_to_use_mmap_io>64Mi</min_bytes_to_use_mmap_io>
<!-- disable jit for perf tests -->
<compile_expressions>0</compile_expressions>
</default>
</profiles>
<users>

19
docker/test/performance-comparison/perf.py
View File

@ -44,7 +44,7 @@ parser.add_argument('--port', nargs='*', default=[9000], help="Space-separated l
parser.add_argument('--runs', type=int, default=1, help='Number of query runs per server.')
parser.add_argument('--max-queries', type=int, default=None, help='Test no more than this number of queries, chosen at random.')
parser.add_argument('--queries-to-run', nargs='*', type=int, default=None, help='Space-separated list of indexes of queries to test.')
parser.add_argument('--max-query-seconds', type=int, default=10, help='For how many seconds at most a query is allowed to run. The script finishes with error if this time is exceeded.')
parser.add_argument('--max-query-seconds', type=int, default=15, help='For how many seconds at most a query is allowed to run. The script finishes with error if this time is exceeded.')
parser.add_argument('--profile-seconds', type=int, default=0, help='For how many seconds to profile a query for which the performance has changed.')
parser.add_argument('--long', action='store_true', help='Do not skip the tests tagged as long.')
parser.add_argument('--print-queries', action='store_true', help='Print test queries and exit.')
@ -273,8 +273,14 @@ for query_index in queries_to_run:
prewarm_id = f'{query_prefix}.prewarm0'
try:
# Will also detect too long queries during warmup stage
res = c.execute(q, query_id = prewarm_id, settings = {'max_execution_time': args.max_query_seconds})
# During the warmup runs, we will also:
# * detect queries that are exceedingly long, to fail fast,
# * collect profiler traces, which might be helpful for analyzing
# test coverage. We disable profiler for normal runs because
# it makes the results unstable.
res = c.execute(q, query_id = prewarm_id,
settings = {'max_execution_time': args.max_query_seconds,
'query_profiler_real_time_period_ns': 10000000})
except clickhouse_driver.errors.Error as e:
# Add query id to the exception to make debugging easier.
e.args = (prewarm_id, *e.args)
@ -359,10 +365,11 @@ for query_index in queries_to_run:
# For very short queries we have a special mode where we run them for at
# least some time. The recommended lower bound of run time for "normal"
# queries is about 0.1 s, and we run them about 10 times, giving the
# time per query per server of about one second. Use this value as a
# reference for "short" queries.
# time per query per server of about one second. Run "short" queries
# for longer time, because they have a high percentage of overhead and
# might give less stable results.
if is_short[query_index]:
if server_seconds >= 2 * len(this_query_connections):
if server_seconds >= 8 * len(this_query_connections):
break
# Also limit the number of runs, so that we don't go crazy processing
# the results -- 'eqmed.sql' is really suboptimal.

23
docker/test/performance-comparison/report.py
View File

@ -446,11 +446,17 @@ if args.report == 'main':
attrs[3] = f'style="background: {color_bad}"'
else:
attrs[3] = ''
# Just don't add the slightly unstable queries we don't consider
# errors. It's not clear what the user should do with them.
continue
text += tableRow(r, attrs, anchor)
text += tableEnd()
tables.append(text)
# Don't add an empty table.
if very_unstable_queries:
tables.append(text)
add_unstable_queries()
@ -549,16 +555,15 @@ if args.report == 'main':
message_array.append(str(slower_queries) + ' slower')
if unstable_partial_queries:
unstable_queries += unstable_partial_queries
error_tests += unstable_partial_queries
very_unstable_queries += unstable_partial_queries
status = 'failure'
if unstable_queries:
message_array.append(str(unstable_queries) + ' unstable')
# Disabled before fix.
# if very_unstable_queries:
# status = 'failure'
# Don't show mildly unstable queries, only the very unstable ones we
# treat as errors.
if very_unstable_queries:
error_tests += very_unstable_queries
status = 'failure'
message_array.append(str(very_unstable_queries) + ' unstable')
error_tests += slow_average_tests
if error_tests:

7
docker/test/testflows/runner/Dockerfile
View File

@ -35,10 +35,10 @@ RUN apt-get update \
ENV TZ=Europe/Moscow
RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
RUN pip3 install urllib3 testflows==1.6.74 docker-compose docker dicttoxml kazoo tzlocal
RUN pip3 install urllib3 testflows==1.6.90 docker-compose==1.29.1 docker==5.0.0 dicttoxml kazoo tzlocal python-dateutil numpy
ENV DOCKER_CHANNEL stable
ENV DOCKER_VERSION 17.09.1-ce
ENV DOCKER_VERSION 20.10.6
RUN set -eux; \
\
@ -73,5 +73,4 @@ RUN set -x \
VOLUME /var/lib/docker
EXPOSE 2375
ENTRYPOINT ["dockerd-entrypoint.sh"]
CMD ["sh", "-c", "python3 regression.py --no-color -o classic --local --clickhouse-binary-path ${CLICKHOUSE_TESTS_SERVER_BIN_PATH} --log test.log ${TESTFLOWS_OPTS}; cat test.log | tfs report results --format json > results.json; /usr/local/bin/process_testflows_result.py || echo -e 'failure\tCannot parse results' > check_status.tsv"]
CMD ["sh", "-c", "python3 regression.py --no-color -o new-fails --local --clickhouse-binary-path ${CLICKHOUSE_TESTS_SERVER_BIN_PATH} --log test.log ${TESTFLOWS_OPTS}; cat test.log | tfs report results --format json > results.json; /usr/local/bin/process_testflows_result.py || echo -e 'failure\tCannot parse results' > check_status.tsv; find * -type f | grep _instances | grep clickhouse-server | xargs -n1 tar -rvf clickhouse_logs.tar; gzip -9 clickhouse_logs.tar"]

12
docs/en/commercial/cloud.md
View File

@ -31,16 +31,24 @@ toc_title: Cloud
## Alibaba Cloud {#alibaba-cloud}
Alibaba Cloud Managed Service for ClickHouse. [China Site](https://www.aliyun.com/product/clickhouse) (will be available at the international site in May 2021). Provides the following key features:
[Alibaba Cloud Managed Service for ClickHouse](https://www.alibabacloud.com/product/clickhouse) provides the following key features:
- Highly reliable cloud disk storage engine based on [Alibaba Cloud Apsara](https://www.alibabacloud.com/product/apsara-stack) distributed system
- Expand capacity on-demand without manual data migration
- Expand capacity on demand without manual data migration
- Support single-node, single-replica, multi-node, and multi-replica architectures, and support hot and cold data tiering
- Support access allow-list, one-key recovery, multi-layer network security protection, cloud disk encryption
- Seamless integration with cloud log systems, databases, and data application tools
- Built-in monitoring and database management platform
- Professional database expert technical support and service
## SberCloud {#sbercloud}
[SberCloud.Advanced](https://sbercloud.ru/en/advanced) provides [MapReduce Service (MRS)](https://docs.sbercloud.ru/mrs/ug/topics/ug__clickhouse.html), a reliable, secure, and easy-to-use enterprise-level platform for storing, processing, and analyzing big data. MRS allows you to quickly create and manage ClickHouse clusters.
- A ClickHouse instance consists of three ZooKeeper nodes and multiple ClickHouse nodes. The Dedicated Replica mode is used to ensure high reliability of dual data copies.
- MRS provides smooth and elastic scaling capabilities to quickly meet service growth requirements in scenarios where the cluster storage capacity or CPU computing resources are not enough. When you expand the capacity of ClickHouse nodes in a cluster, MRS provides a one-click data balancing tool and gives you the initiative to balance data. You can determine the data balancing mode and time based on service characteristics to ensure service availability, implementing smooth scaling.
- MRS uses the Elastic Load Balance ensuring high availability deployment architecture to automatically distribute user access traffic to multiple backend nodes, expanding service capabilities to external systems and improving fault tolerance. With the ELB polling mechanism, data is written to local tables and read from distributed tables on different nodes. In this way, data read/write load and high availability of application access are guaranteed.
## Tencent Cloud {#tencent-cloud}
[Tencent Managed Service for ClickHouse](https://cloud.tencent.com/product/cdwch) provides the following key features:

2
docs/en/commercial/index.md
View File

@ -14,4 +14,4 @@ Service categories:
- [Support](../commercial/support.md)
!!! note "For service providers"
If you happen to represent one of them, feel free to open a pull request adding your company to the respective section (or even adding a new section if the service doesn’t fit into existing categories). The easiest way to open a pull-request for documentation page is by using a “pencil” edit button in the top-right corner. If your service available in some local market, make sure to mention it in a localized documentation page as well (or at least point it out in a pull-request description).
If you happen to represent one of them, feel free to open a pull request adding your company to the respective section (or even adding a new section if the service does not fit into existing categories). The easiest way to open a pull-request for documentation page is by using a “pencil” edit button in the top-right corner. If your service available in some local market, make sure to mention it in a localized documentation page as well (or at least point it out in a pull-request description).

