From 574c445be9368ee481e8d213251106233d417f69 Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Wed, 7 Aug 2024 16:29:07 +0000 Subject: [PATCH] Refactor tests for (experimental) statistics --- docs/en/development/tests.md | 4 +- .../statements/alter/statistics.md | 16 +- src/Storages/MergeTree/MergeTreeData.cpp | 2 +- ...2864_statistics_count_min_sketch.reference | 14 -- .../02864_statistics_count_min_sketch.sql | 70 ------ .../02864_statistics_ddl.reference | 37 +-- .../0_stateless/02864_statistics_ddl.sql | 234 ++++++++++++++---- ...delayed_materialization_in_merge.reference | 12 + ...stics_delayed_materialization_in_merge.sql | 36 +++ .../02864_statistics_exception.reference | 0 .../02864_statistics_exception.sql | 55 ---- ..._statistics_materialize_in_merge.reference | 10 - .../02864_statistics_materialize_in_merge.sql | 52 ---- .../02864_statistics_predicates.reference | 98 ++++++++ .../02864_statistics_predicates.sql | 214 ++++++++++++++++ .../02864_statistics_uniq.reference | 35 --- .../0_stateless/02864_statistics_uniq.sql | 73 ------ .../02864_statistics_usage.reference | 20 ++ .../0_stateless/02864_statistics_usage.sql | 42 ++++ 19 files changed, 625 insertions(+), 399 deletions(-) delete mode 100644 tests/queries/0_stateless/02864_statistics_count_min_sketch.reference delete mode 100644 tests/queries/0_stateless/02864_statistics_count_min_sketch.sql create mode 100644 tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.reference create mode 100644 tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.sql delete mode 100644 tests/queries/0_stateless/02864_statistics_exception.reference delete mode 100644 tests/queries/0_stateless/02864_statistics_exception.sql delete mode 100644 tests/queries/0_stateless/02864_statistics_materialize_in_merge.reference delete mode 100644 tests/queries/0_stateless/02864_statistics_materialize_in_merge.sql create mode 100644 tests/queries/0_stateless/02864_statistics_predicates.reference create mode 100644 tests/queries/0_stateless/02864_statistics_predicates.sql delete mode 100644 tests/queries/0_stateless/02864_statistics_uniq.reference delete mode 100644 tests/queries/0_stateless/02864_statistics_uniq.sql create mode 100644 tests/queries/0_stateless/02864_statistics_usage.reference create mode 100644 tests/queries/0_stateless/02864_statistics_usage.sql diff --git a/docs/en/development/tests.md b/docs/en/development/tests.md index 269995a1a96..6cb36e2049b 100644 --- a/docs/en/development/tests.md +++ b/docs/en/development/tests.md @@ -14,7 +14,7 @@ Each functional test sends one or multiple queries to the running ClickHouse ser Tests are located in `queries` directory. There are two subdirectories: `stateless` and `stateful`. Stateless tests run queries without any preloaded test data - they often create small synthetic datasets on the fly, within the test itself. Stateful tests require preloaded test data from ClickHouse and it is available to general public. -Each test can be one of two types: `.sql` and `.sh`. `.sql` test is the simple SQL script that is piped to `clickhouse-client --multiquery`. `.sh` test is a script that is run by itself. SQL tests are generally preferable to `.sh` tests. You should use `.sh` tests only when you have to test some feature that cannot be exercised from pure SQL, such as piping some input data into `clickhouse-client` or testing `clickhouse-local`. +Each test can be one of two types: `.sql` and `.sh`. `.sql` test is the simple SQL script that is piped to `clickhouse-client`. `.sh` test is a script that is run by itself. SQL tests are generally preferable to `.sh` tests. You should use `.sh` tests only when you have to test some feature that cannot be exercised from pure SQL, such as piping some input data into `clickhouse-client` or testing `clickhouse-local`. :::note A common mistake when testing data types `DateTime` and `DateTime64` is assuming that the server uses a specific time zone (e.g. "UTC"). This is not the case, time zones in CI test runs @@ -38,7 +38,7 @@ For more options, see `tests/clickhouse-test --help`. You can simply run all tes ### Adding a New Test -To add new test, create a `.sql` or `.sh` file in `queries/0_stateless` directory, check it manually and then generate `.reference` file in the following way: `clickhouse-client --multiquery < 00000_test.sql > 00000_test.reference` or `./00000_test.sh > ./00000_test.reference`. +To add new test, create a `.sql` or `.sh` file in `queries/0_stateless` directory, check it manually and then generate `.reference` file in the following way: `clickhouse-client < 00000_test.sql > 00000_test.reference` or `./00000_test.sh > ./00000_test.reference`. Tests should use (create, drop, etc) only tables in `test` database that is assumed to be created beforehand; also tests can use temporary tables. diff --git a/docs/en/sql-reference/statements/alter/statistics.md b/docs/en/sql-reference/statements/alter/statistics.md index 6880cef0e5c..7a1774a01b5 100644 --- a/docs/en/sql-reference/statements/alter/statistics.md +++ b/docs/en/sql-reference/statements/alter/statistics.md @@ -8,26 +8,28 @@ sidebar_label: STATISTICS The following operations are available: -- `ALTER TABLE [db].table ADD STATISTICS (columns list) TYPE (type list)` - Adds statistic description to tables metadata. +- `ALTER TABLE [db].table ADD STATISTICS [IF NOT EXISTS] (column list) TYPE (type list)` - Adds statistic description to tables metadata. -- `ALTER TABLE [db].table MODIFY STATISTICS (columns list) TYPE (type list)` - Modifies statistic description to tables metadata. +- `ALTER TABLE [db].table MODIFY STATISTICS (column list) TYPE (type list)` - Modifies statistic description to tables metadata. -- `ALTER TABLE [db].table DROP STATISTICS (columns list)` - Removes statistics from the metadata of the specified columns and deletes all statistics objects in all parts for the specified columns. +- `ALTER TABLE [db].table DROP STATISTICS [IF EXISTS] (column list)` - Removes statistics from the metadata of the specified columns and deletes all statistics objects in all parts for the specified columns. -- `ALTER TABLE [db].table CLEAR STATISTICS (columns list)` - Deletes all statistics objects in all parts for the specified columns. Statistics objects can be rebuild using `ALTER TABLE MATERIALIZE STATISTICS`. +- `ALTER TABLE [db].table CLEAR STATISTICS [IF EXISTS] (column list)` - Deletes all statistics objects in all parts for the specified columns. Statistics objects can be rebuild using `ALTER TABLE MATERIALIZE STATISTICS`. -- `ALTER TABLE [db.]table MATERIALIZE STATISTICS (columns list)` - Rebuilds the statistic for columns. