diff --git a/.github/workflows/backport_branches.yml b/.github/workflows/backport_branches.yml index 6b05f1fe9f4..51670087ffe 100644 --- a/.github/workflows/backport_branches.yml +++ b/.github/workflows/backport_branches.yml @@ -11,7 +11,7 @@ on: # yamllint disable-line rule:truthy - 'backport/**' jobs: RunConfig: - runs-on: [self-hosted, style-checker] + runs-on: [self-hosted, style-checker-aarch64] outputs: data: ${{ steps.runconfig.outputs.CI_DATA }} steps: diff --git a/.github/workflows/master.yml b/.github/workflows/master.yml index 24daca44da6..7cb5455ed73 100644 --- a/.github/workflows/master.yml +++ b/.github/workflows/master.yml @@ -11,7 +11,7 @@ on: # yamllint disable-line rule:truthy - 'master' jobs: RunConfig: - runs-on: [self-hosted, style-checker] + runs-on: [self-hosted, style-checker-aarch64] outputs: data: ${{ steps.runconfig.outputs.CI_DATA }} steps: diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 770e1ec3789..93ac2be19b4 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -14,7 +14,7 @@ jobs: # The task for having a preserved ENV and event.json for later investigation uses: ./.github/workflows/debug.yml RunConfig: - runs-on: [self-hosted, style-checker] + runs-on: [self-hosted, style-checker-aarch64] outputs: data: ${{ steps.runconfig.outputs.CI_DATA }} steps: diff --git a/.github/workflows/pull_request.yml b/.github/workflows/pull_request.yml index c9cf5ab90dd..1afcdab938b 100644 --- a/.github/workflows/pull_request.yml +++ b/.github/workflows/pull_request.yml @@ -18,7 +18,7 @@ on: # yamllint disable-line rule:truthy ########################################################################################## jobs: RunConfig: - runs-on: [self-hosted, style-checker] + runs-on: [self-hosted, style-checker-aarch64] outputs: data: ${{ steps.runconfig.outputs.CI_DATA }} steps: diff --git a/.github/workflows/release_branches.yml b/.github/workflows/release_branches.yml index c076c2209ec..57e90d79ebd 100644 --- a/.github/workflows/release_branches.yml +++ b/.github/workflows/release_branches.yml @@ -14,7 +14,7 @@ on: # yamllint disable-line rule:truthy jobs: RunConfig: - runs-on: [self-hosted, style-checker] + runs-on: [self-hosted, style-checker-aarch64] outputs: data: ${{ steps.runconfig.outputs.CI_DATA }} steps: diff --git a/base/base/Decimal_fwd.h b/base/base/Decimal_fwd.h index 589d6224917..beb228cea3c 100644 --- a/base/base/Decimal_fwd.h +++ b/base/base/Decimal_fwd.h @@ -1,6 +1,7 @@ #pragma once #include +#include namespace wide { @@ -44,3 +45,8 @@ concept is_over_big_int = || std::is_same_v || std::is_same_v; } + +template <> struct is_signed { static constexpr bool value = true; }; +template <> struct is_signed { static constexpr bool value = true; }; +template <> struct is_signed { static constexpr bool value = true; }; +template <> struct is_signed { static constexpr bool value = true; }; diff --git a/contrib/NuRaft b/contrib/NuRaft index 1278e32bb0d..5bb3a0e8257 160000 --- a/contrib/NuRaft +++ b/contrib/NuRaft @@ -1 +1 @@ -Subproject commit 1278e32bb0d5dc489f947e002bdf8c71b0ddaa63 +Subproject commit 5bb3a0e8257bacd65b099cb1b7239bd6b9a2c477 diff --git a/docker/test/integration/runner/compose/docker_compose_mysql_cluster.yml b/docker/test/integration/runner/compose/docker_compose_mysql_cluster.yml index 73f9e39f0d6..079c451b9d6 100644 --- a/docker/test/integration/runner/compose/docker_compose_mysql_cluster.yml +++ b/docker/test/integration/runner/compose/docker_compose_mysql_cluster.yml @@ -1,7 +1,7 @@ version: '2.3' services: mysql2: - image: mysql:5.7 + image: mysql:8.0 restart: always environment: MYSQL_ROOT_PASSWORD: clickhouse @@ -23,7 +23,7 @@ services: source: ${MYSQL_CLUSTER_LOGS:-} target: /mysql/ mysql3: - image: mysql:5.7 + image: mysql:8.0 restart: always environment: MYSQL_ROOT_PASSWORD: clickhouse @@ -45,7 +45,7 @@ services: source: ${MYSQL_CLUSTER_LOGS:-} target: /mysql/ mysql4: - image: mysql:5.7 + image: mysql:8.0 restart: always environment: MYSQL_ROOT_PASSWORD: clickhouse diff --git a/docker/test/upgrade/run.sh b/docker/test/upgrade/run.sh index aaba5cc6a8c..1aecc7331cd 100644 --- a/docker/test/upgrade/run.sh +++ b/docker/test/upgrade/run.sh @@ -77,6 +77,12 @@ remove_keeper_config "async_replication" "1" # create_if_not_exists feature flag doesn't exist on some older versions remove_keeper_config "create_if_not_exists" "[01]" +# latest_logs_cache_size_threshold setting doesn't exist on some older versions +remove_keeper_config "latest_logs_cache_size_threshold" "[[:digit:]]\+" + +# commit_logs_cache_size_threshold setting doesn't exist on some older versions +remove_keeper_config "commit_logs_cache_size_threshold" "[[:digit:]]\+" + # it contains some new settings, but we can safely remove it rm /etc/clickhouse-server/config.d/merge_tree.xml rm /etc/clickhouse-server/config.d/enable_wait_for_shutdown_replicated_tables.xml @@ -109,6 +115,12 @@ remove_keeper_config "async_replication" "1" # create_if_not_exists feature flag doesn't exist on some older versions remove_keeper_config "create_if_not_exists" "[01]" +# latest_logs_cache_size_threshold setting doesn't exist on some older versions +remove_keeper_config "latest_logs_cache_size_threshold" "[[:digit:]]\+" + +# commit_logs_cache_size_threshold setting doesn't exist on some older versions +remove_keeper_config "commit_logs_cache_size_threshold" "[[:digit:]]\+" + # But we still need default disk because some tables loaded only into it sudo cat /etc/clickhouse-server/config.d/s3_storage_policy_by_default.xml \ | sed "s|
s3
|
s3
default|" \ diff --git a/docs/en/operations/configuration-files.md b/docs/en/operations/configuration-files.md index 81b25a4e897..9f17f4af1e8 100644 --- a/docs/en/operations/configuration-files.md +++ b/docs/en/operations/configuration-files.md @@ -10,11 +10,62 @@ The ClickHouse server can be configured with configuration files in XML or YAML It is possible to mix XML and YAML configuration files, for example you could have a main configuration file `config.xml` and additional configuration files `config.d/network.xml`, `config.d/timezone.yaml` and `config.d/keeper.yaml`. Mixing XML and YAML within a single configuration file is not supported. XML configuration files should use `...` as top-level tag. In YAML configuration files, `clickhouse:` is optional, the parser inserts it implicitly if absent. -## Overriding Configuration {#override} +## Merging Configuration {#merging} -The merge of configuration files behaves as one intuitively expects: The contents of both files are combined recursively, children with the same name are replaced by the element of the more specific configuration file. The merge can be customized using attributes `replace` and `remove`. -- Attribute `replace` means that the element is replaced by the specified one. -- Attribute `remove` means that the element is deleted. +Two configuration files (usually the main configuration file and another configuration files from `config.d/`) are merged as follows: + +- If a node (i.e. a path leading to an element) appears in both files and does not have attributes `replace` or `remove`, it is included in the merged configuration file and children from both nodes are included and merged recursively. +- If one of both nodes contains attribute `replace`, it is included in the merged configuration file but only children from the node with attribute `replace` are included. +- If one of both nodes contains attribute `remove`, the node is not included in the merged configuration file (if it exists already, it is deleted). + +Example: + + +```xml + + + + 1 + + + 2 + + + 3 + + +``` + +and + +```xml + + + + 4 + + + 5 + + + 6 + + +``` + +generates merged configuration file: + +```xml + + + 1 + 4 + + + 5 + + +``` To specify that a value of an element should be replaced by the value of an environment variable, you can use attribute `from_env`. @@ -125,7 +176,7 @@ Users configuration can be split into separate files similar to `config.xml` and Directory name is defined as `users_config` setting without `.xml` postfix concatenated with `.d`. Directory `users.d` is used by default, as `users_config` defaults to `users.xml`. -Note that configuration files are first merged taking into account [Override](#override) settings and includes are processed after that. +Note that configuration files are first [merged](#merging) taking into account settings, and includes are processed after that. ## XML example {#example} diff --git a/docs/en/operations/settings/settings.md b/docs/en/operations/settings/settings.md index 1bdec81ae88..b11a04e10ec 100644 --- a/docs/en/operations/settings/settings.md +++ b/docs/en/operations/settings/settings.md @@ -4279,41 +4279,6 @@ Result: └─────┴─────┴───────┘ ``` -## enable_order_by_all {#enable-order-by-all} - -Enables or disables sorting by `ALL` columns, i.e. [ORDER BY](../../sql-reference/statements/select/order-by.md) - -Possible values: - -- 0 — Disable ORDER BY ALL. -- 1 — Enable ORDER BY ALL. - -Default value: `1`. - -**Example** - -Query: - -```sql -CREATE TABLE TAB(C1 Int, C2 Int, ALL Int) ENGINE=Memory(); - -INSERT INTO TAB VALUES (10, 20, 30), (20, 20, 10), (30, 10, 20); - -SELECT * FROM TAB ORDER BY ALL; -- returns an error that ALL is ambiguous - -SELECT * FROM TAB ORDER BY ALL SETTINGS enable_order_by_all; -``` - -Result: - -```text -┌─C1─┬─C2─┬─ALL─┐ -│ 20 │ 20 │ 10 │ -│ 30 │ 10 │ 20 │ -│ 10 │ 20 │ 30 │ -└────┴────┴─────┘ -``` - ## splitby_max_substrings_includes_remaining_string {#splitby_max_substrings_includes_remaining_string} Controls whether function [splitBy*()](../../sql-reference/functions/splitting-merging-functions.md) with argument `max_substrings` > 0 will include the remaining string in the last element of the result array. diff --git a/docs/en/sql-reference/aggregate-functions/reference/simplelinearregression.md b/docs/en/sql-reference/aggregate-functions/reference/simplelinearregression.md index bcff05ada47..ea3dbff8691 100644 --- a/docs/en/sql-reference/aggregate-functions/reference/simplelinearregression.md +++ b/docs/en/sql-reference/aggregate-functions/reference/simplelinearregression.md @@ -13,8 +13,8 @@ simpleLinearRegression(x, y) Parameters: -- `x` — Column with dependent variable values. -- `y` — Column with explanatory variable values. +- `x` — Column with explanatory variable values. +- `y` — Column with dependent variable values. Returned values: diff --git a/docs/en/sql-reference/functions/distance-functions.md b/docs/en/sql-reference/functions/distance-functions.md index 1774c22014d..e20c35c6b6f 100644 --- a/docs/en/sql-reference/functions/distance-functions.md +++ b/docs/en/sql-reference/functions/distance-functions.md @@ -509,7 +509,7 @@ Result: ## cosineDistance -Calculates the cosine distance between two vectors (the values of the tuples are the coordinates). The less the returned value is, the more similar are the vectors. +Calculates the cosine distance between two vectors (the values of the tuples are the coordinates). The smaller the returned value is, the more similar are the vectors. **Syntax** diff --git a/docs/en/sql-reference/functions/tuple-functions.md b/docs/en/sql-reference/functions/tuple-functions.md index 5930239dc56..b089de67e98 100644 --- a/docs/en/sql-reference/functions/tuple-functions.md +++ b/docs/en/sql-reference/functions/tuple-functions.md @@ -542,7 +542,7 @@ Alias: `scalarProduct`. - Scalar product. -Type: [Int/UInt](../../sql-reference/data-types/int-uint.md), [Float](../../sql-reference/data-types/float.md) or [Decimal](../../sql-reference/data-types/decimal.md). +Type: [Int/UInt](../../sql-reference/data-types/int-uint.md) or [Float](../../sql-reference/data-types/float.md). **Example** diff --git a/docs/en/sql-reference/statements/select/order-by.md b/docs/en/sql-reference/statements/select/order-by.md index d6432a7b4f8..29aca70762e 100644 --- a/docs/en/sql-reference/statements/select/order-by.md +++ b/docs/en/sql-reference/statements/select/order-by.md @@ -9,10 +9,9 @@ The `ORDER BY` clause contains - a list of expressions, e.g. `ORDER BY visits, search_phrase`, - a list of numbers referring to columns in the `SELECT` clause, e.g. `ORDER BY 2, 1`, or -- `ALL` which means all columns of the `SELECT` clause, e.g. `ORDER BY ALL`. +- `*` (without other expressions or numbers) which means all columns of the `SELECT` clause: `ORDER BY *`. To disable sorting by column numbers, set setting [enable_positional_arguments](../../../operations/settings/settings.md#enable-positional-arguments) = 0. -To disable sorting by `ALL`, set setting [enable_order_by_all](../../../operations/settings/settings.md#enable-order-by-all) = 0. The `ORDER BY` clause can be attributed by a `DESC` (descending) or `ASC` (ascending) modifier which determines the sorting direction. Unless an explicit sort order is specified, `ASC` is used by default. diff --git a/docs/zh/sql-reference/statements/select/order-by.md b/docs/zh/sql-reference/statements/select/order-by.md index 3286fc9f9e7..9540c96a10d 100644 --- a/docs/zh/sql-reference/statements/select/order-by.md +++ b/docs/zh/sql-reference/statements/select/order-by.md @@ -61,14 +61,14 @@ sidebar_label: ORDER BY 我们只建议使用 `COLLATE` 对于少量行的最终排序，因为排序与 `COLLATE` 比正常的按字节排序效率低。 -## ORDER BY ALL +## ORDER BY * -`ORDER BY ALL` 对所有选定的列进行升序排序。 +`ORDER BY *` 对所有选定的列进行升序排序。 示例: ``` sql -SELECT a, b, c FROM t ORDER BY ALL +SELECT a, b, c FROM t ORDER BY * ``` 等同于： diff --git a/programs/benchmark/Benchmark.cpp b/programs/benchmark/Benchmark.cpp index 961c678b936..fac88c0621f 100644 --- a/programs/benchmark/Benchmark.cpp +++ b/programs/benchmark/Benchmark.cpp @@ -2,7 +2,6 @@ #include #include #include -#include #include #include #include diff --git a/programs/keeper-converter/KeeperConverter.cpp b/programs/keeper-converter/KeeperConverter.cpp index 92bdea28738..8cd50d0892f 100644 --- a/programs/keeper-converter/KeeperConverter.cpp +++ b/programs/keeper-converter/KeeperConverter.cpp @@ -1,6 +1,7 @@ #include #include +#include #include #include #include @@ -39,7 +40,7 @@ int mainEntryClickHouseKeeperConverter(int argc, char ** argv) try { - auto keeper_context = std::make_shared(true); + auto keeper_context = std::make_shared(true, std::make_shared()); keeper_context->setDigestEnabled(true); keeper_context->setSnapshotDisk(std::make_shared("Keeper-snapshots", options["output-dir"].as())); diff --git a/programs/keeper/CMakeLists.txt b/programs/keeper/CMakeLists.txt index b8a5d9c9c19..70e0f229fd4 100644 --- a/programs/keeper/CMakeLists.txt +++ b/programs/keeper/CMakeLists.txt @@ -41,7 +41,7 @@ if (BUILD_STANDALONE_KEEPER) ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/KeeperStorage.cpp ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/KeeperConstants.cpp ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/KeeperAsynchronousMetrics.cpp - ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/pathUtils.cpp + ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/KeeperCommon.cpp ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/SessionExpiryQueue.cpp ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/SummingStateMachine.cpp ${CMAKE_CURRENT_SOURCE_DIR}/../../src/Coordination/WriteBufferFromNuraftBuffer.cpp diff --git a/programs/keeper/Keeper.cpp b/programs/keeper/Keeper.cpp index 5b844e7d650..8972c82eab8 100644 --- a/programs/keeper/Keeper.cpp +++ b/programs/keeper/Keeper.cpp @@ -560,7 +560,7 @@ try auto main_config_reloader = std::make_unique( config_path, extra_paths, - config().getString("path", ""), + config().getString("path", KEEPER_DEFAULT_PATH), std::move(unused_cache), unused_event, [&](ConfigurationPtr config, bool /* initial_loading */) diff --git a/programs/server/Server.cpp b/programs/server/Server.cpp index 0a3c23d746a..74fcc7326fc 100644 --- a/programs/server/Server.cpp +++ b/programs/server/Server.cpp @@ -1292,7 +1292,7 @@ try auto main_config_reloader = std::make_unique( config_path, extra_paths, - config().getString("path", ""), + config().getString("path", DBMS_DEFAULT_PATH), std::move(main_config_zk_node_cache), main_config_zk_changed_event, [&](ConfigurationPtr config, bool initial_loading) @@ -1391,7 +1391,7 @@ try global_context->setMaxDatabaseNumToWarn(new_server_settings.max_database_num_to_warn); global_context->setMaxPartNumToWarn(new_server_settings.max_part_num_to_warn); - ConcurrencyControl::SlotCount concurrent_threads_soft_limit = ConcurrencyControl::Unlimited; + SlotCount concurrent_threads_soft_limit = UnlimitedSlots; if (new_server_settings.concurrent_threads_soft_limit_num > 0 && new_server_settings.concurrent_threads_soft_limit_num < concurrent_threads_soft_limit) concurrent_threads_soft_limit = new_server_settings.concurrent_threads_soft_limit_num; if (new_server_settings.concurrent_threads_soft_limit_ratio_to_cores > 0) diff --git a/src/AggregateFunctions/AggregateFunctionCount.h b/src/AggregateFunctions/AggregateFunctionCount.h index f5d6030a777..36a8617ba91 100644 --- a/src/AggregateFunctions/AggregateFunctionCount.h +++ b/src/AggregateFunctions/AggregateFunctionCount.h @@ -219,7 +219,7 @@ public: : IAggregateFunctionDataHelper({argument}, params, createResultType()) { if (!argument->isNullable()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Not Nullable data type passed to AggregateFunctionCountNotNullUnary"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: not Nullable data type passed to AggregateFunctionCountNotNullUnary"); } String getName() const override { return "count"; } diff --git a/src/AggregateFunctions/AggregateFunctionFactory.cpp b/src/AggregateFunctions/AggregateFunctionFactory.cpp index 18edb7c8ce0..b6ba562045d 100644 --- a/src/AggregateFunctions/AggregateFunctionFactory.cpp +++ b/src/AggregateFunctions/AggregateFunctionFactory.cpp @@ -100,7 +100,7 @@ AggregateFunctionPtr AggregateFunctionFactory::get( { AggregateFunctionCombinatorPtr combinator = AggregateFunctionCombinatorFactory::instance().tryFindSuffix("Null"); if (!combinator) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot find aggregate function combinator " + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: cannot find aggregate function combinator " "to apply a function to Nullable arguments."); DataTypes nested_types = combinator->transformArguments(types_without_low_cardinality); @@ -123,7 +123,7 @@ AggregateFunctionPtr AggregateFunctionFactory::get( auto with_original_arguments = getImpl(name, action, types_without_low_cardinality, parameters, out_properties, false); if (!with_original_arguments) - throw Exception(ErrorCodes::LOGICAL_ERROR, "AggregateFunctionFactory returned nullptr"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: AggregateFunctionFactory returned nullptr"); return with_original_arguments; } diff --git a/src/AggregateFunctions/AggregateFunctionSum.h b/src/AggregateFunctions/AggregateFunctionSum.h index 5781ab69c6b..58aaddf357a 100644 --- a/src/AggregateFunctions/AggregateFunctionSum.h +++ b/src/AggregateFunctions/AggregateFunctionSum.h @@ -146,9 +146,7 @@ struct AggregateFunctionSumData size_t count = end - start; const auto * end_ptr = ptr + count; - if constexpr ( - (is_integer && !is_big_int_v) - || (is_decimal && !std::is_same_v && !std::is_same_v)) + if constexpr ((is_integer || is_decimal) && !is_over_big_int) { /// For integers we can vectorize the operation if we replace the null check using a multiplication (by 0 for null, 1 for not null) /// https://quick-bench.com/q/MLTnfTvwC2qZFVeWHfOBR3U7a8I @@ -163,8 +161,39 @@ struct AggregateFunctionSumData Impl::add(sum, local_sum); return; } + else if constexpr (is_over_big_int) + { + /// Use a mask to discard or keep the value to reduce branch miss. + /// Notice that for (U)Int128 or Decimal128, MaskType is Int8 instead of Int64, otherwise extra branches will be introduced by compiler (for unknown reason) and performance will be worse. + using MaskType = std::conditional_t; + alignas(64) const MaskType masks[2] = {0, -1}; + T local_sum{}; + while (ptr < end_ptr) + { + Value v = *ptr; + if constexpr (!add_if_zero) + { + if constexpr (is_integer) + v &= masks[!!*condition_map]; + else + v.value &= masks[!!*condition_map]; + } + else + { + if constexpr (is_integer) + v &= masks[!*condition_map]; + else + v.value &= masks[!*condition_map]; + } - if constexpr (std::is_floating_point_v) + Impl::add(local_sum, v); + ++ptr; + ++condition_map; + } + Impl::add(sum, local_sum); + return; + } + else if constexpr (std::is_floating_point_v) { /// For floating point we use a similar trick as above, except that now we reinterpret the floating point number as an unsigned /// integer of the same size and use a mask instead (0 to discard, 0xFF..FF to keep) diff --git a/src/AggregateFunctions/Combinators/AggregateFunctionIf.cpp b/src/AggregateFunctions/Combinators/AggregateFunctionIf.cpp index 9b5ee79a533..20a4bde6bb4 100644 --- a/src/AggregateFunctions/Combinators/AggregateFunctionIf.cpp +++ b/src/AggregateFunctions/Combinators/AggregateFunctionIf.cpp @@ -249,7 +249,7 @@ public: : Base(std::move(nested_function_), arguments, params), number_of_arguments(arguments.size()) { if (number_of_arguments == 1) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Single argument is passed to AggregateFunctionIfNullVariadic"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: single argument is passed to AggregateFunctionIfNullVariadic"); if (number_of_arguments > MAX_ARGS) throw Exception(ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH, diff --git a/src/AggregateFunctions/Combinators/AggregateFunctionNull.h b/src/AggregateFunctions/Combinators/AggregateFunctionNull.h index c8574e82be5..6b6580bf4c4 100644 --- a/src/AggregateFunctions/Combinators/AggregateFunctionNull.h +++ b/src/AggregateFunctions/Combinators/AggregateFunctionNull.h @@ -429,7 +429,7 @@ public: , number_of_arguments(arguments.size()) { if (number_of_arguments == 1) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Single argument is passed to AggregateFunctionNullVariadic"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: single argument is passed to AggregateFunctionNullVariadic"); if (number_of_arguments > MAX_ARGS) throw Exception(ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH, diff --git a/src/Analyzer/Passes/ArrayExistsToHasPass.cpp b/src/Analyzer/Passes/ArrayExistsToHasPass.cpp index 36c3df4d93a..62db502e1dc 100644 --- a/src/Analyzer/Passes/ArrayExistsToHasPass.cpp +++ b/src/Analyzer/Passes/ArrayExistsToHasPass.cpp @@ -1,6 +1,7 @@ #include #include +#include #include @@ -83,7 +84,8 @@ public: return; } - auto has_function = FunctionFactory::instance().get("has", getContext()); + auto has_function = createInternalFunctionHasOverloadResolver(); + array_exists_function_arguments_nodes[0] = std::move(array_exists_function_arguments_nodes[1]); array_exists_function_arguments_nodes[1] = std::move(has_constant_element_argument); array_exists_function_node->resolveAsFunction(has_function->build(array_exists_function_node->getArgumentColumns())); diff --git a/src/Analyzer/Passes/CNF.cpp b/src/Analyzer/Passes/CNF.cpp index aa6ee539934..71549f9e71d 100644 --- a/src/Analyzer/Passes/CNF.cpp +++ b/src/Analyzer/Passes/CNF.cpp @@ -10,6 +10,7 @@ #include #include +#include #include @@ -79,7 +80,7 @@ public: if (name == "and" || name == "or") { - auto function_resolver = FunctionFactory::instance().get(name, current_context); + auto function_resolver = name == "and" ? createInternalFunctionAndOverloadResolver() : createInternalFunctionOrOverloadResolver(); const auto & arguments = function_node->getArguments().getNodes(); if (arguments.size() > 2) @@ -110,10 +111,10 @@ private: class PushNotVisitor { public: - explicit PushNotVisitor(const ContextPtr & context) - : not_function_resolver(FunctionFactory::instance().get("not", context)) - , or_function_resolver(FunctionFactory::instance().get("or", context)) - , and_function_resolver(FunctionFactory::instance().get("and", context)) + explicit PushNotVisitor() + : not_function_resolver(createInternalFunctionNotOverloadResolver()) + , or_function_resolver(createInternalFunctionOrOverloadResolver()) + , and_function_resolver(createInternalFunctionAndOverloadResolver()) {} void visit(QueryTreeNodePtr & node, bool add_negation) @@ -162,10 +163,10 @@ private: class PushOrVisitor { public: - PushOrVisitor(ContextPtr context, size_t max_atoms_) + explicit PushOrVisitor(size_t max_atoms_) : max_atoms(max_atoms_) - , and_resolver(FunctionFactory::instance().get("and", context)) - , or_resolver(FunctionFactory::instance().get("or", context)) + , and_resolver(createInternalFunctionAndOverloadResolver()) + , or_resolver(createInternalFunctionOrOverloadResolver()) {} bool visit(QueryTreeNodePtr & node, size_t num_atoms) @@ -513,11 +514,11 @@ std::optional CNF::tryBuildCNF(const QueryTreeNodePtr & node, ContextPtr co } { - PushNotVisitor visitor(context); + PushNotVisitor visitor; visitor.visit(node_cloned, false); } - if (PushOrVisitor visitor(context, max_atoms); + if (PushOrVisitor visitor(max_atoms); !visitor.visit(node_cloned, atom_count)) return std::nullopt; @@ -542,7 +543,7 @@ CNF CNF::toCNF(const QueryTreeNodePtr & node, ContextPtr context, size_t max_gro return *cnf; } -QueryTreeNodePtr CNF::toQueryTree(ContextPtr context) const +QueryTreeNodePtr CNF::toQueryTree() const { if (statements.empty()) return nullptr; @@ -550,9 +551,9 @@ QueryTreeNodePtr CNF::toQueryTree(ContextPtr context) const QueryTreeNodes and_arguments; and_arguments.reserve(statements.size()); - auto not_resolver = FunctionFactory::instance().get("not", context); - auto or_resolver = FunctionFactory::instance().get("or", context); - auto and_resolver = FunctionFactory::instance().get("and", context); + auto not_resolver = createInternalFunctionNotOverloadResolver(); + auto or_resolver = createInternalFunctionOrOverloadResolver(); + auto and_resolver = createInternalFunctionAndOverloadResolver(); const auto function_node_from_atom = [&](const auto & atom) -> QueryTreeNodePtr { diff --git a/src/Analyzer/Passes/CNF.h b/src/Analyzer/Passes/CNF.h index ec639cd6679..9325d97d2f2 100644 --- a/src/Analyzer/Passes/CNF.h +++ b/src/Analyzer/Passes/CNF.h @@ -54,7 +54,7 @@ public: static std::optional tryBuildCNF(const QueryTreeNodePtr & node, ContextPtr context, size_t max_growth_multiplier = DEFAULT_MAX_GROWTH_MULTIPLIER); static CNF toCNF(const QueryTreeNodePtr & node, ContextPtr context, size_t max_growth_multiplier = DEFAULT_MAX_GROWTH_MULTIPLIER); - QueryTreeNodePtr toQueryTree(ContextPtr context) const; + QueryTreeNodePtr toQueryTree() const; const auto & getStatements() const { diff --git a/src/Analyzer/Passes/ConvertOrLikeChainPass.cpp b/src/Analyzer/Passes/ConvertOrLikeChainPass.cpp index 905819bf49f..eb897ef8746 100644 --- a/src/Analyzer/Passes/ConvertOrLikeChainPass.cpp +++ b/src/Analyzer/Passes/ConvertOrLikeChainPass.cpp @@ -11,6 +11,8 @@ #include #include +#include +#include #include @@ -134,8 +136,10 @@ private: void ConvertOrLikeChainPass::run(QueryTreeNodePtr & query_tree_node, ContextPtr context) { - auto or_function_resolver = FunctionFactory::instance().get("or", context); - auto match_function_resolver = FunctionFactory::instance().get("multiMatchAny", context); + const auto & settings = context->getSettingsRef(); + auto match_function_resolver = createInternalMultiMatchAnyOverloadResolver(settings.allow_hyperscan, settings.max_hyperscan_regexp_length, settings.max_hyperscan_regexp_total_length, settings.reject_expensive_hyperscan_regexps); + auto or_function_resolver = createInternalFunctionOrOverloadResolver(); + ConvertOrLikeChainVisitor visitor(std::move(or_function_resolver), std::move(match_function_resolver), std::move(context)); visitor.visit(query_tree_node); } diff --git a/src/Analyzer/Passes/ConvertQueryToCNFPass.cpp b/src/Analyzer/Passes/ConvertQueryToCNFPass.cpp index 5ce1ea43f2f..96bc62212fd 100644 --- a/src/Analyzer/Passes/ConvertQueryToCNFPass.cpp +++ b/src/Analyzer/Passes/ConvertQueryToCNFPass.cpp @@ -339,7 +339,7 @@ void addIndexConstraint(Analyzer::CNF & cnf, const QueryTreeNodes & table_expres { Analyzer::CNF::OrGroup new_group; auto index_hint_node = std::make_shared("indexHint"); - index_hint_node->getArguments().getNodes().push_back(Analyzer::CNF{std::move(and_group)}.toQueryTree(context)); + index_hint_node->getArguments().getNodes().push_back(Analyzer::CNF{std::move(and_group)}.toQueryTree()); index_hint_node->resolveAsFunction(FunctionFactory::instance().get("indexHint", context)); new_group.insert({false, QueryTreeNodePtrWithHash{std::move(index_hint_node)}}); @@ -676,7 +676,7 @@ void optimizeNode(QueryTreeNodePtr & node, const QueryTreeNodes & table_expressi if (settings.optimize_using_constraints) optimizeWithConstraints(*cnf, table_expressions, context); - auto new_node = cnf->toQueryTree(context); + auto new_node = cnf->toQueryTree(); node = std::move(new_node); } diff --git a/src/Analyzer/Passes/CrossToInnerJoinPass.cpp b/src/Analyzer/Passes/CrossToInnerJoinPass.cpp index 154babf3d9a..d0a5656d334 100644 --- a/src/Analyzer/Passes/CrossToInnerJoinPass.cpp +++ b/src/Analyzer/Passes/CrossToInnerJoinPass.cpp @@ -12,6 +12,7 @@ #include #include +#include #include @@ -256,7 +257,7 @@ private: for (const auto & node : nodes) function_node->getArguments().getNodes().push_back(node); - const auto & function = FunctionFactory::instance().get("and", getContext()); + const auto & function = createInternalFunctionAndOverloadResolver(); function_node->resolveAsFunction(function->build(function_node->getArgumentColumns())); return function_node; } diff --git a/src/Analyzer/Passes/IfChainToMultiIfPass.cpp b/src/Analyzer/Passes/IfChainToMultiIfPass.cpp index 88e350ffa2e..70b717f3108 100644 --- a/src/Analyzer/Passes/IfChainToMultiIfPass.cpp +++ b/src/Analyzer/Passes/IfChainToMultiIfPass.cpp @@ -5,6 +5,7 @@ #include #include #include +#include namespace DB { @@ -75,7 +76,8 @@ private: void IfChainToMultiIfPass::run(QueryTreeNodePtr & query_tree_node, ContextPtr context) { - auto multi_if_function_ptr = FunctionFactory::instance().get("multiIf", context); + const auto & settings = context->getSettingsRef(); + auto multi_if_function_ptr = createInternalMultiIfOverloadResolver(settings.allow_execute_multiif_columnar, settings.allow_experimental_variant_type, settings.use_variant_as_common_type); IfChainToMultiIfPassVisitor visitor(std::move(multi_if_function_ptr), std::move(context)); visitor.visit(query_tree_node); } diff --git a/src/Analyzer/Passes/MultiIfToIfPass.cpp b/src/Analyzer/Passes/MultiIfToIfPass.cpp index 8e09d5cab38..c42ea61b34a 100644 --- a/src/Analyzer/Passes/MultiIfToIfPass.cpp +++ b/src/Analyzer/Passes/MultiIfToIfPass.cpp @@ -3,6 +3,7 @@ #include #include #include +#include namespace DB { @@ -54,7 +55,8 @@ private: void MultiIfToIfPass::run(QueryTreeNodePtr & query_tree_node, ContextPtr context) { - auto if_function_ptr = FunctionFactory::instance().get("if", context); + const auto & settings = context->getSettingsRef(); + auto if_function_ptr = createInternalFunctionIfOverloadResolver(settings.allow_experimental_variant_type, settings.use_variant_as_common_type); MultiIfToIfVisitor visitor(std::move(if_function_ptr), std::move(context)); visitor.visit(query_tree_node); } diff --git a/src/Analyzer/Passes/QueryAnalysisPass.cpp b/src/Analyzer/Passes/QueryAnalysisPass.cpp index 3766d1aa6b9..376701f777e 100644 --- a/src/Analyzer/Passes/QueryAnalysisPass.cpp +++ b/src/Analyzer/Passes/QueryAnalysisPass.cpp @@ -120,7 +120,6 @@ namespace ErrorCodes extern const int NUMBER_OF_COLUMNS_DOESNT_MATCH; extern const int FUNCTION_CANNOT_HAVE_PARAMETERS; extern const int SYNTAX_ERROR; - extern const int UNEXPECTED_EXPRESSION; extern const int INVALID_IDENTIFIER; } @@ -1215,7 +1214,7 @@ private: static void expandGroupByAll(QueryNode & query_tree_node_typed); - void expandOrderByAll(QueryNode & query_tree_node_typed, const Settings & settings); + void expandOrderByAll(QueryNode & query_tree_node_typed); static std::string rewriteAggregateFunctionNameIfNeeded(const std::string & aggregate_function_name, NullsAction action, const ContextPtr & context); @@ -2367,9 +2366,9 @@ void QueryAnalyzer::expandGroupByAll(QueryNode & query_tree_node_typed) query_tree_node_typed.setIsGroupByAll(false); } -void QueryAnalyzer::expandOrderByAll(QueryNode & query_tree_node_typed, const Settings & settings) +void QueryAnalyzer::expandOrderByAll(QueryNode & query_tree_node_typed) { - if (!settings.enable_order_by_all || !query_tree_node_typed.isOrderByAll()) + if (!query_tree_node_typed.isOrderByAll()) return; auto * all_node = query_tree_node_typed.getOrderBy().getNodes()[0]->as(); @@ -2390,9 +2389,6 @@ void QueryAnalyzer::expandOrderByAll(QueryNode & query_tree_node_typed, const Se throw Exception(ErrorCodes::LOGICAL_ERROR, "Expression nodes list expected 1 projection names. Actual {}", projection_names.size()); - if (Poco::toUpper(projection_names[0]) == "ALL") - throw Exception(ErrorCodes::UNEXPECTED_EXPRESSION, - "Cannot use ORDER BY ALL to sort a column with name 'all', please disable setting `enable_order_by_all` and try again"); } auto sort_node = std::make_shared(node, all_node->getSortDirection(), all_node->getNullsSortDirection()); @@ -7559,7 +7555,7 @@ void QueryAnalyzer::resolveQuery(const QueryTreeNodePtr & query_node, Identifier if (settings.enable_positional_arguments) replaceNodesWithPositionalArguments(query_node_typed.getOrderByNode(), query_node_typed.getProjection().getNodes(), scope); - expandOrderByAll(query_node_typed, settings); + expandOrderByAll(query_node_typed); resolveSortNodeList(query_node_typed.getOrderByNode(), scope); } diff --git a/src/Analyzer/QueryNode.h b/src/Analyzer/QueryNode.h index d8b8741afb2..1b389572e42 100644 --- a/src/Analyzer/QueryNode.h +++ b/src/Analyzer/QueryNode.h @@ -219,13 +219,13 @@ public: is_group_by_all = is_group_by_all_value; } - /// Returns true, if query node has ORDER BY ALL modifier, false otherwise + /// Returns true, if query node has ORDER BY * modifier, false otherwise bool isOrderByAll() const { return is_order_by_all; } - /// Set query node ORDER BY ALL modifier value + /// Set query node ORDER BY * modifier value void setIsOrderByAll(bool is_order_by_all_value) { is_order_by_all = is_order_by_all_value; diff --git a/src/Backups/BackupIO_AzureBlobStorage.cpp b/src/Backups/BackupIO_AzureBlobStorage.cpp index 52ce20d5108..b3b92323109 100644 --- a/src/Backups/BackupIO_AzureBlobStorage.cpp +++ b/src/Backups/BackupIO_AzureBlobStorage.cpp @@ -2,7 +2,7 @@ #if USE_AZURE_BLOB_STORAGE #include -#include +#include #include #include #include diff --git a/src/Backups/BackupIO_S3.cpp b/src/Backups/BackupIO_S3.cpp index fa4c1af3698..9359602a651 100644 --- a/src/Backups/BackupIO_S3.cpp +++ b/src/Backups/BackupIO_S3.cpp @@ -2,7 +2,7 @@ #if USE_AWS_S3 #include -#include +#include #include #include #include @@ -127,7 +127,7 @@ BackupReaderS3::BackupReaderS3( : BackupReaderDefault(read_settings_, write_settings_, getLogger("BackupReaderS3")) , s3_uri(s3_uri_) , data_source_description{DataSourceType::ObjectStorage, ObjectStorageType::S3, MetadataStorageType::None, s3_uri.endpoint, false, false} - , s3_settings(context_->getStorageS3Settings().getSettings(s3_uri.uri.toString())) + , s3_settings(context_->getStorageS3Settings().getSettings(s3_uri.uri.toString(), context_->getUserName())) { auto & request_settings = s3_settings.request_settings; request_settings.updateFromSettings(context_->getSettingsRef()); @@ -217,7 +217,7 @@ BackupWriterS3::BackupWriterS3( : BackupWriterDefault(read_settings_, write_settings_, getLogger("BackupWriterS3")) , s3_uri(s3_uri_) , data_source_description{DataSourceType::ObjectStorage, ObjectStorageType::S3, MetadataStorageType::None, s3_uri.endpoint, false, false} - , s3_settings(context_->getStorageS3Settings().getSettings(s3_uri.uri.toString())) + , s3_settings(context_->getStorageS3Settings().getSettings(s3_uri.uri.toString(), context_->getUserName())) { auto & request_settings = s3_settings.request_settings; request_settings.updateFromSettings(context_->getSettingsRef()); diff --git a/src/Backups/BackupOperationInfo.h b/src/Backups/BackupOperationInfo.h index e57b57d75f1..21b5284458c 100644 --- a/src/Backups/BackupOperationInfo.h +++ b/src/Backups/BackupOperationInfo.h @@ -20,6 +20,9 @@ struct BackupOperationInfo /// Base Backup Operation name, a string like "Disk('backups', 'my_base_backup')" String base_backup_name; + /// Query ID of a query that started backup + String query_id; + /// This operation is internal and should not be shown in system.backups bool internal = false; diff --git a/src/Backups/BackupsWorker.cpp b/src/Backups/BackupsWorker.cpp index c19be22c749..5905d723800 100644 --- a/src/Backups/BackupsWorker.cpp +++ b/src/Backups/BackupsWorker.cpp @@ -440,7 +440,13 @@ OperationID BackupsWorker::startMakingBackup(const ASTPtr & query, const Context try { - addInfo(backup_id, backup_name_for_logging, base_backup_name, backup_settings.internal, context->getProcessListElement(), BackupStatus::CREATING_BACKUP); + addInfo(backup_id, + backup_name_for_logging, + base_backup_name, + context->getCurrentQueryId(), + backup_settings.internal, + context->getProcessListElement(), + BackupStatus::CREATING_BACKUP); /// Prepare context to use. ContextPtr context_in_use = context; @@ -823,7 +829,13 @@ OperationID BackupsWorker::startRestoring(const ASTPtr & query, ContextMutablePt if (restore_settings.base_backup_info) base_backup_name = restore_settings.base_backup_info->toStringForLogging(); - addInfo(restore_id, backup_name_for_logging, base_backup_name, restore_settings.internal, context->getProcessListElement(), BackupStatus::RESTORING); + addInfo(restore_id, + backup_name_for_logging, + base_backup_name, + context->getCurrentQueryId(), + restore_settings.internal, + context->getProcessListElement(), + BackupStatus::RESTORING); /// Prepare context to use. ContextMutablePtr context_in_use = context; @@ -1108,13 +1120,15 @@ void BackupsWorker::restoreTablesData(const OperationID & restore_id, BackupPtr } -void BackupsWorker::addInfo(const OperationID & id, const String & name, const String & base_backup_name, bool internal, QueryStatusPtr process_list_element, BackupStatus status) +void BackupsWorker::addInfo(const OperationID & id, const String & name, const String & base_backup_name, const String & query_id, + bool internal, QueryStatusPtr process_list_element, BackupStatus status) { ExtendedOperationInfo extended_info; auto & info = extended_info.info; info.id = id; info.name = name; info.base_backup_name = base_backup_name; + info.query_id = query_id; info.internal = internal; info.status = status; info.start_time = std::chrono::system_clock::now(); @@ -1183,7 +1197,7 @@ void BackupsWorker::setStatus(const String & id, BackupStatus status, bool throw if (isFailedOrCancelled(status)) { - info.error_message = getCurrentExceptionMessage(false); + info.error_message = getCurrentExceptionMessage(true /*with_stacktrace*/); info.exception = std::current_exception(); } diff --git a/src/Backups/BackupsWorker.h b/src/Backups/BackupsWorker.h index 73c8bf19473..ad187552c31 100644 --- a/src/Backups/BackupsWorker.h +++ b/src/Backups/BackupsWorker.h @@ -108,7 +108,8 @@ private: /// Run data restoring tasks which insert data to tables. void restoreTablesData(const BackupOperationID & restore_id, BackupPtr backup, DataRestoreTasks && tasks, ThreadPool & thread_pool, QueryStatusPtr process_list_element); - void addInfo(const BackupOperationID & id, const String & name, const String & base_backup_name, bool internal, QueryStatusPtr process_list_element, BackupStatus status); + void addInfo(const BackupOperationID & id, const String & name, const String & base_backup_name, const String & query_id, + bool internal, QueryStatusPtr process_list_element, BackupStatus status); void setStatus(const BackupOperationID & id, BackupStatus status, bool throw_if_error = true); void setStatusSafe(const String & id, BackupStatus status) { setStatus(id, status, false); } void setNumFilesAndSize(const BackupOperationID & id, size_t num_files, UInt64 total_size, size_t num_entries, diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index 08913ed1b5a..dff70e06ce4 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -506,6 +506,10 @@ if (TARGET ch_contrib::s2) dbms_target_link_libraries (PUBLIC ch_contrib::s2) endif() +if (TARGET ch_contrib::vectorscan) + dbms_target_link_libraries (PRIVATE ch_contrib::vectorscan) +endif() + if (TARGET ch_contrib::brotli) target_link_libraries (clickhouse_common_io PRIVATE ch_contrib::brotli) endif() diff --git a/src/Client/ConnectionEstablisher.h b/src/Client/ConnectionEstablisher.h index a3a01e63246..1fa08d435e9 100644 --- a/src/Client/ConnectionEstablisher.h +++ b/src/Client/ConnectionEstablisher.h @@ -3,6 +3,7 @@ #include #include #include +#include #include #include #include diff --git a/src/Client/ConnectionPool.h b/src/Client/ConnectionPool.h index 8e707e8190f..574c4992d75 100644 --- a/src/Client/ConnectionPool.h +++ b/src/Client/ConnectionPool.h @@ -28,7 +28,10 @@ public: using Entry = PoolBase::Entry; IConnectionPool() = default; - IConnectionPool(String host_, UInt16 port_) : host(host_), port(port_), address(host + ":" + toString(port_)) {} + IConnectionPool(String host_, UInt16 port_, Priority config_priority_) + : host(host_), port(port_), address(host + ":" + toString(port_)), config_priority(config_priority_) + { + } virtual ~IConnectionPool() = default; @@ -42,12 +45,13 @@ public: const std::string & getHost() const { return host; } UInt16 getPort() const { return port; } const String & getAddress() const { return address; } - virtual Priority getPriority() const { return Priority{1}; } + Priority getConfigPriority() const { return config_priority; } protected: const String host; const UInt16 port = 0; const String address; + const Priority config_priority; }; using ConnectionPoolPtr = std::shared_ptr; @@ -61,32 +65,31 @@ public: using Entry = IConnectionPool::Entry; using Base = PoolBase; - ConnectionPool(unsigned max_connections_, - const String & host_, - UInt16 port_, - const String & default_database_, - const String & user_, - const String & password_, - const String & quota_key_, - const String & cluster_, - const String & cluster_secret_, - const String & client_name_, - Protocol::Compression compression_, - Protocol::Secure secure_, - Priority priority_ = Priority{1}) - : IConnectionPool(host_, port_), - Base(max_connections_, - getLogger("ConnectionPool (" + host_ + ":" + toString(port_) + ")")), - default_database(default_database_), - user(user_), - password(password_), - quota_key(quota_key_), - cluster(cluster_), - cluster_secret(cluster_secret_), - client_name(client_name_), - compression(compression_), - secure(secure_), - priority(priority_) + ConnectionPool( + unsigned max_connections_, + const String & host_, + UInt16 port_, + const String & default_database_, + const String & user_, + const String & password_, + const String & quota_key_, + const String & cluster_, + const String & cluster_secret_, + const String & client_name_, + Protocol::Compression compression_, + Protocol::Secure secure_, + Priority config_priority_ = Priority{1}) + : IConnectionPool(host_, port_, config_priority_) + , Base(max_connections_, getLogger("ConnectionPool (" + host_ + ":" + toString(port_) + ")")) + , default_database(default_database_) + , user(user_) + , password(password_) + , quota_key(quota_key_) + , cluster(cluster_) + , cluster_secret(cluster_secret_) + , client_name(client_name_) + , compression(compression_) + , secure(secure_) { } @@ -114,11 +117,6 @@ public: return host + ":" + toString(port); } - Priority getPriority() const override - { - return priority; - } - protected: /** Creates a new object to put in the pool. */ ConnectionPtr allocObject() override @@ -143,7 +141,6 @@ private: String client_name; Protocol::Compression compression; /// Whether to compress data when interacting with the server. Protocol::Secure secure; /// Whether to encrypt data when interacting with the server. - Priority priority; /// priority from }; /** diff --git a/src/Client/ConnectionPoolWithFailover.cpp b/src/Client/ConnectionPoolWithFailover.cpp index fdc0a11e533..492fd4ae9e2 100644 --- a/src/Client/ConnectionPoolWithFailover.cpp +++ b/src/Client/ConnectionPoolWithFailover.cpp @@ -79,14 +79,6 @@ IConnectionPool::Entry ConnectionPoolWithFailover::get(const ConnectionTimeouts return Base::get(max_ignored_errors, fallback_to_stale_replicas, try_get_entry, get_priority); } -Priority ConnectionPoolWithFailover::getPriority() const -{ - return (*std::max_element(nested_pools.begin(), nested_pools.end(), [](const auto & a, const auto & b) - { - return a->getPriority() < b->getPriority(); - }))->getPriority(); -} - ConnectionPoolWithFailover::Status ConnectionPoolWithFailover::getStatus() const { const auto [states, pools, error_decrease_time] = getPoolExtendedStates(); @@ -253,13 +245,13 @@ ConnectionPoolWithFailover::tryGetEntry( } std::vector -ConnectionPoolWithFailover::getShuffledPools(const Settings & settings, GetPriorityForLoadBalancing::Func priority_func) +ConnectionPoolWithFailover::getShuffledPools(const Settings & settings, GetPriorityForLoadBalancing::Func priority_func, bool use_slowdown_count) { if (!priority_func) priority_func = makeGetPriorityFunc(settings); UInt64 max_ignored_errors = settings.distributed_replica_max_ignored_errors.value; - return Base::getShuffledPools(max_ignored_errors, priority_func); + return Base::getShuffledPools(max_ignored_errors, priority_func, use_slowdown_count); } } diff --git a/src/Client/ConnectionPoolWithFailover.h b/src/Client/ConnectionPoolWithFailover.h index 7ccdd4787a4..edfcbe6e4df 100644 --- a/src/Client/ConnectionPoolWithFailover.h +++ b/src/Client/ConnectionPoolWithFailover.h @@ -49,8 +49,6 @@ public: const Settings & settings, bool force_connected) override; /// From IConnectionPool - Priority getPriority() const override; /// From IConnectionPool - /** Allocates up to the specified number of connections to work. * Connections provide access to different replicas of one shard. */ @@ -83,15 +81,15 @@ public: struct NestedPoolStatus { const Base::NestedPoolPtr pool; - size_t error_count; - size_t slowdown_count; + size_t error_count = 0; + size_t slowdown_count = 0; std::chrono::seconds estimated_recovery_time; }; using Status = std::vector; Status getStatus() const; - std::vector getShuffledPools(const Settings & settings, GetPriorityFunc priority_func = {}); + std::vector getShuffledPools(const Settings & settings, GetPriorityFunc priority_func = {}, bool use_slowdown_count = false); size_t getMaxErrorCup() const { return Base::max_error_cap; } diff --git a/src/Client/HedgedConnectionsFactory.cpp b/src/Client/HedgedConnectionsFactory.cpp index 82bacece415..f5b074a0257 100644 --- a/src/Client/HedgedConnectionsFactory.cpp +++ b/src/Client/HedgedConnectionsFactory.cpp @@ -40,7 +40,8 @@ HedgedConnectionsFactory::HedgedConnectionsFactory( , max_parallel_replicas(max_parallel_replicas_) , skip_unavailable_shards(skip_unavailable_shards_) { - shuffled_pools = pool->getShuffledPools(settings_, priority_func); + shuffled_pools = pool->getShuffledPools(settings_, priority_func, /* use_slowdown_count */ true); + for (const auto & shuffled_pool : shuffled_pools) replicas.emplace_back( std::make_unique(shuffled_pool.pool, &timeouts, settings_, log, table_to_check.get())); diff --git a/src/Client/MultiplexedConnections.cpp b/src/Client/MultiplexedConnections.cpp index 8433c8afe9f..c7d7d0143c8 100644 --- a/src/Client/MultiplexedConnections.cpp +++ b/src/Client/MultiplexedConnections.cpp @@ -320,7 +320,7 @@ Packet MultiplexedConnections::receivePacketUnlocked(AsyncCallback async_callbac ReplicaState & state = getReplicaForReading(); current_connection = state.connection; if (current_connection == nullptr) - throw Exception(ErrorCodes::NO_AVAILABLE_REPLICA, "No available replica"); + throw Exception(ErrorCodes::NO_AVAILABLE_REPLICA, "Logical error: no available replica"); Packet packet; try diff --git a/src/Client/PacketReceiver.h b/src/Client/PacketReceiver.h index 6b3da659290..deedf5cccdc 100644 --- a/src/Client/PacketReceiver.h +++ b/src/Client/PacketReceiver.h @@ -5,6 +5,7 @@ #include #include +#include #include #include #include diff --git a/src/Columns/ColumnArray.cpp b/src/Columns/ColumnArray.cpp index b3376b35b2e..6f60ec0e642 100644 --- a/src/Columns/ColumnArray.cpp +++ b/src/Columns/ColumnArray.cpp @@ -810,7 +810,7 @@ ColumnPtr ColumnArray::filterTuple(const Filter & filt, ssize_t result_size_hint size_t tuple_size = tuple.tupleSize(); if (tuple_size == 0) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Empty tuple"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: empty tuple"); Columns temporary_arrays(tuple_size); for (size_t i = 0; i < tuple_size; ++i) @@ -1263,7 +1263,7 @@ ColumnPtr ColumnArray::replicateTuple(const Offsets & replicate_offsets) const size_t tuple_size = tuple.tupleSize(); if (tuple_size == 0) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Empty tuple"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: empty tuple"); Columns temporary_arrays(tuple_size); for (size_t i = 0; i < tuple_size; ++i) diff --git a/src/Columns/ColumnNullable.cpp b/src/Columns/ColumnNullable.cpp index ddf5fc696fb..93638371b84 100644 --- a/src/Columns/ColumnNullable.cpp +++ b/src/Columns/ColumnNullable.cpp @@ -1,5 +1,7 @@ #include #include +#include +#include #include #include #include @@ -26,6 +28,7 @@ namespace ErrorCodes { extern const int LOGICAL_ERROR; extern const int ILLEGAL_COLUMN; + extern const int SIZES_OF_NESTED_COLUMNS_ARE_INCONSISTENT; extern const int NOT_IMPLEMENTED; } @@ -826,7 +829,8 @@ void ColumnNullable::applyNullMap(const ColumnNullable & other) void ColumnNullable::checkConsistency() const { if (null_map->size() != getNestedColumn().size()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Sizes of nested column and null map of Nullable column are not equal"); + throw Exception(ErrorCodes::SIZES_OF_NESTED_COLUMNS_ARE_INCONSISTENT, + "Logical error: Sizes of nested column and null map of Nullable column are not equal"); } ColumnPtr ColumnNullable::createWithOffsets(const IColumn::Offsets & offsets, const ColumnConst & column_with_default_value, size_t total_rows, size_t shift) const diff --git a/src/Columns/getLeastSuperColumn.cpp b/src/Columns/getLeastSuperColumn.cpp index 4f4a5f2b9b8..6ec5ca7a9c1 100644 --- a/src/Columns/getLeastSuperColumn.cpp +++ b/src/Columns/getLeastSuperColumn.cpp @@ -21,7 +21,7 @@ static bool sameConstants(const IColumn & a, const IColumn & b) ColumnWithTypeAndName getLeastSuperColumn(const std::vector & columns) { if (columns.empty()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No src columns for supercolumn"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no src columns for supercolumn"); ColumnWithTypeAndName result = *columns[0]; diff --git a/src/Common/CPUID.h b/src/Common/CPUID.h index b47e7e808d7..d7a714ec5af 100644 --- a/src/Common/CPUID.h +++ b/src/Common/CPUID.h @@ -57,6 +57,249 @@ inline bool cpuid(UInt32 op, UInt32 * res) noexcept /// NOLINT #endif } +union CPUInfo +{ + UInt32 info[4]; + + struct Registers + { + UInt32 eax; + UInt32 ebx; + UInt32 ecx; + UInt32 edx; + } registers; + + inline explicit CPUInfo(UInt32 op) noexcept { cpuid(op, info); } + + inline CPUInfo(UInt32 op, UInt32 sub_op) noexcept { cpuid(op, sub_op, info); } +}; + +inline bool haveRDTSCP() noexcept +{ + return (CPUInfo(0x80000001).registers.edx >> 27) & 1u; +} + +inline bool haveSSE() noexcept +{ + return (CPUInfo(0x1).registers.edx >> 25) & 1u; +} + +inline bool haveSSE2() noexcept +{ + return (CPUInfo(0x1).registers.edx >> 26) & 1u; +} + +inline bool haveSSE3() noexcept +{ + return CPUInfo(0x1).registers.ecx & 1u; +} + +inline bool havePCLMUL() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 1) & 1u; +} + +inline bool haveSSSE3() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 9) & 1u; +} + +inline bool haveSSE41() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 19) & 1u; +} + +inline bool haveSSE42() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 20) & 1u; +} + +inline bool haveF16C() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 29) & 1u; +} + +inline bool havePOPCNT() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 23) & 1u; +} + +inline bool haveAES() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 25) & 1u; +} + +inline bool haveXSAVE() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 26) & 1u; +} + +inline bool haveOSXSAVE() noexcept +{ + return (CPUInfo(0x1).registers.ecx >> 27) & 1u; +} + +inline bool haveAVX() noexcept +{ +#if defined(__x86_64__) + // http://www.intel.com/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-optimization-manual.pdf + // https://bugs.chromium.org/p/chromium/issues/detail?id=375968 + return haveOSXSAVE() // implies haveXSAVE() + && (our_xgetbv(0) & 6u) == 6u // XMM state and YMM state are enabled by OS + && ((CPUInfo(0x1).registers.ecx >> 28) & 1u); // AVX bit +#else + return false; +#endif +} + +inline bool haveFMA() noexcept +{ + return haveAVX() && ((CPUInfo(0x1).registers.ecx >> 12) & 1u); +} + +inline bool haveAVX2() noexcept +{ + return haveAVX() && ((CPUInfo(0x7, 0).registers.ebx >> 5) & 1u); +} + +inline bool haveBMI1() noexcept +{ + return (CPUInfo(0x7, 0).registers.ebx >> 3) & 1u; +} + +inline bool haveBMI2() noexcept +{ + return (CPUInfo(0x7, 0).registers.ebx >> 8) & 1u; +} + +inline bool haveAVX512F() noexcept +{ +#if defined(__x86_64__) + // https://software.intel.com/en-us/articles/how-to-detect-knl-instruction-support + return haveOSXSAVE() // implies haveXSAVE() + && (our_xgetbv(0) & 6u) == 6u // XMM state and YMM state are enabled by OS + && ((our_xgetbv(0) >> 5) & 7u) == 7u // ZMM state is enabled by OS + && CPUInfo(0x0).registers.eax >= 0x7 // leaf 7 is present + && ((CPUInfo(0x7, 0).registers.ebx >> 16) & 1u); // AVX512F bit +#else + return false; +#endif +} + +inline bool haveAVX512DQ() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 17) & 1u); +} + +inline bool haveRDSEED() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 18) & 1u); +} + +inline bool haveADX() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 19) & 1u); +} + +inline bool haveAVX512IFMA() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 21) & 1u); +} + +inline bool havePCOMMIT() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 22) & 1u); +} + +inline bool haveCLFLUSHOPT() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 23) & 1u); +} + +inline bool haveCLWB() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 24) & 1u); +} + +inline bool haveAVX512PF() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 26) & 1u); +} + +inline bool haveAVX512ER() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 27) & 1u); +} + +inline bool haveAVX512CD() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 28) & 1u); +} + +inline bool haveSHA() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 29) & 1u); +} + +inline bool haveAVX512BW() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 30) & 1u); +} + +inline bool haveAVX512VL() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 31) & 1u); +} + +inline bool havePREFETCHWT1() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ecx >> 0) & 1u); +} + +inline bool haveAVX512VBMI() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ecx >> 1) & 1u); +} + +inline bool haveAVX512VBMI2() noexcept +{ + return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ecx >> 6) & 1u); +} + +inline bool haveRDRAND() noexcept +{ + return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x1).registers.ecx >> 30) & 1u); +} + +inline bool haveAMX() noexcept +{ +#if defined(__x86_64__) + // http://www.intel.com/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-optimization-manual.pdf + return haveOSXSAVE() // implies haveXSAVE() + && ((our_xgetbv(0) >> 17) & 0x3) == 0x3; // AMX state are enabled by OS +#else + return false; +#endif +} + +inline bool haveAMXBF16() noexcept +{ + return haveAMX() + && ((CPUInfo(0x7, 0).registers.edx >> 22) & 1u); // AMX-BF16 bit +} + +inline bool haveAMXTILE() noexcept +{ + return haveAMX() + && ((CPUInfo(0x7, 0).registers.edx >> 24) & 1u); // AMX-TILE bit +} + +inline bool haveAMXINT8() noexcept +{ + return haveAMX() + && ((CPUInfo(0x7, 0).registers.edx >> 25) & 1u); // AMX-INT8 bit +} + #define CPU_ID_ENUMERATE(OP) \ OP(SSE) \ OP(SSE2) \ @@ -98,253 +341,6 @@ inline bool cpuid(UInt32 op, UInt32 * res) noexcept /// NOLINT OP(AMXTILE) \ OP(AMXINT8) -union CPUInfo -{ - UInt32 info[4]; - - struct Registers - { - UInt32 eax; - UInt32 ebx; - UInt32 ecx; - UInt32 edx; - } registers; - - inline explicit CPUInfo(UInt32 op) noexcept { cpuid(op, info); } - - inline CPUInfo(UInt32 op, UInt32 sub_op) noexcept { cpuid(op, sub_op, info); } -}; - -#define DEF_NAME(X) inline bool have##X() noexcept; - CPU_ID_ENUMERATE(DEF_NAME) -#undef DEF_NAME - -bool haveRDTSCP() noexcept -{ - return (CPUInfo(0x80000001).registers.edx >> 27) & 1u; -} - -bool haveSSE() noexcept -{ - return (CPUInfo(0x1).registers.edx >> 25) & 1u; -} - -bool haveSSE2() noexcept -{ - return (CPUInfo(0x1).registers.edx >> 26) & 1u; -} - -bool haveSSE3() noexcept -{ - return CPUInfo(0x1).registers.ecx & 1u; -} - -bool havePCLMUL() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 1) & 1u; -} - -bool haveSSSE3() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 9) & 1u; -} - -bool haveSSE41() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 19) & 1u; -} - -bool haveSSE42() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 20) & 1u; -} - -bool haveF16C() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 29) & 1u; -} - -bool havePOPCNT() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 23) & 1u; -} - -bool haveAES() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 25) & 1u; -} - -bool haveXSAVE() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 26) & 1u; -} - -bool haveOSXSAVE() noexcept -{ - return (CPUInfo(0x1).registers.ecx >> 27) & 1u; -} - -bool haveAVX() noexcept -{ -#if defined(__x86_64__) - // http://www.intel.com/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-optimization-manual.pdf - // https://bugs.chromium.org/p/chromium/issues/detail?id=375968 - return haveOSXSAVE() // implies haveXSAVE() - && (our_xgetbv(0) & 6u) == 6u // XMM state and YMM state are enabled by OS - && ((CPUInfo(0x1).registers.ecx >> 28) & 1u); // AVX bit -#else - return false; -#endif -} - -bool haveFMA() noexcept -{ - return haveAVX() && ((CPUInfo(0x1).registers.ecx >> 12) & 1u); -} - -bool haveAVX2() noexcept -{ - return haveAVX() && ((CPUInfo(0x7, 0).registers.ebx >> 5) & 1u); -} - -bool haveBMI1() noexcept -{ - return (CPUInfo(0x7, 0).registers.ebx >> 3) & 1u; -} - -bool haveBMI2() noexcept -{ - return (CPUInfo(0x7, 0).registers.ebx >> 8) & 1u; -} - -bool haveAVX512F() noexcept -{ -#if defined(__x86_64__) - // https://software.intel.com/en-us/articles/how-to-detect-knl-instruction-support - return haveOSXSAVE() // implies haveXSAVE() - && (our_xgetbv(0) & 6u) == 6u // XMM state and YMM state are enabled by OS - && ((our_xgetbv(0) >> 5) & 7u) == 7u // ZMM state is enabled by OS - && CPUInfo(0x0).registers.eax >= 0x7 // leaf 7 is present - && ((CPUInfo(0x7, 0).registers.ebx >> 16) & 1u); // AVX512F bit -#else - return false; -#endif -} - -bool haveAVX512DQ() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 17) & 1u); -} - -bool haveRDSEED() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 18) & 1u); -} - -bool haveADX() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 19) & 1u); -} - -bool haveAVX512IFMA() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 21) & 1u); -} - -bool havePCOMMIT() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 22) & 1u); -} - -bool haveCLFLUSHOPT() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 23) & 1u); -} - -bool haveCLWB() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 24) & 1u); -} - -bool haveAVX512PF() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 26) & 1u); -} - -bool haveAVX512ER() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 27) & 1u); -} - -bool haveAVX512CD() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 28) & 1u); -} - -bool haveSHA() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ebx >> 29) & 1u); -} - -bool haveAVX512BW() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 30) & 1u); -} - -bool haveAVX512VL() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ebx >> 31) & 1u); -} - -bool havePREFETCHWT1() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x7, 0).registers.ecx >> 0) & 1u); -} - -bool haveAVX512VBMI() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ecx >> 1) & 1u); -} - -bool haveAVX512VBMI2() noexcept -{ - return haveAVX512F() && ((CPUInfo(0x7, 0).registers.ecx >> 6) & 1u); -} - -bool haveRDRAND() noexcept -{ - return CPUInfo(0x0).registers.eax >= 0x7 && ((CPUInfo(0x1).registers.ecx >> 30) & 1u); -} - -inline bool haveAMX() noexcept -{ -#if defined(__x86_64__) - // http://www.intel.com/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-optimization-manual.pdf - return haveOSXSAVE() // implies haveXSAVE() - && ((our_xgetbv(0) >> 17) & 0x3) == 0x3; // AMX state are enabled by OS -#else - return false; -#endif -} - -bool haveAMXBF16() noexcept -{ - return haveAMX() - && ((CPUInfo(0x7, 0).registers.edx >> 22) & 1u); // AMX-BF16 bit -} - -bool haveAMXTILE() noexcept -{ - return haveAMX() - && ((CPUInfo(0x7, 0).registers.edx >> 24) & 1u); // AMX-TILE bit -} - -bool haveAMXINT8() noexcept -{ - return haveAMX() - && ((CPUInfo(0x7, 0).registers.edx >> 25) & 1u); // AMX-INT8 bit -} - struct CPUFlagsCache { #define DEF_NAME(X) static inline bool have_##X = have##X(); diff --git a/src/Common/ConcurrencyControl.cpp b/src/Common/ConcurrencyControl.cpp index c9fe51550dc..0893cfce955 100644 --- a/src/Common/ConcurrencyControl.cpp +++ b/src/Common/ConcurrencyControl.cpp @@ -12,10 +12,10 @@ namespace ErrorCodes ConcurrencyControl::Slot::~Slot() { - allocation->release(); + static_cast(*allocation).release(); } -ConcurrencyControl::Slot::Slot(AllocationPtr && allocation_) +ConcurrencyControl::Slot::Slot(SlotAllocationPtr && allocation_) : allocation(std::move(allocation_)) { } @@ -27,7 +27,7 @@ ConcurrencyControl::Allocation::~Allocation() parent.free(this); } -[[nodiscard]] ConcurrencyControl::SlotPtr ConcurrencyControl::Allocation::tryAcquire() +[[nodiscard]] AcquiredSlotPtr ConcurrencyControl::Allocation::tryAcquire() { SlotCount value = granted.load(); while (value) @@ -35,15 +35,21 @@ ConcurrencyControl::Allocation::~Allocation() if (granted.compare_exchange_strong(value, value - 1)) { std::unique_lock lock{mutex}; - return SlotPtr(new Slot(shared_from_this())); // can't use std::make_shared due to private ctor + return AcquiredSlotPtr(new Slot(shared_from_this())); // can't use std::make_shared due to private ctor } } return {}; // avoid unnecessary locking } -ConcurrencyControl::SlotCount ConcurrencyControl::Allocation::grantedCount() const +SlotCount ConcurrencyControl::Allocation::grantedCount() const { - return granted; + return granted.load(); +} + +SlotCount ConcurrencyControl::Allocation::allocatedCount() const +{ + std::unique_lock lock{mutex}; + return allocated; } ConcurrencyControl::Allocation::Allocation(ConcurrencyControl & parent_, SlotCount limit_, SlotCount granted_, Waiters::iterator waiter_) @@ -87,7 +93,7 @@ ConcurrencyControl::~ConcurrencyControl() abort(); } -[[nodiscard]] ConcurrencyControl::AllocationPtr ConcurrencyControl::allocate(SlotCount min, SlotCount max) +[[nodiscard]] SlotAllocationPtr ConcurrencyControl::allocate(SlotCount min, SlotCount max) { if (min > max) throw Exception(ErrorCodes::LOGICAL_ERROR, "ConcurrencyControl: invalid allocation requirements"); @@ -100,13 +106,13 @@ ConcurrencyControl::~ConcurrencyControl() // Create allocation and start waiting if more slots are required if (granted < max) - return AllocationPtr(new Allocation(*this, max, granted, + return SlotAllocationPtr(new Allocation(*this, max, granted, waiters.insert(cur_waiter, nullptr /* pointer is set by Allocation ctor */))); else - return AllocationPtr(new Allocation(*this, max, granted)); + return SlotAllocationPtr(new Allocation(*this, max, granted)); } -void ConcurrencyControl::setMaxConcurrency(ConcurrencyControl::SlotCount value) +void ConcurrencyControl::setMaxConcurrency(SlotCount value) { std::unique_lock lock{mutex}; max_concurrency = std::max(1, value); // never allow max_concurrency to be zero @@ -162,7 +168,7 @@ void ConcurrencyControl::schedule(std::unique_lock &) } } -ConcurrencyControl::SlotCount ConcurrencyControl::available(std::unique_lock &) const +SlotCount ConcurrencyControl::available(std::unique_lock &) const { if (cur_concurrency < max_concurrency) return max_concurrency - cur_concurrency; diff --git a/src/Common/ConcurrencyControl.h b/src/Common/ConcurrencyControl.h index 7e20384aa2a..ba94502962c 100644 --- a/src/Common/ConcurrencyControl.h +++ b/src/Common/ConcurrencyControl.h @@ -7,6 +7,7 @@ #include #include +#include namespace DB { @@ -34,41 +35,35 @@ namespace DB * Oversubscription is possible: total amount of allocated slots can exceed `setMaxConcurrency(limit)` * because `min` amount of slots is allocated for each query unconditionally. */ -class ConcurrencyControl : boost::noncopyable +class ConcurrencyControl : public ISlotControl { public: struct Allocation; - using AllocationPtr = std::shared_ptr; - using SlotCount = UInt64; using Waiters = std::list; - static constexpr SlotCount Unlimited = std::numeric_limits::max(); - // Scoped guard for acquired slot, see Allocation::tryAcquire() - struct Slot : boost::noncopyable + struct Slot : public IAcquiredSlot { - ~Slot(); + ~Slot() override; private: friend struct Allocation; // for ctor - explicit Slot(AllocationPtr && allocation_); + explicit Slot(SlotAllocationPtr && allocation_); - AllocationPtr allocation; + SlotAllocationPtr allocation; }; - // FIXME: have to be unique_ptr, but ThreadFromGlobalPool does not support move semantics yet - using SlotPtr = std::shared_ptr; - // Manages group of slots for a single query, see ConcurrencyControl::allocate(min, max) - struct Allocation : std::enable_shared_from_this, boost::noncopyable + struct Allocation : public ISlotAllocation { - ~Allocation(); + ~Allocation() override; // Take one already granted slot if available. Lock-free iff there is no granted slot. - [[nodiscard]] SlotPtr tryAcquire(); + [[nodiscard]] AcquiredSlotPtr tryAcquire() override; - SlotCount grantedCount() const; + SlotCount grantedCount() const override; + SlotCount allocatedCount() const override; private: friend struct Slot; // for release() @@ -94,7 +89,7 @@ public: ConcurrencyControl & parent; const SlotCount limit; - std::mutex mutex; // the following values must be accessed under this mutex + mutable std::mutex mutex; // the following values must be accessed under this mutex SlotCount allocated; // allocated total (including already `released`) SlotCount released = 0; @@ -103,17 +98,16 @@ public: const Waiters::iterator waiter; // iterator to itself in Waiters list; valid iff allocated < limit }; -public: ConcurrencyControl(); // WARNING: all Allocation objects MUST be destructed before ConcurrencyControl // NOTE: Recommended way to achieve this is to use `instance()` and do graceful shutdown of queries - ~ConcurrencyControl(); + ~ConcurrencyControl() override; // Allocate at least `min` and at most `max` slots. // If not all `max` slots were successfully allocated, a subscription for later allocation is created // Use `Allocation::tryAcquire()` to acquire allocated slot, before running a thread. - [[nodiscard]] AllocationPtr allocate(SlotCount min, SlotCount max); + [[nodiscard]] SlotAllocationPtr allocate(SlotCount min, SlotCount max) override; void setMaxConcurrency(SlotCount value); @@ -134,7 +128,7 @@ private: std::mutex mutex; Waiters waiters; Waiters::iterator cur_waiter; // round-robin pointer - SlotCount max_concurrency = Unlimited; + SlotCount max_concurrency = UnlimitedSlots; SlotCount cur_concurrency = 0; }; diff --git a/src/Common/CurrentMetrics.cpp b/src/Common/CurrentMetrics.cpp index c6fbafa8dc3..6931001202d 100644 --- a/src/Common/CurrentMetrics.cpp +++ b/src/Common/CurrentMetrics.cpp @@ -262,6 +262,9 @@ M(ActiveTimersInQueryProfiler, "Number of Active thread local timers in QueryProfiler") \ M(RefreshableViews, "Number materialized views with periodic refreshing (REFRESH)") \ M(RefreshingViews, "Number of materialized views currently executing a refresh") \ + M(StorageBufferFlushThreads, "Number of threads for background flushes in StorageBuffer") \ + M(StorageBufferFlushThreadsActive, "Number of threads for background flushes in StorageBuffer running a task") \ + M(StorageBufferFlushThreadsScheduled, "Number of queued or active threads for background flushes in StorageBuffer") #ifdef APPLY_FOR_EXTERNAL_METRICS #define APPLY_FOR_METRICS(M) APPLY_FOR_BUILTIN_METRICS(M) APPLY_FOR_EXTERNAL_METRICS(M) diff --git a/src/Common/Fiber.h b/src/Common/Fiber.h index 8b88bd323ef..f48ace149f4 100644 --- a/src/Common/Fiber.h +++ b/src/Common/Fiber.h @@ -17,7 +17,7 @@ private: template friend class FiberLocal; public: - template + template< typename StackAlloc, typename Fn> Fiber(StackAlloc && salloc, Fn && fn) : impl(std::allocator_arg_t(), std::forward(salloc), RoutineImpl(std::forward(fn))) { } @@ -46,12 +46,6 @@ public: current_fiber = parent_fiber; } - static FiberPtr & getCurrentFiber() - { - thread_local static FiberPtr current_fiber; - return current_fiber; - } - private: template struct RoutineImpl @@ -80,6 +74,12 @@ private: Fn fn; }; + static FiberPtr & getCurrentFiber() + { + thread_local static FiberPtr current_fiber; + return current_fiber; + } + /// Special wrapper to store data in uniquer_ptr. struct DataWrapper { @@ -146,3 +146,4 @@ private: T main_instance; }; + diff --git a/src/Common/ISlotControl.h b/src/Common/ISlotControl.h new file mode 100644 index 00000000000..daeb956f5a8 --- /dev/null +++ b/src/Common/ISlotControl.h @@ -0,0 +1,76 @@ +#pragma once + +#include +#include +#include +#include + + +namespace DB +{ + +// Interfaces for abstract "slot" allocation and control. +// Slot is a virtual entity existing in a limited amount (CPUs or memory chunks, etc). +// +// Every slot can be in one of the following states: +// * free: slot is available to be allocated. +// * allocated: slot is allocated to a specific ISlotAllocation. +// +// Allocated slots can be in one of the following states: +// * granted: allocated, but not yet acquired. +// * acquired: a granted slot becomes acquired by using IAcquiredSlot. +// +// Example for CPU (see ConcurrencyControl.h). Every slot represents one CPU in the system. +// Slot allocation is a request to allocate specific number of CPUs for a specific query. +// Acquired slot is an entity that is held by a thread as long as it is running. This allows +// total number of threads in the system to be limited and the distribution process to be controlled. +// +// TODO: +// - for preemption - ability to return granted slot back and reacquire it later. +// - for memory allocations - variable size of slots (in bytes). + +/// Number of slots +using SlotCount = UInt64; + +/// Unlimited number of slots +constexpr SlotCount UnlimitedSlots = std::numeric_limits::max(); + +/// Acquired slot holder. Slot is considered to be acquired as long as the object exists. +class IAcquiredSlot : public std::enable_shared_from_this, boost::noncopyable +{ +public: + virtual ~IAcquiredSlot() = default; +}; + +using AcquiredSlotPtr = std::shared_ptr; + +/// Request for allocation of slots from ISlotControl. +/// Allows for more slots to be acquired and the whole request to be canceled. +class ISlotAllocation : public std::enable_shared_from_this, boost::noncopyable +{ +public: + virtual ~ISlotAllocation() = default; + + /// Take one already granted slot if available. + [[nodiscard]] virtual AcquiredSlotPtr tryAcquire() = 0; + + /// Returns the number of granted slots for given allocation (i.e. available to be acquired) + virtual SlotCount grantedCount() const = 0; + + /// Returns the total number of slots allocated at the moment (acquired and granted) + virtual SlotCount allocatedCount() const = 0; +}; + +using SlotAllocationPtr = std::shared_ptr; + +class ISlotControl : boost::noncopyable +{ +public: + virtual ~ISlotControl() = default; + + // Allocate at least `min` and at most `max` slots. + // If not all `max` slots were successfully allocated, a "subscription" for later allocation is created + [[nodiscard]] virtual SlotAllocationPtr allocate(SlotCount min, SlotCount max) = 0; +}; + +} diff --git a/src/Common/PoolWithFailoverBase.h b/src/Common/PoolWithFailoverBase.h index 8fd83300eff..cf270c9dad0 100644 --- a/src/Common/PoolWithFailoverBase.h +++ b/src/Common/PoolWithFailoverBase.h @@ -66,7 +66,7 @@ public: , log(log_) { for (size_t i = 0;i < nested_pools.size(); ++i) - shared_pool_states[i].config_priority = nested_pools[i]->getPriority(); + shared_pool_states[i].config_priority = nested_pools[i]->getConfigPriority(); } struct TryResult @@ -133,7 +133,7 @@ protected: void updateErrorCounts(PoolStates & states, time_t & last_decrease_time) const; - std::vector getShuffledPools(size_t max_ignored_errors, const GetPriorityFunc & get_priority); + std::vector getShuffledPools(size_t max_ignored_errors, const GetPriorityFunc & get_priority, bool use_slowdown_count = false); inline void updateSharedErrorCounts(std::vector & shuffled_pools); @@ -160,7 +160,7 @@ protected: template std::vector::ShuffledPool> PoolWithFailoverBase::getShuffledPools( - size_t max_ignored_errors, const PoolWithFailoverBase::GetPriorityFunc & get_priority) + size_t max_ignored_errors, const PoolWithFailoverBase::GetPriorityFunc & get_priority, bool use_slowdown_count) { /// Update random numbers and error counts. PoolStates pool_states = updatePoolStates(max_ignored_errors); @@ -175,13 +175,13 @@ PoolWithFailoverBase::getShuffledPools( std::vector shuffled_pools; shuffled_pools.reserve(nested_pools.size()); for (size_t i = 0; i < nested_pools.size(); ++i) - shuffled_pools.push_back(ShuffledPool{nested_pools[i], &pool_states[i], i, /* error_count = */ 0, /* slowdown_count = */ 0}); + shuffled_pools.emplace_back(ShuffledPool{.pool = nested_pools[i], .state = &pool_states[i], .index = i}); ::sort( shuffled_pools.begin(), shuffled_pools.end(), - [](const ShuffledPool & lhs, const ShuffledPool & rhs) + [use_slowdown_count](const ShuffledPool & lhs, const ShuffledPool & rhs) { - return PoolState::compare(*lhs.state, *rhs.state); + return PoolState::compare(*lhs.state, *rhs.state, use_slowdown_count); }); return shuffled_pools; @@ -344,10 +344,14 @@ struct PoolWithFailoverBase::PoolState random = rng(); } - static bool compare(const PoolState & lhs, const PoolState & rhs) + static bool compare(const PoolState & lhs, const PoolState & rhs, bool use_slowdown_count) { - return std::forward_as_tuple(lhs.error_count, lhs.slowdown_count, lhs.config_priority, lhs.priority, lhs.random) - < std::forward_as_tuple(rhs.error_count, rhs.slowdown_count, rhs.config_priority, rhs.priority, rhs.random); + if (use_slowdown_count) + return std::forward_as_tuple(lhs.error_count, lhs.slowdown_count, lhs.config_priority, lhs.priority, lhs.random) + < std::forward_as_tuple(rhs.error_count, rhs.slowdown_count, rhs.config_priority, rhs.priority, rhs.random); + else + return std::forward_as_tuple(lhs.error_count, lhs.config_priority, lhs.priority, lhs.random) + < std::forward_as_tuple(rhs.error_count, rhs.config_priority, rhs.priority, rhs.random); } private: diff --git a/src/Common/ProfileEvents.cpp b/src/Common/ProfileEvents.cpp index f14223ec644..d8ca1ab9e93 100644 --- a/src/Common/ProfileEvents.cpp +++ b/src/Common/ProfileEvents.cpp @@ -632,6 +632,12 @@ The server successfully detected this situation and will download merged part fr M(InterfacePostgreSQLReceiveBytes, "Number of bytes received through PostgreSQL interfaces") \ \ M(ParallelReplicasUsedCount, "Number of replicas used to execute a query with task-based parallel replicas") \ + \ + M(KeeperLogsEntryReadFromLatestCache, "Number of log entries in Keeper being read from latest logs cache") \ + M(KeeperLogsEntryReadFromCommitCache, "Number of log entries in Keeper being read from commit logs cache") \ + M(KeeperLogsEntryReadFromFile, "Number of log entries in Keeper being read directly from the changelog file") \ + M(KeeperLogsPrefetchedEntries, "Number of log entries in Keeper being prefetched from the changelog file") \ + \ M(ParallelReplicasAvailableCount, "Number of replicas available to execute a query with task-based parallel replicas") \ M(ParallelReplicasUnavailableCount, "Number of replicas which was chosen, but found to be unavailable during query execution with task-based parallel replicas") \ diff --git a/src/Common/SensitiveDataMasker.cpp b/src/Common/SensitiveDataMasker.cpp index 28eae6f451d..70346919f65 100644 --- a/src/Common/SensitiveDataMasker.cpp +++ b/src/Common/SensitiveDataMasker.cpp @@ -91,7 +91,7 @@ void SensitiveDataMasker::setInstance(std::unique_ptr&& sen { if (!sensitive_data_masker_) - throw Exception(ErrorCodes::LOGICAL_ERROR, "The 'sensitive_data_masker' is not set"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: the 'sensitive_data_masker' is not set"); if (sensitive_data_masker_->rulesCount() > 0) { diff --git a/src/Common/SipHash.h b/src/Common/SipHash.h index 729fb76a573..5f27fdaa4b6 100644 --- a/src/Common/SipHash.h +++ b/src/Common/SipHash.h @@ -209,7 +209,7 @@ public: { if (!is_reference_128) throw DB::Exception( - DB::ErrorCodes::LOGICAL_ERROR, "Can't call get128Reference when is_reference_128 is not set"); + DB::ErrorCodes::LOGICAL_ERROR, "Logical error: can't call get128Reference when is_reference_128 is not set"); finalize(); const auto lo = v0 ^ v1 ^ v2 ^ v3; v1 ^= 0xdd; diff --git a/src/Common/StackTrace.cpp b/src/Common/StackTrace.cpp index 436b85ff30b..7e683ae91de 100644 --- a/src/Common/StackTrace.cpp +++ b/src/Common/StackTrace.cpp @@ -448,6 +448,9 @@ toStringEveryLineImpl([[maybe_unused]] bool fatal, const StackTraceRefTriple & s DB::writePointerHex(frame.physical_addr, out); } + if (frame.object.has_value()) + out << " in " << *frame.object; + callback(out.str()); }; #else diff --git a/src/Common/checkStackSize.cpp b/src/Common/checkStackSize.cpp index c88554ca8fe..8c2a0aaed7f 100644 --- a/src/Common/checkStackSize.cpp +++ b/src/Common/checkStackSize.cpp @@ -1,8 +1,8 @@ -#include -#include /// THREAD_SANITIZER #include #include -#include +#include +#include +#include /// THREAD_SANITIZER #include #include #include @@ -114,10 +114,6 @@ __attribute__((__weak__)) void checkStackSize() { using namespace DB; - /// Not implemented for coroutines. - if (Fiber::getCurrentFiber()) - return; - if (!stack_address) max_stack_size = getStackSize(&stack_address); @@ -140,7 +136,7 @@ __attribute__((__weak__)) void checkStackSize() /// We assume that stack grows towards lower addresses. And that it starts to grow from the end of a chunk of memory of max_stack_size. if (int_frame_address > int_stack_address + max_stack_size) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Frame address is greater than stack begin address"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: frame address is greater than stack begin address"); size_t stack_size = int_stack_address + max_stack_size - int_frame_address; size_t max_stack_size_allowed = static_cast(max_stack_size * STACK_SIZE_FREE_RATIO); diff --git a/src/Common/tests/gtest_async_loader.cpp b/src/Common/tests/gtest_async_loader.cpp index 950c7bbab76..fc2537abcfc 100644 --- a/src/Common/tests/gtest_async_loader.cpp +++ b/src/Common/tests/gtest_async_loader.cpp @@ -427,9 +427,7 @@ TEST(AsyncLoader, CancelExecutingTask) } } -// This test is disabled due to `MemorySanitizer: use-of-uninitialized-value` issue in `collectSymbolsFromProgramHeaders` function -// More details: https://github.com/ClickHouse/ClickHouse/pull/48923#issuecomment-1545415482 -TEST(AsyncLoader, DISABLED_JobFailure) +TEST(AsyncLoader, JobFailure) { AsyncLoaderTest t; t.loader.start(); diff --git a/src/Common/tests/gtest_concurrency_control.cpp b/src/Common/tests/gtest_concurrency_control.cpp index 8e5b89a72a0..5e579317ade 100644 --- a/src/Common/tests/gtest_concurrency_control.cpp +++ b/src/Common/tests/gtest_concurrency_control.cpp @@ -15,7 +15,7 @@ struct ConcurrencyControlTest { ConcurrencyControl cc; - explicit ConcurrencyControlTest(ConcurrencyControl::SlotCount limit = ConcurrencyControl::Unlimited) + explicit ConcurrencyControlTest(SlotCount limit = UnlimitedSlots) { cc.setMaxConcurrency(limit); } @@ -25,7 +25,7 @@ TEST(ConcurrencyControl, Unlimited) { ConcurrencyControlTest t; // unlimited number of slots auto slots = t.cc.allocate(0, 100500); - std::vector acquired; + std::vector acquired; while (auto slot = slots->tryAcquire()) acquired.emplace_back(std::move(slot)); ASSERT_TRUE(acquired.size() == 100500); @@ -34,14 +34,14 @@ TEST(ConcurrencyControl, Unlimited) TEST(ConcurrencyControl, Fifo) { ConcurrencyControlTest t(1); // use single slot - std::vector allocations; + std::vector allocations; constexpr int count = 42; allocations.reserve(count); for (int i = 0; i < count; i++) allocations.emplace_back(t.cc.allocate(0, 1)); for (int i = 0; i < count; i++) { - ConcurrencyControl::SlotPtr holder; + AcquiredSlotPtr holder; for (int j = 0; j < count; j++) { auto slot = allocations[j]->tryAcquire(); @@ -60,11 +60,11 @@ TEST(ConcurrencyControl, Fifo) TEST(ConcurrencyControl, Oversubscription) { ConcurrencyControlTest t(10); - std::vector allocations; + std::vector allocations; allocations.reserve(10); for (int i = 0; i < 10; i++) allocations.emplace_back(t.cc.allocate(1, 2)); - std::vector slots; + std::vector slots; // Normal allocation using maximum amount of slots for (int i = 0; i < 5; i++) { @@ -90,7 +90,7 @@ TEST(ConcurrencyControl, ReleaseUnacquiredSlots) { ConcurrencyControlTest t(10); { - std::vector allocations; + std::vector allocations; allocations.reserve(10); for (int i = 0; i < 10; i++) allocations.emplace_back(t.cc.allocate(1, 2)); @@ -98,7 +98,7 @@ TEST(ConcurrencyControl, ReleaseUnacquiredSlots) } // Check that slots were actually released auto allocation = t.cc.allocate(0, 20); - std::vector acquired; + std::vector acquired; while (auto slot = allocation->tryAcquire()) acquired.emplace_back(std::move(slot)); ASSERT_TRUE(acquired.size() == 10); @@ -110,7 +110,7 @@ TEST(ConcurrencyControl, DestroyNotFullyAllocatedAllocation) for (int i = 0; i < 3; i++) { auto allocation = t.cc.allocate(5, 20); - std::vector acquired; + std::vector acquired; while (auto slot = allocation->tryAcquire()) acquired.emplace_back(std::move(slot)); ASSERT_TRUE(acquired.size() == 10); @@ -122,7 +122,7 @@ TEST(ConcurrencyControl, DestroyAllocationBeforeSlots) ConcurrencyControlTest t(10); for (int i = 0; i < 3; i++) { - std::vector acquired; + std::vector acquired; auto allocation = t.cc.allocate(5, 20); while (auto slot = allocation->tryAcquire()) acquired.emplace_back(std::move(slot)); @@ -135,7 +135,7 @@ TEST(ConcurrencyControl, GrantReleasedToTheSameAllocation) { ConcurrencyControlTest t(3); auto allocation = t.cc.allocate(0, 10); - std::list acquired; + std::list acquired; while (auto slot = allocation->tryAcquire()) acquired.emplace_back(std::move(slot)); ASSERT_TRUE(acquired.size() == 