10
docs/en/development/adding_test_queries.md
View File

@ -1,6 +1,6 @@
# How to add test queries to ClickHouse CI
ClickHouse has hundreds (or even thousands) of features. Every commit get checked by a complex set of tests containing many thousands of test cases.
ClickHouse has hundreds (or even thousands) of features. Every commit gets checked by a complex set of tests containing many thousands of test cases.
The core functionality is very well tested, but some corner-cases and different combinations of features can be uncovered with ClickHouse CI.
@ -105,13 +105,13 @@ clickhouse-client -nmT < tests/queries/0_stateless/01521_dummy_test.sql | tee te
5) ensure everything is correct, if the test output is incorrect (due to some bug for example), adjust the reference file using text editor.
#### How create good test
#### How to create good test
- test should be
- minimal - create only tables related to tested functionality, remove unrelated columns and parts of query
- fast - should not take longer than few seconds (better subseconds)
- correct - fails then feature is not working
- deteministic
- deterministic
- isolated / stateless
- don't rely on some environment things
- don't rely on timing when possible
@ -120,11 +120,11 @@ clickhouse-client -nmT < tests/queries/0_stateless/01521_dummy_test.sql | tee te
- don't switch databases (unless necessary)
- you can create several table replicas on the same node if needed
- you can use one of the test cluster definitions when needed (see system.clusters)
- use `number` / `numbers_mt` / `zeros` / `zeros_mt` and similar for queries / to initialize data when appliable
- use `number` / `numbers_mt` / `zeros` / `zeros_mt` and similar for queries / to initialize data when applicable
- clean up the created objects after test and before the test (DROP IF EXISTS) - in case of some dirty state
- prefer sync mode of operations (mutations, merges, etc.)
- use other SQL files in the `0_stateless` folder as an example
- ensure the feature / feature combination you want to tests is not covered yet with existsing tests
- ensure the feature / feature combination you want to test is not yet covered with existing tests
#### Commit / push / create PR.

20
docs/en/development/architecture.md
View File

@ -21,11 +21,11 @@ Various `IColumn` implementations (`ColumnUInt8`, `ColumnString`, and so on) are
Nevertheless, it is possible to work with individual values as well. To represent an individual value, the `Field` is used. `Field` is just a discriminated union of `UInt64`, `Int64`, `Float64`, `String` and `Array`. `IColumn` has the `operator []` method to get the n-th value as a `Field`, and the `insert` method to append a `Field` to the end of a column. These methods are not very efficient, because they require dealing with temporary `Field` objects representing an individual value. There are more efficient methods, such as `insertFrom`, `insertRangeFrom`, and so on.
`Field` doesn’t have enough information about a specific data type for a table. For example, `UInt8`, `UInt16`, `UInt32`, and `UInt64` are all represented as `UInt64` in a `Field`.
`Field` does not have enough information about a specific data type for a table. For example, `UInt8`, `UInt16`, `UInt32`, and `UInt64` are all represented as `UInt64` in a `Field`.
## Leaky Abstractions {#leaky-abstractions}
`IColumn` has methods for common relational transformations of data, but they don’t meet all needs. For example, `ColumnUInt64` doesn’t have a method to calculate the sum of two columns, and `ColumnString` doesn’t have a method to run a substring search. These countless routines are implemented outside of `IColumn`.
`IColumn` has methods for common relational transformations of data, but they do not meet all needs. For example, `ColumnUInt64` does not have a method to calculate the sum of two columns, and `ColumnString` does not have a method to run a substring search. These countless routines are implemented outside of `IColumn`.
Various functions on columns can be implemented in a generic, non-efficient way using `IColumn` methods to extract `Field` values, or in a specialized way using knowledge of inner memory layout of data in a specific `IColumn` implementation. It is implemented by casting functions to a specific `IColumn` type and deal with internal representation directly. For example, `ColumnUInt64` has the `getData` method that returns a reference to an internal array, then a separate routine reads or fills that array directly. We have “leaky abstractions” to allow efficient specializations of various routines.
@ -35,7 +35,7 @@ Various functions on columns can be implemented in a generic, non-efficient way
`IDataType` and `IColumn` are only loosely related to each other. Different data types can be represented in memory by the same `IColumn` implementations. For example, `DataTypeUInt32` and `DataTypeDateTime` are both represented by `ColumnUInt32` or `ColumnConstUInt32`. In addition, the same data type can be represented by different `IColumn` implementations. For example, `DataTypeUInt8` can be represented by `ColumnUInt8` or `ColumnConstUInt8`.
`IDataType` only stores metadata. For instance, `DataTypeUInt8` doesn’t store anything at all (except virtual pointer `vptr`) and `DataTypeFixedString` stores just `N` (the size of fixed-size strings).
`IDataType` only stores metadata. For instance, `DataTypeUInt8` does not store anything at all (except virtual pointer `vptr`) and `DataTypeFixedString` stores just `N` (the size of fixed-size strings).
`IDataType` has helper methods for various data formats. Examples are methods to serialize a value with possible quoting, to serialize a value for JSON, and to serialize a value as part of the XML format. There is no direct correspondence to data formats. For example, the different data formats `Pretty` and `TabSeparated` can use the same `serializeTextEscaped` helper method from the `IDataType` interface.
@ -43,7 +43,7 @@ Various functions on columns can be implemented in a generic, non-efficient way
A `Block` is a container that represents a subset (chunk) of a table in memory. It is just a set of triples: `(IColumn, IDataType, column name)`. During query execution, data is processed by `Block`s. If we have a `Block`, we have data (in the `IColumn` object), we have information about its type (in `IDataType`) that tells us how to deal with that column, and we have the column name. It could be either the original column name from the table or some artificial name assigned for getting temporary results of calculations.
When we calculate some function over columns in a block, we add another column with its result to the block, and we don’t touch columns for arguments of the function because operations are immutable. Later, unneeded columns can be removed from the block, but not modified. It is convenient for the elimination of common subexpressions.
When we calculate some function over columns in a block, we add another column with its result to the block, and we do not touch columns for arguments of the function because operations are immutable. Later, unneeded columns can be removed from the block, but not modified. It is convenient for the elimination of common subexpressions.
Blocks are created for every processed chunk of data. Note that for the same type of calculation, the column names and types remain the same for different blocks, and only column data changes. It is better to split block data from the block header because small block sizes have a high overhead of temporary strings for copying shared_ptrs and column names.
@ -118,11 +118,11 @@ Interpreters are responsible for creating the query execution pipeline from an `
There are ordinary functions and aggregate functions. For aggregate functions, see the next section.
Ordinary functions don’t change the number of rows – they work as if they are processing each row independently. In fact, functions are not called for individual rows, but for `Block`’s of data to implement vectorized query execution.
Ordinary functions do not change the number of rows – they work as if they are processing each row independently. In fact, functions are not called for individual rows, but for `Block`’s of data to implement vectorized query execution.
There are some miscellaneous functions, like [blockSize](../sql-reference/functions/other-functions.md#function-blocksize), [rowNumberInBlock](../sql-reference/functions/other-functions.md#function-rownumberinblock), and [runningAccumulate](../sql-reference/functions/other-functions.md#runningaccumulate), that exploit block processing and violate the independence of rows.
ClickHouse has strong typing, so there’s no implicit type conversion. If a function doesn’t support a specific combination of types, it throws an exception. But functions can work (be overloaded) for many different combinations of types. For example, the `plus` function (to implement the `+` operator) works for any combination of numeric types: `UInt8` + `Float32`, `UInt16` + `Int8`, and so on. Also, some variadic functions can accept any number of arguments, such as the `concat` function.
ClickHouse has strong typing, so there’s no implicit type conversion. If a function does not support a specific combination of types, it throws an exception. But functions can work (be overloaded) for many different combinations of types. For example, the `plus` function (to implement the `+` operator) works for any combination of numeric types: `UInt8` + `Float32`, `UInt16` + `Int8`, and so on. Also, some variadic functions can accept any number of arguments, such as the `concat` function.
Implementing a function may be slightly inconvenient because a function explicitly dispatches supported data types and supported `IColumns`. For example, the `plus` function has code generated by instantiation of a C++ template for each combination of numeric types, and constant or non-constant left and right arguments.
@ -152,7 +152,7 @@ Internally, it is just a primitive multithreaded server without coroutines or fi
The server initializes the `Context` class with the necessary environment for query execution: the list of available databases, users and access rights, settings, clusters, the process list, the query log, and so on. Interpreters use this environment.
We maintain full backward and forward compatibility for the server TCP protocol: old clients can talk to new servers, and new clients can talk to old servers. But we don’t want to maintain it eternally, and we are removing support for old versions after about one year.
We maintain full backward and forward compatibility for the server TCP protocol: old clients can talk to new servers, and new clients can talk to old servers. But we do not want to maintain it eternally, and we are removing support for old versions after about one year.
!!! note "Note"
For most external applications, we recommend using the HTTP interface because it is simple and easy to use. The TCP protocol is more tightly linked to internal data structures: it uses an internal format for passing blocks of data, and it uses custom framing for compressed data. We haven’t released a C library for that protocol because it requires linking most of the ClickHouse codebase, which is not practical.
@ -169,13 +169,13 @@ There is no global query plan for distributed query execution. Each node has its
`MergeTree` is a family of storage engines that supports indexing by primary key. The primary key can be an arbitrary tuple of columns or expressions. Data in a `MergeTree` table is stored in “parts”. Each part stores data in the primary key order, so data is ordered lexicographically by the primary key tuple. All the table columns are stored in separate `column.bin` files in these parts. The files consist of compressed blocks. Each block is usually from 64 KB to 1 MB of uncompressed data, depending on the average value size. The blocks consist of column values placed contiguously one after the other. Column values are in the same order for each column (the primary key defines the order), so when you iterate by many columns, you get values for the corresponding rows.
The primary key itself is “sparse”. It doesn’t address every single row, but only some ranges of data. A separate `primary.idx` file has the value of the primary key for each N-th row, where N is called `index_granularity` (usually, N = 8192). Also, for each column, we have `column.mrk` files with “marks,” which are offsets to each N-th row in the data file. Each mark is a pair: the offset in the file to the beginning of the compressed block, and the offset in the decompressed block to the beginning of data. Usually, compressed blocks are aligned by marks, and the offset in the decompressed block is zero. Data for `primary.idx` always resides in memory, and data for `column.mrk` files is cached.
The primary key itself is “sparse”. It does not address every single row, but only some ranges of data. A separate `primary.idx` file has the value of the primary key for each N-th row, where N is called `index_granularity` (usually, N = 8192). Also, for each column, we have `column.mrk` files with “marks,” which are offsets to each N-th row in the data file. Each mark is a pair: the offset in the file to the beginning of the compressed block, and the offset in the decompressed block to the beginning of data. Usually, compressed blocks are aligned by marks, and the offset in the decompressed block is zero. Data for `primary.idx` always resides in memory, and data for `column.mrk` files is cached.
When we are going to read something from a part in `MergeTree`, we look at `primary.idx` data and locate ranges that could contain requested data, then look at `column.mrk` data and calculate offsets for where to start reading those ranges. Because of sparseness, excess data may be read. ClickHouse is not suitable for a high load of simple point queries, because the entire range with `index_granularity` rows must be read for each key, and the entire compressed block must be decompressed for each column. We made the index sparse because we must be able to maintain trillions of rows per single server without noticeable memory consumption for the index. Also, because the primary key is sparse, it is not unique: it cannot check the existence of the key in the table at INSERT time. You could have many rows with the same key in a table.
When you `INSERT` a bunch of data into `MergeTree`, that bunch is sorted by primary key order and forms a new part. There are background threads that periodically select some parts and merge them into a single sorted part to keep the number of parts relatively low. That’s why it is called `MergeTree`. Of course, merging leads to “write amplification”. All parts are immutable: they are only created and deleted, but not modified. When SELECT is executed, it holds a snapshot of the table (a set of parts). After merging, we also keep old parts for some time to make a recovery after failure easier, so if we see that some merged part is probably broken, we can replace it with its source parts.
`MergeTree` is not an LSM tree because it doesn’t contain “memtable” and “log”: inserted data is written directly to the filesystem. This makes it suitable only to INSERT data in batches, not by individual row and not very frequently – about once per second is ok, but a thousand times a second is not. We did it this way for simplicity’s sake, and because we are already inserting data in batches in our applications.
`MergeTree` is not an LSM tree because it does not contain “memtable” and “log”: inserted data is written directly to the filesystem. This makes it suitable only to INSERT data in batches, not by individual row and not very frequently – about once per second is ok, but a thousand times a second is not. We did it this way for simplicity’s sake, and because we are already inserting data in batches in our applications.
There are MergeTree engines that are doing additional work during background merges. Examples are `CollapsingMergeTree` and `AggregatingMergeTree`. This could be treated as special support for updates. Keep in mind that these are not real updates because users usually have no control over the time when background merges are executed, and data in a `MergeTree` table is almost always stored in more than one part, not in completely merged form.
@ -185,7 +185,7 @@ Replication in ClickHouse can be configured on a per-table basis. You could have
Replication is implemented in the `ReplicatedMergeTree` storage engine. The path in `ZooKeeper` is specified as a parameter for the storage engine. All tables with the same path in `ZooKeeper` become replicas of each other: they synchronize their data and maintain consistency. Replicas can be added and removed dynamically simply by creating or dropping a table.
Replication uses an asynchronous multi-master scheme. You can insert data into any replica that has a session with `ZooKeeper`, and data is replicated to all other replicas asynchronously. Because ClickHouse doesn’t support UPDATEs, replication is conflict-free. As there is no quorum acknowledgment of inserts, just-inserted data might be lost if one node fails.
Replication uses an asynchronous multi-master scheme. You can insert data into any replica that has a session with `ZooKeeper`, and data is replicated to all other replicas asynchronously. Because ClickHouse does not support UPDATEs, replication is conflict-free. As there is no quorum acknowledgment of inserts, just-inserted data might be lost if one node fails.
Metadata for replication is stored in ZooKeeper. There is a replication log that lists what actions to do. Actions are: get part; merge parts; drop a partition, and so on. Each replica copies the replication log to its queue and then executes the actions from the queue. For example, on insertion, the “get the part” action is created in the log, and every replica downloads that part. Merges are coordinated between replicas to get byte-identical results. All parts are merged in the same way on all replicas. One of the leaders initiates a new merge first and writes “merge parts” actions to the log. Multiple replicas (or all) can be leaders at the same time. A replica can be prevented from becoming a leader using the `merge_tree` setting `replicated_can_become_leader`. The leaders are responsible for scheduling background merges.