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). +- `ALTER TABLE [db.]table MATERIALIZE STATISTICS [IF EXISTS] (column list)` - Rebuilds the statistic for columns. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). The first two commands are lightweight in a sense that they only change metadata or remove files. Also, they are replicated, syncing statistics metadata via ZooKeeper. -There is an example adding two statistics types to two columns: +## Example: + +Adding two statistics types to two columns: ``` ALTER TABLE t1 MODIFY STATISTICS c, d TYPE TDigest, Uniq; ``` :::note -Statistic manipulation is supported only for tables with [`*MergeTree`](../../../engines/table-engines/mergetree-family/mergetree.md) engine (including [replicated](../../../engines/table-engines/mergetree-family/replication.md) variants). +Statistic are supported only for [`*MergeTree`](../../../engines/table-engines/mergetree-family/mergetree.md) engine tables (including [replicated](../../../engines/table-engines/mergetree-family/replication.md) variants). ::: diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 625b1281c61..c7101021f02 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -3517,7 +3517,7 @@ void MergeTreeData::checkAlterIsPossible(const AlterCommands & commands, Context const auto & new_column = new_metadata.getColumns().get(command.column_name); if (!old_column.type->equals(*new_column.type)) throw Exception(ErrorCodes::ALTER_OF_COLUMN_IS_FORBIDDEN, - "ALTER types of column {} with statistics is not not safe " + "ALTER types of column {} with statistics is not safe " "because it can change the representation of statistics", backQuoteIfNeed(command.column_name)); } diff --git a/tests/queries/0_stateless/02864_statistics_count_min_sketch.reference b/tests/queries/0_stateless/02864_statistics_count_min_sketch.reference deleted file mode 100644 index 02c41656a36..00000000000 --- a/tests/queries/0_stateless/02864_statistics_count_min_sketch.reference +++ /dev/null @@ -1,14 +0,0 @@ -CREATE TABLE default.tab\n(\n `a` String,\n `b` UInt64,\n `c` Int64,\n `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS min_bytes_for_wide_part = 0, index_granularity = 8192 -Test statistics count_min: - Prewhere info - Prewhere filter - Prewhere filter column: and(equals(a, \'0\'), equals(b, 0), equals(c, 0)) (removed) -Test statistics multi-types: - Prewhere info - Prewhere filter - Prewhere filter column: and(equals(a, \'0\'), less(c, -90), greater(b, 900)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(equals(a, \'10000\'), equals(b, 0), less(c, 0)) (removed) -Test LowCardinality and Nullable data type: -tab2 diff --git a/tests/queries/0_stateless/02864_statistics_count_min_sketch.sql b/tests/queries/0_stateless/02864_statistics_count_min_sketch.sql deleted file mode 100644 index c730aa7b4a7..00000000000 --- a/tests/queries/0_stateless/02864_statistics_count_min_sketch.sql +++ /dev/null @@ -1,70 +0,0 @@ --- Tags: no-fasttest - -DROP TABLE IF EXISTS tab SYNC; - -SET allow_experimental_statistics = 1; -SET allow_statistics_optimize = 1; -SET allow_suspicious_low_cardinality_types=1; -SET mutations_sync = 2; - -CREATE TABLE tab -( - a String, - b UInt64, - c Int64, - pk String, -) Engine = MergeTree() ORDER BY pk -SETTINGS min_bytes_for_wide_part = 0; - -SHOW CREATE TABLE tab; - -INSERT INTO tab select toString(number % 10000), number % 1000, -(number % 100), generateUUIDv4() FROM system.numbers LIMIT 10000; - -SELECT 'Test statistics count_min:'; - -ALTER TABLE tab ADD STATISTICS a TYPE count_min; -ALTER TABLE tab ADD STATISTICS b TYPE count_min; -ALTER TABLE tab ADD STATISTICS c TYPE count_min; -ALTER TABLE tab MATERIALIZE STATISTICS a, b, c; - -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8|_UInt16|_String', '') -FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE c = 0/*100*/ and b = 0/*10*/ and a = '0'/*1*/) xx -WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; - -ALTER TABLE tab DROP STATISTICS a, b, c; - - -SELECT 'Test statistics multi-types:'; - -ALTER TABLE tab ADD STATISTICS a TYPE count_min; -ALTER TABLE tab ADD STATISTICS b TYPE count_min, uniq, tdigest; -ALTER TABLE tab ADD STATISTICS c TYPE count_min, uniq, tdigest; -ALTER TABLE tab MATERIALIZE STATISTICS a, b, c; - -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8|_UInt16|_String', '') -FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE c < -90/*900*/ and b > 900/*990*/ and a = '0'/*1*/) -WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; - -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8|_UInt16|_String', '') -FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE c < 0/*9900*/ and b = 0/*10*/ and a = '10000'/*0*/) -WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; - -ALTER TABLE tab DROP STATISTICS a, b, c; - -DROP TABLE IF EXISTS tab SYNC; - - -SELECT 'Test LowCardinality and Nullable data type:'; -DROP TABLE IF EXISTS tab2 SYNC; -SET allow_suspicious_low_cardinality_types=1; -CREATE TABLE tab2 -( - a LowCardinality(Int64) STATISTICS(count_min), - b Nullable(Int64) STATISTICS(count_min), - c LowCardinality(Nullable(Int64)) STATISTICS(count_min), - pk String, -) Engine = MergeTree() ORDER BY pk; - -select name from system.tables where name = 'tab2' and database = currentDatabase(); - -DROP TABLE IF EXISTS tab2 SYNC; diff --git a/tests/queries/0_stateless/02864_statistics_ddl.reference b/tests/queries/0_stateless/02864_statistics_ddl.reference index a7ff5caa0b0..0e453b0ee8a 100644 --- a/tests/queries/0_stateless/02864_statistics_ddl.reference +++ b/tests/queries/0_stateless/02864_statistics_ddl.reference @@ -1,31 +1,6 @@ -CREATE TABLE default.tab\n(\n `a` Float64 STATISTICS(tdigest),\n `b` Int64 STATISTICS(tdigest),\n `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS min_bytes_for_wide_part = 0, index_granularity = 8192 -After insert - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), less(b, 10)) (removed) -10 -0 -After drop statistic - Prewhere info - Prewhere filter - Prewhere filter column: and(less(b, 10), less(a, 10)) (removed) -10 -CREATE TABLE default.tab\n(\n `a` Float64,\n `b` Int64,\n `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS min_bytes_for_wide_part = 0, index_granularity = 8192 -After add statistic -CREATE TABLE default.tab\n(\n `a` Float64 STATISTICS(tdigest),\n `b` Int64 STATISTICS(tdigest),\n `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS min_bytes_for_wide_part = 0, index_granularity = 8192 -After materialize statistic - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), less(b, 10)) (removed) -20 -After merge - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), less(b, 10)) (removed) -20 -CREATE TABLE default.tab\n(\n `a` Float64 STATISTICS(tdigest),\n `c` Int64 STATISTICS(tdigest),\n `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS min_bytes_for_wide_part = 0, index_granularity = 8192 -After rename - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), less(c, 10)) (removed) -20 +CREATE TABLE default.tab\n(\n `f64` Float64,\n `f64_tdigest` Float64 STATISTICS(tdigest),\n `f32` Float32,\n `s` String,\n `a` Array(Float64)\n)\nENGINE = MergeTree\nORDER BY tuple()\nSETTINGS index_granularity = 8192 +CREATE TABLE default.tab\n(\n `f64` Float64 STATISTICS(tdigest, uniq),\n `f64_tdigest` Float64 STATISTICS(tdigest),\n `f32` Float32 STATISTICS(tdigest, uniq),\n `s` String,\n `a` Array(Float64)\n)\nENGINE = MergeTree\nORDER BY tuple()\nSETTINGS index_granularity = 8192 +CREATE TABLE default.tab\n(\n `f64` Float64 STATISTICS(tdigest, uniq),\n `f64_tdigest` Float64 STATISTICS(tdigest),\n `f32` Float32 STATISTICS(tdigest, uniq),\n `s` String,\n `a` Array(Float64)\n)\nENGINE = MergeTree\nORDER BY tuple()\nSETTINGS index_granularity = 8192 +CREATE TABLE default.tab\n(\n `f64` Float64 STATISTICS(tdigest, uniq),\n `f64_tdigest` Float64 STATISTICS(tdigest),\n `f32` Float32 STATISTICS(tdigest, uniq),\n `s` String,\n `a` Array(Float64)\n)\nENGINE = MergeTree\nORDER BY tuple()\nSETTINGS index_granularity = 8192 +CREATE TABLE default.tab\n(\n `f64` Float64 STATISTICS(tdigest, uniq),\n `f64_tdigest` Float64 STATISTICS(tdigest),\n `f32` Float32 STATISTICS(tdigest, uniq),\n `s` String,\n `a` Array(Float64)\n)\nENGINE = MergeTree\nORDER BY tuple()\nSETTINGS index_granularity = 8192 +CREATE TABLE default.tab\n(\n `f64` Float64,\n `f64_tdigest` Float64 STATISTICS(tdigest),\n `f32` Float32,\n `s` String,\n `a` Array(Float64)\n)\nENGINE = MergeTree\nORDER BY tuple()\nSETTINGS index_granularity = 8192 diff --git a/tests/queries/0_stateless/02864_statistics_ddl.sql b/tests/queries/0_stateless/02864_statistics_ddl.sql index fe612efe2ac..32b56a842b7 100644 --- a/tests/queries/0_stateless/02864_statistics_ddl.sql +++ b/tests/queries/0_stateless/02864_statistics_ddl.sql @@ -1,59 +1,195 @@ --- Tests that various DDL statements create/drop/materialize statistics +-- Tags: no-fasttest +-- no-fasttest: 'count_min' sketches need a 3rd party library + +-- Tests that DDL statements which create / drop / materialize statistics + +SET mutations_sync = 1; DROP TABLE IF EXISTS tab; +-- Error case: Can't create statistics when allow_experimental_statistics = 0 +CREATE TABLE tab (col Float64 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError INCORRECT_QUERY } + SET allow_experimental_statistics = 1; -SET allow_statistics_optimize = 1; + +-- Error case: Unknown statistics types are rejected +CREATE TABLE tab (col Float64 STATISTICS(no_statistics_type)) Engine = MergeTree() ORDER BY tuple(); -- { serverError INCORRECT_QUERY } + +-- Error case: The same statistics type can't exist more than once on a column +CREATE TABLE tab (col Float64 STATISTICS(tdigest, tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError INCORRECT_QUERY } + +SET allow_suspicious_low_cardinality_types = 1; + +-- Statistics can only be created on columns of specific data types (depending on the statistics kind), (*) + +-- tdigest requires data_type.isValueRepresentedByInteger +-- These types work: +CREATE TABLE tab (col UInt8 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col UInt256 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Float32 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Decimal32(3) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Date STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Date32 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col DateTime STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col DateTime64 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Enum('hello', 'world') STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col IPv4 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Nullable(UInt8) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col LowCardinality(UInt8) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col LowCardinality(Nullable(UInt8)) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +-- These types don't work: +CREATE TABLE tab (col String STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col FixedString(1) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Array(Float64) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Tuple(Float64, Float64) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Map(UInt64, UInt64) STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col UUID STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col IPv6 STATISTICS(tdigest)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } + +-- uniq requires data_type.isValueRepresentedByInteger +-- These types work: +CREATE TABLE tab (col UInt8 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col UInt256 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Float32 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Decimal32(3) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Date STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Date32 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col DateTime STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col DateTime64 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Enum('hello', 'world') STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col IPv4 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Nullable(UInt8) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col LowCardinality(UInt8) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col LowCardinality(Nullable(UInt8)) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +-- These types don't work: +CREATE TABLE tab (col String STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col FixedString(1) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Array(Float64) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Tuple(Float64, Float64) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Map(UInt64, UInt64) STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col UUID STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col IPv6 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } + +-- count_min