3); // 0 1 2 @@ -183,7 +183,7 @@ TEST(ConcurrencyControl, SetSlotCount) { ConcurrencyControlTest t(10); auto allocation = t.cc.allocate(5, 30); - std::vector acquired; + std::vector acquired; while (auto slot = allocation->tryAcquire()) acquired.emplace_back(std::move(slot)); ASSERT_TRUE(acquired.size() == 10); @@ -200,7 +200,7 @@ TEST(ConcurrencyControl, SetSlotCount) ASSERT_TRUE(acquired.size() == 5); // Check that newly added slots are equally distributed over waiting allocations - std::vector acquired2; + std::vector acquired2; auto allocation2 = t.cc.allocate(0, 30); ASSERT_TRUE(!allocation->tryAcquire()); t.cc.setMaxConcurrency(15); // 10 slots added: 5 to the first allocation and 5 to the second one @@ -224,7 +224,7 @@ TEST(ConcurrencyControl, MultipleThreads) auto run_query = [&] (size_t max_threads) { - ConcurrencyControl::AllocationPtr slots = t.cc.allocate(1, max_threads); + SlotAllocationPtr slots = t.cc.allocate(1, max_threads); std::mutex threads_mutex; std::vector threads; threads.reserve(max_threads); diff --git a/src/Interpreters/threadPoolCallbackRunner.h b/src/Common/threadPoolCallbackRunner.h similarity index 100% rename from src/Interpreters/threadPoolCallbackRunner.h rename to src/Common/threadPoolCallbackRunner.h diff --git a/src/Compression/CompressionCodecT64.cpp b/src/Compression/CompressionCodecT64.cpp index bf9a9414bc1..3ddc56fe4f6 100644 --- a/src/Compression/CompressionCodecT64.cpp +++ b/src/Compression/CompressionCodecT64.cpp @@ -91,6 +91,7 @@ enum class MagicNumber : uint8_t Decimal32 = 19, Decimal64 = 20, IPv4 = 21, + Date32 = 22, }; MagicNumber serializeTypeId(std::optional type_id) @@ -109,6 +110,7 @@ MagicNumber serializeTypeId(std::optional type_id) case TypeIndex::Int32: return MagicNumber::Int32; case TypeIndex::Int64: return MagicNumber::Int64; case TypeIndex::Date: return MagicNumber::Date; + case TypeIndex::Date32: return MagicNumber::Date32; case TypeIndex::DateTime: return MagicNumber::DateTime; case TypeIndex::DateTime64: return MagicNumber::DateTime64; case TypeIndex::Enum8: return MagicNumber::Enum8; @@ -137,6 +139,7 @@ TypeIndex deserializeTypeId(uint8_t serialized_type_id) case MagicNumber::Int32: return TypeIndex::Int32; case MagicNumber::Int64: return TypeIndex::Int64; case MagicNumber::Date: return TypeIndex::Date; + case MagicNumber::Date32: return TypeIndex::Date32; case MagicNumber::DateTime: return TypeIndex::DateTime; case MagicNumber::DateTime64: return TypeIndex::DateTime64; case MagicNumber::Enum8: return TypeIndex::Enum8; @@ -165,6 +168,7 @@ TypeIndex baseType(TypeIndex type_idx) return TypeIndex::Int16; case TypeIndex::Int32: case TypeIndex::Decimal32: + case TypeIndex::Date32: return TypeIndex::Int32; case TypeIndex::Int64: case TypeIndex::Decimal64: @@ -205,6 +209,7 @@ TypeIndex typeIdx(const IDataType * data_type) case TypeIndex::UInt16: case TypeIndex::Enum16: case TypeIndex::Date: + case TypeIndex::Date32: case TypeIndex::Int32: case TypeIndex::UInt32: case TypeIndex::IPv4: diff --git a/src/Coordination/Changelog.cpp b/src/Coordination/Changelog.cpp index 5a58932606e..469244605e0 100644 --- a/src/Coordination/Changelog.cpp +++ b/src/Coordination/Changelog.cpp @@ -1,5 +1,12 @@ +#include #include +#include +#include #include +#include +#include +#include +#include #include #include #include @@ -14,8 +21,19 @@ #include #include #include +#include +#include #include +#include +#include +namespace ProfileEvents +{ + extern const Event KeeperLogsEntryReadFromLatestCache; + extern const Event KeeperLogsEntryReadFromCommitCache; + extern const Event KeeperLogsEntryReadFromFile; + extern const Event KeeperLogsPrefetchedEntries; +} namespace DB { @@ -33,25 +51,31 @@ namespace ErrorCodes namespace { -constexpr std::string_view tmp_prefix = "tmp_"; - -void moveFileBetweenDisks(DiskPtr disk_from, ChangelogFileDescriptionPtr description, DiskPtr disk_to, const std::string & path_to) +void moveChangelogBetweenDisks( + DiskPtr disk_from, + ChangelogFileDescriptionPtr description, + DiskPtr disk_to, + const std::string & path_to, + const KeeperContextPtr & keeper_context) { - /// we use empty file with prefix tmp_ to detect incomplete copies - /// if a copy is complete we don't care from which disk we use the same file - /// so it's okay if a failure happens after removing of tmp file but before we remove - /// the changelog from the source disk - auto from_path = fs::path(description->path); - auto tmp_changelog_name = from_path.parent_path() / (std::string{tmp_prefix} + from_path.filename().string()); - { - auto buf = disk_to->writeFile(tmp_changelog_name); - buf->finalize(); - } - disk_from->copyFile(from_path, *disk_to, path_to, {}); - disk_to->removeFile(tmp_changelog_name); - disk_from->removeFile(description->path); - description->path = path_to; - description->disk = disk_to; + moveFileBetweenDisks( + disk_from, + description->path, + disk_to, + path_to, + [&] + { + /// a different thread could be trying to read from the file + /// we should make sure the source disk contains the file while read is in progress + description->withLock( + [&] + { + description->disk = disk_to; + description->path = path_to; + }); + }, + getLogger("Changelog"), + keeper_context); } constexpr auto DEFAULT_PREFIX = "changelog"; @@ -111,9 +135,11 @@ class ChangelogWriter public: ChangelogWriter( std::map & existing_changelogs_, + LogEntryStorage & entry_storage_, KeeperContextPtr keeper_context_, LogFileSettings log_file_settings_) : existing_changelogs(existing_changelogs_) + , entry_storage(entry_storage_) , log_file_settings(log_file_settings_) , keeper_context(std::move(keeper_context_)) , log(getLogger("Changelog")) @@ -173,15 +199,15 @@ public: } else { - moveFileBetweenDisks(log_disk, current_file_description, disk, new_path); + moveChangelogBetweenDisks(log_disk, current_file_description, disk, new_path, keeper_context); } } } auto latest_log_disk = getLatestLogDisk(); - assert(file_description->disk == latest_log_disk); + chassert(file_description->disk == latest_log_disk); file_buf = latest_log_disk->writeFile(file_description->path, DBMS_DEFAULT_BUFFER_SIZE, mode); - assert(file_buf); + chassert(file_buf); last_index_written.reset(); current_file_description = std::move(file_description); @@ -196,7 +222,7 @@ public: } catch (...) { - tryLogCurrentException(log); + tryLogCurrentException(log, "While setting new changelog file"); throw; } } @@ -238,6 +264,7 @@ public: } auto & write_buffer = getBuffer(); + auto current_position = initial_file_size + write_buffer.count(); writeIntBinary(computeRecordChecksum(record), write_buffer); writeIntBinary(record.header.version, write_buffer); @@ -255,6 +282,15 @@ public: /// Flush compressed data to file buffer compressed_buffer->next(); } + else + { + unflushed_indices_with_log_location.emplace_back( + record.header.index, + LogLocation{ + .file_description = current_file_description, + .position = current_position, + .size = record.header.blob_size}); + } last_index_written = record.header.index; @@ -272,6 +308,8 @@ public: else file_buffer->next(); } + entry_storage.addLogLocations(std::move(unflushed_indices_with_log_location)); + unflushed_indices_with_log_location.clear(); } uint64_t getStartIndex() const @@ -314,9 +352,9 @@ public: private: void finalizeCurrentFile() { - assert(prealloc_done); + chassert(prealloc_done); - assert(current_file_description); + chassert(current_file_description); // compact can delete the file and we don't need to do anything if (current_file_description->deleted) { @@ -400,9 +438,11 @@ private: { const auto * file_buffer = tryGetFileBuffer(); + if (file_buffer) + initial_file_size = getSizeFromFileDescriptor(file_buffer->getFD()); + if (log_file_settings.max_size == 0 || !file_buffer) { - initial_file_size = 0; prealloc_done = true; return; } @@ -428,7 +468,6 @@ private: } } #endif - initial_file_size = getSizeFromFileDescriptor(file_buffer->getFD()); prealloc_done = true; } @@ -441,6 +480,10 @@ private: std::map & existing_changelogs; + LogEntryStorage & entry_storage; + + std::vector> unflushed_indices_with_log_location; + ChangelogFileDescriptionPtr current_file_description{nullptr}; std::unique_ptr file_buf; std::optional last_index_written; @@ -457,22 +500,25 @@ private: LoggerPtr const log; }; +namespace +{ + struct ChangelogReadResult { /// Total entries read from log including skipped. /// Useful when we decide to continue to write in the same log and want to know /// how many entries was already written in it. - uint64_t total_entries_read_from_log; + uint64_t total_entries_read_from_log{0}; /// First index in log - uint64_t log_start_index; + uint64_t log_start_index{0}; /// First entry actually read log (not including skipped) - uint64_t first_read_index; + uint64_t first_read_index{0}; /// Last entry read from log (last entry in log) /// When we don't skip anything last_read_index - first_read_index = total_entries_read_from_log. /// But when some entries from the start of log can be skipped because they are not required. - uint64_t last_read_index; + uint64_t last_read_index{0}; /// last offset we were able to read from log off_t last_position; @@ -482,69 +528,99 @@ struct ChangelogReadResult bool error; }; +ChangelogRecord readChangelogRecord(ReadBuffer & read_buf, const std::string & filepath) +{ + /// Read checksum + Checksum record_checksum; + readIntBinary(record_checksum, read_buf); + + /// Read header + ChangelogRecord record; + readIntBinary(record.header.version, read_buf); + readIntBinary(record.header.index, read_buf); + readIntBinary(record.header.term, read_buf); + readIntBinary(record.header.value_type, read_buf); + readIntBinary(record.header.blob_size, read_buf); + + if (record.header.version > CURRENT_CHANGELOG_VERSION) + throw Exception( + ErrorCodes::UNKNOWN_FORMAT_VERSION, "Unsupported changelog version {} on path {}", static_cast(record.header.version), filepath); + + /// Read data + if (record.header.blob_size != 0) + { + auto buffer = nuraft::buffer::alloc(record.header.blob_size); + auto * buffer_begin = reinterpret_cast(buffer->data_begin()); + read_buf.readStrict(buffer_begin, record.header.blob_size); + record.blob = buffer; + } + else + record.blob = nullptr; + + /// Compare checksums + Checksum checksum = computeRecordChecksum(record); + if (checksum != record_checksum) + { + throw Exception( + ErrorCodes::CHECKSUM_DOESNT_MATCH, + "Checksums doesn't match for log {} (version {}), index {}, blob_size {}", + filepath, + record.header.version, + record.header.index, + record.header.blob_size); + } + + return record; +} + +LogEntryPtr logEntryFromRecord(const ChangelogRecord & record) +{ + return nuraft::cs_new(record.header.term, record.blob, static_cast(record.header.value_type)); +} + +size_t logEntrySize(const LogEntryPtr & log_entry) +{ + return log_entry->get_buf().size(); +} + +LogEntryPtr getLogEntry(const CacheEntry & cache_entry) +{ + if (const auto * log_entry = std::get_if(&cache_entry)) + return *log_entry; + + const auto & prefetched_log_entry = std::get(cache_entry); + return prefetched_log_entry.getLogEntry(); +} + +} + class ChangelogReader { public: - explicit ChangelogReader(DiskPtr disk_, const std::string & filepath_) : disk(disk_), filepath(filepath_) + explicit ChangelogReader(ChangelogFileDescriptionPtr changelog_description_) : changelog_description(changelog_description_) { - compression_method = chooseCompressionMethod(filepath, ""); - auto read_buffer_from_file = disk->readFile(filepath); + compression_method = chooseCompressionMethod(changelog_description->path, ""); + auto read_buffer_from_file = changelog_description->disk->readFile(changelog_description->path); read_buf = wrapReadBufferWithCompressionMethod(std::move(read_buffer_from_file), compression_method); } /// start_log_index -- all entries with index < start_log_index will be skipped, but accounted into total_entries_read_from_log - ChangelogReadResult readChangelog(IndexToLogEntry & logs, uint64_t start_log_index, LoggerPtr log) + ChangelogReadResult readChangelog(LogEntryStorage & entry_storage, uint64_t start_log_index, LoggerPtr log) { ChangelogReadResult result{}; result.compressed_log = compression_method != CompressionMethod::None; + const auto & filepath = changelog_description->path; try { while (!read_buf->eof()) { result.last_position = read_buf->count(); - /// Read checksum - Checksum record_checksum; - readIntBinary(record_checksum, *read_buf); - /// Read header - ChangelogRecord record; - readIntBinary(record.header.version, *read_buf); - readIntBinary(record.header.index, *read_buf); - readIntBinary(record.header.term, *read_buf); - readIntBinary(record.header.value_type, *read_buf); - readIntBinary(record.header.blob_size, *read_buf); - - if (record.header.version > CURRENT_CHANGELOG_VERSION) - throw Exception( - ErrorCodes::UNKNOWN_FORMAT_VERSION, "Unsupported changelog version {} on path {}", static_cast(record.header.version), filepath); - - /// Read data - if (record.header.blob_size != 0) - { - auto buffer = nuraft::buffer::alloc(record.header.blob_size); - auto * buffer_begin = reinterpret_cast(buffer->data_begin()); - read_buf->readStrict(buffer_begin, record.header.blob_size); - record.blob = buffer; - } - else - record.blob = nullptr; - - /// Compare checksums - Checksum checksum = computeRecordChecksum(record); - if (checksum != record_checksum) - { - throw Exception( - ErrorCodes::CHECKSUM_DOESNT_MATCH, - "Checksums doesn't match for log {} (version {}), index {}, blob_size {}", - filepath, - record.header.version, - record.header.index, - record.header.blob_size); - } + auto record = readChangelogRecord(*read_buf, filepath); /// Check for duplicated changelog ids - if (logs.contains(record.header.index)) - std::erase_if(logs, [&record](const auto & item) { return item.first >= record.header.index; }); + if (entry_storage.contains(record.header.index)) + entry_storage.cleanAfter(record.header.index - 1); result.total_entries_read_from_log += 1; @@ -553,12 +629,18 @@ public: continue; /// Create log entry for read data - auto log_entry = nuraft::cs_new(record.header.term, record.blob, static_cast(record.header.value_type)); + auto log_entry = logEntryFromRecord(record); if (result.first_read_index == 0) result.first_read_index = record.header.index; /// Put it into in memory structure - logs.emplace(record.header.index, log_entry); + entry_storage.addEntryWithLocation( + record.header.index, + log_entry, + LogLocation{ + .file_description = changelog_description, + .position = static_cast(result.last_position), + .size = record.header.blob_size}); result.last_read_index = record.header.index; if (result.total_entries_read_from_log % 50000 == 0) @@ -585,131 +667,971 @@ public: } private: - DiskPtr disk; - std::string filepath; + ChangelogFileDescriptionPtr changelog_description; CompressionMethod compression_method; std::unique_ptr read_buf; }; +PrefetchedCacheEntry::PrefetchedCacheEntry() + : log_entry(log_entry_resolver.get_future()) +{} + +const LogEntryPtr & PrefetchedCacheEntry::getLogEntry() const +{ + return log_entry.get(); +} + +void PrefetchedCacheEntry::resolve(std::exception_ptr exception) +{ + log_entry_resolver.set_exception(exception); +} + +void PrefetchedCacheEntry::resolve(LogEntryPtr log_entry_) +{ + log_entry_resolver.set_value(std::move(log_entry_)); +} + +LogEntryStorage::LogEntryStorage(const LogFileSettings & log_settings, KeeperContextPtr keeper_context_) + : latest_logs_cache(log_settings.latest_logs_cache_size_threshold) + , commit_logs_cache(log_settings.commit_logs_cache_size_threshold) + , prefetch_queue(std::numeric_limits::max()) + , keeper_context(std::move(keeper_context_)) + , log(getLogger("Changelog")) +{ + commit_logs_prefetcher = std::make_unique([this] { prefetchCommitLogs(); }); +} + +LogEntryStorage::~LogEntryStorage() +{ + shutdown(); +} + +void LogEntryStorage::prefetchCommitLogs() +{ + std::shared_ptr prefetch_info; + while (prefetch_queue.pop(prefetch_info)) + { + if (prefetch_info->cancel) + { + prefetch_info->done = true; + prefetch_info->done.notify_all(); + continue; + } + + auto current_index = prefetch_info->commit_prefetch_index_range.first; + try + { + for (const auto & prefetch_file_info : prefetch_info->file_infos) + { + prefetch_file_info.file_description->withLock( + [&] + { + const auto & [changelog_description, position, count] = prefetch_file_info; + auto file = changelog_description->disk->readFile(changelog_description->path, ReadSettings()); + file->seek(position, SEEK_SET); + LOG_TRACE( + log, "Prefetching {} log entries from path {}, from position {}", count, changelog_description->path, position); + ProfileEvents::increment(ProfileEvents::KeeperLogsPrefetchedEntries, count); + + for (size_t i = 0; i < count; ++i) + { + if (prefetch_info->cancel) + break; + + auto record = readChangelogRecord(*file, changelog_description->path); + auto entry = logEntryFromRecord(record); + if (current_index != record.header.index) + throw Exception( + ErrorCodes::LOGICAL_ERROR, + "Invalid index prefetched, expected {}, actual {}", + current_index, + record.header.index); + + commit_logs_cache.getPrefetchedCacheEntry(record.header.index).resolve(std::move(entry)); + ++current_index; + } + }); + + if (prefetch_info->cancel) + break; + } + } + catch (...) + { + tryLogCurrentException(log, "While prefetching log entries"); + auto exception = std::current_exception(); + + for (; current_index <= prefetch_info->commit_prefetch_index_range.second; ++current_index) + commit_logs_cache.getPrefetchedCacheEntry(current_index).resolve(exception); + } + + prefetch_info->done = true; + prefetch_info->done.notify_all(); + } +} + +void LogEntryStorage::startCommitLogsPrefetch(uint64_t last_committed_index) const +{ + if (keeper_context->isShutdownCalled()) + return; + + /// commit logs is not empty and it's not next log + if (!commit_logs_cache.empty() && commit_logs_cache.max_index_in_cache != last_committed_index) + return; + + if (logs_location.empty()) + return; + + /// we are already prefetching some logs for commit + if (current_prefetch_info && !current_prefetch_info->done) + return; + + auto new_prefetch_info = std::make_shared(); + auto & [prefetch_from, prefetch_to] = new_prefetch_info->commit_prefetch_index_range; + + /// if there are no entries in commit cache we will start from the next log that will be committed + /// otherwise we continue appending the commit cache from the latest entry stored in it + size_t current_index = commit_logs_cache.empty() ? last_committed_index + 1 : commit_logs_cache.max_index_in_cache + 1; + + prefetch_from = current_index; + + size_t total_size = 0; + std::vector file_infos; + FileReadInfo * current_file_info = nullptr; + + size_t max_index_for_prefetch = 0; + if (!latest_logs_cache.empty()) + max_index_for_prefetch = latest_logs_cache.min_index_in_cache - 1; + else + max_index_for_prefetch = max_index_with_location; + + for (; current_index <= max_index_for_prefetch; ++current_index) + { + const auto & [changelog_description, position, size] = logs_location.at(current_index); + if (total_size == 0) + current_file_info = &file_infos.emplace_back(changelog_description, position, /* count */ 1); + else if (total_size + size > commit_logs_cache.size_threshold) + break; + else if (changelog_description == current_file_info->file_description) + ++current_file_info->count; + else + current_file_info = &file_infos.emplace_back(changelog_description, position, /* count */ 1); + + total_size += size; + commit_logs_cache.addEntry(current_index, size, PrefetchedCacheEntry()); + } + + if (!file_infos.empty()) + { + current_prefetch_info = std::move(new_prefetch_info); + prefetch_to = current_index - 1; + LOG_TRACE(log, "Will prefetch {} commit log entries [{} - {}]", prefetch_to - prefetch_from + 1, prefetch_from, prefetch_to); + + current_prefetch_info->file_infos = std::move(file_infos); + auto inserted = prefetch_queue.push(current_prefetch_info); + chassert(inserted); + } +} + +LogEntryStorage::InMemoryCache::InMemoryCache(size_t size_threshold_) + : size_threshold(size_threshold_) +{} + +void LogEntryStorage::InMemoryCache::updateStatsWithNewEntry(uint64_t index, size_t size) +{ + cache_size += size; + + if (cache.size() == 1) + { + min_index_in_cache = index; + max_index_in_cache = index; + } + else + { + chassert(index > max_index_in_cache); + max_index_in_cache = index; + } +} + +void LogEntryStorage::InMemoryCache::addEntry(uint64_t index, size_t size, CacheEntry log_entry) +{ + auto [_, inserted] = cache.emplace(index, std::move(log_entry)); + if (!inserted) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Trying to insert log with index {} which is already present in cache", index); + + updateStatsWithNewEntry(index, size); +} + +void LogEntryStorage::InMemoryCache::addEntry(IndexToCacheEntryNode && node) +{ + auto index = node.key(); + auto entry_size = logEntrySize(getLogEntry(node.mapped())); + + auto result = cache.insert(std::move(node)); + if (!result.inserted) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Trying to insert log with index {} which is already present in cache", index); + + updateStatsWithNewEntry(index, entry_size); +} + +IndexToCacheEntryNode LogEntryStorage::InMemoryCache::popOldestEntry() +{ + auto node = cache.extract(min_index_in_cache); + if (node.empty()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Couldn't find the oldest entry of index {} in logs cache", min_index_in_cache); + ++min_index_in_cache; + cache_size -= logEntrySize(getLogEntry(node.mapped())); + return node; +} + +bool LogEntryStorage::InMemoryCache::containsEntry(uint64_t index) const +{ + return !cache.empty() && index >= min_index_in_cache && index <= max_index_in_cache; +} + +CacheEntry * LogEntryStorage::InMemoryCache::getCacheEntry(uint64_t index) +{ + if (!containsEntry(index)) + return nullptr; + + auto it = cache.find(index); + if (it == cache.end()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Index {} missing from cache while it should be present", index); + + return &it->second; +} + +const CacheEntry * LogEntryStorage::InMemoryCache::getCacheEntry(uint64_t index) const +{ + return const_cast(*this).getCacheEntry(index); +} + +PrefetchedCacheEntry & LogEntryStorage::InMemoryCache::getPrefetchedCacheEntry(uint64_t index) +{ + auto * cache_entry = getCacheEntry(index); + if (cache_entry == nullptr) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Missing expected index {} in cache", index); + + return std::get(*cache_entry); +} + + +LogEntryPtr LogEntryStorage::InMemoryCache::getEntry(uint64_t index) const +{ + const auto * cache_entry = getCacheEntry(index); + if (cache_entry == nullptr) + return nullptr; + + return getLogEntry(*cache_entry); +} + +void LogEntryStorage::InMemoryCache::cleanUpTo(uint64_t index) +{ + if (empty() || index <= min_index_in_cache) + return; + + if (index > max_index_in_cache) + { + cache.clear(); + cache_size = 0; + return; + } + + for (size_t i = min_index_in_cache; i < index; ++i) + { + auto it = cache.find(i); + if (it == cache.end()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Log entry with index {} unexpectedly missing from cache", i); + + cache_size -= logEntrySize(getLogEntry(it->second)); + cache.erase(it); + } + min_index_in_cache = index; +} + +void LogEntryStorage::InMemoryCache::cleanAfter(uint64_t index) +{ + if (empty() || index >= max_index_in_cache) + return; + + if (index < min_index_in_cache) + { + cache.clear(); + cache_size = 0; + return; + } + + for (size_t i = index + 1; i <= max_index_in_cache; ++i) + { + auto it = cache.find(i); + if (it == cache.end()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Log entry with index {} unexpectedly missing from cache", i); + + cache_size -= logEntrySize(getLogEntry(it->second)); + cache.erase(it); + } + + max_index_in_cache = index; +} + +void LogEntryStorage::InMemoryCache::clear() +{ + cache.clear(); + cache_size = 0; +} + +bool LogEntryStorage::InMemoryCache::empty() const +{ + return cache.empty(); +} + +size_t LogEntryStorage::InMemoryCache::numberOfEntries() const +{ + return cache.size(); +} + +bool LogEntryStorage::InMemoryCache::hasSpaceAvailable(size_t log_entry_size) const +{ + return size_threshold == 0 || empty() || cache_size + log_entry_size < size_threshold; +} + +void LogEntryStorage::addEntry(uint64_t index, const LogEntryPtr & log_entry) +{ + /// we update the cache for added entries on refreshCache call + latest_logs_cache.addEntry(index, logEntrySize(log_entry), log_entry); + + if (log_entry->get_val_type() == nuraft::conf) + { + latest_config = log_entry; + latest_config_index = index; + logs_with_config_changes.insert(index); + } + + updateTermInfoWithNewEntry(index, log_entry->get_term()); +} + +bool LogEntryStorage::shouldMoveLogToCommitCache(uint64_t index, size_t log_entry_size) +{ + /// if commit logs cache is empty, we need it only if it's the next log to commit + if (commit_logs_cache.empty()) + return keeper_context->lastCommittedIndex() + 1 == index; + + return commit_logs_cache.max_index_in_cache == index - 1 && commit_logs_cache.hasSpaceAvailable(log_entry_size); +} + +void LogEntryStorage::updateTermInfoWithNewEntry(uint64_t index, uint64_t term) +{ + if (!log_term_infos.empty() && log_term_infos.back().term == term) + return; + + log_term_infos.push_back(LogTermInfo{.term = term, .first_index = index}); +} + +void LogEntryStorage::addEntryWithLocation(uint64_t index, const LogEntryPtr & log_entry, LogLocation log_location) +{ + auto entry_size = logEntrySize(log_entry); + while (!latest_logs_cache.hasSpaceAvailable(entry_size)) + { + auto entry_handle = latest_logs_cache.popOldestEntry(); + size_t removed_entry_size = logEntrySize(getLogEntry(entry_handle.mapped())); + if (shouldMoveLogToCommitCache(entry_handle.key(), removed_entry_size)) + commit_logs_cache.addEntry(std::move(entry_handle)); + } + latest_logs_cache.addEntry(index, entry_size, CacheEntry(log_entry)); + + logs_location.emplace(index, std::move(log_location)); + + if (logs_location.size() == 1) + min_index_with_location = index; + + max_index_with_location = index; + + if (log_entry->get_val_type() == nuraft::conf) + { + latest_config = log_entry; + latest_config_index = index; + logs_with_config_changes.insert(index); + } + + updateTermInfoWithNewEntry(index, log_entry->get_term()); +} + +void LogEntryStorage::cleanUpTo(uint64_t index) +{ + latest_logs_cache.cleanUpTo(index); + + if (!logs_location.empty() && index > min_index_with_location) + { + if (index > max_index_with_location) + { + logs_location.clear(); + } + else + { + for (size_t i = min_index_with_location; i < index; ++i) + { + auto it = logs_location.find(i); + if (it == logs_location.end()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Log entry with index {} unexpectedly missing from logs location", i); + + logs_location.erase(it); + } + + min_index_with_location = index; + + } + } + + { + std::lock_guard lock(logs_location_mutex); + if (!unapplied_indices_with_log_locations.empty()) + { + auto last = std::ranges::lower_bound( + unapplied_indices_with_log_locations, + index, + std::ranges::less{}, + [](const auto & index_with_location) { return index_with_location.first; }); + + unapplied_indices_with_log_locations.erase(unapplied_indices_with_log_locations.begin(), last); + } + } + + /// uncommitted logs should be compacted only if we received snapshot from leader + if (current_prefetch_info && !current_prefetch_info->done) + { + auto [prefetch_from, prefetch_to] = current_prefetch_info->commit_prefetch_index_range; + /// if we will clean some logs that are currently prefetched, stop prefetching + /// and clean all logs from it + if (index > prefetch_from) + { + current_prefetch_info->cancel = true; + current_prefetch_info->done.wait(false); + commit_logs_cache.clear(); + } + + /// start prefetching logs for committing at the current index + /// the last log index in the snapshot should be the + /// last log we cleaned up + startCommitLogsPrefetch(index - 1); + } + else + commit_logs_cache.cleanUpTo(index); + + std::erase_if(logs_with_config_changes, [&](const auto conf_index) { return conf_index < index; }); + if (auto it = std::max_element(logs_with_config_changes.begin(), logs_with_config_changes.end()); it != logs_with_config_changes.end()) + { + latest_config_index = *it; + latest_config = getEntry(latest_config_index); + } + else + latest_config = nullptr; + + if (first_log_index < index) + first_log_entry = nullptr; + + /// remove all the term infos we don't need (all terms that start before index) + uint64_t last_removed_term = 0; + while (!log_term_infos.empty() && log_term_infos.front().first_index < index) + { + last_removed_term = log_term_infos.front().term; + log_term_infos.pop_front(); + } + + /// the last removed term info could contain terms for some indices we didn't cleanup + /// so we add the last removed term info back but with new first index + if (last_removed_term != 0 && (log_term_infos.empty() || log_term_infos.front().first_index > index)) + log_term_infos.push_front(LogTermInfo{.term = last_removed_term, .first_index = index}); +} + +void LogEntryStorage::cleanAfter(uint64_t index) +{ + latest_logs_cache.cleanAfter(index); + + if (!logs_location.empty() && index < max_index_with_location) + { + if (index < min_index_with_location) + { + logs_location.clear(); + } + else + { + for (size_t i = index + 1; i <= max_index_with_location; ++i) + { + auto it = logs_location.find(i); + if (it == logs_location.end()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Log entry with index {} unexpectedly missing from logs location", i); + + logs_location.erase(it); + } + + max_index_with_location = index; + } + } + + { + std::lock_guard lock(logs_location_mutex); + if (!unapplied_indices_with_log_locations.empty()) + { + auto first = std::ranges::upper_bound( + unapplied_indices_with_log_locations, + index, + std::ranges::less{}, + [](const auto & index_with_location) { return index_with_location.first; }); + + unapplied_indices_with_log_locations.erase(first, unapplied_indices_with_log_locations.end()); + } + } + + /// if we cleared all latest logs, there is a possibility we would need to clear commit logs + if (latest_logs_cache.empty()) + { + /// we will clean everything after the index, if there is a prefetch in progress + /// wait until we fetch everything until index + /// afterwards we can stop prefetching of newer logs because they will be cleaned up + commit_logs_cache.getEntry(index); + if (current_prefetch_info && !current_prefetch_info->done) + { + auto [prefetch_from, prefetch_to] = current_prefetch_info->commit_prefetch_index_range; + /// if we will clean some logs that are currently prefetched, stop prefetching + if (index < prefetch_to) + { + current_prefetch_info->cancel = true; + current_prefetch_info->done.wait(false); + } + } + + commit_logs_cache.cleanAfter(index); + startCommitLogsPrefetch(keeper_context->lastCommittedIndex()); + } + + if (empty() || first_log_index > index) + { + /// if we don't store any logs or if the first log index changed, reset first log cache + first_log_entry = nullptr; + } + + std::erase_if(logs_with_config_changes, [&](const auto conf_index) { return conf_index > index; }); + if (auto it = std::max_element(logs_with_config_changes.begin(), logs_with_config_changes.end()); it != logs_with_config_changes.end()) + { + latest_config_index = *it; + latest_config = getEntry(latest_config_index); + } + else + latest_config = nullptr; + + /// remove all the term infos we don't need (all terms that start after index) + while (!log_term_infos.empty() && log_term_infos.back().first_index > index) + log_term_infos.pop_back(); +} + +bool LogEntryStorage::contains(uint64_t index) const +{ + return logs_location.contains(index) || latest_logs_cache.containsEntry(index); +} + +LogEntryPtr LogEntryStorage::getEntry(uint64_t index) const +{ + auto last_committed_index = keeper_context->lastCommittedIndex(); + commit_logs_cache.cleanUpTo(last_committed_index); + startCommitLogsPrefetch(last_committed_index); + + LogEntryPtr entry = nullptr; + + if (latest_config != nullptr && index == latest_config_index) + return latest_config; + + if (first_log_entry != nullptr && index == first_log_index) + return first_log_entry; + + if (auto entry_from_latest_cache = latest_logs_cache.getEntry(index)) + { + ProfileEvents::increment(ProfileEvents::KeeperLogsEntryReadFromLatestCache); + return entry_from_latest_cache; + } + + if (auto entry_from_commit_cache = commit_logs_cache.getEntry(index)) + { + ProfileEvents::increment(ProfileEvents::KeeperLogsEntryReadFromCommitCache); + return entry_from_commit_cache; + } + + if (auto it = logs_location.find(index); it != logs_location.end()) + { + it->second.file_description->withLock( + [&] + { + const auto & [changelog_description, position, size] = it->second; + auto file = changelog_description->disk->readFile(changelog_description->path, ReadSettings()); + file->seek(position, SEEK_SET); + LOG_TRACE( + log, + "Reading log entry at index {} from path {}, position {}, size {}", + index, + changelog_description->path, + position, + size); + + auto record = readChangelogRecord(*file, changelog_description->path); + entry = logEntryFromRecord(record); + }); + + /// if we fetched the first log entry, we will cache it because it's often accessed + if (first_log_entry == nullptr && index == getFirstIndex()) + { + first_log_index = index; + first_log_entry = entry; + } + + ProfileEvents::increment(ProfileEvents::KeeperLogsEntryReadFromFile); + } + + return entry; +} + +void LogEntryStorage::clear() +{ + latest_logs_cache.clear(); + commit_logs_cache.clear(); + logs_location.clear(); +} + +LogEntryPtr LogEntryStorage::getLatestConfigChange() const +{ + return latest_config; +} + +uint64_t LogEntryStorage::termAt(uint64_t index) const +{ + uint64_t term_for_index = 0; + for (const auto [term, first_index] : log_term_infos) + { + if (index < first_index) + return term_for_index; + + term_for_index = term; + } + + return term_for_index; +} + +void LogEntryStorage::addLogLocations(std::vector> && indices_with_log_locations) +{ + /// if we have unlimited space in latest logs cache we don't need log location + if (latest_logs_cache.size_threshold == 0) + return; + + std::lock_guard lock(logs_location_mutex); + unapplied_indices_with_log_locations.insert( + unapplied_indices_with_log_locations.end(), + std::make_move_iterator(indices_with_log_locations.begin()), + std::make_move_iterator(indices_with_log_locations.end())); +} + +void LogEntryStorage::refreshCache() +{ + /// if we have unlimited space in latest logs cache we don't need log location + if (latest_logs_cache.size_threshold == 0) + return; + + std::vector new_unapplied_indices_with_log_locations; + { + std::lock_guard lock(logs_location_mutex); + new_unapplied_indices_with_log_locations.swap(unapplied_indices_with_log_locations); + } + + for (auto & [index, log_location] : new_unapplied_indices_with_log_locations) + { + if (logs_location.empty()) + min_index_with_location = index; + + logs_location.emplace(index, std::move(log_location)); + max_index_with_location = index; + } + + if (logs_location.empty()) + return; + + while (latest_logs_cache.numberOfEntries() > 1 && latest_logs_cache.min_index_in_cache <= max_index_with_location + && latest_logs_cache.cache_size > latest_logs_cache.size_threshold) + { + auto node = latest_logs_cache.popOldestEntry(); + auto log_entry_size = logEntrySize(getLogEntry(node.mapped())); + if (shouldMoveLogToCommitCache(node.key(), log_entry_size)) + commit_logs_cache.addEntry(std::move(node)); + } +} + +LogEntriesPtr LogEntryStorage::getLogEntriesBetween(uint64_t start, uint64_t end) const +{ + LogEntriesPtr ret = nuraft::cs_new>>(); + ret->reserve(end - start); + + /// we rely on fact that changelogs need to be written sequentially with + /// no other writes between + std::optional read_info; + const auto set_new_file = [&](const auto & log_location) + { + read_info.emplace(); + read_info->file_description = log_location.file_description; + read_info->position = log_location.position; + read_info->count = 1; + }; + + const auto flush_file = [&] + { + if (!read_info) + return; + + LOG_TRACE(log, "Reading from path {} {} entries", read_info->file_description->path, read_info->count); + read_info->file_description->withLock( + [&] + { + const auto & [file_description, start_position, count] = *read_info; + auto file = file_description->disk->readFile(file_description->path); + file->seek(start_position, SEEK_SET); + + for (size_t i = 0; i < count; ++i) + { + auto record = readChangelogRecord(*file, file_description->path); + ret->push_back(logEntryFromRecord(record)); + ProfileEvents::increment(ProfileEvents::KeeperLogsEntryReadFromFile); + } + }); + + read_info.reset(); + }; + + for (size_t i = start; i < end; ++i) + { + if (auto commit_cache_entry = commit_logs_cache.getEntry(i)) + { + flush_file(); + ret->push_back(std::move(commit_cache_entry)); + } + else if (auto latest_cache_entry = latest_logs_cache.getEntry(i)) + { + flush_file(); + ret->push_back(std::move(latest_cache_entry)); + } + else + { + const auto & log_location = logs_location.at(i); + + if (!read_info) + set_new_file(log_location); + else if (read_info->file_description == log_location.file_description) + ++read_info->count; + else + { + flush_file(); + set_new_file(log_location); + } + } + } + + flush_file(); + return ret; +} + +void LogEntryStorage::getKeeperLogInfo(KeeperLogInfo & log_info) const +{ + log_info.latest_logs_cache_entries = latest_logs_cache.numberOfEntries(); + log_info.latest_logs_cache_size = latest_logs_cache.cache_size; + + log_info.commit_logs_cache_entries = commit_logs_cache.numberOfEntries(); + log_info.commit_logs_cache_size = commit_logs_cache.cache_size; +} + +bool LogEntryStorage::isConfigLog(uint64_t index) const +{ + return logs_with_config_changes.contains(index); +} + +size_t LogEntryStorage::empty() const +{ + return logs_location.empty() && latest_logs_cache.empty(); +} + +size_t LogEntryStorage::size() const +{ + if (empty()) + return 0; + + size_t min_index = 0; + size_t max_index = 0; + + if (!logs_location.empty()) + { + min_index = min_index_with_location; + max_index = max_index_with_location; + } + else + min_index = latest_logs_cache.min_index_in_cache; + + if (!latest_logs_cache.empty()) + max_index = latest_logs_cache.max_index_in_cache; + + return max_index - min_index + 1; +} + +size_t LogEntryStorage::getFirstIndex() const +{ + if (!logs_location.empty()) + return min_index_with_location; + + if (!latest_logs_cache.empty()) + return latest_logs_cache.min_index_in_cache; + + return 0; +} + +void LogEntryStorage::shutdown() +{ + if (std::exchange(is_shutdown, true)) + return; + + if (!prefetch_queue.isFinished()) + prefetch_queue.finish(); + + if (current_prefetch_info) + { + current_prefetch_info->cancel = true; + current_prefetch_info->done.wait(false); + } + + if (commit_logs_prefetcher->joinable()) + commit_logs_prefetcher->join(); +} + Changelog::Changelog( LoggerPtr log_, LogFileSettings log_file_settings, FlushSettings flush_settings_, KeeperContextPtr keeper_context_) : changelogs_detached_dir("detached") , rotate_interval(log_file_settings.rotate_interval) , compress_logs(log_file_settings.compress_logs) , log(log_) + , entry_storage(log_file_settings, keeper_context_) , write_operations(std::numeric_limits::max()) , append_completion_queue(std::numeric_limits::max()) , keeper_context(std::move(keeper_context_)) , flush_settings(flush_settings_) { - if (auto latest_log_disk = getLatestLogDisk(); - log_file_settings.force_sync && dynamic_cast(latest_log_disk.get()) == nullptr) + try { - throw DB::Exception( - DB::ErrorCodes::BAD_ARGUMENTS, - "force_sync is set to true for logs but disk '{}' cannot satisfy such guarantee because it's not of type DiskLocal.\n" - "If you want to use force_sync and same disk for all logs, please set keeper_server.log_storage_disk to a local disk.\n" - "If you want to use force_sync and different disk only for old logs, please set 'keeper_server.log_storage_disk' to any " - "supported disk and 'keeper_server.latest_log_storage_disk' to a local disk.\n" - "Otherwise, disable force_sync", - latest_log_disk->getName()); - } - - /// Load all files on changelog disks - - std::unordered_set read_disks; - - const auto load_from_disk = [&](const auto & disk) - { - if (read_disks.contains(disk)) - return; - - LOG_TRACE(log, "Reading from disk {}", disk->getName()); - std::unordered_map incomplete_files; - - const auto clean_incomplete_file = [&](const auto & file_path) + if (auto latest_log_disk = getLatestLogDisk(); + log_file_settings.force_sync && dynamic_cast(latest_log_disk.get()) == nullptr) { - if (auto incomplete_it = incomplete_files.find(fs::path(file_path).filename()); incomplete_it != incomplete_files.end()) + throw DB::Exception( + DB::ErrorCodes::BAD_ARGUMENTS, + "force_sync is set to true for logs but disk '{}' cannot satisfy such guarantee because it's not of type DiskLocal.\n" + "If you want to use force_sync and same disk for all logs, please set keeper_server.log_storage_disk to a local disk.\n" + "If you want to use force_sync and different disk only for old logs, please set 'keeper_server.log_storage_disk' to any " + "supported disk and 'keeper_server.latest_log_storage_disk' to a local disk.\n" + "Otherwise, disable force_sync", + latest_log_disk->getName()); + } + + /// Load all files on changelog disks + + std::unordered_set read_disks; + + const auto load_from_disk = [&](const auto & disk) + { + if (read_disks.contains(disk)) + return; + + LOG_TRACE(log, "Reading from disk {}", disk->getName()); + std::unordered_map incomplete_files; + + const auto clean_incomplete_file = [&](const auto & file_path) { - LOG_TRACE(log, "Removing {} from {}", file_path, disk->getName()); - disk->removeFile(file_path); - disk->removeFile(incomplete_it->second); - incomplete_files.erase(incomplete_it); - return true; + if (auto incomplete_it = incomplete_files.find(fs::path(file_path).filename()); incomplete_it != incomplete_files.end()) + { + LOG_TRACE(log, "Removing {} from {}", file_path, disk->getName()); + disk->removeFile(file_path); + disk->removeFile(incomplete_it->second); + incomplete_files.erase(incomplete_it); + return true; + } + + return false; + }; + + std::vector changelog_files; + for (auto it = disk->iterateDirectory(""); it->isValid(); it->next()) + { + const auto & file_name = it->name(); + if (file_name == changelogs_detached_dir) + continue; + + if (file_name.starts_with(tmp_keeper_file_prefix)) + { + incomplete_files.emplace(file_name.substr(tmp_keeper_file_prefix.size()), it->path()); + continue; + } + + if (file_name.starts_with(DEFAULT_PREFIX)) + { + if (!clean_incomplete_file(it->path())) + changelog_files.push_back(it->path()); + } + else + { + LOG_WARNING(log, "Unknown file found in log directory: {}", file_name); + } } - return false; + for (const auto & changelog_file : changelog_files) + { + if (clean_incomplete_file(fs::path(changelog_file).filename())) + continue; + + auto file_description = getChangelogFileDescription(changelog_file); + file_description->disk = disk; + + LOG_TRACE(log, "Found {} on {}", changelog_file, disk->getName()); + auto [changelog_it, inserted] = existing_changelogs.insert_or_assign(file_description->from_log_index, std::move(file_description)); + + if (!inserted) + LOG_WARNING(log, "Found duplicate entries for {}, will use the entry from {}", changelog_it->second->path, disk->getName()); + } + + for (const auto & [name, path] : incomplete_files) + disk->removeFile(path); + + read_disks.insert(disk); }; - std::vector changelog_files; - for (auto it = disk->iterateDirectory(""); it->isValid(); it->next()) - { - const auto & file_name = it->name(); - if (file_name == changelogs_detached_dir) - continue; + /// Load all files from old disks + for (const auto & disk : keeper_context->getOldLogDisks()) + load_from_disk(disk); - if (file_name.starts_with(tmp_prefix)) - { - incomplete_files.emplace(file_name.substr(tmp_prefix.size()), it->path()); - continue; - } - - if (file_name.starts_with(DEFAULT_PREFIX)) - { - if (!clean_incomplete_file(it->path())) - changelog_files.push_back(it->path()); - } - else - { - LOG_WARNING(log, "Unknown file found in log directory: {}", file_name); - } - } - - for (const auto & changelog_file : changelog_files) - { - if (clean_incomplete_file(fs::path(changelog_file).filename())) - continue; - - auto file_description = getChangelogFileDescription(changelog_file); - file_description->disk = disk; - - LOG_TRACE(log, "Found {} on {}", changelog_file, disk->getName()); - auto [changelog_it, inserted] = existing_changelogs.insert_or_assign(file_description->from_log_index, std::move(file_description)); - - if (!inserted) - LOG_WARNING(log, "Found duplicate entries for {}, will use the entry from {}", changelog_it->second->path, disk->getName()); - } - - for (const auto & [name, path] : incomplete_files) - disk->removeFile(path); - - read_disks.insert(disk); - }; - - /// Load all files from old disks - for (const auto & disk : keeper_context->getOldLogDisks()) + auto disk = getDisk(); load_from_disk(disk); - auto disk = getDisk(); - load_from_disk(disk); + auto latest_log_disk = getLatestLogDisk(); + if (disk != latest_log_disk) + load_from_disk(latest_log_disk); - auto latest_log_disk = getLatestLogDisk(); - if (disk != latest_log_disk) - load_from_disk(latest_log_disk); + if (existing_changelogs.empty()) + LOG_WARNING(log, "No logs exists in {}. It's Ok if it's the first run of clickhouse-keeper.", disk->getPath()); - if (existing_changelogs.empty()) - LOG_WARNING(log, "No logs exists in {}. It's Ok if it's the first run of clickhouse-keeper.", disk->getPath()); + clean_log_thread = std::make_unique([this] { cleanLogThread(); }); - clean_log_thread = ThreadFromGlobalPool([this] { cleanLogThread(); }); + write_thread = std::make_unique([this] { writeThread(); }); - write_thread = ThreadFromGlobalPool([this] { writeThread(); }); + append_completion_thread = std::make_unique([this] { appendCompletionThread(); }); - append_completion_thread = ThreadFromGlobalPool([this] { appendCompletionThread(); }); - - current_writer = std::make_unique(existing_changelogs, keeper_context, log_file_settings); + current_writer = std::make_unique(existing_changelogs, entry_storage, keeper_context, log_file_settings); + } + catch (...) + { + tryLogCurrentException(log); + throw; + } } void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uint64_t logs_to_keep) +try { std::lock_guard writer_lock(writer_mutex); std::optional last_log_read_result; @@ -751,7 +1673,6 @@ void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uin changelog_description.from_log_index); /// Nothing to do with our more fresh log, leader will overwrite them, so remove everything and just start from last_commited_index removeAllLogs(); - min_log_id = last_commited_log_index; max_log_id = last_commited_log_index == 0 ? 0 : last_commited_log_index - 1; current_writer->rotate(max_log_id + 1); initialized = true; @@ -783,18 +1704,14 @@ void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uin break; } - ChangelogReader reader(changelog_description.disk, changelog_description.path); - last_log_read_result = reader.readChangelog(logs, start_to_read_from, log); + ChangelogReader reader(changelog_description_ptr); + last_log_read_result = reader.readChangelog(entry_storage, start_to_read_from, log); if (last_log_read_result->last_read_index != 0) last_read_index = last_log_read_result->last_read_index; last_log_read_result->log_start_index = changelog_description.from_log_index; - /// Otherwise we have already initialized it - if (min_log_id == 0) - min_log_id = last_log_read_result->first_read_index; - if (last_log_read_result->last_read_index != 0) max_log_id = last_log_read_result->last_read_index; @@ -813,16 +1730,14 @@ void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uin auto disk = getDisk(); if (latest_log_disk != disk && latest_log_disk == description->disk) - moveFileBetweenDisks(latest_log_disk, description, disk, description->path); + moveChangelogBetweenDisks(latest_log_disk, description, disk, description->path, keeper_context); }; /// we can have empty log (with zero entries) and last_log_read_result will be initialized - if (!last_log_read_result || min_log_id == 0) /// We just may have no logs (only snapshot or nothing) + if (!last_log_read_result || entry_storage.empty()) /// We just may have no logs (only snapshot or nothing) { /// Just to be sure they don't exist removeAllLogs(); - - min_log_id = last_commited_log_index; max_log_id = last_commited_log_index == 0 ? 0 : last_commited_log_index - 1; } else if (last_commited_log_index != 0 && max_log_id < last_commited_log_index - 1) /// If we have more fresh snapshot than our logs @@ -834,7 +1749,6 @@ void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uin last_commited_log_index - 1); removeAllLogs(); - min_log_id = last_commited_log_index; max_log_id = last_commited_log_index - 1; } else if (last_log_is_not_complete) /// if it's complete just start new one @@ -861,13 +1775,13 @@ void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uin remove_invalid_logs(); description->disk->removeFile(description->path); existing_changelogs.erase(last_log_read_result->log_start_index); - std::erase_if(logs, [last_log_read_result](const auto & item) { return item.first >= last_log_read_result->log_start_index; }); + entry_storage.cleanAfter(last_log_read_result->log_start_index - 1); } else if (last_log_read_result->error) { - LOG_INFO(log, "Chagelog {} read finished with error but some logs were read from it, file will not be removed", description->path); + LOG_INFO(log, "Changelog {} read finished with error but some logs were read from it, file will not be removed", description->path); remove_invalid_logs(); - std::erase_if(logs, [last_log_read_result](const auto & item) { return item.first > last_log_read_result->last_read_index; }); + entry_storage.cleanAfter(last_log_read_result->last_read_index); move_from_latest_logs_disks(existing_changelogs.at(last_log_read_result->log_start_index)); } /// don't mix compressed and uncompressed writes @@ -899,12 +1813,15 @@ void Changelog::readChangelogAndInitWriter(uint64_t last_commited_log_index, uin } if (description->disk != disk) - moveFileBetweenDisks(description->disk, description, disk, description->path); + moveChangelogBetweenDisks(description->disk, description, disk, description->path, keeper_context); } - initialized = true; } +catch (...) +{ + tryLogCurrentException(__PRETTY_FUNCTION__); +} void Changelog::initWriter(ChangelogFileDescriptionPtr description) @@ -921,7 +1838,7 @@ void Changelog::initWriter(ChangelogFileDescriptionPtr description) auto log_disk = description->disk; auto latest_log_disk = getLatestLogDisk(); if (log_disk != latest_log_disk) - moveFileBetweenDisks(log_disk, description, latest_log_disk, description->path); + moveChangelogBetweenDisks(log_disk, description, latest_log_disk, description->path, keeper_context); current_writer->setFile(std::move(description), WriteMode::Append); } @@ -984,11 +1901,11 @@ void Changelog::removeExistingLogs(ChangelogIter begin, ChangelogIter end) catch (const DB::Exception & e) { if (e.code() == DB::ErrorCodes::NOT_IMPLEMENTED) - moveFileBetweenDisks(changelog_disk, changelog_description, disk, new_path); + moveChangelogBetweenDisks(changelog_disk, changelog_description, disk, new_path, keeper_context); } } else - moveFileBetweenDisks(changelog_disk, changelog_description, disk, new_path); + moveChangelogBetweenDisks(changelog_disk, changelog_description, disk, new_path, keeper_context); itr = existing_changelogs.erase(itr); } @@ -1006,14 +1923,14 @@ void Changelog::removeAllLogsAfter(uint64_t remove_after_log_start_index) LOG_WARNING(log, "Removing changelogs that go after broken changelog entry"); removeExistingLogs(start_to_remove_from_itr, existing_changelogs.end()); - std::erase_if(logs, [start_to_remove_from_log_id](const auto & item) { return item.first >= start_to_remove_from_log_id; }); + entry_storage.cleanAfter(start_to_remove_from_log_id - 1); } void Changelog::removeAllLogs() { LOG_WARNING(log, "Removing all changelogs"); removeExistingLogs(existing_changelogs.begin(), existing_changelogs.end()); - logs.clear(); + entry_storage.clear(); } ChangelogRecord Changelog::buildRecord(uint64_t index, const LogEntryPtr & log_entry) @@ -1045,7 +1962,7 @@ void Changelog::appendCompletionThread() if (auto raft_server_locked = raft_server.lock()) raft_server_locked->notify_log_append_completion(append_ok); else - LOG_WARNING(log, "Raft server is not set in LogStore."); + LOG_INFO(log, "Raft server is not set in LogStore."); } } @@ -1085,70 +2002,78 @@ void Changelog::writeThread() LOG_WARNING(log, "Changelog is shut down"); }; - /// NuRaft writes a batch of request by first calling multiple store requests, i.e. AppendLog - /// finished by a flush request - /// We assume that after some number of appends, we always get flush request - while (true) + try { - if (try_batch_flush) + /// NuRaft writes a batch of request by first calling multiple store requests, i.e. AppendLog + /// finished by a flush request + /// We assume that after some number of appends, we always get flush request + while (true) { - try_batch_flush = false; - /// we have Flush request stored in write operation - /// but we try to get new append operations - /// if there are none, we apply the currently set Flush - chassert(std::holds_alternative(write_operation)); - if (!write_operations.tryPop(write_operation)) + if (try_batch_flush) { - chassert(batch_append_ok); - const auto & flush = std::get(write_operation); - flush_logs(flush); - notify_append_completion(); - if (!write_operations.pop(write_operation)) - break; - } - } - else if (!write_operations.pop(write_operation)) - { - break; - } - - assert(initialized); - - if (auto * append_log = std::get_if(&write_operation)) - { - if (!batch_append_ok) - continue; - - std::lock_guard writer_lock(writer_mutex); - assert(current_writer); - - batch_append_ok = current_writer->appendRecord(buildRecord(append_log->index, append_log->log_entry)); - ++pending_appends; - } - else - { - const auto & flush = std::get(write_operation); - - if (batch_append_ok) - { - /// we can try batching more logs for flush - if (pending_appends < flush_settings.max_flush_batch_size) + try_batch_flush = false; + /// we have Flush request stored in write operation + /// but we try to get new append operations + /// if there are none, we apply the currently set Flush + chassert(std::holds_alternative(write_operation)); + if (!write_operations.tryPop(write_operation)) { - try_batch_flush = true; - continue; + chassert(batch_append_ok); + const auto & flush = std::get(write_operation); + flush_logs(flush); + notify_append_completion(); + if (!write_operations.pop(write_operation)) + break; } - /// we need to flush because we have maximum allowed pending records - flush_logs(flush); + } + else if (!write_operations.pop(write_operation)) + { + break; + } + + assert(initialized); + + if (auto * append_log = std::get_if(&write_operation)) + { + if (!batch_append_ok) + continue; + + std::lock_guard writer_lock(writer_mutex); + assert(current_writer); + + batch_append_ok = current_writer->appendRecord(buildRecord(append_log->index, append_log->log_entry)); + ++pending_appends; } else { - std::lock_guard lock{durable_idx_mutex}; - *flush.failed = true; + const auto & flush = std::get(write_operation); + + if (batch_append_ok) + { + /// we can try batching more logs for flush + if (pending_appends < flush_settings.max_flush_batch_size) + { + try_batch_flush = true; + continue; + } + /// we need to flush because we have maximum allowed pending records + flush_logs(flush); + } + else + { + std::lock_guard lock{durable_idx_mutex}; + *flush.failed = true; + } + notify_append_completion(); + batch_append_ok = true; } - notify_append_completion(); - batch_append_ok = true; } } + catch (...) + { + tryLogCurrentException(log, "Write thread failed, aborting"); + std::abort(); + } } @@ -1157,10 +2082,7 @@ void Changelog::appendEntry(uint64_t index, const LogEntryPtr & log_entry) if (!initialized) throw Exception(ErrorCodes::LOGICAL_ERROR, "Changelog must be initialized before appending records"); - if (logs.empty()) - min_log_id = index; - - logs[index] = log_entry; + entry_storage.addEntry(index, log_entry); max_log_id = index; if (!write_operations.push(AppendLog{index, log_entry})) @@ -1191,7 +2113,7 @@ void Changelog::writeAt(uint64_t index, const LogEntryPtr & log_entry) auto log_disk = description->disk; auto latest_log_disk = getLatestLogDisk(); if (log_disk != latest_log_disk) - moveFileBetweenDisks(log_disk, description, latest_log_disk, description->path); + moveChangelogBetweenDisks(log_disk, description, latest_log_disk, description->path, keeper_context); current_writer->setFile(std::move(description), WriteMode::Append); @@ -1207,7 +2129,7 @@ void Changelog::writeAt(uint64_t index, const LogEntryPtr & log_entry) /// Remove redundant logs from memory /// Everything >= index must be removed - std::erase_if(logs, [index](const auto & item) { return item.first >= index; }); + entry_storage.cleanAfter(index - 1); /// Now we can actually override entry at index appendEntry(index, log_entry); @@ -1274,14 +2196,23 @@ void Changelog::compact(uint64_t up_to_log_index) else /// Files are ordered, so all subsequent should exist break; } - /// Compaction from the past is possible, so don't make our min_log_id smaller. - min_log_id = std::max(min_log_id, up_to_log_index + 1); - std::erase_if(logs, [up_to_log_index](const auto & item) { return item.first <= up_to_log_index; }); + + entry_storage.cleanUpTo(up_to_log_index + 1); if (need_rotate) current_writer->rotate(up_to_log_index + 1); - LOG_INFO(log, "Compaction up to {} finished new min index {}, new max index {}", up_to_log_index, min_log_id, max_log_id); + LOG_INFO(log, "Compaction up to {} finished new min index {}, new max index {}", up_to_log_index, getStartIndex(), max_log_id); +} + +uint64_t Changelog::getNextEntryIndex() const +{ + return max_log_id + 1; +} + +uint64_t Changelog::getStartIndex() const +{ + return entry_storage.empty() ? max_log_id + 1 : entry_storage.getFirstIndex(); } LogEntryPtr Changelog::getLastEntry() const @@ -1289,46 +2220,26 @@ LogEntryPtr Changelog::getLastEntry() const /// This entry treaded in special way by NuRaft static LogEntryPtr fake_entry = nuraft::cs_new(0, nuraft::buffer::alloc(sizeof(uint64_t))); - auto entry = logs.find(max_log_id); - if (entry == logs.end()) - { + auto entry = entry_storage.getEntry(max_log_id); + if (entry == nullptr) return fake_entry; - } - return entry->second; + return entry; } LogEntriesPtr Changelog::getLogEntriesBetween(uint64_t start, uint64_t end) { - LogEntriesPtr ret = nuraft::cs_new>>(); - - ret->resize(end - start); - uint64_t result_pos = 0; - for (uint64_t i = start; i < end; ++i) - { - (*ret)[result_pos] = entryAt(i); - result_pos++; - } - return ret; + return entry_storage.getLogEntriesBetween(start, end); } -LogEntryPtr Changelog::entryAt(uint64_t index) +LogEntryPtr Changelog::entryAt(uint64_t index) const { - nuraft::ptr src = nullptr; - auto entry = logs.find(index); - if (entry == logs.end()) - return nullptr; - - src = entry->second; - return src; + return entry_storage.getEntry(index); } LogEntryPtr Changelog::getLatestConfigChange() const { - for (const auto & [_, entry] : logs) - if (entry->get_val_type() == nuraft::conf) - return entry; - return nullptr; + return entry_storage.getLatestConfigChange(); } nuraft::ptr Changelog::serializeEntriesToBuffer(uint64_t index, int32_t count) @@ -1339,11 +2250,11 @@ nuraft::ptr Changelog::serializeEntriesToBuffer(uint64_t index, uint64_t size_total = 0; for (uint64_t i = index; i < index + count; ++i) { - auto entry = logs.find(i); - if (entry == logs.end()) + auto entry = entry_storage.getEntry(i); + if (entry == nullptr) throw Exception(ErrorCodes::LOGICAL_ERROR, "Don't have log entry {}", i); - nuraft::ptr buf = entry->second->serialize(); + nuraft::ptr buf = entry->serialize(); size_total += buf->size(); returned_logs.push_back(std::move(buf)); } @@ -1374,13 +2285,23 @@ void Changelog::applyEntriesFromBuffer(uint64_t index, nuraft::buffer & buffer) buffer.get(buf_local); LogEntryPtr log_entry = nuraft::log_entry::deserialize(*buf_local); - if (i == 0 && logs.contains(cur_index)) + if (i == 0 && cur_index >= entry_storage.getFirstIndex() && cur_index <= max_log_id) writeAt(cur_index, log_entry); else appendEntry(cur_index, log_entry); } } +bool Changelog::isConfigLog(uint64_t index) const +{ + return entry_storage.isConfigLog(index); +} + +uint64_t Changelog::termAt(uint64_t index) const +{ + return entry_storage.termAt(index); +} + bool Changelog::flush() { if (auto failed_ptr = flushAsync()) @@ -1406,37 +2327,47 @@ std::shared_ptr Changelog::flushAsync() if (!pushed) { - LOG_WARNING(log, "Changelog is shut down"); + LOG_INFO(log, "Changelog is shut down"); return nullptr; } + + entry_storage.refreshCache(); return failed; } +uint64_t Changelog::size() const +{ + return entry_storage.size(); +} + void Changelog::shutdown() { + LOG_DEBUG(log, "Shutting down Changelog"); if (!log_files_to_delete_queue.isFinished()) log_files_to_delete_queue.finish(); - if (clean_log_thread.joinable()) - clean_log_thread.join(); + if (clean_log_thread->joinable()) + clean_log_thread->join(); if (!write_operations.isFinished()) write_operations.finish(); - if (write_thread.joinable()) - write_thread.join(); + if (write_thread->joinable()) + write_thread->join(); if (!append_completion_queue.isFinished()) append_completion_queue.finish(); - if (append_completion_thread.joinable()) - append_completion_thread.join(); + if (append_completion_thread->joinable()) + append_completion_thread->join(); if (current_writer) { current_writer->finalize(); current_writer.reset(); } + + entry_storage.shutdown(); } Changelog::~Changelog() @@ -1485,4 +2416,18 @@ bool Changelog::isInitialized() const return initialized; } +void Changelog::getKeeperLogInfo(KeeperLogInfo & log_info) const +{ + if (!entry_storage.empty()) + { + log_info.first_log_idx = getStartIndex(); + log_info.first_log_term = termAt(log_info.first_log_idx); + + log_info.last_log_idx = max_log_id; + log_info.last_log_term = termAt(log_info.last_log_idx); + } + + entry_storage.getKeeperLogInfo(log_info); +} + } diff --git a/src/Coordination/Changelog.h b/src/Coordination/Changelog.h index 612c68ab733..2e8dbe75e90 100644 --- a/src/Coordination/Changelog.h +++ b/src/Coordination/Changelog.h @@ -1,17 +1,26 @@ #pragma once -#include -#include -#include -#include -#include -#include -#include -#include -#include +#include +#include #include -#include -#include + +#include +#include +#include + +namespace nuraft +{ + struct log_entry; + struct buffer; + struct raft_server; +} + +namespace Poco +{ + class Logger; +} + +using LoggerPtr = std::shared_ptr; namespace DB { @@ -23,8 +32,11 @@ using LogEntries = std::vector; using LogEntriesPtr = nuraft::ptr; using BufferPtr = nuraft::ptr; -using IndexToOffset = std::unordered_map; -using IndexToLogEntry = std::unordered_map; +struct KeeperLogInfo; +class KeeperContext; +using KeeperContextPtr = std::shared_ptr; +class IDisk; +using DiskPtr = std::shared_ptr; enum class ChangelogVersion : uint8_t { @@ -63,10 +75,19 @@ struct ChangelogFileDescription DiskPtr disk; std::string path; + std::mutex file_mutex; + bool deleted = false; /// How many entries should be stored in this log uint64_t expectedEntriesCountInLog() const { return to_log_index - from_log_index + 1; } + + template + void withLock(TFunction && fn) + { + std::lock_guard lock(file_mutex); + fn(); + } }; using ChangelogFileDescriptionPtr = std::shared_ptr; @@ -80,6 +101,8 @@ struct LogFileSettings uint64_t rotate_interval = 100000; uint64_t max_size = 0; uint64_t overallocate_size = 0; + uint64_t latest_logs_cache_size_threshold = 0; + uint64_t commit_logs_cache_size_threshold = 0; }; struct FlushSettings @@ -87,6 +110,191 @@ struct FlushSettings uint64_t max_flush_batch_size = 1000; }; +struct LogLocation +{ + ChangelogFileDescriptionPtr file_description; + size_t position; + size_t size; +}; + +struct PrefetchedCacheEntry +{ + explicit PrefetchedCacheEntry(); + + const LogEntryPtr & getLogEntry() const; + void resolve(std::exception_ptr exception); + void resolve(LogEntryPtr log_entry_); +private: + std::promise log_entry_resolver; + mutable std::shared_future log_entry; +}; + +using CacheEntry = std::variant; +using IndexToCacheEntry = std::unordered_map; +using IndexToCacheEntryNode = typename IndexToCacheEntry::node_type; + +/** + * Storage for storing and handling deserialized entries from disk. + * It consists of 2 in-memory caches that rely heavily on the way + * entries are used in Raft. + * Random and repeated access to certain entries is almost never done so we can't implement a solution + * like LRU/SLRU cache because entries would be cached and never read again. + * Entries are often read sequentially for 2 cases: + * - for replication + * - for committing + * + * First cache will store latest logs in memory, limited by the latest_logs_cache_size_threshold coordination setting. + * Once the log is persisted to the disk, we store it's location in the file and allow the storage + * to evict that log from cache if it's needed. + * Latest logs cache should have a high hit rate in "normal" operation for both replication and committing. + * + * As we commit (and read) logs sequentially, we will try to read from latest logs cache. + * In some cases, latest logs could be ahead from last committed log by more than latest_logs_cache_size_threshold + * which means that for each commit we would need to read the log from disk. + * In case latest logs cache hits the threshold we have a second cache called commit logs cache limited by commit_logs_cache_size_threshold. + * If a log is evicted from the latest logs cache, we check if we can move it to commit logs cache to avoid re-reading the log from disk. + * If latest logs cache moves ahead of the commit log by a lot or commit log hits the threshold + * we cannot move the entries from latest logs and we will need to refill the commit cache from disk. + * To avoid reading entry by entry (which can have really bad effect on performance because we support disks based on S3), + * we try to prefetch multiple entries ahead of time because we know that they will be read by commit thread + * in the future. + * Commit logs cache should have a high hit rate if we start with a lot of unprocessed logs that cannot fit in the + * latest logs cache. + */ +struct LogEntryStorage +{ + LogEntryStorage(const LogFileSettings & log_settings, KeeperContextPtr keeper_context_); + + ~LogEntryStorage(); + + void addEntry(uint64_t index, const LogEntryPtr & log_entry); + void addEntryWithLocation(uint64_t index, const LogEntryPtr & log_entry, LogLocation log_location); + /// clean all logs up to (but not including) index + void cleanUpTo(uint64_t index); + /// clean all logs after (but not including) index + void cleanAfter(uint64_t index); + bool contains(uint64_t index) const; + LogEntryPtr getEntry(uint64_t index) const; + void clear(); + LogEntryPtr getLatestConfigChange() const; + uint64_t termAt(uint64_t index) const; + + using IndexWithLogLocation = std::pair; + + void addLogLocations(std::vector && indices_with_log_locations); + + void refreshCache(); + + LogEntriesPtr getLogEntriesBetween(uint64_t start, uint64_t end) const; + + void getKeeperLogInfo(KeeperLogInfo & log_info) const; + + bool isConfigLog(uint64_t index) const; + + size_t empty() const; + size_t size() const; + size_t getFirstIndex() const; + + void shutdown(); +private: + void prefetchCommitLogs(); + + void startCommitLogsPrefetch(uint64_t last_committed_index) const; + + bool shouldMoveLogToCommitCache(uint64_t index, size_t log_entry_size); + + void updateTermInfoWithNewEntry(uint64_t index, uint64_t term); + + struct InMemoryCache + { + explicit InMemoryCache(size_t size_threshold_); + + void addEntry(uint64_t index, size_t size, CacheEntry log_entry); + void addEntry(IndexToCacheEntryNode && node); + + void updateStatsWithNewEntry(uint64_t index, size_t size); + + IndexToCacheEntryNode popOldestEntry(); + + bool containsEntry(uint64_t index) const; + + LogEntryPtr getEntry(uint64_t index) const; + + CacheEntry * getCacheEntry(uint64_t index); + const CacheEntry * getCacheEntry(uint64_t index) const; + PrefetchedCacheEntry & getPrefetchedCacheEntry(uint64_t index); + + void cleanUpTo(uint64_t index); + void cleanAfter(uint64_t index); + + bool empty() const; + size_t numberOfEntries() const; + bool hasSpaceAvailable(size_t log_entry_size) const; + void clear(); + + /// Mapping log_id -> log_entry + mutable IndexToCacheEntry cache; + size_t cache_size = 0; + size_t min_index_in_cache = 0; + size_t max_index_in_cache = 0; + + const size_t size_threshold; + }; + + InMemoryCache latest_logs_cache; + mutable InMemoryCache commit_logs_cache; + + LogEntryPtr latest_config; + uint64_t latest_config_index = 0; + + mutable LogEntryPtr first_log_entry; + mutable uint64_t first_log_index = 0; + + std::unique_ptr commit_logs_prefetcher; + + struct FileReadInfo + { + ChangelogFileDescriptionPtr file_description; + size_t position; + size_t count; + }; + + struct PrefetchInfo + { + std::vector file_infos; + std::pair commit_prefetch_index_range; + std::atomic cancel; + std::atomic done = false; + }; + + mutable ConcurrentBoundedQueue> prefetch_queue; + mutable std::shared_ptr current_prefetch_info; + + mutable std::mutex logs_location_mutex; + std::vector unapplied_indices_with_log_locations; + std::unordered_map logs_location; + size_t max_index_with_location = 0; + size_t min_index_with_location = 0; + + /// store indices of logs that contain config changes + std::unordered_set logs_with_config_changes; + + struct LogTermInfo + { + uint64_t term = 0; + uint64_t first_index = 0; + }; + + /// store first index of each term + /// so we don't have to fetch log to return that information + /// terms are monotonically increasing so first index is enough + std::deque log_term_infos; + + bool is_shutdown = false; + KeeperContextPtr keeper_context; + LoggerPtr log; +}; + /// Simplest changelog with files rotation. /// No compression, no metadata, just entries with headers one by one. /// Able to read broken files/entries and discard them. Not thread safe. @@ -114,9 +322,9 @@ public: /// Remove log files with to_log_index <= up_to_log_index. void compact(uint64_t up_to_log_index); - uint64_t getNextEntryIndex() const { return max_log_id + 1; } + uint64_t getNextEntryIndex() const; - uint64_t getStartIndex() const { return min_log_id; } + uint64_t getStartIndex() const; /// Last entry in log, or fake entry with term 0 if log is empty LogEntryPtr getLastEntry() const; @@ -128,7 +336,7 @@ public: LogEntriesPtr getLogEntriesBetween(uint64_t start_index, uint64_t end_index); /// Return entry at position index - LogEntryPtr entryAt(uint64_t index); + LogEntryPtr entryAt(uint64_t index) const; /// Serialize entries from index into buffer BufferPtr serializeEntriesToBuffer(uint64_t index, int32_t count); @@ -136,6 +344,9 @@ public: /// Apply entries from buffer overriding existing entries void applyEntriesFromBuffer(uint64_t index, nuraft::buffer & buffer); + bool isConfigLog(uint64_t index) const; + uint64_t termAt(uint64_t index) const; + /// Fsync latest log to disk and flush buffer bool flush(); @@ -143,7 +354,7 @@ public: void shutdown(); - uint64_t size() const { return logs.size(); } + uint64_t size() const; uint64_t lastDurableIndex() const { @@ -155,6 +366,8 @@ public: bool isInitialized() const; + void getKeeperLogInfo(KeeperLogInfo & log_info) const; + /// Fsync log to disk ~Changelog(); @@ -190,16 +403,14 @@ private: std::mutex writer_mutex; /// Current writer for changelog file std::unique_ptr current_writer; - /// Mapping log_id -> log_entry - IndexToLogEntry logs; - /// Start log_id which exists in all "active" logs - /// min_log_id + 1 == max_log_id means empty log storage for NuRaft - uint64_t min_log_id = 0; + + LogEntryStorage entry_storage; + uint64_t max_log_id = 0; /// For compaction, queue of delete not used logs /// 128 is enough, even if log is not removed, it's not a problem ConcurrentBoundedQueue> log_files_to_delete_queue{128}; - ThreadFromGlobalPool clean_log_thread; + std::unique_ptr clean_log_thread; struct AppendLog { @@ -217,7 +428,7 @@ private: void writeThread(); - ThreadFromGlobalPool write_thread; + std::unique_ptr write_thread; ConcurrentBoundedQueue write_operations; /// Append log completion callback tries to acquire NuRaft's global lock @@ -226,7 +437,7 @@ private: /// For those reasons we call the completion callback in a different thread void appendCompletionThread(); - ThreadFromGlobalPool append_completion_thread; + std::unique_ptr append_completion_thread; ConcurrentBoundedQueue append_completion_queue; // last_durable_index needs to be exposed through const getter so we make mutex mutable diff --git a/src/Coordination/CoordinationSettings.cpp b/src/Coordination/CoordinationSettings.cpp index 2436d730ae4..ea1acf02450 100644 --- a/src/Coordination/CoordinationSettings.cpp +++ b/src/Coordination/CoordinationSettings.cpp @@ -34,6 +34,11 @@ void CoordinationSettings::loadFromConfig(const String & config_elem, const Poco e.addMessage("in Coordination settings config"); throw; } + + /// for backwards compatibility we set max_requests_append_size to max_requests_batch_size + /// if max_requests_append_size was not changed + if (!max_requests_append_size.changed) + max_requests_append_size = max_requests_batch_size; } @@ -41,7 +46,7 @@ const String KeeperConfigurationAndSettings::DEFAULT_FOUR_LETTER_WORD_CMD = #if USE_JEMALLOC "jmst,jmfp,jmep,jmdp," #endif -"conf,cons,crst,envi,ruok,srst,srvr,stat,wchs,dirs,mntr,isro,rcvr,apiv,csnp,lgif,rqld,rclc,clrs,ftfl,ydld"; +"conf,cons,crst,envi,ruok,srst,srvr,stat,wchs,dirs,mntr,isro,rcvr,apiv,csnp,lgif,rqld,rclc,clrs,ftfl,ydld,pfev"; KeeperConfigurationAndSettings::KeeperConfigurationAndSettings() : server_id(NOT_EXIST) diff --git a/src/Coordination/CoordinationSettings.h b/src/Coordination/CoordinationSettings.h index a58f2b04797..ed0490681b7 100644 --- a/src/Coordination/CoordinationSettings.h +++ b/src/Coordination/CoordinationSettings.h @@ -41,6 +41,7 @@ struct Settings; M(UInt64, max_request_queue_size, 100000, "Maximum number of request that can be in queue for processing", 0) \ M(UInt64, max_requests_batch_size, 100, "Max size of batch of requests that can be sent to RAFT", 0) \ M(UInt64, max_requests_batch_bytes_size, 100*1024, "Max size in bytes of batch of requests that can be sent to RAFT", 0) \ + M(UInt64, max_requests_append_size, 100, "Max size of batch of requests that can be sent to replica in append request", 0) \ M(UInt64, max_flush_batch_size, 1000, "Max size of batch of requests that can be flushed together", 0) \ M(UInt64, max_requests_quick_batch_size, 100, "Max size of batch of requests to try to get before proceeding with RAFT. Keeper will not wait for requests but take only requests that are already in queue" , 0) \ M(Bool, quorum_reads, false, "Execute read requests as writes through whole RAFT consesus with similar speed", 0) \ @@ -52,7 +53,11 @@ struct Settings; M(UInt64, log_file_overallocate_size, 50 * 1024 * 1024, "If max_log_file_size is not set to 0, this value will be added to it for preallocating bytes on disk. If a log record is larger than this value, it could lead to uncaught out-of-space issues so a larger value is preferred", 0) \ M(UInt64, min_request_size_for_cache, 50 * 1024, "Minimal size of the request to cache the deserialization result. Caching can have negative effect on latency for smaller requests, set to 0 to disable", 0) \ M(UInt64, raft_limits_reconnect_limit, 50, "If connection to a peer is silent longer than this limit * (multiplied by heartbeat interval), we re-establish the connection.", 0) \ - M(Bool, async_replication, false, "Enable async replication. All write and read guarantees are preserved while better performance is achieved. Settings is disabled by default to not break backwards compatibility.", 0) + M(Bool, async_replication, false, "Enable async replication. All write and read guarantees are preserved while better performance is achieved. Settings is disabled by default to not break backwards compatibility.", 0) \ + M(UInt64, latest_logs_cache_size_threshold, 1 * 1024 * 1024 * 1024, "Maximum total size of in-memory cache of latest log entries.", 0) \ + M(UInt64, commit_logs_cache_size_threshold, 500 * 1024 * 1024, "Maximum total size of in-memory cache of log entries needed next for commit.", 0) \ + M(UInt64, disk_move_retries_wait_ms, 1000, "How long to wait between retries after a failure which happened while a file was being moved between disks.", 0) \ + M(UInt64, disk_move_retries_during_init, 100, "The amount of retries after a failure which happened while a file was being moved between disks during initialization.", 0) DECLARE_SETTINGS_TRAITS(CoordinationSettingsTraits, LIST_OF_COORDINATION_SETTINGS) diff --git a/src/Coordination/FourLetterCommand.cpp b/src/Coordination/FourLetterCommand.cpp index 4862acd448f..d7fa5abe742 100644 --- a/src/Coordination/FourLetterCommand.cpp +++ b/src/Coordination/FourLetterCommand.cpp @@ -9,6 +9,7 @@ #include #include #include +#include #include "Coordination/KeeperFeatureFlags.h" #include #include @@ -37,6 +38,12 @@ String formatZxid(int64_t zxid) } +#if USE_NURAFT +namespace ProfileEvents +{ + extern const std::vector keeper_profile_events; +} +#endif namespace DB { @@ -193,6 +200,8 @@ void FourLetterCommandFactory::registerCommands(KeeperDispatcher & keeper_dispat FourLetterCommandPtr jemalloc_disable_profile = std::make_shared(keeper_dispatcher); factory.registerCommand(jemalloc_disable_profile); #endif + FourLetterCommandPtr profile_events_command = std::make_shared(keeper_dispatcher); + factory.registerCommand(profile_events_command); factory.initializeAllowList(keeper_dispatcher); factory.setInitialize(true); @@ -561,6 +570,12 @@ String LogInfoCommand::run() append("leader_committed_log_idx", log_info.leader_committed_log_idx); append("target_committed_log_idx", log_info.target_committed_log_idx); append("last_snapshot_idx", log_info.last_snapshot_idx); + + append("latest_logs_cache_entries", log_info.latest_logs_cache_entries); + append("latest_logs_cache_size", log_info.latest_logs_cache_size); + + append("commit_logs_cache_entries", log_info.commit_logs_cache_entries); + append("commit_logs_cache_size", log_info.commit_logs_cache_size); return ret.str(); } @@ -644,4 +659,31 @@ String JemallocDisableProfile::run() } #endif +String ProfileEventsCommand::run() +{ + StringBuffer ret; + +#if USE_NURAFT + auto append = [&ret] (const String & metric, uint64_t value, const String & docs) -> void + { + writeText(metric, ret); + writeText('\t', ret); + writeText(std::to_string(value), ret); + writeText('\t', ret); + writeText(docs, ret); + writeText('\n', ret); + }; + + for (auto i : ProfileEvents::keeper_profile_events) + { + const auto counter = ProfileEvents::global_counters[i].load(std::memory_order_relaxed); + std::string metric_name{ProfileEvents::getName(static_cast(i))}; + std::string metric_doc{ProfileEvents::getDocumentation(static_cast(i))}; + append(metric_name, counter, metric_doc); + } +#endif + + return ret.str(); +} + } diff --git a/src/Coordination/FourLetterCommand.h b/src/Coordination/FourLetterCommand.h index 7fc044881cf..82b30a0b5f6 100644 --- a/src/Coordination/FourLetterCommand.h +++ b/src/Coordination/FourLetterCommand.h @@ -1,18 +1,19 @@ #pragma once -#include -#include +#include "config.h" + #include - -#include -#include - -#include - +#include +#include namespace DB { +class WriteBufferFromOwnString; +class KeeperDispatcher; + +using String = std::string; + struct IFourLetterCommand; using FourLetterCommandPtr = std::shared_ptr; @@ -479,4 +480,16 @@ struct JemallocDisableProfile : public IFourLetterCommand }; #endif +struct ProfileEventsCommand : public IFourLetterCommand +{ + explicit ProfileEventsCommand(KeeperDispatcher & keeper_dispatcher_) + : IFourLetterCommand(keeper_dispatcher_) + { + } + + String name() override { return "pfev"; } + String run() override; + ~ProfileEventsCommand() override = default; +}; + } diff --git a/src/Coordination/InMemoryLogStore.cpp b/src/Coordination/InMemoryLogStore.cpp index ca240584a54..ee93c02b4b0 100644 --- a/src/Coordination/InMemoryLogStore.cpp +++ b/src/Coordination/InMemoryLogStore.cpp @@ -191,4 +191,10 @@ bool InMemoryLogStore::compact(uint64_t last_log_index) return true; } +bool InMemoryLogStore::is_conf(uint64_t index) +{ + auto entry = entry_at(index); + return entry != nullptr && entry->get_val_type() == nuraft::conf; +} + } diff --git a/src/Coordination/InMemoryLogStore.h b/src/Coordination/InMemoryLogStore.h index fc56826c81b..82c676639d5 100644 --- a/src/Coordination/InMemoryLogStore.h +++ b/src/Coordination/InMemoryLogStore.h @@ -39,6 +39,8 @@ public: bool flush() override { return true; } + bool is_conf(uint64_t index) override; + private: std::map> logs TSA_GUARDED_BY(logs_lock); mutable std::mutex logs_lock; diff --git a/src/Coordination/Keeper4LWInfo.h b/src/Coordination/Keeper4LWInfo.h index f99be0682ce..80b00b3f36e 100644 --- a/src/Coordination/Keeper4LWInfo.h +++ b/src/Coordination/Keeper4LWInfo.h @@ -52,16 +52,16 @@ struct Keeper4LWInfo struct KeeperLogInfo { /// My first log index in log store. - uint64_t first_log_idx; + uint64_t first_log_idx{0}; /// My first log term. - uint64_t first_log_term; + uint64_t first_log_term{0}; /// My last log index in log store. - uint64_t last_log_idx; + uint64_t last_log_idx{0}; /// My last log term. - uint64_t last_log_term; + uint64_t last_log_term{0}; /// My last committed log index in state machine. uint64_t last_committed_log_idx; @@ -74,6 +74,12 @@ struct KeeperLogInfo /// The largest committed log index in last snapshot. uint64_t last_snapshot_idx; + + uint64_t latest_logs_cache_entries; + uint64_t latest_logs_cache_size; + + uint64_t commit_logs_cache_entries; + uint64_t commit_logs_cache_size; }; } diff --git a/src/Coordination/KeeperAsynchronousMetrics.cpp b/src/Coordination/KeeperAsynchronousMetrics.cpp index 8f6e1dec6c1..a5b4bc4af97 100644 --- a/src/Coordination/KeeperAsynchronousMetrics.cpp +++ b/src/Coordination/KeeperAsynchronousMetrics.cpp @@ -20,7 +20,6 @@ void updateKeeperInformation(KeeperDispatcher & keeper_dispatcher, AsynchronousM size_t ephemerals_count = 0; size_t approximate_data_size = 0; size_t key_arena_size = 0; - size_t latest_snapshot_size = 0; size_t open_file_descriptor_count = 0; std::optional max_file_descriptor_count = 0; size_t followers = 0; @@ -46,11 +45,8 @@ void updateKeeperInformation(KeeperDispatcher & keeper_dispatcher, AsynchronousM ephemerals_count = state_machine.getTotalEphemeralNodesCount(); approximate_data_size = state_machine.getApproximateDataSize(); key_arena_size = state_machine.getKeyArenaSize(); - latest_snapshot_size = state_machine.getLatestSnapshotBufSize(); session_with_watches = state_machine.getSessionsWithWatchesCount(); paths_watched = state_machine.getWatchedPathsCount(); - //snapshot_dir_size = keeper_dispatcher.getSnapDirSize(); - //log_dir_size = keeper_dispatcher.getLogDirSize(); # if defined(__linux__) || defined(__APPLE__) open_file_descriptor_count = getCurrentProcessFDCount(); @@ -76,7 +72,9 @@ void updateKeeperInformation(KeeperDispatcher & keeper_dispatcher, AsynchronousM new_values["KeeperApproximateDataSize"] = { approximate_data_size, "The approximate data size of ClickHouse Keeper, in bytes." }; new_values["KeeperKeyArenaSize"] = { key_arena_size, "The size in bytes of the memory arena for keys in ClickHouse Keeper." }; - new_values["KeeperLatestSnapshotSize"] = { latest_snapshot_size, "The uncompressed size in bytes of the latest snapshot created by ClickHouse Keeper." }; + /// TODO: value was incorrectly set to 0 previously for local snapshots + /// it needs to be fixed and it needs to be atomic to avoid deadlock + ///new_values["KeeperLatestSnapshotSize"] = { latest_snapshot_size, "The uncompressed size in bytes of the latest snapshot created by ClickHouse Keeper." }; new_values["KeeperOpenFileDescriptorCount"] = { open_file_descriptor_count, "The number of open file descriptors in ClickHouse Keeper." }; if (max_file_descriptor_count.has_value()) @@ -99,6 +97,12 @@ void updateKeeperInformation(KeeperDispatcher & keeper_dispatcher, AsynchronousM new_values["KeeperTargetCommitLogIdx"] = { keeper_log_info.target_committed_log_idx, "Index until which logs can be committed in ClickHouse Keeper." }; new_values["KeeperLastSnapshotIdx"] = { keeper_log_info.last_snapshot_idx, "Index of the last log present in the last created snapshot." }; + new_values["KeeperLatestLogsCacheEntries"] = {keeper_log_info.latest_logs_cache_entries, "Number of entries stored in the in-memory cache for latest logs"}; + new_values["KeeperLatestLogsCacheSize"] = {keeper_log_info.latest_logs_cache_size, "Total size of in-memory cache for latest logs"}; + + new_values["KeeperCommitLogsCacheEntries"] = {keeper_log_info.commit_logs_cache_entries, "Number of entries stored in the in-memory cache for next logs to be committed"}; + new_values["KeeperCommitLogsCacheSize"] = {keeper_log_info.commit_logs_cache_size, "Total size of in-memory cache for next logs to be committed"}; + auto & keeper_connection_stats = keeper_dispatcher.getKeeperConnectionStats(); new_values["KeeperMinLatency"] = { keeper_connection_stats.getMinLatency(), "Minimal request latency of ClickHouse Keeper." }; diff --git a/src/Coordination/KeeperCommon.cpp b/src/Coordination/KeeperCommon.cpp new file mode 100644 index 00000000000..0245dbf28a2 --- /dev/null +++ b/src/Coordination/KeeperCommon.cpp @@ -0,0 +1,122 @@ +#include + +#include +#include + +#include +#include +#include +#include + +namespace DB +{ + +static size_t findLastSlash(StringRef path) +{ + if (path.size == 0) + return std::string::npos; + + for (size_t i = path.size - 1; i > 0; --i) + { + if (path.data[i] == '/') + return i; + } + + if (path.data[0] == '/') + return 0; + + return std::string::npos; +} + +StringRef parentNodePath(StringRef path) +{ + auto rslash_pos = findLastSlash(path); + if (rslash_pos > 0) + return StringRef{path.data, rslash_pos}; + return "/"; +} + +StringRef getBaseNodeName(StringRef path) +{ + size_t basename_start = findLastSlash(path); + return StringRef{path.data + basename_start + 1, path.size - basename_start - 1}; +} + +void moveFileBetweenDisks( + DiskPtr disk_from, + const std::string & path_from, + DiskPtr disk_to, + const std::string & path_to, + std::function before_file_remove_op, + LoggerPtr logger, + const KeeperContextPtr & keeper_context) +{ + LOG_TRACE(logger, "Moving {} to {} from disk {} to disk {}", path_from, path_to, disk_from->getName(), disk_to->getName()); + /// we use empty file with prefix tmp_ to detect incomplete copies + /// if a copy is complete we don't care from which disk we use the same file + /// so it's okay if a failure happens after removing of tmp file but before we remove + /// the file from the source disk + auto from_path = fs::path(path_from); + auto tmp_file_name = from_path.parent_path() / (std::string{tmp_keeper_file_prefix} + from_path.filename().string()); + + const auto & coordination_settings = keeper_context->getCoordinationSettings(); + auto max_retries_on_init = coordination_settings->disk_move_retries_during_init.value; + auto retries_sleep = std::chrono::milliseconds(coordination_settings->disk_move_retries_wait_ms); + auto run_with_retries = [&](const auto & op, std::string_view operation_description) + { + size_t retry_num = 0; + do + { + try + { + op(); + return true; + } + catch (...) + { + tryLogCurrentException( + logger, + fmt::format( + "While moving file {} to disk {} and running '{}'", path_from, disk_to->getName(), operation_description)); + std::this_thread::sleep_for(retries_sleep); + } + + ++retry_num; + if (keeper_context->getServerState() == KeeperContext::Phase::INIT && retry_num == max_retries_on_init) + { + LOG_ERROR(logger, "Operation '{}' failed too many times", operation_description); + break; + } + } while (!keeper_context->isShutdownCalled()); + + LOG_ERROR( + logger, + "Failed to run '{}' while moving file {} to disk {}", + operation_description, + path_from, + disk_to->getName()); + return false; + }; + + if (!run_with_retries( + [&] + { + auto buf = disk_to->writeFile(tmp_file_name); + buf->finalize(); + }, + "creating temporary file")) + return; + + if (!run_with_retries([&] { disk_from->copyFile(from_path, *disk_to, path_to, {}); }, "copying file")) + return; + + if (!run_with_retries([&] { disk_to->removeFileIfExists(tmp_file_name); }, "removing temporary file")) + return; + + if (before_file_remove_op) + before_file_remove_op(); + + if (!run_with_retries([&] { disk_from->removeFileIfExists(path_from); }, "removing file from source disk")) + return; +} +} diff --git a/src/Coordination/KeeperCommon.h b/src/Coordination/KeeperCommon.h new file mode 100644 index 00000000000..02f71a04d06 --- /dev/null +++ b/src/Coordination/KeeperCommon.h @@ -0,0 +1,29 @@ +#pragma once + +#include +#include "Common/Logger.h" + +namespace DB +{ + +class IDisk; +using DiskPtr = std::shared_ptr; +class KeeperContext; +using KeeperContextPtr = std::shared_ptr; + +StringRef parentNodePath(StringRef path); + +StringRef getBaseNodeName(StringRef path); + +inline static constexpr std::string_view tmp_keeper_file_prefix = "tmp_"; + +void moveFileBetweenDisks( + DiskPtr disk_from, + const std::string & path_from, + DiskPtr disk_to, + const std::string & path_to, + std::function before_file_remove_op, + LoggerPtr logger, + const KeeperContextPtr & keeper_context); + +} diff --git a/src/Coordination/KeeperConstants.cpp b/src/Coordination/KeeperConstants.cpp index f788095334e..aea2391cf13 100644 --- a/src/Coordination/KeeperConstants.cpp +++ b/src/Coordination/KeeperConstants.cpp @@ -284,7 +284,12 @@ M(InterfaceMySQLSendBytes) \ M(InterfaceMySQLReceiveBytes) \ M(InterfacePostgreSQLSendBytes) \ - M(InterfacePostgreSQLReceiveBytes) + M(InterfacePostgreSQLReceiveBytes) \ +\ + M(KeeperLogsEntryReadFromLatestCache) \ + M(KeeperLogsEntryReadFromCommitCache) \ + M(KeeperLogsEntryReadFromFile) \ + M(KeeperLogsPrefetchedEntries) \ namespace ProfileEvents { diff --git a/src/Coordination/KeeperContext.cpp b/src/Coordination/KeeperContext.cpp index 374571bae7e..b06e321aeec 100644 --- a/src/Coordination/KeeperContext.cpp +++ b/src/Coordination/KeeperContext.cpp @@ -1,14 +1,16 @@ #include #include -#include -#include -#include -#include #include -#include #include #include +#include +#include +#include +#include +#include +#include + #include namespace DB @@ -21,9 +23,10 @@ extern const int BAD_ARGUMENTS; } -KeeperContext::KeeperContext(bool standalone_keeper_) +KeeperContext::KeeperContext(bool standalone_keeper_, CoordinationSettingsPtr coordination_settings_) : disk_selector(std::make_shared()) , standalone_keeper(standalone_keeper_) + , coordination_settings(std::move(coordination_settings_)) { /// enable by default some feature flags feature_flags.enableFeatureFlag(KeeperFeatureFlag::FILTERED_LIST); @@ -402,4 +405,9 @@ void KeeperContext::waitLocalLogsPreprocessedOrShutdown() local_logs_preprocessed_cv.wait(lock, [this]{ return shutdown_called || local_logs_preprocessed; }); } +const CoordinationSettingsPtr & KeeperContext::getCoordinationSettings() const +{ + return coordination_settings; +} + } diff --git a/src/Coordination/KeeperContext.h b/src/Coordination/KeeperContext.h index 891bef00446..a7169e64387 100644 --- a/src/Coordination/KeeperContext.h +++ b/src/Coordination/KeeperContext.h @@ -1,8 +1,7 @@ #pragma once #include -#include -#include #include +#include #include #include #include @@ -12,10 +11,19 @@ namespace DB class KeeperDispatcher; +struct CoordinationSettings; +using CoordinationSettingsPtr = std::shared_ptr; + +class DiskSelector; +class IDisk; +using DiskPtr = std::shared_ptr; + +class WriteBufferFromOwnString; + class KeeperContext { public: - explicit KeeperContext(bool standalone_keeper_); + KeeperContext(bool standalone_keeper_, CoordinationSettingsPtr coordination_settings_); enum class Phase : uint8_t { @@ -68,6 +76,24 @@ public: void waitLocalLogsPreprocessedOrShutdown(); + uint64_t lastCommittedIndex() const + { + return last_committed_log_idx.load(std::memory_order_relaxed); + } + + void setLastCommitIndex(uint64_t commit_index) + { + last_committed_log_idx.store(commit_index, std::memory_order_relaxed); + last_committed_log_idx.notify_all(); + } + + void waitLastCommittedIndexUpdated(uint64_t current_last_committed_idx) + { + last_committed_log_idx.wait(current_last_committed_idx, std::memory_order_relaxed); + } + + const CoordinationSettingsPtr & getCoordinationSettings() const; + private: /// local disk defined using path or disk name using Storage = std::variant; @@ -89,7 +115,7 @@ private: std::atomic local_logs_preprocessed = false; std::atomic shutdown_called = false; - Phase server_state{Phase::INIT}; + std::atomic server_state{Phase::INIT}; bool ignore_system_path_on_startup{false}; bool digest_enabled{true}; @@ -113,6 +139,10 @@ private: KeeperDispatcher * dispatcher{nullptr}; std::atomic memory_soft_limit = 0; + + std::atomic last_committed_log_idx = 0; + + CoordinationSettingsPtr coordination_settings; }; using KeeperContextPtr = std::shared_ptr; diff --git a/src/Coordination/KeeperDispatcher.cpp b/src/Coordination/KeeperDispatcher.cpp index 35bc953a705..959ded47f27 100644 --- a/src/Coordination/KeeperDispatcher.cpp +++ b/src/Coordination/KeeperDispatcher.cpp @@ -256,11 +256,11 @@ void KeeperDispatcher::requestThread() if (shutdown_called) return; - auto current_last_committed_idx = our_last_committed_log_idx.load(std::memory_order_relaxed); + auto current_last_committed_idx = keeper_context->lastCommittedIndex(); if (current_last_committed_idx >= log_idx) break; - our_last_committed_log_idx.wait(current_last_committed_idx); + keeper_context->waitLastCommittedIndexUpdated(current_last_committed_idx); } } } @@ -414,8 +414,8 @@ void KeeperDispatcher::initialize(const Poco::Util::AbstractConfiguration & conf { LOG_DEBUG(log, "Initializing storage dispatcher"); - keeper_context = std::make_shared(standalone_keeper); configuration_and_settings = KeeperConfigurationAndSettings::loadFromConfig(config, standalone_keeper); + keeper_context = std::make_shared(standalone_keeper, configuration_and_settings->coordination_settings); keeper_context->initialize(config, this); @@ -433,7 +433,7 @@ void KeeperDispatcher::initialize(const Poco::Util::AbstractConfiguration & conf snapshots_queue, keeper_context, snapshot_s3, - [this](uint64_t log_idx, const KeeperStorage::RequestForSession & request_for_session) + [this](uint64_t /*log_idx*/, const KeeperStorage::RequestForSession & request_for_session) { { /// check if we have queue of read requests depending on this request to be committed @@ -457,9 +457,6 @@ void KeeperDispatcher::initialize(const Poco::Util::AbstractConfiguration & conf } } } - - our_last_committed_log_idx.store(log_idx, std::memory_order_relaxed); - our_last_committed_log_idx.notify_all(); }); try @@ -504,8 +501,9 @@ void KeeperDispatcher::shutdown() LOG_DEBUG(log, "Shutting down storage dispatcher"); - our_last_committed_log_idx = std::numeric_limits::max(); - our_last_committed_log_idx.notify_all(); + /// some threads can be waiting for certain commits, so we set value + /// of the last commit index to something that will always unblock + keeper_context->setLastCommitIndex(std::numeric_limits::max()); if (session_cleaner_thread.joinable()) session_cleaner_thread.join(); diff --git a/src/Coordination/KeeperDispatcher.h b/src/Coordination/KeeperDispatcher.h index db41fb2ea26..e8ee486be88 100644 --- a/src/Coordination/KeeperDispatcher.h +++ b/src/Coordination/KeeperDispatcher.h @@ -105,8 +105,6 @@ private: public: std::mutex read_request_queue_mutex; - std::atomic our_last_committed_log_idx = 0; - /// queue of read requests that can be processed after a request with specific session ID and XID is committed std::unordered_map> read_request_queue; diff --git a/src/Coordination/KeeperLogStore.cpp b/src/Coordination/KeeperLogStore.cpp index ce7c715237e..820039d8a8f 100644 --- a/src/Coordination/KeeperLogStore.cpp +++ b/src/Coordination/KeeperLogStore.cpp @@ -66,13 +66,16 @@ nuraft::ptr KeeperLogStore::entry_at(uint64_t index) return changelog.entryAt(index); } +bool KeeperLogStore::is_conf(uint64_t index) +{ + std::lock_guard lock(changelog_lock); + return changelog.isConfigLog(index); +} + uint64_t KeeperLogStore::term_at(uint64_t index) { std::lock_guard lock(changelog_lock); - auto entry = changelog.entryAt(index); - if (entry) - return entry->get_term(); - return 0; + return changelog.termAt(index); } nuraft::ptr KeeperLogStore::pack(uint64_t index, int32_t cnt) @@ -145,4 +148,10 @@ void KeeperLogStore::setRaftServer(const nuraft::ptr & raft return changelog.setRaftServer(raft_server); } +void KeeperLogStore::getKeeperLogInfo(KeeperLogInfo & log_info) const +{ + std::lock_guard lock(changelog_lock); + changelog.getKeeperLogInfo(log_info); +} + } diff --git a/src/Coordination/KeeperLogStore.h b/src/Coordination/KeeperLogStore.h index aa277f19d88..21d9479ee47 100644 --- a/src/Coordination/KeeperLogStore.h +++ b/src/Coordination/KeeperLogStore.h @@ -1,10 +1,10 @@ #pragma once #include -#include #include #include #include #include +#include #include namespace DB @@ -38,6 +38,8 @@ public: /// Return entry at index nuraft::ptr entry_at(uint64_t index) override; + bool is_conf(uint64_t index) override; + /// Term if the index uint64_t term_at(uint64_t index) override; @@ -72,6 +74,8 @@ public: void setRaftServer(const nuraft::ptr & raft_server); + void getKeeperLogInfo(KeeperLogInfo & log_info) const; + private: mutable std::mutex changelog_lock; LoggerPtr log; diff --git a/src/Coordination/KeeperServer.cpp b/src/Coordination/KeeperServer.cpp index 722b1303cc8..ca7fdfe9ee9 100644 --- a/src/Coordination/KeeperServer.cpp +++ b/src/Coordination/KeeperServer.cpp @@ -6,6 +6,7 @@ #include #include #include +#include #include #include #include @@ -119,22 +120,20 @@ KeeperServer::KeeperServer( KeeperSnapshotManagerS3 & snapshot_manager_s3, KeeperStateMachine::CommitCallback commit_callback) : server_id(configuration_and_settings_->server_id) - , coordination_settings(configuration_and_settings_->coordination_settings) , log(getLogger("KeeperServer")) , is_recovering(config.getBool("keeper_server.force_recovery", false)) , keeper_context{std::move(keeper_context_)} , create_snapshot_on_exit(config.getBool("keeper_server.create_snapshot_on_exit", true)) , enable_reconfiguration(config.getBool("keeper_server.enable_reconfiguration", false)) { - if (coordination_settings->quorum_reads) + if (keeper_context->getCoordinationSettings()->quorum_reads) LOG_WARNING(log, "Quorum reads enabled, Keeper will work slower."); state_machine = nuraft::cs_new( responses_queue_, snapshots_queue_, - coordination_settings, keeper_context, - config.getBool("keeper_server.upload_snapshot_on_exit", true) ? &snapshot_manager_s3 : nullptr, + config.getBool("keeper_server.upload_snapshot_on_exit", false) ? &snapshot_manager_s3 : nullptr, commit_callback, checkAndGetSuperdigest(configuration_and_settings_->super_digest)); @@ -143,7 +142,6 @@ KeeperServer::KeeperServer( "keeper_server", "state", config, - coordination_settings, keeper_context); } @@ -226,7 +224,7 @@ void KeeperServer::loadLatestConfig() { auto latest_snapshot_config = state_machine->getClusterConfig(); auto latest_log_store_config = state_manager->getLatestConfigFromLogStore(); - auto async_replication = coordination_settings->async_replication; + auto async_replication = keeper_context->getCoordinationSettings()->async_replication; if (latest_snapshot_config && latest_log_store_config) { @@ -293,6 +291,8 @@ void KeeperServer::forceRecovery() void KeeperServer::launchRaftServer(const Poco::Util::AbstractConfiguration & config, bool enable_ipv6) { + const auto & coordination_settings = keeper_context->getCoordinationSettings(); + nuraft::raft_params params; params.parallel_log_appending_ = true; params.heart_beat_interval_ @@ -332,7 +332,7 @@ void KeeperServer::launchRaftServer(const Poco::Util::AbstractConfiguration & co params.auto_forwarding_req_timeout_ = getValueOrMaxInt32AndLogWarning(coordination_settings->operation_timeout_ms.totalMilliseconds() * 2, "operation_timeout_ms", log); params.max_append_size_ - = getValueOrMaxInt32AndLogWarning(coordination_settings->max_requests_batch_size, "max_requests_batch_size", log); + = getValueOrMaxInt32AndLogWarning(coordination_settings->max_requests_append_size, "max_requests_append_size", log); params.return_method_ = nuraft::raft_params::async_handler; @@ -427,6 +427,10 @@ void KeeperServer::startup(const Poco::Util::AbstractConfiguration & config, boo { state_machine->init(); + keeper_context->setLastCommitIndex(state_machine->last_commit_index()); + + const auto & coordination_settings = keeper_context->getCoordinationSettings(); + state_manager->loadLogStore(state_machine->last_commit_index() + 1, coordination_settings->reserved_log_items); auto log_store = state_manager->load_log_store(); @@ -446,7 +450,7 @@ void KeeperServer::startup(const Poco::Util::AbstractConfiguration & config, boo void KeeperServer::shutdownRaftServer() { - size_t timeout = coordination_settings->shutdown_timeout.totalSeconds(); + size_t timeout = keeper_context->getCoordinationSettings()->shutdown_timeout.totalSeconds(); if (!raft_instance) { @@ -870,7 +874,7 @@ nuraft::cb_func::ReturnCode KeeperServer::callbackFunc(nuraft::cb_func::Type typ /// Node first became leader, and after that some other node became leader. /// BecameFresh for this node will not be called because it was already fresh /// when it was leader. - if (leader_index < our_index + coordination_settings->fresh_log_gap) + if (leader_index < our_index + keeper_context->getCoordinationSettings()->fresh_log_gap) set_initialized(); } return nuraft::cb_func::ReturnCode::Ok; @@ -905,7 +909,7 @@ void KeeperServer::waitInit() { std::unique_lock lock(initialized_mutex); - int64_t timeout = coordination_settings->startup_timeout.totalMilliseconds(); + int64_t timeout = keeper_context->getCoordinationSettings()->startup_timeout.totalMilliseconds(); if (!initialized_cv.wait_for(lock, std::chrono::milliseconds(timeout), [&] { return initialized_flag.load(); })) LOG_WARNING(log, "Failed to wait for RAFT initialization in {}ms, will continue in background", timeout); } @@ -977,6 +981,7 @@ KeeperServer::ConfigUpdateState KeeperServer::applyConfigUpdate( ClusterUpdateActions KeeperServer::getRaftConfigurationDiff(const Poco::Util::AbstractConfiguration & config) { + const auto & coordination_settings = keeper_context->getCoordinationSettings(); auto diff = state_manager->getRaftConfigurationDiff(config, coordination_settings); if (!diff.empty()) @@ -1004,6 +1009,7 @@ void KeeperServer::applyConfigUpdateWithReconfigDisabled(const ClusterUpdateActi std::this_thread::sleep_for(sleep_time * (i + 1)); }; + const auto & coordination_settings = keeper_context->getCoordinationSettings(); if (const auto * add = std::get_if(&action)) { for (size_t i = 0; i < coordination_settings->configuration_change_tries_count && !is_recovering; ++i) @@ -1059,6 +1065,7 @@ bool KeeperServer::waitForConfigUpdateWithReconfigDisabled(const ClusterUpdateAc auto became_leader = [&] { LOG_INFO(log, "Became leader, aborting"); return false; }; auto backoff = [&](size_t i) { std::this_thread::sleep_for(sleep_time * (i + 1)); }; + const auto & coordination_settings = keeper_context->getCoordinationSettings(); if (const auto* add = std::get_if(&action)) { for (size_t i = 0; i < coordination_settings->configuration_change_tries_count && !is_recovering; ++i) @@ -1125,14 +1132,12 @@ KeeperLogInfo KeeperServer::getKeeperLogInfo() auto log_store = state_manager->load_log_store(); if (log_store) { - log_info.first_log_idx = log_store->start_index(); - log_info.first_log_term = log_store->term_at(log_info.first_log_idx); + const auto & keeper_log_storage = static_cast(*log_store); + keeper_log_storage.getKeeperLogInfo(log_info); } if (raft_instance) { - log_info.last_log_idx = raft_instance->get_last_log_idx(); - log_info.last_log_term = raft_instance->get_last_log_term(); log_info.last_committed_log_idx = raft_instance->get_committed_log_idx(); log_info.leader_committed_log_idx = raft_instance->get_leader_committed_log_idx(); log_info.target_committed_log_idx = raft_instance->get_target_committed_log_idx(); diff --git a/src/Coordination/KeeperServer.h b/src/Coordination/KeeperServer.h index ef298df3efc..dd54539a92b 100644 --- a/src/Coordination/KeeperServer.h +++ b/src/Coordination/KeeperServer.h @@ -22,8 +22,6 @@ class KeeperServer private: const int server_id; - CoordinationSettingsPtr coordination_settings; - nuraft::ptr state_machine; nuraft::ptr state_manager; diff --git a/src/Coordination/KeeperSnapshotManager.cpp b/src/Coordination/KeeperSnapshotManager.cpp index 091571b4a1a..be783d826b6 100644 --- a/src/Coordination/KeeperSnapshotManager.cpp +++ b/src/Coordination/KeeperSnapshotManager.cpp @@ -3,6 +3,7 @@ #include #include #include +#include #include #include #include @@ -13,7 +14,7 @@ #include #include #include -#include +#include #include #include #include @@ -32,23 +33,21 @@ namespace ErrorCodes namespace { - constexpr std::string_view tmp_prefix = "tmp_"; - - void moveFileBetweenDisks(DiskPtr disk_from, const std::string & path_from, DiskPtr disk_to, const std::string & path_to) + void moveSnapshotBetweenDisks( + DiskPtr disk_from, + const std::string & path_from, + DiskPtr disk_to, + const std::string & path_to, + const KeeperContextPtr & keeper_context) { - /// we use empty file with prefix tmp_ to detect incomplete copies - /// if a copy is complete we don't care from which disk we use the same file - /// so it's okay if a failure happens after removing of tmp file but before we remove - /// the snapshot from the source disk - auto from_path = fs::path(path_from); - auto tmp_snapshot_name = from_path.parent_path() / (std::string{tmp_prefix} + from_path.filename().string()); - { - auto buf = disk_to->writeFile(tmp_snapshot_name); - buf->finalize(); - } - disk_from->copyFile(from_path, *disk_to, path_to, {}); - disk_to->removeFile(tmp_snapshot_name); - disk_from->removeFile(path_from); + moveFileBetweenDisks( + std::move(disk_from), + path_from, + std::move(disk_to), + path_to, + /*before_file_remove_op=*/{}, + getLogger("KeeperSnapshotManager"), + keeper_context); } uint64_t getSnapshotPathUpToLogIdx(const String & snapshot_path) @@ -582,9 +581,9 @@ KeeperSnapshotManager::KeeperSnapshotManager( std::vector snapshot_files; for (auto it = disk->iterateDirectory(""); it->isValid(); it->next()) { - if (it->name().starts_with(tmp_prefix)) + if (it->name().starts_with(tmp_keeper_file_prefix)) { - incomplete_files.emplace(it->name().substr(tmp_prefix.size()), it->path()); + incomplete_files.emplace(it->name().substr(tmp_keeper_file_prefix.size()), it->path()); continue; } @@ -603,7 +602,7 @@ KeeperSnapshotManager::KeeperSnapshotManager( if (!inserted) LOG_WARNING( - getLogger("KeeperSnapshotManager"), + log, "Found another snapshots with last log idx {}, will use snapshot from disk {}", snapshot_up_to, disk->getName()); @@ -612,6 +611,9 @@ KeeperSnapshotManager::KeeperSnapshotManager( for (const auto & [name, path] : incomplete_files) disk->removeFile(path); + if (snapshot_files.empty()) + LOG_TRACE(log, "No snapshots were found on {}", disk->getName()); + read_disks.insert(disk); }; @@ -774,7 +776,7 @@ void KeeperSnapshotManager::moveSnapshotsIfNeeded() { if (file_info.disk != latest_snapshot_disk) { - moveFileBetweenDisks(file_info.disk, file_info.path, latest_snapshot_disk, file_info.path); + moveSnapshotBetweenDisks(file_info.disk, file_info.path, latest_snapshot_disk, file_info.path, keeper_context); file_info.disk = latest_snapshot_disk; } } @@ -782,7 +784,7 @@ void KeeperSnapshotManager::moveSnapshotsIfNeeded() { if (file_info.disk != disk) { - moveFileBetweenDisks(file_info.disk, file_info.path, disk, file_info.path); + moveSnapshotBetweenDisks(file_info.disk, file_info.path, disk, file_info.path, keeper_context); file_info.disk = disk; } } diff --git a/src/Coordination/KeeperStateMachine.cpp b/src/Coordination/KeeperStateMachine.cpp index c82f8301eff..f1426fd4e11 100644 --- a/src/Coordination/KeeperStateMachine.cpp +++ b/src/Coordination/KeeperStateMachine.cpp @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -42,23 +43,20 @@ namespace ErrorCodes KeeperStateMachine::KeeperStateMachine( ResponsesQueue & responses_queue_, SnapshotsQueue & snapshots_queue_, - const CoordinationSettingsPtr & coordination_settings_, const KeeperContextPtr & keeper_context_, KeeperSnapshotManagerS3 * snapshot_manager_s3_, CommitCallback commit_callback_, const std::string & superdigest_) : commit_callback(commit_callback_) - , coordination_settings(coordination_settings_) , snapshot_manager( - coordination_settings->snapshots_to_keep, + keeper_context_->getCoordinationSettings()->snapshots_to_keep, keeper_context_, - coordination_settings->compress_snapshots_with_zstd_format, + keeper_context_->getCoordinationSettings()->compress_snapshots_with_zstd_format, superdigest_, - coordination_settings->dead_session_check_period_ms.totalMilliseconds()) + keeper_context_->getCoordinationSettings()->dead_session_check_period_ms.totalMilliseconds()) , responses_queue(responses_queue_) , snapshots_queue(snapshots_queue_) - , min_request_size_to_cache(coordination_settings_->min_request_size_for_cache) - , last_committed_idx(0) + , min_request_size_to_cache(keeper_context_->getCoordinationSettings()->min_request_size_for_cache) , log(getLogger("KeeperStateMachine")) , superdigest(superdigest_) , keeper_context(keeper_context_) @@ -100,7 +98,7 @@ void KeeperStateMachine::init() storage = std::move(snapshot_deserialization_result.storage); latest_snapshot_meta = snapshot_deserialization_result.snapshot_meta; cluster_config = snapshot_deserialization_result.cluster_config; - last_committed_idx = latest_snapshot_meta->get_last_log_idx(); + keeper_context->setLastCommitIndex(latest_snapshot_meta->get_last_log_idx()); loaded = true; break; } @@ -115,6 +113,7 @@ void KeeperStateMachine::init() } } + auto last_committed_idx = keeper_context->lastCommittedIndex(); if (has_snapshots) { if (loaded) @@ -129,7 +128,7 @@ void KeeperStateMachine::init() if (!storage) storage = std::make_unique( - coordination_settings->dead_session_check_period_ms.totalMilliseconds(), superdigest, keeper_context); + keeper_context->getCoordinationSettings()->dead_session_check_period_ms.totalMilliseconds(), superdigest, keeper_context); } namespace @@ -139,16 +138,18 @@ void assertDigest( const KeeperStorage::Digest & expected, const KeeperStorage::Digest & actual, const Coordination::ZooKeeperRequest & request, + uint64_t log_idx, bool committing) { if (!KeeperStorage::checkDigest(expected, actual)) { LOG_FATAL( getLogger("KeeperStateMachine"), - "Digest for nodes is not matching after {} request of type '{}'.