9
docs/en/development/build-osx.md
View File

@ -20,7 +20,7 @@ Install the latest [Xcode](https://apps.apple.com/am/app/xcode/id497799835?mt=12
Open it at least once to accept the end-user license agreement and automatically install the required components.
Then, make sure that the latest Comman Line Tools are installed and selected in the system:
Then, make sure that the latest Command Line Tools are installed and selected in the system:
``` bash
sudo rm -rf /Library/Developer/CommandLineTools
@ -124,4 +124,11 @@ Reboot.
To check if it’s working, you can use `ulimit -n` command.
## Run ClickHouse server:
```
cd ClickHouse
./build/programs/clickhouse-server --config-file ./programs/server/config.xml
```
[Original article](https://clickhouse.tech/docs/en/development/build_osx/) <!--hide-->

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

@ -134,7 +134,7 @@ $ ./release
## Faster builds for development
Normally all tools of the ClickHouse bundle, such as `clickhouse-server`, `clickhouse-client` etc., are linked into a single static executable, `clickhouse`. This executable must be re-linked on every change, which might be slow. Two common ways to improve linking time are to use `lld` linker, and use the 'split' build configuration, which builds a separate binary for every tool, and further splits the code into serveral shared libraries. To enable these tweaks, pass the following flags to `cmake`:
Normally all tools of the ClickHouse bundle, such as `clickhouse-server`, `clickhouse-client` etc., are linked into a single static executable, `clickhouse`. This executable must be re-linked on every change, which might be slow. Two common ways to improve linking time are to use `lld` linker, and use the 'split' build configuration, which builds a separate binary for every tool, and further splits the code into several shared libraries. To enable these tweaks, pass the following flags to `cmake`:
```
-DCMAKE_C_FLAGS="--ld-path=lld" -DCMAKE_CXX_FLAGS="--ld-path=lld" -DUSE_STATIC_LIBRARIES=0 -DSPLIT_SHARED_LIBRARIES=1 -DCLICKHOUSE_SPLIT_BINARY=1

2
docs/en/development/developer-instruction.md
View File

@ -15,7 +15,7 @@ ClickHouse cannot work or build on a 32-bit system. You should acquire access to
To start working with ClickHouse repository you will need a GitHub account.
You probably already have one, but if you don’t, please register at https://github.com. In case you do not have SSH keys, you should generate them and then upload them on GitHub. It is required for sending over your patches. It is also possible to use the same SSH keys that you use with any other SSH servers - probably you already have those.
You probably already have one, but if you do not, please register at https://github.com. In case you do not have SSH keys, you should generate them and then upload them on GitHub. It is required for sending over your patches. It is also possible to use the same SSH keys that you use with any other SSH servers - probably you already have those.
Create a fork of ClickHouse repository. To do that please click on the “fork” button in the upper right corner at https://github.com/ClickHouse/ClickHouse. It will fork your own copy of ClickHouse/ClickHouse to your account.