requires data_type.isValueRepresentedByInteger or data_type = (Fixed)String +-- These types work: +CREATE TABLE tab (col UInt8 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col UInt256 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Float32 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Decimal32(3) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Date STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Date32 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col DateTime STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col DateTime64 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Enum('hello', 'world') STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col IPv4 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col Nullable(UInt8) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col LowCardinality(UInt8) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col LowCardinality(Nullable(UInt8)) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col String STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +CREATE TABLE tab (col FixedString(1) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab; +-- These types don't work: +CREATE TABLE tab (col Array(Float64) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Tuple(Float64, Float64) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col Map(UInt64, UInt64) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col UUID STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } +CREATE TABLE tab (col IPv6 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } + +-- CREATE TABLE was easy, ALTER is more fun CREATE TABLE tab ( - a Float64 STATISTICS(tdigest), - b Int64 STATISTICS(tdigest), - pk String, -) Engine = MergeTree() ORDER BY pk -SETTINGS min_bytes_for_wide_part = 0; + f64 Float64, + f64_tdigest Float64 STATISTICS(tdigest), + f32 Float32, + s String, + a Array(Float64) +) +Engine = MergeTree() +ORDER BY tuple(); +-- Error case: Unknown statistics types are rejected +-- (relevant for ADD and MODIFY) +ALTER TABLE tab ADD STATISTICS f64 TYPE no_statistics_type; -- { serverError INCORRECT_QUERY } +ALTER TABLE tab ADD STATISTICS IF NOT EXISTS f64 TYPE no_statistics_type; -- { serverError INCORRECT_QUERY } +ALTER TABLE tab MODIFY STATISTICS f64 TYPE no_statistics_type; -- { serverError INCORRECT_QUERY } +-- for some reason, ALTER TABLE tab MODIFY STATISTICS IF EXISTS is not supported + +-- Error case: The same statistics type can't exist more than once on a column +-- (relevant for ADD and MODIFY) +-- Create the same statistics object twice +ALTER TABLE tab ADD STATISTICS f64 TYPE tdigest, tdigest; -- { serverError INCORRECT_QUERY } +ALTER TABLE tab ADD STATISTICS IF NOT EXISTS f64 TYPE tdigest, tdigest; -- { serverError INCORRECT_QUERY } +ALTER TABLE tab MODIFY STATISTICS f64 TYPE tdigest, tdigest; -- { serverError INCORRECT_QUERY } +-- Create an statistics which exists already +ALTER TABLE tab ADD STATISTICS f64_tdigest TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab ADD STATISTICS IF NOT EXISTS f64_tdigest TYPE tdigest; -- no-op +ALTER TABLE tab MODIFY STATISTICS f64_tdigest TYPE tdigest; -- no-op + +-- Error case: Column does not exist +-- (relevant for ADD, MODIFY, DROP, CLEAR, and MATERIALIZE) +-- Note that the results are unfortunately quite inconsistent ... +ALTER TABLE tab ADD STATISTICS no_such_column TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab ADD STATISTICS IF NOT EXISTS no_such_column TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab MODIFY STATISTICS no_such_column TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab DROP STATISTICS no_such_column; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab DROP STATISTICS IF EXISTS no_such_column; -- no-op +ALTER TABLE tab CLEAR STATISTICS no_such_column; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab CLEAR STATISTICS IF EXISTS no_such_column; -- no-op +ALTER TABLE tab MATERIALIZE STATISTICS no_such_column; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab MATERIALIZE STATISTICS IF EXISTS no_such_column; -- { serverError ILLEGAL_STATISTICS } + +-- Error case: Column exists but has no statistics +-- (relevant for MODIFY, DROP, CLEAR, and MATERIALIZE) +-- Note that the results are unfortunately quite inconsistent ... +ALTER TABLE tab MODIFY STATISTICS s TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab DROP STATISTICS s; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab DROP STATISTICS IF EXISTS s; -- no-op +ALTER TABLE tab CLEAR STATISTICS s; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab CLEAR STATISTICS IF EXISTS s; -- no-op +ALTER TABLE tab MATERIALIZE STATISTICS s; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab MATERIALIZE STATISTICS IF EXISTS s; -- { serverError ILLEGAL_STATISTICS } + +-- We don't check systematically that that statistics can only be created via ALTER ADD STATISTICS on columns of specific data types (the +-- internal type validation code is tested already above, (*)). Only do a rudimentary check for each statistics type with a data type that +-- works and one that doesn't work. +-- tdigest +-- Works: +ALTER TABLE tab ADD STATISTICS f64 TYPE tdigest; ALTER TABLE tab DROP STATISTICS f64; +ALTER TABLE tab MODIFY STATISTICS f64 TYPE tdigest; ALTER TABLE tab DROP STATISTICS f64; +-- Doesn't work: +ALTER TABLE tab ADD STATISTICS a TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab MODIFY STATISTICS a TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } +-- uniq +-- Works: +ALTER TABLE tab ADD STATISTICS f64 TYPE uniq; ALTER TABLE tab DROP STATISTICS f64; +ALTER TABLE tab MODIFY STATISTICS f64 TYPE count_min; ALTER TABLE tab DROP STATISTICS f64; +-- Doesn't work: +ALTER TABLE tab ADD STATISTICS a TYPE uniq; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab MODIFY STATISTICS a TYPE uniq; -- { serverError ILLEGAL_STATISTICS } +-- count_min +-- Works: +ALTER TABLE tab ADD STATISTICS f64 TYPE count_min; ALTER TABLE tab DROP STATISTICS f64; +ALTER TABLE tab MODIFY STATISTICS f64 TYPE count_min; ALTER TABLE tab DROP STATISTICS f64; +-- Doesn't work: +ALTER TABLE tab ADD STATISTICS a TYPE count_min; -- { serverError ILLEGAL_STATISTICS } +ALTER TABLE tab MODIFY STATISTICS a TYPE count_min; -- { serverError ILLEGAL_STATISTICS } + +-- Any data type changes on columns with statistics are disallowed, for simplicity even if the new data type is compatible with all existing +-- statistics objects (e.g. tdigest can be created on Float64 and UInt64) +ALTER TABLE tab MODIFY COLUMN f64_tdigest UInt64; -- { serverError ALTER_OF_COLUMN_IS_FORBIDDEN } + +-- Finally, do a full-circle test of a good case. Print table definition after each step. +-- Intentionally specifying _two_ columns and _two_ statistics types to have that also tested. +SHOW CREATE TABLE tab; +ALTER TABLE tab ADD STATISTICS f64, f32 TYPE tdigest, uniq; +SHOW CREATE TABLE tab; +ALTER TABLE tab MODIFY STATISTICS f64, f32 TYPE tdigest, uniq; +SHOW CREATE TABLE tab; +ALTER TABLE tab CLEAR STATISTICS f64, f32; +SHOW CREATE TABLE tab; +ALTER TABLE tab MATERIALIZE STATISTICS f64, f32; +SHOW CREATE TABLE tab; +ALTER TABLE tab DROP STATISTICS f64, f32; SHOW CREATE TABLE tab; -INSERT INTO tab select number, -number, generateUUIDv4() FROM system.numbers LIMIT 10000; - -SELECT 'After insert'; -SELECT replaceRegexpAll(explain, '__table1\.