10
docs/en/development/style.md
View File

@ -195,7 +195,7 @@ std::cerr << static_cast<int>(c) << std::endl;
The same is true for small methods in any classes or structs.
For templated classes and structs, don’t separate the method declarations from the implementation (because otherwise they must be defined in the same translation unit).
For templated classes and structs, do not separate the method declarations from the implementation (because otherwise they must be defined in the same translation unit).
**31.** You can wrap lines at 140 characters, instead of 80.
@ -442,7 +442,7 @@ Use `RAII` and see above.
**3.** Error handling.
Use exceptions. In most cases, you only need to throw an exception, and don’t need to catch it (because of `RAII`).
Use exceptions. In most cases, you only need to throw an exception, and do not need to catch it (because of `RAII`).
In offline data processing applications, it’s often acceptable to not catch exceptions.
@ -599,7 +599,7 @@ public:
There is no need to use a separate `namespace` for application code.
Small libraries don’t need this, either.
Small libraries do not need this, either.
For medium to large libraries, put everything in a `namespace`.
@ -755,9 +755,9 @@ If there is a good solution already available, then use it, even if it means you
(But be prepared to remove bad libraries from code.)
**3.** You can install a library that isn’t in the packages, if the packages don’t have what you need or have an outdated version or the wrong type of compilation.
**3.** You can install a library that isn’t in the packages, if the packages do not have what you need or have an outdated version or the wrong type of compilation.
**4.** If the library is small and doesn’t have its own complex build system, put the source files in the `contrib` folder.
**4.** If the library is small and does not have its own complex build system, put the source files in the `contrib` folder.
**5.** Preference is always given to libraries that are already in use.

18
docs/en/development/tests.md
View File

@ -35,7 +35,7 @@ Tests should use (create, drop, etc) only tables in `test` database that is assu
### Choosing the Test Name
The name of the test starts with a five-digit prefix followed by a descriptive name, such as `00422_hash_function_constexpr.sql`. To choose the prefix, find the largest prefix already present in the directory, and increment it by one. In the meantime, some other tests might be added with the same numeric prefix, but this is OK and doesn't lead to any problems, you don't have to change it later.
The name of the test starts with a five-digit prefix followed by a descriptive name, such as `00422_hash_function_constexpr.sql`. To choose the prefix, find the largest prefix already present in the directory, and increment it by one. In the meantime, some other tests might be added with the same numeric prefix, but this is OK and does not lead to any problems, you don't have to change it later.
Some tests are marked with `zookeeper`, `shard` or `long` in their names. `zookeeper` is for tests that are using ZooKeeper. `shard` is for tests that requires server to listen `127.0.0.*`; `distributed` or `global` have the same meaning. `long` is for tests that run slightly longer that one second. You can disable these groups of tests using `--no-zookeeper`, `--no-shard` and `--no-long` options, respectively. Make sure to add a proper prefix to your test name if it needs ZooKeeper or distributed queries.
@ -51,7 +51,7 @@ Do not check for a particular wording of error message, it may change in the fut
### Testing a Distributed Query
If you want to use distributed queries in functional tests, you can leverage `remote` table function with `127.0.0.{1..2}` addresses for the server to query itself; or you can use predefined test clusters in server configuration file like `test_shard_localhost`. Remember to add the words `shard` or `distributed` to the test name, so that it is ran in CI in correct configurations, where the server is configured to support distributed queries.
If you want to use distributed queries in functional tests, you can leverage `remote` table function with `127.0.0.{1..2}` addresses for the server to query itself; or you can use predefined test clusters in server configuration file like `test_shard_localhost`. Remember to add the words `shard` or `distributed` to the test name, so that it is run in CI in correct configurations, where the server is configured to support distributed queries.
## Known Bugs {#known-bugs}
@ -60,11 +60,11 @@ If we know some bugs that can be easily reproduced by functional tests, we place
## Integration Tests {#integration-tests}
Integration tests allow to test ClickHouse in clustered configuration and ClickHouse interaction with other servers like MySQL, Postgres, MongoDB. They are useful to emulate network splits, packet drops, etc. These tests are run under Docker and create multiple containers with various software.
Integration tests allow testing ClickHouse in clustered configuration and ClickHouse interaction with other servers like MySQL, Postgres, MongoDB. They are useful to emulate network splits, packet drops, etc. These tests are run under Docker and create multiple containers with various software.
See `tests/integration/README.md` on how to run these tests.
Note that integration of ClickHouse with third-party drivers is not tested. Also we currently don’t have integration tests with our JDBC and ODBC drivers.
Note that integration of ClickHouse with third-party drivers is not tested. Also, we currently do not have integration tests with our JDBC and ODBC drivers.
## Unit Tests {#unit-tests}
@ -123,7 +123,7 @@ Example with gdb:
$ sudo -u clickhouse gdb --args /usr/bin/clickhouse server --config-file /etc/clickhouse-server/config.xml
```
If the system clickhouse-server is already running and you don’t want to stop it, you can change port numbers in your `config.xml` (or override them in a file in `config.d` directory), provide appropriate data path, and run it.
If the system clickhouse-server is already running and you do not want to stop it, you can change port numbers in your `config.xml` (or override them in a file in `config.d` directory), provide appropriate data path, and run it.
`clickhouse` binary has almost no dependencies and works across wide range of Linux distributions. To quick and dirty test your changes on a server, you can simply `scp` your fresh built `clickhouse` binary to your server and then run it as in examples above.
@ -161,7 +161,7 @@ $ clickhouse benchmark --concurrency 16 < queries.tsv
Then leave it for a night or weekend and go take a rest.
You should check that `clickhouse-server` doesn’t crash, memory footprint is bounded and performance not degrading over time.
You should check that `clickhouse-server` does not crash, memory footprint is bounded and performance not degrading over time.
Precise query execution timings are not recorded and not compared due to high variability of queries and environment.
@ -230,7 +230,7 @@ Fuzzers are not built by default. To build fuzzers both `-DENABLE_FUZZING=1` and
We recommend to disable Jemalloc while building fuzzers. Configuration used to integrate ClickHouse fuzzing to
Google OSS-Fuzz can be found at `docker/fuzz`.
We also use simple fuzz test to generate random SQL queries and to check that the server doesn’t die executing them.
We also use simple fuzz test to generate random SQL queries and to check that the server does not die executing them.
You can find it in `00746_sql_fuzzy.pl`. This test should be run continuously (overnight and longer).
We also use sophisticated AST-based query fuzzer that is able to find huge amount of corner cases. It does random permutations and substitutions in queries AST. It remembers AST nodes from previous tests to use them for fuzzing of subsequent tests while processing them in random order. You can learn more about this fuzzer in [this blog article](https://clickhouse.tech/blog/en/2021/fuzzing-clickhouse/).
@ -332,7 +332,7 @@ We run tests with Yandex internal CI and job automation system named “Sandbox
Build jobs and tests are run in Sandbox on per commit basis. Resulting packages and test results are published in GitHub and can be downloaded by direct links. Artifacts are stored for several months. When you send a pull request on GitHub, we tag it as “can be tested” and our CI system will build ClickHouse packages (release, debug, with address sanitizer, etc) for you.
We don’t use Travis CI due to the limit on time and computational power.
We don’t use Jenkins. It was used before and now we are happy we are not using Jenkins.
We do not use Travis CI due to the limit on time and computational power.
We do not use Jenkins. It was used before and now we are happy we are not using Jenkins.
[Original article](https://clickhouse.tech/docs/en/development/tests/) <!--hide-->

2
docs/en/engines/database-engines/atomic.md
View File

@ -47,7 +47,7 @@ EXCHANGE TABLES new_table AND old_table;
### ReplicatedMergeTree in Atomic Database {#replicatedmergetree-in-atomic-database}
For [ReplicatedMergeTree](../table-engines/mergetree-family/replication.md#table_engines-replication) tables is recomended do not specify parameters of engine - path in ZooKeeper and replica name. In this case will be used parameters of the configuration [default_replica_path](../../operations/server-configuration-parameters/settings.md#default_replica_path) and [default_replica_name](../../operations/server-configuration-parameters/settings.md#default_replica_name). If you want specify parameters of engine explicitly than recomended to use {uuid} macros. This is useful so that unique paths are automatically generated for each table in the ZooKeeper.
For [ReplicatedMergeTree](../table-engines/mergetree-family/replication.md#table_engines-replication) tables, it is recommended to not specify engine parameters - path in ZooKeeper and replica name. In this case, configuration parameters will be used [default_replica_path](../../operations/server-configuration-parameters/settings.md#default_replica_path) and [default_replica_name](../../operations/server-configuration-parameters/settings.md#default_replica_name). If you want to specify engine parameters explicitly, it is recommended to use {uuid} macros. This is useful so that unique paths are automatically generated for each table in ZooKeeper.
## See Also

4
docs/en/engines/table-engines/index.md
View File

@ -82,8 +82,8 @@ Virtual column is an integral table engine attribute that is defined in the engi
You shouldn’t specify virtual columns in the `CREATE TABLE` query and you can’t see them in `SHOW CREATE TABLE` and `DESCRIBE TABLE` query results. Virtual columns are also read-only, so you can’t insert data into virtual columns.
To select data from a virtual column, you must specify its name in the `SELECT` query. `SELECT *` doesn’t return values from virtual columns.
To select data from a virtual column, you must specify its name in the `SELECT` query. `SELECT *` does not return values from virtual columns.
If you create a table with a column that has the same name as one of the table virtual columns, the virtual column becomes inaccessible. We don’t recommend doing this. To help avoid conflicts, virtual column names are usually prefixed with an underscore.
If you create a table with a column that has the same name as one of the table virtual columns, the virtual column becomes inaccessible. We do not recommend doing this. To help avoid conflicts, virtual column names are usually prefixed with an underscore.
[Original article](https://clickhouse.tech/docs/en/engines/table-engines/) <!--hide-->

2
docs/en/engines/table-engines/integrations/kafka.md
View File

@ -40,7 +40,7 @@ Required parameters:
- `kafka_broker_list` — A comma-separated list of brokers (for example, `localhost:9092`).
- `kafka_topic_list` — A list of Kafka topics.
- `kafka_group_name` — A group of Kafka consumers. Reading margins are tracked for each group separately. If you don’t want messages to be duplicated in the cluster, use the same group name everywhere.
- `kafka_group_name` — A group of Kafka consumers. Reading margins are tracked for each group separately. If you do not want messages to be duplicated in the cluster, use the same group name everywhere.
- `kafka_format` — Message format. Uses the same notation as the SQL `FORMAT` function, such as `JSONEachRow`. For more information, see the [Formats](../../../interfaces/formats.md) section.
Optional parameters:

7
docs/en/engines/table-engines/integrations/mysql.md
View File

@ -15,7 +15,12 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster]
name1 [type1] [DEFAULT|MATERIALIZED|ALIAS expr1] [TTL expr1],
name2 [type2] [DEFAULT|MATERIALIZED|ALIAS expr2] [TTL expr2],
...
) ENGINE = MySQL('host:port', 'database', 'table', 'user', 'password'[, replace_query, 'on_duplicate_clause']);
) ENGINE = MySQL('host:port', 'database', 'table', 'user', 'password'[, replace_query, 'on_duplicate_clause'])
SETTINGS
[connection_pool_size=16, ]
[connection_max_tries=3, ]
[connection_auto_close=true ]
;
```
See a detailed description of the [CREATE TABLE](../../../sql-reference/statements/create/table.md#create-table-query) query.

4
docs/en/engines/table-engines/integrations/s3.md
View File

@ -139,7 +139,8 @@ The following settings can be specified in configuration file for given endpoint
- `endpoint` — Specifies prefix of an endpoint. Mandatory.
- `access_key_id` and `secret_access_key` — Specifies credentials to use with given endpoint. Optional.
- `use_environment_credentials` — If set to `true`, S3 client will try to obtain credentials from environment variables and Amazon EC2 metadata for given endpoint. Optional, default value is `false`.
- `use_environment_credentials` — If set to `true`, S3 client will try to obtain credentials from environment variables and [Amazon EC2](https://en.wikipedia.org/wiki/Amazon_Elastic_Compute_Cloud) metadata for given endpoint. Optional, default value is `false`.
- `region` — Specifies S3 region name. Optional.
- `use_insecure_imds_request` — If set to `true`, S3 client will use insecure IMDS request while obtaining credentials from Amazon EC2 metadata. Optional, default value is `false`.
- `header` — Adds specified HTTP header to a request to given endpoint. Optional, can be speficied multiple times.
- `server_side_encryption_customer_key_base64` — If specified, required headers for accessing S3 objects with SSE-C encryption will be set. Optional.
@ -152,6 +153,7 @@ The following settings can be specified in configuration file for given endpoint
<endpoint>https://storage.yandexcloud.net/my-test-bucket-768/</endpoint>
<!-- <access_key_id>ACCESS_KEY_ID</access_key_id> -->
<!-- <secret_access_key>SECRET_ACCESS_KEY</secret_access_key> -->
<!-- <region>us-west-1</region> -->
<!-- <use_environment_credentials>false</use_environment_credentials> -->
<!-- <use_insecure_imds_request>false</use_insecure_imds_request> -->
<!-- <header>Authorization: Bearer SOME-TOKEN</header> -->

2
docs/en/engines/table-engines/log-family/index.md
View File

@ -38,7 +38,7 @@ Engines:
## Differences {#differences}
The `TinyLog` engine is the simplest in the family and provides the poorest functionality and lowest efficiency. The `TinyLog` engine doesn’t support parallel data reading by several threads in a single query. It reads data slower than other engines in the family that support parallel reading from a single query and it uses almost as many file descriptors as the `Log` engine because it stores each column in a separate file. Use it only in simple scenarios.
The `TinyLog` engine is the simplest in the family and provides the poorest functionality and lowest efficiency. The `TinyLog` engine does not support parallel data reading by several threads in a single query. It reads data slower than other engines in the family that support parallel reading from a single query and it uses almost as many file descriptors as the `Log` engine because it stores each column in a separate file. Use it only in simple scenarios.
The `Log` and `StripeLog` engines support parallel data reading. When reading data, ClickHouse uses multiple threads. Each thread processes a separate data block. The `Log` engine uses a separate file for each column of the table. `StripeLog` stores all the data in one file. As a result, the `StripeLog` engine uses fewer file descriptors, but the `Log` engine provides higher efficiency when reading data.

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

@ -126,7 +126,7 @@ Also when there are at least 2 more “state” rows than “cancel” rows, or
Thus, collapsing should not change the results of calculating statistics.
Changes gradually collapsed so that in the end only the last state of almost every object left.
The `Sign` is required because the merging algorithm doesn’t guarantee that all of the rows with the same sorting key will be in the same resulting data part and even on the same physical server. ClickHouse process `SELECT` queries with multiple threads, and it can not predict the order of rows in the result. The aggregation is required if there is a need to get completely “collapsed” data from `CollapsingMergeTree` table.
The `Sign` is required because the merging algorithm does not guarantee that all of the rows with the same sorting key will be in the same resulting data part and even on the same physical server. ClickHouse process `SELECT` queries with multiple threads, and it can not predict the order of rows in the result. The aggregation is required if there is a need to get completely “collapsed” data from `CollapsingMergeTree` table.
To finalize collapsing, write a query with `GROUP BY` clause and aggregate functions that account for the sign. For example, to calculate quantity, use `sum(Sign)` instead of `count()`. To calculate the sum of something, use `sum(Sign * x)` instead of `sum(x)`, and so on, and also add `HAVING sum(Sign) > 0`.

2
docs/en/engines/table-engines/mergetree-family/graphitemergetree.md
View File

@ -7,7 +7,7 @@ toc_title: GraphiteMergeTree
This engine is designed for thinning and aggregating/averaging (rollup) [Graphite](http://graphite.readthedocs.io/en/latest/index.html) data. It may be helpful to developers who want to use ClickHouse as a data store for Graphite.
You can use any ClickHouse table engine to store the Graphite data if you don’t need rollup, but if you need a rollup use `GraphiteMergeTree`. The engine reduces the volume of storage and increases the efficiency of queries from Graphite.
You can use any ClickHouse table engine to store the Graphite data if you do not need rollup, but if you need a rollup use `GraphiteMergeTree`. The engine reduces the volume of storage and increases the efficiency of queries from Graphite.
The engine inherits properties from [MergeTree](../../../engines/table-engines/mergetree-family/mergetree.md).