|_UInt8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT count(*) FROM tab WHERE b < 10 and a < 10; -SELECT count(*) FROM tab WHERE b < NULL and a < '10'; - -ALTER TABLE tab DROP STATISTICS a, b; - -SELECT 'After drop statistic'; -SELECT replaceRegexpAll(explain, '__table1\.|_UInt8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT count(*) FROM tab WHERE b < 10 and a < 10; - -SHOW CREATE TABLE tab; - -ALTER TABLE tab ADD STATISTICS a, b TYPE tdigest; - -SELECT 'After add statistic'; - -SHOW CREATE TABLE tab; - -ALTER TABLE tab MATERIALIZE STATISTICS a, b; -INSERT INTO tab select number, -number, generateUUIDv4() FROM system.numbers LIMIT 10000; - -SELECT 'After materialize statistic'; -SELECT replaceRegexpAll(explain, '__table1\.|_UInt8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT count(*) FROM tab WHERE b < 10 and a < 10; - -OPTIMIZE TABLE tab FINAL; - -SELECT 'After merge'; -SELECT replaceRegexpAll(explain, '__table1\.|_UInt8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT count(*) FROM tab WHERE b < 10 and a < 10; - -ALTER TABLE tab RENAME COLUMN b TO c; -SHOW CREATE TABLE tab; - -SELECT 'After rename'; -SELECT replaceRegexpAll(explain, '__table1\.|_UInt8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE c < 10 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT count(*) FROM tab WHERE c < 10 and a < 10; - -DROP TABLE IF EXISTS tab; +DROP TABLE tab; diff --git a/tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.reference b/tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.reference new file mode 100644 index 00000000000..eb5e685597c --- /dev/null +++ b/tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.reference @@ -0,0 +1,12 @@ +After insert + Prewhere info + Prewhere filter + Prewhere filter column: and(less(b, 10_UInt8), less(a, 10_UInt8)) (removed) +After merge + Prewhere info + Prewhere filter + Prewhere filter column: and(less(a, 10_UInt8), less(b, 10_UInt8)) (removed) +After truncate, insert, and materialize + Prewhere info + Prewhere filter + Prewhere filter column: and(less(a, 10_UInt8), less(b, 10_UInt8)) (removed) diff --git a/tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.sql b/tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.sql new file mode 100644 index 00000000000..d469a4c2036 --- /dev/null +++ b/tests/queries/0_stateless/02864_statistics_delayed_materialization_in_merge.sql @@ -0,0 +1,36 @@ +-- Tests delayed materialization of statistics in merge instead of during insert (setting 'materialize_statistics_on_insert = 0'). +-- (The concrete statistics type, column data type and predicate type don't matter) + +-- Checks by the predicate evaluation order in EXPLAIN. This is quite fragile, a better approach would be helpful (maybe 'send_logs_level'?) + +DROP TABLE IF EXISTS tab; + +SET allow_experimental_statistics = 1; +SET allow_statistics_optimize = 1; +SET enable_analyzer = 1; + +SET materialize_statistics_on_insert = 0; + +CREATE TABLE tab +( + a Int64 STATISTICS(tdigest), + b Int16 STATISTICS(tdigest), +) ENGINE = MergeTree() ORDER BY tuple() +SETTINGS min_bytes_for_wide_part = 0, enable_vertical_merge_algorithm = 0; -- TODO: there is a bug in vertical merge with statistics. + +INSERT INTO tab SELECT number, -number FROM system.numbers LIMIT 10000; +SELECT 'After insert'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks b first, then a (statistics not used) + +OPTIMIZE TABLE tab FINAL; +SELECT 'After merge'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks a first, then b (statistics used) + +TRUNCATE TABLE tab; +SET mutations_sync = 2; +INSERT INTO tab SELECT number, -number FROM system.numbers LIMIT 10000; +ALTER TABLE tab MATERIALIZE STATISTICS a, b; +SELECT 'After truncate, insert, and materialize'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks a first, then b (statistics used) + +DROP TABLE tab; diff --git a/tests/queries/0_stateless/02864_statistics_exception.reference b/tests/queries/0_stateless/02864_statistics_exception.reference deleted file mode 100644 index e69de29bb2d..00000000000 diff --git a/tests/queries/0_stateless/02864_statistics_exception.sql b/tests/queries/0_stateless/02864_statistics_exception.sql deleted file mode 100644 index 289ffee6600..00000000000 --- a/tests/queries/0_stateless/02864_statistics_exception.sql +++ /dev/null @@ -1,55 +0,0 @@ --- Tests creating/dropping/materializing statistics produces the right exceptions. - -DROP TABLE IF EXISTS tab; - --- Can't create statistics when allow_experimental_statistics = 0 -CREATE TABLE tab -( - a Float64 STATISTICS(tdigest) -) Engine = MergeTree() ORDER BY tuple(); -- { serverError INCORRECT_QUERY } - -SET allow_experimental_statistics = 1; - --- The same type of statistics can't exist more than once on a column -CREATE TABLE tab -( - a Float64 STATISTICS(tdigest, tdigest) -) Engine = MergeTree() ORDER BY tuple(); -- { serverError INCORRECT_QUERY } - --- Unknown statistics types are rejected -CREATE TABLE tab -( - a Float64 STATISTICS(no_statistics_type) -) Engine = MergeTree() ORDER BY tuple(); -- { serverError INCORRECT_QUERY } - --- tDigest statistics can only be created on numeric columns -CREATE TABLE tab -( - a String STATISTICS(tdigest), -) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS } - -CREATE TABLE tab -( - a Float64, - b String -) Engine = MergeTree() ORDER BY tuple(); - -ALTER TABLE tab ADD STATISTICS a TYPE no_statistics_type; -- { serverError INCORRECT_QUERY } -ALTER TABLE tab ADD STATISTICS a TYPE tdigest; -ALTER TABLE tab ADD STATISTICS IF NOT EXISTS a TYPE tdigest; -ALTER TABLE tab ADD STATISTICS a TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } -ALTER TABLE tab