10
docs/en/engines/table-engines/mergetree-family/mergetree.md
View File

@ -64,7 +64,7 @@ For a description of parameters, see the [CREATE query description](../../../sql
ClickHouse uses the sorting key as a primary key if the primary key is not defined obviously by the `PRIMARY KEY` clause.
Use the `ORDER BY tuple()` syntax, if you don’t need sorting. See [Selecting the Primary Key](#selecting-the-primary-key).
Use the `ORDER BY tuple()` syntax, if you do not need sorting. See [Selecting the Primary Key](#selecting-the-primary-key).
- `PARTITION BY` — The [partitioning key](../../../engines/table-engines/mergetree-family/custom-partitioning-key.md). Optional.
@ -162,7 +162,7 @@ Data parts can be stored in `Wide` or `Compact` format. In `Wide` format each co
Data storing format is controlled by the `min_bytes_for_wide_part` and `min_rows_for_wide_part` settings of the table engine. If the number of bytes or rows in a data part is less then the corresponding setting's value, the part is stored in `Compact` format. Otherwise it is stored in `Wide` format. If none of these settings is set, data parts are stored in `Wide` format.
Each data part is logically divided into granules. A granule is the smallest indivisible data set that ClickHouse reads when selecting data. ClickHouse doesn’t split rows or values, so each granule always contains an integer number of rows. The first row of a granule is marked with the value of the primary key for the row. For each data part, ClickHouse creates an index file that stores the marks. For each column, whether it’s in the primary key or not, ClickHouse also stores the same marks. These marks let you find data directly in column files.
Each data part is logically divided into granules. A granule is the smallest indivisible data set that ClickHouse reads when selecting data. ClickHouse does not split rows or values, so each granule always contains an integer number of rows. The first row of a granule is marked with the value of the primary key for the row. For each data part, ClickHouse creates an index file that stores the marks. For each column, whether it’s in the primary key or not, ClickHouse also stores the same marks. These marks let you find data directly in column files.
The granule size is restricted by the `index_granularity` and `index_granularity_bytes` settings of the table engine. The number of rows in a granule lays in the `[1, index_granularity]` range, depending on the size of the rows. The size of a granule can exceed `index_granularity_bytes` if the size of a single row is greater than the value of the setting. In this case, the size of the granule equals the size of the row.
@ -227,7 +227,7 @@ This feature is helpful when using the [SummingMergeTree](../../../engines/table
In this case it makes sense to leave only a few columns in the primary key that will provide efficient range scans and add the remaining dimension columns to the sorting key tuple.
[ALTER](../../../sql-reference/statements/alter/index.md) of the sorting key is a lightweight operation because when a new column is simultaneously added to the table and to the sorting key, existing data parts don’t need to be changed. Since the old sorting key is a prefix of the new sorting key and there is no data in the newly added column, the data is sorted by both the old and new sorting keys at the moment of table modification.
[ALTER](../../../sql-reference/statements/alter/index.md) of the sorting key is a lightweight operation because when a new column is simultaneously added to the table and to the sorting key, existing data parts do not need to be changed. Since the old sorting key is a prefix of the new sorting key and there is no data in the newly added column, the data is sorted by both the old and new sorting keys at the moment of table modification.
### Use of Indexes and Partitions in Queries {#use-of-indexes-and-partitions-in-queries}
@ -265,7 +265,7 @@ The key for partitioning by month allows reading only those data blocks which co
Consider, for example, the days of the month. They form a [monotonic sequence](https://en.wikipedia.org/wiki/Monotonic_function) for one month, but not monotonic for more extended periods. This is a partially-monotonic sequence. If a user creates the table with partially-monotonic primary key, ClickHouse creates a sparse index as usual. When a user selects data from this kind of table, ClickHouse analyzes the query conditions. If the user wants to get data between two marks of the index and both these marks fall within one month, ClickHouse can use the index in this particular case because it can calculate the distance between the parameters of a query and index marks.
ClickHouse cannot use an index if the values of the primary key in the query parameter range don’t represent a monotonic sequence. In this case, ClickHouse uses the full scan method.
ClickHouse cannot use an index if the values of the primary key in the query parameter range do not represent a monotonic sequence. In this case, ClickHouse uses the full scan method.
ClickHouse uses this logic not only for days of the month sequences, but for any primary key that represents a partially-monotonic sequence.
@ -739,6 +739,7 @@ Configuration markup:
<endpoint>https://storage.yandexcloud.net/my-bucket/root-path/</endpoint>
<access_key_id>your_access_key_id</access_key_id>
<secret_access_key>your_secret_access_key</secret_access_key>
<region></region>
<server_side_encryption_customer_key_base64>your_base64_encoded_customer_key</server_side_encryption_customer_key_base64>
<proxy>
<uri>http://proxy1</uri>
@ -764,6 +765,7 @@ Required parameters:
- `secret_access_key` — S3 secret access key.
Optional parameters:
- `region` — S3 region name.
- `use_environment_credentials` — Reads AWS credentials from the Environment variables AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY and AWS_SESSION_TOKEN if they exist. Default value is `false`.
- `use_insecure_imds_request` — If set to `true`, S3 client will use insecure IMDS request while obtaining credentials from Amazon EC2 metadata. Default value is `false`.
- `proxy` — Proxy configuration for S3 endpoint. Each `uri` element inside `proxy` block should contain a proxy URL.

8
docs/en/engines/table-engines/mergetree-family/replacingmergetree.md
View File

@ -7,9 +7,9 @@ toc_title: ReplacingMergeTree
The engine differs from [MergeTree](../../../engines/table-engines/mergetree-family/mergetree.md#table_engines-mergetree) in that it removes duplicate entries with the same [sorting key](../../../engines/table-engines/mergetree-family/mergetree.md) value (`ORDER BY` table section, not `PRIMARY KEY`).
Data deduplication occurs only during a merge. Merging occurs in the background at an unknown time, so you can’t plan for it. Some of the data may remain unprocessed. Although you can run an unscheduled merge using the `OPTIMIZE` query, don’t count on using it, because the `OPTIMIZE` query will read and write a large amount of data.
Data deduplication occurs only during a merge. Merging occurs in the background at an unknown time, so you can’t plan for it. Some of the data may remain unprocessed. Although you can run an unscheduled merge using the `OPTIMIZE` query, do not count on using it, because the `OPTIMIZE` query will read and write a large amount of data.
Thus, `ReplacingMergeTree` is suitable for clearing out duplicate data in the background in order to save space, but it doesn’t guarantee the absence of duplicates.
Thus, `ReplacingMergeTree` is suitable for clearing out duplicate data in the background in order to save space, but it does not guarantee the absence of duplicates.
## Creating a Table {#creating-a-table}
@ -34,7 +34,7 @@ For a description of request parameters, see [statement description](../../../sq
**ReplacingMergeTree Parameters**
- `ver` — column with version. Type `UInt*`, `Date` or `DateTime`. Optional parameter.
- `ver` — column with the version number. Type `UInt*`, `Date`, `DateTime` or `DateTime64`. Optional parameter.
When merging, `ReplacingMergeTree` from all the rows with the same sorting key leaves only one:
@ -66,5 +66,3 @@ All of the parameters excepting `ver` have the same meaning as in `MergeTree`.
- `ver` - column with the version. Optional parameter. For a description, see the text above.
</details>
[Original article](https://clickhouse.tech/docs/en/operations/table_engines/replacingmergetree/) <!--hide-->