MODIFY STATISTICS a TYPE tdigest; --- Statistics can be created only on integer columns -ALTER TABLE tab ADD STATISTICS b TYPE tdigest; -- { serverError ILLEGAL_STATISTICS } -ALTER TABLE tab DROP STATISTICS b; -- { serverError ILLEGAL_STATISTICS } -ALTER TABLE tab DROP STATISTICS a; -ALTER TABLE tab DROP STATISTICS IF EXISTS a; -ALTER TABLE tab CLEAR STATISTICS a; -- { serverError ILLEGAL_STATISTICS } -ALTER TABLE tab CLEAR STATISTICS IF EXISTS a; -ALTER TABLE tab MATERIALIZE STATISTICS b; -- { serverError ILLEGAL_STATISTICS } - -ALTER TABLE tab ADD STATISTICS a TYPE tdigest; -ALTER TABLE tab MODIFY COLUMN a Float64 TTL toDateTime(b) + INTERVAL 1 MONTH; -ALTER TABLE tab MODIFY COLUMN a Int64; -- { serverError ALTER_OF_COLUMN_IS_FORBIDDEN } - -DROP TABLE tab; diff --git a/tests/queries/0_stateless/02864_statistics_materialize_in_merge.reference b/tests/queries/0_stateless/02864_statistics_materialize_in_merge.reference deleted file mode 100644 index 5e969cf41cb..00000000000 --- a/tests/queries/0_stateless/02864_statistics_materialize_in_merge.reference +++ /dev/null @@ -1,10 +0,0 @@ -10 -10 -10 -statistics not used Condition less(b, 10_UInt8) moved to PREWHERE -statistics not used Condition less(a, 10_UInt8) moved to PREWHERE -statistics used after merge Condition less(a, 10_UInt8) moved to PREWHERE -statistics used after merge Condition less(b, 10_UInt8) moved to PREWHERE -statistics used after materialize Condition less(a, 10_UInt8) moved to PREWHERE -statistics used after materialize Condition less(b, 10_UInt8) moved to PREWHERE -2 0 diff --git a/tests/queries/0_stateless/02864_statistics_materialize_in_merge.sql b/tests/queries/0_stateless/02864_statistics_materialize_in_merge.sql deleted file mode 100644 index 6606cff263f..00000000000 --- a/tests/queries/0_stateless/02864_statistics_materialize_in_merge.sql +++ /dev/null @@ -1,52 +0,0 @@ --- Tests delayed materialization of statistics in merge instead of during insert (setting 'materialize_statistics_on_insert = 0'). - -DROP TABLE IF EXISTS tab; - -SET enable_analyzer = 1; -SET allow_experimental_statistics = 1; -SET allow_statistics_optimize = 1; - -SET materialize_statistics_on_insert = 0; - -CREATE TABLE tab -( - a Int64 STATISTICS(tdigest), - b Int16 STATISTICS(tdigest), -) ENGINE = MergeTree() ORDER BY tuple() -SETTINGS min_bytes_for_wide_part = 0, enable_vertical_merge_algorithm = 0; -- TODO: there is a bug in vertical merge with statistics. - -INSERT INTO tab SELECT number, -number FROM system.numbers LIMIT 10000; - -SELECT count(*) FROM tab WHERE b < 10 and a < 10 SETTINGS log_comment = 'statistics not used'; - -OPTIMIZE TABLE tab FINAL; - -SELECT count(*) FROM tab WHERE b < 10 and a < 10 SETTINGS log_comment = 'statistics used after merge'; - -TRUNCATE TABLE tab; -SET mutations_sync = 2; - -INSERT INTO tab SELECT number, -number FROM system.numbers LIMIT 10000; -ALTER TABLE tab MATERIALIZE STATISTICS a, b; - -SELECT count(*) FROM tab WHERE b < 10 and a < 10 SETTINGS log_comment = 'statistics used after materialize'; - -DROP TABLE tab; - -SYSTEM FLUSH LOGS; - -SELECT log_comment, message FROM system.text_log JOIN -( - SELECT Settings['log_comment'] AS log_comment, query_id FROM system.query_log - WHERE current_database = currentDatabase() - AND query LIKE 'SELECT count(*) FROM tab%' - AND type = 'QueryFinish' -) AS query_log USING (query_id) -WHERE message LIKE '%moved to PREWHERE%' -ORDER BY event_time_microseconds; - -SELECT count(), sum(ProfileEvents['MergeTreeDataWriterStatisticsCalculationMicroseconds']) -FROM system.query_log -WHERE current_database = currentDatabase() - AND query LIKE 'INSERT INTO tab SELECT%' - AND type = 'QueryFinish'; diff --git a/tests/queries/0_stateless/02864_statistics_predicates.reference b/tests/queries/0_stateless/02864_statistics_predicates.reference new file mode 100644 index 00000000000..ffbd7269e05 --- /dev/null +++ b/tests/queries/0_stateless/02864_statistics_predicates.reference @@ -0,0 +1,98 @@ +u64 and = +10 +10 +10 +10 +0 +0 +0 +0 +10 +10 +10 +10 +u64 and < +70 +70 +70 +70 +80 +80 +80 +80 +70 +70 +70 +70 +f64 and = +10 +10 +10 +10 +0 +0 +0 +0 +10 +10 +10 +10 +0 +0 +0 +0 +f64 and < +70 +70 +70 +70 +80 +80 +80 +80 +70 +70 +70 +70 +80 +80 +80 +80 +dt and = +0 +0 +0 +0 +10 +10 +10 +10 +dt and < +10000 +10000 +10000 +10000 +70 +70 +70 +70 +b and = +5000 +5000 +5000 +5000 +5000 +5000 +5000 +5000 +5000 +5000 +5000 +5000 +0 +0 +0 +0 +s and = +10 +10 diff --git a/tests/queries/0_stateless/02864_statistics_predicates.sql b/tests/queries/0_stateless/02864_statistics_predicates.sql new file mode 100644 index 00000000000..779116cf19a --- /dev/null +++ b/tests/queries/0_stateless/02864_statistics_predicates.sql @@ -0,0 +1,214 @@ +-- Tags: no-fasttest +-- no-fasttest: 'count_min' sketches need a 3rd party library + +-- Tests the cross product of all predicates with all right-hand sides on all data types and all statistics types. + +SET allow_experimental_statistics = 1; +SET allow_statistics_optimize = 1; + +DROP TABLE IF EXISTS tab; + +CREATE TABLE tab +( + u64 UInt64, + u64_tdigest UInt64 STATISTICS(tdigest), + u64_count_min UInt64 STATISTICS(count_min), + u64_uniq UInt64 STATISTICS(uniq), + f64 Float64, + f64_tdigest Float64 STATISTICS(tdigest), + f64_count_min Float64 STATISTICS(count_min), + f64_uniq Float64 STATISTICS(uniq), + dt DateTime, + dt_tdigest DateTime STATISTICS(tdigest), + dt_count_min DateTime STATISTICS(count_min), + dt_uniq DateTime STATISTICS(uniq), + b Bool, + b_tdigest Bool STATISTICS(tdigest), + b_count_min Bool STATISTICS(count_min), + b_uniq Bool STATISTICS(uniq), + s String, + -- s_tdigest String STATISTICS(tdigest), -- not supported by tdigest + s_count_min String STATISTICS(count_min) + -- s_uniq String STATISTICS(uniq), -- not supported by uniq +) Engine = MergeTree() ORDER BY tuple() +SETTINGS min_bytes_for_wide_part = 0; + +INSERT INTO tab +-- SELECT number % 10000, number % 1000, -(number % 100) FROM system.numbers LIMIT 10000; +SELECT number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 1000, + number % 2, + number % 2, + number % 2, + number % 2, + toString(number % 1000), + toString(number % 1000) +FROM system.numbers LIMIT 10000; + +-- u64 ---------------------------------------------------- + +SELECT 'u64 and ='; + +SELECT count(*) FROM tab WHERE u64 = 7; +SELECT count(*) FROM tab WHERE u64_tdigest = 7; +SELECT count(*) FROM tab WHERE u64_count_min = 7; +SELECT count(*) FROM tab WHERE u64_uniq = 7; + +SELECT count(*) FROM tab WHERE u64 = 7.7; +SELECT count(*) FROM tab WHERE u64_tdigest = 7.7; +SELECT count(*) FROM tab WHERE u64_count_min = 7.7; +SELECT count(*) FROM tab WHERE u64_uniq = 7.7; + +SELECT count(*) FROM tab WHERE u64 = '7'; +SELECT count(*) FROM tab WHERE u64_tdigest = '7'; +SELECT count(*) FROM tab WHERE u64_count_min = '7'; +SELECT count(*) FROM tab WHERE u64_uniq = '7'; + +SELECT count(*) FROM tab WHERE u64 = '7.7'; -- { serverError TYPE_MISMATCH } +SELECT count(*) FROM tab WHERE u64_tdigest = '7.7'; -- { serverError TYPE_MISMATCH } +SELECT count(*) FROM tab WHERE u64_count_min = '7.7'; -- { serverError TYPE_MISMATCH } +SELECT count(*) FROM tab WHERE u64_uniq = '7.7'; -- { serverError TYPE_MISMATCH } + +SELECT 'u64 and <'; + +SELECT count(*) FROM tab WHERE u64 < 7; +SELECT count(*) FROM tab WHERE u64_tdigest < 7; +SELECT count(*) FROM tab WHERE u64_count_min < 7; +SELECT count(*) FROM tab WHERE u64_uniq < 7; + +SELECT count(*) FROM tab WHERE u64 < 7.7; +SELECT count(*) FROM tab WHERE u64_tdigest < 7.7; +SELECT count(*) FROM tab WHERE u64_count_min < 7.7; +SELECT count(*) FROM tab WHERE u64_uniq < 7.7; + +SELECT count(*) FROM tab WHERE u64 < '7'; +SELECT count(*) FROM tab WHERE u64_tdigest < '7'; +SELECT count(*) FROM tab WHERE u64_count_min < '7'; +SELECT count(*) FROM tab WHERE u64_uniq < '7'; + +SELECT count(*) FROM tab WHERE u64 < '7.7'; -- { serverError TYPE_MISMATCH } +SELECT count(*) FROM tab WHERE u64_tdigest < '7.7'; -- { serverError TYPE_MISMATCH } +SELECT count(*) FROM tab WHERE u64_count_min < '7.7'; -- { serverError TYPE_MISMATCH } +SELECT count(*) FROM tab WHERE u64_uniq < '7.7'; -- { serverError TYPE_MISMATCH } + +-- f64 ---------------------------------------------------- + +SELECT 'f64 and ='; + +SELECT count(*) FROM tab WHERE f64 = 7; +SELECT count(*) FROM tab WHERE f64_tdigest = 7; +SELECT count(*) FROM tab WHERE f64_count_min = 7; +SELECT count(*) FROM tab WHERE f64_uniq = 7; + +SELECT count(*) FROM tab WHERE f64 = 7.7; +SELECT count(*) FROM tab WHERE f64_tdigest = 7.7; +SELECT count(*) FROM tab WHERE f64_count_min = 7.7; +SELECT count(*) FROM tab WHERE f64_uniq = 7.7; + +SELECT count(*) FROM tab WHERE f64 = '7'; +SELECT count(*) FROM tab WHERE f64_tdigest = '7'; +SELECT count(*) FROM tab WHERE f64_count_min = '7'; +SELECT count(*) FROM tab WHERE f64_uniq = '7'; + +SELECT count(*) FROM tab WHERE f64 = '7.7'; +SELECT count(*) FROM tab WHERE f64_tdigest = '7.7'; +SELECT count(*) FROM tab WHERE f64_count_min = '7.7'; +SELECT count(*) FROM tab WHERE f64_uniq = '7.7'; + +SELECT 'f64 and <'; + +SELECT count(*) FROM tab WHERE f64 < 7; +SELECT count(*) FROM tab WHERE f64_tdigest < 7; +SELECT count(*) FROM tab WHERE f64_count_min < 7; +SELECT count(*) FROM tab WHERE f64_uniq < 7; + +SELECT count(*) FROM tab WHERE f64 < 7.7; +SELECT count(*) FROM tab WHERE f64_tdigest < 7.7; +SELECT count(*) FROM tab WHERE f64_count_min < 7.7; +SELECT count(*) FROM tab WHERE f64_uniq < 7.7; + +SELECT count(*) FROM tab WHERE f64 < '7'; +SELECT count(*) FROM tab WHERE f64_tdigest < '7'; +SELECT count(*) FROM tab WHERE f64_count_min < '7'; +SELECT count(*) FROM tab WHERE f64_uniq < '7'; + +SELECT count(*) FROM tab WHERE f64 < '7.7'; +SELECT count(*) FROM tab WHERE f64_tdigest < '7.7'; +SELECT count(*) FROM tab WHERE f64_count_min < '7.7'; +SELECT count(*) FROM tab WHERE f64_uniq < '7.7'; + +-- dt ---------------------------------------------------- + +SELECT 'dt and ='; + +SELECT count(*) FROM tab WHERE dt = '2024-08-08 11:12:13'; +SELECT count(*) FROM tab WHERE dt_tdigest = '2024-08-08 11:12:13'; +SELECT count(*) FROM tab WHERE dt_count_min = '2024-08-08 11:12:13'; +SELECT count(*) FROM tab WHERE dt_uniq = '2024-08-08 11:12:13'; + +SELECT count(*) FROM tab WHERE dt = 7; +SELECT count(*) FROM tab WHERE dt_tdigest = 7; +SELECT count(*) FROM tab WHERE dt_count_min = 7; +SELECT count(*) FROM tab WHERE dt_uniq = 7; + +SELECT 'dt and <'; + +SELECT count(*) FROM tab WHERE dt < '2024-08-08 11:12:13'; +SELECT count(*) FROM tab WHERE dt_tdigest < '2024-08-08 11:12:13'; +SELECT count(*) FROM tab WHERE dt_count_min < '2024-08-08 11:12:13'; +SELECT count(*) FROM tab WHERE dt_uniq < '2024-08-08 11:12:13'; + +SELECT count(*) FROM tab WHERE dt < 7; +SELECT count(*) FROM tab WHERE dt_tdigest < 7; +SELECT count(*) FROM tab WHERE dt_count_min < 7; +SELECT count(*) FROM tab WHERE dt_uniq < 7; + +-- b ---------------------------------------------------- + +SELECT 'b and ='; + +SELECT count(*) FROM tab WHERE b = true; +SELECT count(*) FROM tab WHERE b_tdigest = true; +SELECT count(*) FROM tab WHERE b_count_min = true; +SELECT count(*) FROM tab WHERE b_uniq = true; + +SELECT count(*) FROM tab WHERE b = 'true'; +SELECT count(*) FROM tab WHERE b_tdigest = 'true'; +SELECT count(*) FROM tab WHERE b_count_min = 'true'; +SELECT count(*) FROM tab WHERE b_uniq = 'true'; + +SELECT count(*) FROM tab WHERE b = 1; +SELECT count(*) FROM tab WHERE b_tdigest = 1; +SELECT count(*) FROM tab WHERE b_count_min = 1; +SELECT count(*) FROM tab WHERE b_uniq = 1; + +SELECT count(*) FROM tab WHERE b = 1.1; +SELECT count(*) FROM tab WHERE b_tdigest = 1.1; +SELECT count(*) FROM tab WHERE b_count_min = 1.1; +SELECT count(*) FROM tab WHERE b_uniq = 1.1; + +-- s ---------------------------------------------------- + +SELECT 's and ='; + +SELECT count(*) FROM tab WHERE s = 7; -- { serverError NO_COMMON_TYPE } +-- SELECT count(*) FROM tab WHERE s_tdigest = 7; -- not supported +SELECT count(*) FROM tab WHERE s_count_min = 7; -- { serverError NO_COMMON_TYPE } +-- SELECT count(*) FROM tab WHERE s_uniq = 7; -- not supported + +SELECT count(*) FROM tab WHERE s = '7'; +-- SELECT count(*) FROM tab WHERE s_tdigest = '7'; -- not supported +SELECT count(*) FROM tab WHERE s_count_min = '7'; +-- SELECT count(*) FROM tab WHERE s_uniq = '7'; -- not supported + +DROP TABLE tab; diff --git a/tests/queries/0_stateless/02864_statistics_uniq.reference b/tests/queries/0_stateless/02864_statistics_uniq.reference deleted file mode 100644 index 77786dbdd8c..00000000000 --- a/tests/queries/0_stateless/02864_statistics_uniq.reference +++ /dev/null @@ -1,35 +0,0 @@ -CREATE TABLE default.t1\n(\n `a` Float64 STATISTICS(tdigest),\n `b` Int64 STATISTICS(tdigest),\n `c` Int64 STATISTICS(tdigest, uniq),\n `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS min_bytes_for_wide_part = 0, index_granularity = 8192 -After insert - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), equals(c, 0), less(b, 10)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(equals(c, 11), less(a, 10), less(b, 10)) (removed) -After merge - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), equals(c, 