9
docs/en/engines/table-engines/mergetree-family/replication.md
View File

@ -95,17 +95,19 @@ If ZooKeeper isn’t set in the config file, you can’t create replicated table
ZooKeeper is not used in `SELECT` queries because replication does not affect the performance of `SELECT` and queries run just as fast as they do for non-replicated tables. When querying distributed replicated tables, ClickHouse behavior is controlled by the settings [max_replica_delay_for_distributed_queries](../../../operations/settings/settings.md#settings-max_replica_delay_for_distributed_queries) and [fallback_to_stale_replicas_for_distributed_queries](../../../operations/settings/settings.md#settings-fallback_to_stale_replicas_for_distributed_queries).
For each `INSERT` query, approximately ten entries are added to ZooKeeper through several transactions. (To be more precise, this is for each inserted block of data; an INSERT query contains one block or one block per `max_insert_block_size = 1048576` rows.) This leads to slightly longer latencies for `INSERT` compared to non-replicated tables. But if you follow the recommendations to insert data in batches of no more than one `INSERT` per second, it doesn’t create any problems. The entire ClickHouse cluster used for coordinating one ZooKeeper cluster has a total of several hundred `INSERTs` per second. The throughput on data inserts (the number of rows per second) is just as high as for non-replicated data.
For each `INSERT` query, approximately ten entries are added to ZooKeeper through several transactions. (To be more precise, this is for each inserted block of data; an INSERT query contains one block or one block per `max_insert_block_size = 1048576` rows.) This leads to slightly longer latencies for `INSERT` compared to non-replicated tables. But if you follow the recommendations to insert data in batches of no more than one `INSERT` per second, it does not create any problems. The entire ClickHouse cluster used for coordinating one ZooKeeper cluster has a total of several hundred `INSERTs` per second. The throughput on data inserts (the number of rows per second) is just as high as for non-replicated data.
For very large clusters, you can use different ZooKeeper clusters for different shards. However, this hasn’t proven necessary on the Yandex.Metrica cluster (approximately 300 servers).
Replication is asynchronous and multi-master. `INSERT` queries (as well as `ALTER`) can be sent to any available server. Data is inserted on the server where the query is run, and then it is copied to the other servers. Because it is asynchronous, recently inserted data appears on the other replicas with some latency. If part of the replicas are not available, the data is written when they become available. If a replica is available, the latency is the amount of time it takes to transfer the block of compressed data over the network. The number of threads performing background tasks for replicated tables can be set by [background_schedule_pool_size](../../../operations/settings/settings.md#background_schedule_pool_size) setting.
`ReplicatedMergeTree` engine uses a separate thread pool for replicated fetches. Size of the pool is limited by the [background_fetches_pool_size](../../../operations/settings/settings.md#background_fetches_pool_size) setting which can be tuned with a server restart.
By default, an INSERT query waits for confirmation of writing the data from only one replica. If the data was successfully written to only one replica and the server with this replica ceases to exist, the stored data will be lost. To enable getting confirmation of data writes from multiple replicas, use the `insert_quorum` option.
Each block of data is written atomically. The INSERT query is divided into blocks up to `max_insert_block_size = 1048576` rows. In other words, if the `INSERT` query has less than 1048576 rows, it is made atomically.
Data blocks are deduplicated. For multiple writes of the same data block (data blocks of the same size containing the same rows in the same order), the block is only written once. The reason for this is in case of network failures when the client application doesn’t know if the data was written to the DB, so the `INSERT` query can simply be repeated. It doesn’t matter which replica INSERTs were sent to with identical data. `INSERTs` are idempotent. Deduplication parameters are controlled by [merge_tree](../../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-merge_tree) server settings.
Data blocks are deduplicated. For multiple writes of the same data block (data blocks of the same size containing the same rows in the same order), the block is only written once. The reason for this is in case of network failures when the client application does not know if the data was written to the DB, so the `INSERT` query can simply be repeated. It does not matter which replica INSERTs were sent to with identical data. `INSERTs` are idempotent. Deduplication parameters are controlled by [merge_tree](../../../operations/server-configuration-parameters/settings.md#server_configuration_parameters-merge_tree) server settings.
During replication, only the source data to insert is transferred over the network. Further data transformation (merging) is coordinated and performed on all the replicas in the same way. This minimizes network usage, which means that replication works well when replicas reside in different datacenters. (Note that duplicating data in different datacenters is the main goal of replication.)
@ -172,7 +174,7 @@ In this case, the path consists of the following parts:
`{layer}-{shard}` is the shard identifier. In this example it consists of two parts, since the Yandex.Metrica cluster uses bi-level sharding. For most tasks, you can leave just the {shard} substitution, which will be expanded to the shard identifier.
`table_name` is the name of the node for the table in ZooKeeper. It is a good idea to make it the same as the table name. It is defined explicitly, because in contrast to the table name, it doesn’t change after a RENAME query.
`table_name` is the name of the node for the table in ZooKeeper. It is a good idea to make it the same as the table name. It is defined explicitly, because in contrast to the table name, it does not change after a RENAME query.
*HINT*: you could add a database name in front of `table_name` as well. E.g. `db_name.table_name`
The two built-in substitutions `{database}` and `{table}` can be used, they expand into the table name and the database name respectively (unless these macros are defined in the `macros` section). So the zookeeper path can be specified as `'/clickhouse/tables/{layer}-{shard}/{database}/{table}'`.
@ -284,6 +286,7 @@ If the data in ZooKeeper was lost or damaged, you can save data by moving it to
**See Also**
- [background_schedule_pool_size](../../../operations/settings/settings.md#background_schedule_pool_size)
- [background_fetches_pool_size](../../../operations/settings/settings.md#background_fetches_pool_size)
- [execute_merges_on_single_replica_time_threshold](../../../operations/settings/settings.md#execute-merges-on-single-replica-time-threshold)
[Original article](https://clickhouse.tech/docs/en/operations/table_engines/replication/) <!--hide-->

2
docs/en/engines/table-engines/mergetree-family/summingmergetree.md
View File

@ -7,7 +7,7 @@ toc_title: SummingMergeTree
The engine inherits from [MergeTree](../../../engines/table-engines/mergetree-family/mergetree.md#table_engines-mergetree). The difference is that when merging data parts for `SummingMergeTree` tables ClickHouse replaces all the rows with the same primary key (or more accurately, with the same [sorting key](../../../engines/table-engines/mergetree-family/mergetree.md)) with one row which contains summarized values for the columns with the numeric data type. If the sorting key is composed in a way that a single key value corresponds to large number of rows, this significantly reduces storage volume and speeds up data selection.
We recommend to use the engine together with `MergeTree`. Store complete data in `MergeTree` table, and use `SummingMergeTree` for aggregated data storing, for example, when preparing reports. Such an approach will prevent you from losing valuable data due to an incorrectly composed primary key.
We recommend using the engine together with `MergeTree`. Store complete data in `MergeTree` table, and use `SummingMergeTree` for aggregated data storing, for example, when preparing reports. Such an approach will prevent you from losing valuable data due to an incorrectly composed primary key.
## Creating a Table {#creating-a-table}

4
docs/en/engines/table-engines/mergetree-family/versionedcollapsingmergetree.md
View File

@ -133,7 +133,7 @@ When ClickHouse inserts data, it orders rows by the primary key. If the `Version
## Selecting Data {#selecting-data}
ClickHouse doesn’t guarantee that all of the rows with the same primary key will be in the same resulting data part or even on the same physical server. This is true both for writing the data and for subsequent merging of the data parts. In addition, ClickHouse processes `SELECT` queries with multiple threads, and it cannot predict the order of rows in the result. This means that aggregation is required if there is a need to get completely “collapsed” data from a `VersionedCollapsingMergeTree` table.
ClickHouse does not guarantee that all of the rows with the same primary key will be in the same resulting data part or even on the same physical server. This is true both for writing the data and for subsequent merging of the data parts. In addition, ClickHouse processes `SELECT` queries with multiple threads, and it cannot predict the order of rows in the result. This means that aggregation is required if there is a need to get completely “collapsed” data from a `VersionedCollapsingMergeTree` table.
To finalize collapsing, write a query with a `GROUP BY` clause and aggregate functions that account for the sign. For example, to calculate quantity, use `sum(Sign)` instead of `count()`. To calculate the sum of something, use `sum(Sign * x)` instead of `sum(x)`, and add `HAVING sum(Sign) > 0`.
@ -219,7 +219,7 @@ HAVING sum(Sign) > 0
└─────────────────────┴───────────┴──────────┴─────────┘
```
If we don’t need aggregation and want to force collapsing, we can use the `FINAL` modifier for the `FROM` clause.
If we do not need aggregation and want to force collapsing, we can use the `FINAL` modifier for the `FROM` clause.
``` sql
SELECT * FROM UAct FINAL

14
docs/en/engines/table-engines/special/buffer.md
View File

@ -20,11 +20,11 @@ Engine parameters:
Optional engine parameters:
- `flush_time`, `flush_rows`, `flush_bytes` – Conditions for flushing data from the buffer, that will happen only in background (ommited or zero means no `flush*` parameters).
- `flush_time`, `flush_rows`, `flush_bytes` – Conditions for flushing data from the buffer, that will happen only in background (omitted or zero means no `flush*` parameters).
Data is flushed from the buffer and written to the destination table if all the `min*` conditions or at least one `max*` condition are met.
Also if at least one `flush*` condition are met flush initiated in background, this is different from `max*`, since `flush*` allows you to configure background flushes separately to avoid adding latency for `INSERT` (into `Buffer`) queries.
Also, if at least one `flush*` condition are met flush initiated in background, this is different from `max*`, since `flush*` allows you to configure background flushes separately to avoid adding latency for `INSERT` (into `Buffer`) queries.
- `min_time`, `max_time`, `flush_time` – Condition for the time in seconds from the moment of the first write to the buffer.
- `min_rows`, `max_rows`, `flush_rows` – Condition for the number of rows in the buffer.
@ -49,12 +49,12 @@ You can set empty strings in single quotation marks for the database and table n
When reading from a Buffer table, data is processed both from the buffer and from the destination table (if there is one).
Note that the Buffer tables does not support an index. In other words, data in the buffer is fully scanned, which might be slow for large buffers. (For data in a subordinate table, the index that it supports will be used.)
If the set of columns in the Buffer table doesn’t match the set of columns in a subordinate table, a subset of columns that exist in both tables is inserted.
If the set of columns in the Buffer table does not match the set of columns in a subordinate table, a subset of columns that exist in both tables is inserted.
If the types don’t match for one of the columns in the Buffer table and a subordinate table, an error message is entered in the server log and the buffer is cleared.
The same thing happens if the subordinate table doesn’t exist when the buffer is flushed.
If the types do not match for one of the columns in the Buffer table and a subordinate table, an error message is entered in the server log, and the buffer is cleared.
The same thing happens if the subordinate table does not exist when the buffer is flushed.
If you need to run ALTER for a subordinate table and the Buffer table, we recommend first deleting the Buffer table, running ALTER for the subordinate table, then creating the Buffer table again.
If you need to run ALTER for a subordinate table, and the Buffer table, we recommend first deleting the Buffer table, running ALTER for the subordinate table, then creating the Buffer table again.
If the server is restarted abnormally, the data in the buffer is lost.
@ -70,6 +70,6 @@ Due to these disadvantages, we can only recommend using a Buffer table in rare c
A Buffer table is used when too many INSERTs are received from a large number of servers over a unit of time and data can’t be buffered before insertion, which means the INSERTs can’t run fast enough.
Note that it doesn’t make sense to insert data one row at a time, even for Buffer tables. This will only produce a speed of a few thousand rows per second, while inserting larger blocks of data can produce over a million rows per second (see the section “Performance”).
Note that it does not make sense to insert data one row at a time, even for Buffer tables. This will only produce a speed of a few thousand rows per second, while inserting larger blocks of data can produce over a million rows per second (see the section “Performance”).
[Original article](https://clickhouse.tech/docs/en/operations/table_engines/buffer/) <!--hide-->

10
docs/en/engines/table-engines/special/distributed.md
View File

@ -25,7 +25,7 @@ The Distributed engine accepts parameters:
- [insert_distributed_sync](../../../operations/settings/settings.md#insert_distributed_sync) setting
- [MergeTree](../../../engines/table-engines/mergetree-family/mergetree.md#table_engine-mergetree-multiple-volumes) for the examples
Also it accept the following settings:
Also, it accepts the following settings:
- `fsync_after_insert` - do the `fsync` for the file data after asynchronous insert to Distributed. Guarantees that the OS flushed the whole inserted data to a file **on the initiator node** disk.
@ -124,7 +124,7 @@ Replicas are duplicating servers (in order to read all the data, you can access
Cluster names must not contain dots.
The parameters `host`, `port`, and optionally `user`, `password`, `secure`, `compression` are specified for each server:
- `host` – The address of the remote server. You can use either the domain or the IPv4 or IPv6 address. If you specify the domain, the server makes a DNS request when it starts, and the result is stored as long as the server is running. If the DNS request fails, the server doesn’t start. If you change the DNS record, restart the server.
- `host` – The address of the remote server. You can use either the domain or the IPv4 or IPv6 address. If you specify the domain, the server makes a DNS request when it starts, and the result is stored as long as the server is running. If the DNS request fails, the server does not start. If you change the DNS record, restart the server.
- `port` – The TCP port for messenger activity (`tcp_port` in the config, usually set to 9000). Do not confuse it with http_port.
- `user` – Name of the user for connecting to a remote server. Default value: default. This user must have access to connect to the specified server. Access is configured in the users.xml file. For more information, see the section [Access rights](../../../operations/access-rights.md).
- `password` – The password for connecting to a remote server (not masked). Default value: empty string.
@ -143,13 +143,13 @@ To view your clusters, use the `system.clusters` table.
The Distributed engine allows working with a cluster like a local server. However, the cluster is inextensible: you must write its configuration in the server config file (even better, for all the cluster’s servers).
The Distributed engine requires writing clusters to the config file. Clusters from the config file are updated on the fly, without restarting the server. If you need to send a query to an unknown set of shards and replicas each time, you don’t need to create a Distributed table – use the `remote` table function instead. See the section [Table functions](../../../sql-reference/table-functions/index.md).
The Distributed engine requires writing clusters to the config file. Clusters from the config file are updated on the fly, without restarting the server. If you need to send a query to an unknown set of shards and replicas each time, you do not need to create a Distributed table – use the `remote` table function instead. See the section [Table functions](../../../sql-reference/table-functions/index.md).
There are two methods for writing data to a cluster:
First, you can define which servers to write which data to and perform the write directly on each shard. In other words, perform INSERT in the tables that the distributed table “looks at”. This is the most flexible solution as you can use any sharding scheme, which could be non-trivial due to the requirements of the subject area. This is also the most optimal solution since data can be written to different shards completely independently.
Second, you can perform INSERT in a Distributed table. In this case, the table will distribute the inserted data across the servers itself. In order to write to a Distributed table, it must have a sharding key set (the last parameter). In addition, if there is only one shard, the write operation works without specifying the sharding key, since it doesn’t mean anything in this case.
Second, you can perform INSERT in a Distributed table. In this case, the table will distribute the inserted data across the servers itself. In order to write to a Distributed table, it must have a sharding key set (the last parameter). In addition, if there is only one shard, the write operation works without specifying the sharding key, since it does not mean anything in this case.
Each shard can have a weight defined in the config file. By default, the weight is equal to one. Data is distributed across shards in the amount proportional to the shard weight. For example, if there are two shards and the first has a weight of 9 while the second has a weight of 10, the first will be sent 9 / 19 parts of the rows, and the second will be sent 10 / 19.
@ -165,7 +165,7 @@ The sharding expression can be any expression from constants and table columns t
A simple reminder from the division is a limited solution for sharding and isn’t always appropriate. It works for medium and large volumes of data (dozens of servers), but not for very large volumes of data (hundreds of servers or more). In the latter case, use the sharding scheme required by the subject area, rather than using entries in Distributed tables.
SELECT queries are sent to all the shards and work regardless of how data is distributed across the shards (they can be distributed completely randomly). When you add a new shard, you don’t have to transfer the old data to it. You can write new data with a heavier weight – the data will be distributed slightly unevenly, but queries will work correctly and efficiently.
SELECT queries are sent to all the shards and work regardless of how data is distributed across the shards (they can be distributed completely randomly). When you add a new shard, you do not have to transfer the old data to it. You can write new data with a heavier weight – the data will be distributed slightly unevenly, but queries will work correctly and efficiently.
You should be concerned about the sharding scheme in the following cases:

2
docs/en/engines/table-engines/special/external-data.md
View File

@ -9,7 +9,7 @@ ClickHouse allows sending a server the data that is needed for processing a quer
For example, if you have a text file with important user identifiers, you can upload it to the server along with a query that uses filtration by this list.
If you need to run more than one query with a large volume of external data, don’t use this feature. It is better to upload the data to the DB ahead of time.
If you need to run more than one query with a large volume of external data, do not use this feature. It is better to upload the data to the DB ahead of time.
External data can be uploaded using the command-line client (in non-interactive mode), or using the HTTP interface.

2
docs/en/engines/table-engines/special/file.md
View File

@ -24,7 +24,7 @@ The `Format` parameter specifies one of the available file formats. To perform
`INSERT` queries – for output. The available formats are listed in the
[Formats](../../../interfaces/formats.md#formats) section.
ClickHouse does not allow to specify filesystem path for`File`. It will use folder defined by [path](../../../operations/server-configuration-parameters/settings.md) setting in server configuration.
ClickHouse does not allow specifying filesystem path for`File`. It will use folder defined by [path](../../../operations/server-configuration-parameters/settings.md) setting in server configuration.
When creating table using `File(Format)` it creates empty subdirectory in that folder. When data is written to that table, it’s put into `data.Format` file in that subdirectory.

2
docs/en/engines/table-engines/special/join.md
View File

@ -28,7 +28,7 @@ See the detailed description of the [CREATE TABLE](../../../sql-reference/statem
- `join_type` – [JOIN type](../../../sql-reference/statements/select/join.md#select-join-types).
- `k1[, k2, ...]` – Key columns from the `USING` clause that the `JOIN` operation is made with.
Enter `join_strictness` and `join_type` parameters without quotes, for example, `Join(ANY, LEFT, col1)`. They must match the `JOIN` operation that the table will be used for. If the parameters don’t match, ClickHouse doesn’t throw an exception and may return incorrect data.
Enter `join_strictness` and `join_type` parameters without quotes, for example, `Join(ANY, LEFT, col1)`. They must match the `JOIN` operation that the table will be used for. If the parameters do not match, ClickHouse does not throw an exception and may return incorrect data.
## Table Usage {#table-usage}

2
docs/en/engines/table-engines/special/memory.md
View File

@ -6,7 +6,7 @@ toc_title: Memory
# Memory Table Engine {#memory}
The Memory engine stores data in RAM, in uncompressed form. Data is stored in exactly the same form as it is received when read. In other words, reading from this table is completely free.
Concurrent data access is synchronized. Locks are short: read and write operations don’t block each other.
Concurrent data access is synchronized. Locks are short: read and write operations do not block each other.
Indexes are not supported. Reading is parallelized.
Maximal productivity (over 10 GB/sec) is reached on simple queries, because there is no reading from the disk, decompressing, or deserializing data. (We should note that in many cases, the productivity of the MergeTree engine is almost as high.)

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