0), less(b, 10)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(equals(c, 11), less(a, 10), less(b, 10)) (removed) -After modify TDigest - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), equals(c, 11), less(b, 10)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), equals(c, 0), less(b, 10)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(less(c, -1), less(a, 10), less(b, 10)) (removed) -After drop - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), equals(c, 11), less(b, 10)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), equals(c, 0), less(b, 10)) (removed) - Prewhere info - Prewhere filter - Prewhere filter column: and(less(a, 10), less(c, -1), less(b, 10)) (removed) diff --git a/tests/queries/0_stateless/02864_statistics_uniq.sql b/tests/queries/0_stateless/02864_statistics_uniq.sql deleted file mode 100644 index 0f5f353c045..00000000000 --- a/tests/queries/0_stateless/02864_statistics_uniq.sql +++ /dev/null @@ -1,73 +0,0 @@ -DROP TABLE IF EXISTS t1; - -SET allow_experimental_statistics = 1; -SET allow_statistics_optimize = 1; -SET mutations_sync = 1; - -CREATE TABLE t1 -( - a Float64 STATISTICS(tdigest), - b Int64 STATISTICS(tdigest), - c Int64 STATISTICS(tdigest, uniq), - pk String, -) Engine = MergeTree() ORDER BY pk -SETTINGS min_bytes_for_wide_part = 0; - -SHOW CREATE TABLE t1; - -INSERT INTO t1 select number, -number, number/1000, generateUUIDv4() FROM system.numbers LIMIT 10000; -INSERT INTO t1 select 0, 0, 11, generateUUIDv4(); - -SELECT 'After insert'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 0 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 11 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -OPTIMIZE TABLE t1 FINAL; - -SELECT 'After merge'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 0 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 11 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; - -SELECT 'After modify TDigest'; -ALTER TABLE t1 MODIFY STATISTICS c TYPE TDigest; -ALTER TABLE t1 MATERIALIZE STATISTICS c; - -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 11 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 0 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c < -1 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; - - -ALTER TABLE t1 DROP STATISTICS c; - -SELECT 'After drop'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 11 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c = 0 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; -SELECT replaceRegexpAll(explain, '__table1.|_UInt8|_Int8', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM t1 WHERE b < 10 and c < -1 and a < 10) WHERE explain LIKE '%Prewhere%' OR explain LIKE '%Filter column%'; - -DROP TABLE IF EXISTS t1; -DROP TABLE IF EXISTS t2; -SET allow_suspicious_low_cardinality_types=1; -CREATE TABLE t2 -( - a Float64 STATISTICS(tdigest), - b Int64 STATISTICS(tdigest), - c LowCardinality(Int64) STATISTICS(tdigest, uniq), - pk String, -) Engine = MergeTree() ORDER BY pk -SETTINGS min_bytes_for_wide_part = 0; -INSERT INTO t2 select number, -number, number/1000, generateUUIDv4() FROM system.numbers LIMIT 10000; - -DROP TABLE IF EXISTS t2; -DROP TABLE IF EXISTS t3; - -CREATE TABLE t3 -( - a Float64 STATISTICS(tdigest), - b Int64 STATISTICS(tdigest), - c Nullable(Int64) STATISTICS(tdigest, uniq), - pk String, -) Engine = MergeTree() ORDER BY pk -SETTINGS min_bytes_for_wide_part = 0; -INSERT INTO t3 select number, -number, number/1000, generateUUIDv4() FROM system.numbers LIMIT 10000; - -DROP TABLE IF EXISTS t3; - diff --git a/tests/queries/0_stateless/02864_statistics_usage.reference b/tests/queries/0_stateless/02864_statistics_usage.reference new file mode 100644 index 00000000000..a9f669b88c1 --- /dev/null +++ b/tests/queries/0_stateless/02864_statistics_usage.reference @@ -0,0 +1,20 @@ +After insert + Prewhere info + Prewhere filter + Prewhere filter column: and(less(a, 10_UInt8), less(b, 10_UInt8)) (removed) +After drop statistic + Prewhere info + Prewhere filter + Prewhere filter column: and(less(b, 10_UInt8), less(a, 10_UInt8)) (removed) +After add and materialize statistic + Prewhere info + Prewhere filter + Prewhere filter column: and(less(a, 10_UInt8), less(b, 10_UInt8)) (removed) +After merge + Prewhere info + Prewhere filter + Prewhere filter column: and(less(a, 10_UInt8), less(b, 10_UInt8)) (removed) +After rename + Prewhere info + Prewhere filter + Prewhere filter column: and(less(a, 10_UInt8), less(c, 10_UInt8)) (removed) diff --git a/tests/queries/0_stateless/02864_statistics_usage.sql b/tests/queries/0_stateless/02864_statistics_usage.sql new file mode 100644 index 00000000000..4956bd27e87 --- /dev/null +++ b/tests/queries/0_stateless/02864_statistics_usage.sql @@ -0,0 +1,42 @@ +-- Test that the optimizer picks up column statistics +-- (The concrete statistics type, column data type and predicate type don't matter) + +-- Checks by the predicate evaluation order in EXPLAIN. This is quite fragile, a better approach would be helpful (maybe 'send_logs_level'?) + +SET allow_experimental_statistics = 1; +SET allow_statistics_optimize = 1; +SET mutations_sync = 1; +SET enable_analyzer = 1; + +DROP TABLE IF EXISTS tab; + +CREATE TABLE tab +( + a Float64 STATISTICS(tdigest), + b Int64 STATISTICS(tdigest) +) Engine = MergeTree() ORDER BY tuple() +SETTINGS min_bytes_for_wide_part = 0; + +INSERT INTO tab select number, -number FROM system.numbers LIMIT 10000; +SELECT 'After insert'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks a first, then b (statistics used) + +ALTER TABLE tab DROP STATISTICS a, b; +SELECT 'After drop statistic'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks b first, then a (statistics not used) + +ALTER TABLE tab ADD STATISTICS a, b TYPE tdigest; +ALTER TABLE tab MATERIALIZE STATISTICS a, b; +INSERT INTO tab select number, -number FROM system.numbers LIMIT 10000; +SELECT 'After add and materialize statistic'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks a first, then b (statistics used) + +OPTIMIZE TABLE tab FINAL; +SELECT 'After merge'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE b < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks a first, then b (statistics used) + +ALTER TABLE tab RENAME COLUMN b TO c; +SELECT 'After rename'; +SELECT replaceRegexpAll(explain, '__table1\.', '') FROM (EXPLAIN actions=1 SELECT count(*) FROM tab WHERE c < 10 and a < 10) WHERE explain LIKE '%Prewhere%'; -- checks a first, then c (statistics used) + +DROP TABLE IF EXISTS tab;