\nExpected digest - {}, actual digest - {} (digest " - "{}). Keeper will terminate to avoid inconsistencies.\nExtra information about the request:\n{}", + "Digest for nodes is not matching after {} request of type '{}' at log index {}.\nExpected digest - {}, actual digest - {} " + "(digest {}). Keeper will terminate to avoid inconsistencies.\nExtra information about the request:\n{}", committing ? "committing" : "preprocessing", request.getOpNum(), + log_idx, expected.value, actual.value, expected.version, @@ -296,12 +297,12 @@ bool KeeperStateMachine::preprocess(const KeeperStorage::RequestForSession & req } catch (...) { - tryLogCurrentException(__PRETTY_FUNCTION__, "Failed to preprocess stored log, aborting to avoid inconsistent state"); + tryLogCurrentException(__PRETTY_FUNCTION__, fmt::format("Failed to preprocess stored log at index {}, aborting to avoid inconsistent state", request_for_session.log_idx)); std::abort(); } if (keeper_context->digestEnabled() && request_for_session.digest) - assertDigest(*request_for_session.digest, storage->getNodesDigest(false), *request_for_session.request, false); + assertDigest(*request_for_session.digest, storage->getNodesDigest(false), *request_for_session.request, request_for_session.log_idx, false); return true; } @@ -408,48 +409,57 @@ nuraft::ptr KeeperStateMachine::commit(const uint64_t log_idx, n } }; - const auto op_num = request_for_session->request->getOpNum(); - if (op_num == Coordination::OpNum::SessionID) + try { - const Coordination::ZooKeeperSessionIDRequest & session_id_request - = dynamic_cast(*request_for_session->request); - int64_t session_id; - std::shared_ptr response = std::make_shared(); - response->internal_id = session_id_request.internal_id; - response->server_id = session_id_request.server_id; - KeeperStorage::ResponseForSession response_for_session; - response_for_session.session_id = -1; - response_for_session.response = response; - - std::lock_guard lock(storage_and_responses_lock); - session_id = storage->getSessionID(session_id_request.session_timeout_ms); - LOG_DEBUG(log, "Session ID response {} with timeout {}", session_id, session_id_request.session_timeout_ms); - response->session_id = session_id; - try_push(response_for_session); - } - else - { - if (op_num == Coordination::OpNum::Close) + const auto op_num = request_for_session->request->getOpNum(); + if (op_num == Coordination::OpNum::SessionID) { - std::lock_guard lock(request_cache_mutex); - parsed_request_cache.erase(request_for_session->session_id); + const Coordination::ZooKeeperSessionIDRequest & session_id_request + = dynamic_cast(*request_for_session->request); + int64_t session_id; + std::shared_ptr response = std::make_shared(); + response->internal_id = session_id_request.internal_id; + response->server_id = session_id_request.server_id; + KeeperStorage::ResponseForSession response_for_session; + response_for_session.session_id = -1; + response_for_session.response = response; + + std::lock_guard lock(storage_and_responses_lock); + session_id = storage->getSessionID(session_id_request.session_timeout_ms); + LOG_DEBUG(log, "Session ID response {} with timeout {}", session_id, session_id_request.session_timeout_ms); + response->session_id = session_id; + try_push(response_for_session); + } + else + { + if (op_num == Coordination::OpNum::Close) + { + std::lock_guard lock(request_cache_mutex); + parsed_request_cache.erase(request_for_session->session_id); + } + + std::lock_guard lock(storage_and_responses_lock); + KeeperStorage::ResponsesForSessions responses_for_sessions + = storage->processRequest(request_for_session->request, request_for_session->session_id, request_for_session->zxid); + for (auto & response_for_session : responses_for_sessions) + try_push(response_for_session); + + if (keeper_context->digestEnabled() && request_for_session->digest) + assertDigest(*request_for_session->digest, storage->getNodesDigest(true), *request_for_session->request, request_for_session->log_idx, true); } - std::lock_guard lock(storage_and_responses_lock); - KeeperStorage::ResponsesForSessions responses_for_sessions - = storage->processRequest(request_for_session->request, request_for_session->session_id, request_for_session->zxid); - for (auto & response_for_session : responses_for_sessions) - try_push(response_for_session); + ProfileEvents::increment(ProfileEvents::KeeperCommits); + keeper_context->setLastCommitIndex(log_idx); - if (keeper_context->digestEnabled() && request_for_session->digest) - assertDigest(*request_for_session->digest, storage->getNodesDigest(true), *request_for_session->request, true); + if (commit_callback) + commit_callback(log_idx, *request_for_session); + } + catch (...) + { + tryLogCurrentException(log, fmt::format("Failed to commit stored log at index {}", log_idx)); + throw; } - ProfileEvents::increment(ProfileEvents::KeeperCommits); - last_committed_idx = log_idx; - - if (commit_callback) - commit_callback(log_idx, *request_for_session); return nullptr; } @@ -496,7 +506,7 @@ bool KeeperStateMachine::apply_snapshot(nuraft::snapshot & s) } ProfileEvents::increment(ProfileEvents::KeeperSnapshotApplys); - last_committed_idx = s.get_last_log_idx(); + keeper_context->setLastCommitIndex(s.get_last_log_idx()); return true; } @@ -506,7 +516,7 @@ void KeeperStateMachine::commit_config(const uint64_t log_idx, nuraft::ptrserialize(); cluster_config = ClusterConfig::deserialize(*tmp); - last_committed_idx = log_idx; + keeper_context->setLastCommitIndex(log_idx); } void KeeperStateMachine::rollback(uint64_t log_idx, nuraft::buffer & data) diff --git a/src/Coordination/KeeperStateMachine.h b/src/Coordination/KeeperStateMachine.h index b11cd53c00e..3727beadb98 100644 --- a/src/Coordination/KeeperStateMachine.h +++ b/src/Coordination/KeeperStateMachine.h @@ -25,7 +25,6 @@ public: KeeperStateMachine( ResponsesQueue & responses_queue_, SnapshotsQueue & snapshots_queue_, - const CoordinationSettingsPtr & coordination_settings_, const KeeperContextPtr & keeper_context_, KeeperSnapshotManagerS3 * snapshot_manager_s3_, CommitCallback commit_callback_ = {}, @@ -70,7 +69,7 @@ public: const KeeperStorage::RequestForSession & request_for_session, bool allow_missing) TSA_NO_THREAD_SAFETY_ANALYSIS; - uint64_t last_commit_index() override { return last_committed_idx; } + uint64_t last_commit_index() override { return keeper_context->lastCommittedIndex(); } /// Apply preliminarily saved (save_logical_snp_obj) snapshot to our state. bool apply_snapshot(nuraft::snapshot & s) override; @@ -139,8 +138,6 @@ private: SnapshotFileInfo latest_snapshot_info; nuraft::ptr latest_snapshot_buf = nullptr; - CoordinationSettingsPtr coordination_settings; - /// Main state machine logic KeeperStoragePtr storage TSA_PT_GUARDED_BY(storage_and_responses_lock); @@ -170,9 +167,6 @@ private: /// can be processed only in 1 thread at any point std::mutex request_cache_mutex; - /// Last committed Raft log number. - std::atomic last_committed_idx; - LoggerPtr log; /// Cluster config for our quorum. diff --git a/src/Coordination/KeeperStateManager.cpp b/src/Coordination/KeeperStateManager.cpp index 4fbb9b52e6e..87c56909387 100644 --- a/src/Coordination/KeeperStateManager.cpp +++ b/src/Coordination/KeeperStateManager.cpp @@ -241,23 +241,25 @@ KeeperStateManager::KeeperStateManager( const std::string & config_prefix_, const std::string & server_state_file_name_, const Poco::Util::AbstractConfiguration & config, - const CoordinationSettingsPtr & coordination_settings, KeeperContextPtr keeper_context_) : my_server_id(my_server_id_) , secure(config.getBool(config_prefix_ + ".raft_configuration.secure", false)) , config_prefix(config_prefix_) - , configuration_wrapper(parseServersConfiguration(config, false, coordination_settings->async_replication)) + , configuration_wrapper(parseServersConfiguration(config, false, keeper_context_->getCoordinationSettings()->async_replication)) , log_store(nuraft::cs_new( LogFileSettings { - .force_sync = coordination_settings->force_sync, - .compress_logs = coordination_settings->compress_logs, - .rotate_interval = coordination_settings->rotate_log_storage_interval, - .max_size = coordination_settings->max_log_file_size, - .overallocate_size = coordination_settings->log_file_overallocate_size}, + .force_sync = keeper_context_->getCoordinationSettings()->force_sync, + .compress_logs = keeper_context_->getCoordinationSettings()->compress_logs, + .rotate_interval = keeper_context_->getCoordinationSettings()->rotate_log_storage_interval, + .max_size = keeper_context_->getCoordinationSettings()->max_log_file_size, + .overallocate_size = keeper_context_->getCoordinationSettings()->log_file_overallocate_size, + .latest_logs_cache_size_threshold = keeper_context_->getCoordinationSettings()->latest_logs_cache_size_threshold, + .commit_logs_cache_size_threshold = keeper_context_->getCoordinationSettings()->commit_logs_cache_size_threshold + }, FlushSettings { - .max_flush_batch_size = coordination_settings->max_flush_batch_size, + .max_flush_batch_size = keeper_context_->getCoordinationSettings()->max_flush_batch_size, }, keeper_context_)) , server_state_file_name(server_state_file_name_) diff --git a/src/Coordination/KeeperStateManager.h b/src/Coordination/KeeperStateManager.h index 02dd6b2ff53..60f6dbe7b62 100644 --- a/src/Coordination/KeeperStateManager.h +++ b/src/Coordination/KeeperStateManager.h @@ -23,7 +23,6 @@ public: const std::string & config_prefix_, const std::string & server_state_file_name_, const Poco::Util::AbstractConfiguration & config, - const CoordinationSettingsPtr & coordination_settings, KeeperContextPtr keeper_context_); /// Constructor for tests diff --git a/src/Coordination/KeeperStorage.cpp b/src/Coordination/KeeperStorage.cpp index eaa0c3c9e68..d3101543362 100644 --- a/src/Coordination/KeeperStorage.cpp +++ b/src/Coordination/KeeperStorage.cpp @@ -18,7 +18,7 @@ #include #include -#include +#include #include #include #include @@ -26,7 +26,6 @@ #include #include -#include namespace ProfileEvents { @@ -1583,7 +1582,7 @@ struct KeeperStorageListRequestProcessor final : public KeeperStorageRequestProc { auto path_prefix = request.path; if (path_prefix.empty()) - throw DB::Exception(ErrorCodes::LOGICAL_ERROR, "Path cannot be empty"); + throw DB::Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: path cannot be empty"); const auto & children = node_it->value.getChildren(); response.names.reserve(children.size()); diff --git a/src/Coordination/ZooKeeperDataReader.cpp b/src/Coordination/ZooKeeperDataReader.cpp index c7b1abf1d83..c205db942b9 100644 --- a/src/Coordination/ZooKeeperDataReader.cpp +++ b/src/Coordination/ZooKeeperDataReader.cpp @@ -8,7 +8,7 @@ #include #include #include -#include +#include namespace DB diff --git a/src/Coordination/pathUtils.cpp b/src/Coordination/pathUtils.cpp deleted file mode 100644 index 25f8e25cf06..00000000000 --- a/src/Coordination/pathUtils.cpp +++ /dev/null @@ -1,37 +0,0 @@ -#include - -namespace DB -{ - -static size_t findLastSlash(StringRef path) -{ - if (path.size == 0) - return std::string::npos; - - for (size_t i = path.size - 1; i > 0; --i) - { - if (path.data[i] == '/') - return i; - } - - if (path.data[0] == '/') - return 0; - - return std::string::npos; -} - -StringRef parentNodePath(StringRef path) -{ - auto rslash_pos = findLastSlash(path); - if (rslash_pos > 0) - return StringRef{path.data, rslash_pos}; - return "/"; -} - -StringRef getBaseNodeName(StringRef path) -{ - size_t basename_start = findLastSlash(path); - return StringRef{path.data + basename_start + 1, path.size - basename_start - 1}; -} - -} diff --git a/src/Coordination/pathUtils.h b/src/Coordination/pathUtils.h deleted file mode 100644 index b2b79b14110..00000000000 --- a/src/Coordination/pathUtils.h +++ /dev/null @@ -1,13 +0,0 @@ -#pragma once - -#include -#include - -namespace DB -{ - -StringRef parentNodePath(StringRef path); - -StringRef getBaseNodeName(StringRef path); - -} diff --git a/src/Coordination/tests/gtest_coordination.cpp b/src/Coordination/tests/gtest_coordination.cpp index bd9dc4c3fd3..806c0a88edf 100644 --- a/src/Coordination/tests/gtest_coordination.cpp +++ b/src/Coordination/tests/gtest_coordination.cpp @@ -1,8 +1,6 @@ #include #include -#include "Common/ZooKeeper/IKeeper.h" -#include "Core/Defines.h" #include "config.h" #if USE_NURAFT @@ -22,7 +20,7 @@ #include #include #include -#include +#include #include #include #include @@ -65,7 +63,7 @@ struct CompressionParam class CoordinationTest : public ::testing::TestWithParam { protected: - DB::KeeperContextPtr keeper_context = std::make_shared(true); + DB::KeeperContextPtr keeper_context = std::make_shared(true, std::make_shared()); LoggerPtr log{getLogger("CoordinationTest")}; void SetUp() override @@ -558,6 +556,7 @@ TEST_P(CoordinationTest, ChangelogTestCompaction) EXPECT_EQ(changelog.size(), 3); + keeper_context->setLastCommitIndex(2); changelog.compact(2); EXPECT_EQ(changelog.size(), 1); @@ -582,6 +581,7 @@ TEST_P(CoordinationTest, ChangelogTestCompaction) EXPECT_TRUE(fs::exists("./logs/changelog_1_5.bin" + params.extension)); EXPECT_TRUE(fs::exists("./logs/changelog_6_10.bin" + params.extension)); + keeper_context->setLastCommitIndex(6); changelog.compact(6); std::this_thread::sleep_for(std::chrono::microseconds(1000)); @@ -1758,22 +1758,30 @@ getLogEntryFromZKRequest(size_t term, int64_t session_id, int64_t zxid, const Co } void testLogAndStateMachine( - Coordination::CoordinationSettingsPtr settings, + DB::CoordinationSettingsPtr settings, uint64_t total_logs, - bool enable_compression, - Coordination::KeeperContextPtr keeper_context) + bool enable_compression) { using namespace Coordination; using namespace DB; ChangelogDirTest snapshots("./snapshots"); - keeper_context->setSnapshotDisk(std::make_shared("SnapshotDisk", "./snapshots")); ChangelogDirTest logs("./logs"); - keeper_context->setLogDisk(std::make_shared("LogDisk", "./logs")); + + auto get_keeper_context = [&] + { + auto local_keeper_context = std::make_shared(true, settings); + local_keeper_context->setSnapshotDisk(std::make_shared("SnapshotDisk", "./snapshots")); + local_keeper_context->setLogDisk(std::make_shared("LogDisk", "./logs")); + return local_keeper_context; + }; ResponsesQueue queue(std::numeric_limits::max()); SnapshotsQueue snapshots_queue{1}; - auto state_machine = std::make_shared(queue, snapshots_queue, settings, keeper_context, nullptr); + + auto keeper_context = get_keeper_context(); + auto state_machine = std::make_shared(queue, snapshots_queue, keeper_context, nullptr); + state_machine->init(); DB::KeeperLogStore changelog( DB::LogFileSettings{ @@ -1811,12 +1819,14 @@ void testLogAndStateMachine( snapshot_task.create_snapshot(std::move(snapshot_task.snapshot)); } + if (snapshot_created && changelog.size() > settings->reserved_log_items) changelog.compact(i - settings->reserved_log_items); } SnapshotsQueue snapshots_queue1{1}; - auto restore_machine = std::make_shared(queue, snapshots_queue1, settings, keeper_context, nullptr); + keeper_context = get_keeper_context(); + auto restore_machine = std::make_shared(queue, snapshots_queue1, keeper_context, nullptr); restore_machine->init(); EXPECT_EQ(restore_machine->last_commit_index(), total_logs - total_logs % settings->snapshot_distance); @@ -1863,63 +1873,64 @@ TEST_P(CoordinationTest, TestStateMachineAndLogStore) settings->snapshot_distance = 10; settings->reserved_log_items = 10; settings->rotate_log_storage_interval = 10; - testLogAndStateMachine(settings, 37, params.enable_compression, keeper_context); + + testLogAndStateMachine(settings, 37, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 10; settings->reserved_log_items = 10; settings->rotate_log_storage_interval = 10; - testLogAndStateMachine(settings, 11, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 11, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 10; settings->reserved_log_items = 10; settings->rotate_log_storage_interval = 10; - testLogAndStateMachine(settings, 40, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 40, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 10; settings->reserved_log_items = 20; settings->rotate_log_storage_interval = 30; - testLogAndStateMachine(settings, 40, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 40, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 10; settings->reserved_log_items = 0; settings->rotate_log_storage_interval = 10; - testLogAndStateMachine(settings, 40, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 40, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 1; settings->reserved_log_items = 1; settings->rotate_log_storage_interval = 32; - testLogAndStateMachine(settings, 32, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 32, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 10; settings->reserved_log_items = 7; settings->rotate_log_storage_interval = 1; - testLogAndStateMachine(settings, 33, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 33, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 37; settings->reserved_log_items = 1000; settings->rotate_log_storage_interval = 5000; - testLogAndStateMachine(settings, 33, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 33, params.enable_compression); } { CoordinationSettingsPtr settings = std::make_shared(); settings->snapshot_distance = 37; settings->reserved_log_items = 1000; settings->rotate_log_storage_interval = 5000; - testLogAndStateMachine(settings, 45, params.enable_compression, keeper_context); + testLogAndStateMachine(settings, 45, params.enable_compression); } } @@ -1931,11 +1942,10 @@ TEST_P(CoordinationTest, TestEphemeralNodeRemove) ChangelogDirTest snapshots("./snapshots"); setSnapshotDirectory("./snapshots"); - CoordinationSettingsPtr settings = std::make_shared(); - ResponsesQueue queue(std::numeric_limits::max()); SnapshotsQueue snapshots_queue{1}; - auto state_machine = std::make_shared(queue, snapshots_queue, settings, keeper_context, nullptr); + + auto state_machine = std::make_shared(queue, snapshots_queue, keeper_context, nullptr); state_machine->init(); std::shared_ptr request_c = std::make_shared(); @@ -1965,11 +1975,10 @@ TEST_P(CoordinationTest, TestCreateNodeWithAuthSchemeForAclWhenAuthIsPrecommitte ChangelogDirTest snapshots("./snapshots"); setSnapshotDirectory("./snapshots"); - CoordinationSettingsPtr settings = std::make_shared(); ResponsesQueue queue(std::numeric_limits::max()); SnapshotsQueue snapshots_queue{1}; - auto state_machine = std::make_shared(queue, snapshots_queue, settings, keeper_context, nullptr); + auto state_machine = std::make_shared(queue, snapshots_queue, keeper_context, nullptr); state_machine->init(); String user_auth_data = "test_user:test_password"; @@ -2017,11 +2026,10 @@ TEST_P(CoordinationTest, TestSetACLWithAuthSchemeForAclWhenAuthIsPrecommitted) ChangelogDirTest snapshots("./snapshots"); setSnapshotDirectory("./snapshots"); - CoordinationSettingsPtr settings = std::make_shared(); ResponsesQueue queue(std::numeric_limits::max()); SnapshotsQueue snapshots_queue{1}; - auto state_machine = std::make_shared(queue, snapshots_queue, settings, keeper_context, nullptr); + auto state_machine = std::make_shared(queue, snapshots_queue, keeper_context, nullptr); state_machine->init(); String user_auth_data = "test_user:test_password"; @@ -2132,6 +2140,7 @@ TEST_P(CoordinationTest, TestRotateIntervalChanges) waitDurableLogs(changelog_2); + keeper_context->setLastCommitIndex(105); changelog_2.compact(105); std::this_thread::sleep_for(std::chrono::microseconds(1000)); @@ -2157,6 +2166,7 @@ TEST_P(CoordinationTest, TestRotateIntervalChanges) waitDurableLogs(changelog_3); + keeper_context->setLastCommitIndex(125); changelog_3.compact(125); std::this_thread::sleep_for(std::chrono::microseconds(1000)); assertFileDeleted("./logs/changelog_101_110.bin" + params.extension); diff --git a/src/Core/MySQL/PacketEndpoint.cpp b/src/Core/MySQL/PacketEndpoint.cpp index 085d7595167..97b5d3b4d11 100644 --- a/src/Core/MySQL/PacketEndpoint.cpp +++ b/src/Core/MySQL/PacketEndpoint.cpp @@ -40,7 +40,7 @@ bool PacketEndpoint::tryReceivePacket(IMySQLReadPacket & packet, UInt64 millisec ReadBufferFromPocoSocket * socket_in = typeid_cast(in); if (!socket_in) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Attempt to pull the duration in a non socket stream"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: Attempt to pull the duration in a non socket stream"); if (!socket_in->poll(millisecond * 1000)) return false; diff --git a/src/Core/Settings.h b/src/Core/Settings.h index 53ba18a3001..433195af9c3 100644 --- a/src/Core/Settings.h +++ b/src/Core/Settings.h @@ -872,7 +872,6 @@ class IColumn; M(UInt64, cache_warmer_threads, 4, "Only available in ClickHouse Cloud", 0) \ M(Int64, ignore_cold_parts_seconds, 0, "Only available in ClickHouse Cloud", 0) \ M(Int64, prefer_warmed_unmerged_parts_seconds, 0, "Only available in ClickHouse Cloud", 0) \ - M(Bool, enable_order_by_all, true, "Enable sorting expression ORDER BY ALL.", 0) \ M(Bool, iceberg_engine_ignore_schema_evolution, false, "Ignore schema evolution in Iceberg table engine and read all data using latest schema saved on table creation. Note that it can lead to incorrect result", 0) \ // End of COMMON_SETTINGS @@ -940,6 +939,7 @@ class IColumn; MAKE_OBSOLETE(M, Bool, allow_experimental_undrop_table_query, true) \ MAKE_OBSOLETE(M, Bool, allow_experimental_s3queue, true) \ MAKE_OBSOLETE(M, Bool, query_plan_optimize_primary_key, true) \ + MAKE_OBSOLETE(M, Bool, enable_order_by_all, true) \ /** The section above is for obsolete settings. Do not add anything there. */ @@ -1019,6 +1019,7 @@ class IColumn; M(Bool, input_format_try_infer_integers, true, "Try to infer integers instead of floats while schema inference in text formats", 0) \ M(Bool, input_format_try_infer_dates, true, "Try to infer dates from string fields while schema inference in text formats", 0) \ M(Bool, input_format_try_infer_datetimes, true, "Try to infer datetimes from string fields while schema inference in text formats", 0) \ + M(Bool, input_format_try_infer_exponent_floats, false, "Try to infer floats in exponential notation while schema inference in text formats", 0) \ M(Bool, output_format_markdown_escape_special_characters, false, "Escape special characters in Markdown", 0) \ M(Bool, input_format_protobuf_flatten_google_wrappers, false, "Enable Google wrappers for regular non-nested columns, e.g. google.protobuf.StringValue 'str' for String column 'str'. For Nullable columns empty wrappers are recognized as defaults, and missing as nulls", 0) \ M(Bool, output_format_protobuf_nullables_with_google_wrappers, false, "When serializing Nullable columns with Google wrappers, serialize default values as empty wrappers. If turned off, default and null values are not serialized", 0) \ diff --git a/src/Core/SettingsChangesHistory.h b/src/Core/SettingsChangesHistory.h index 64650bf0ef5..e97a411e2c1 100644 --- a/src/Core/SettingsChangesHistory.h +++ b/src/Core/SettingsChangesHistory.h @@ -85,6 +85,8 @@ namespace SettingsChangesHistory static std::map settings_changes_history = { {"24.2", { + {"output_format_values_escape_quote_with_quote", false, false, "If true escape ' with '', otherwise quoted with \\'"}, + {"input_format_try_infer_exponent_floats", true, false, "Don't infer floats in exponential notation by default"}, {"async_insert_max_data_size", 1000000, 10485760, "The previous value appeared to be too small."}, {"async_insert_poll_timeout_ms", 10, 10, "Timeout in milliseconds for polling data from asynchronous insert queue"}, {"async_insert_use_adaptive_busy_timeout", true, true, "Use adaptive asynchronous insert timeout"}, diff --git a/src/Daemon/SentryWriter.cpp b/src/Daemon/SentryWriter.cpp index ebfd18abeee..192e9952b9a 100644 --- a/src/Daemon/SentryWriter.cpp +++ b/src/Daemon/SentryWriter.cpp @@ -78,7 +78,7 @@ void SentryWriter::initialize(Poco::Util::LayeredConfiguration & config) if (enabled) { - server_data_path = config.getString("path", ""); + server_data_path = config.getString("path", DB::DBMS_DEFAULT_PATH); const std::filesystem::path & default_tmp_path = fs::path(config.getString("tmp_path", fs::temp_directory_path())) / "sentry"; const std::string & endpoint = config.getString("send_crash_reports.endpoint"); diff --git a/src/DataTypes/DataTypeAggregateFunction.cpp b/src/DataTypes/DataTypeAggregateFunction.cpp index 14a3c6a4248..7dc036cafa4 100644 --- a/src/DataTypes/DataTypeAggregateFunction.cpp +++ b/src/DataTypes/DataTypeAggregateFunction.cpp @@ -239,7 +239,7 @@ static DataTypePtr create(const ASTPtr & arguments) argument_types.push_back(DataTypeFactory::instance().get(arguments->children[i])); if (function_name.empty()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Empty name of aggregate function passed"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: empty name of aggregate function passed"); AggregateFunctionProperties properties; AggregateFunctionPtr function = AggregateFunctionFactory::instance().get(function_name, action, argument_types, params_row, properties); diff --git a/src/DataTypes/DataTypeArray.cpp b/src/DataTypes/DataTypeArray.cpp index 24cd759e2a5..6e5760933eb 100644 --- a/src/DataTypes/DataTypeArray.cpp +++ b/src/DataTypes/DataTypeArray.cpp @@ -69,6 +69,11 @@ String DataTypeArray::doGetPrettyName(size_t indent) const return s.str(); } +void DataTypeArray::forEachChild(const ChildCallback & callback) const +{ + callback(*nested); + nested->forEachChild(callback); +} static DataTypePtr create(const ASTPtr & arguments) { diff --git a/src/DataTypes/DataTypeArray.h b/src/DataTypes/DataTypeArray.h index 6a09b3b530d..4423f137e1a 100644 --- a/src/DataTypes/DataTypeArray.h +++ b/src/DataTypes/DataTypeArray.h @@ -43,6 +43,7 @@ public: MutableColumnPtr createColumn() const override; + void forEachChild(const ChildCallback & callback) const override; Field getDefault() const override; diff --git a/src/DataTypes/DataTypeCustomSimpleAggregateFunction.cpp b/src/DataTypes/DataTypeCustomSimpleAggregateFunction.cpp index ee9870eb0ef..aa3b154e49b 100644 --- a/src/DataTypes/DataTypeCustomSimpleAggregateFunction.cpp +++ b/src/DataTypes/DataTypeCustomSimpleAggregateFunction.cpp @@ -141,7 +141,7 @@ static std::pair create(const ASTPtr & argum argument_types.push_back(DataTypeFactory::instance().get(arguments->children[i])); if (function_name.empty()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Empty name of aggregate function passed"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: empty name of aggregate function passed"); AggregateFunctionProperties properties; /// NullsAction is not part of the type definition, instead it will have transformed the function into a different one diff --git a/src/DataTypes/DataTypeLowCardinality.cpp b/src/DataTypes/DataTypeLowCardinality.cpp index 3e94b533c7a..5af1f28cbad 100644 --- a/src/DataTypes/DataTypeLowCardinality.cpp +++ b/src/DataTypes/DataTypeLowCardinality.cpp @@ -153,6 +153,12 @@ SerializationPtr DataTypeLowCardinality::doGetDefaultSerialization() const return std::make_shared(dictionary_type); } +void DataTypeLowCardinality::forEachChild(const ChildCallback & callback) const +{ + callback(*dictionary_type); + dictionary_type->forEachChild(callback); +} + static DataTypePtr create(const ASTPtr & arguments) { diff --git a/src/DataTypes/DataTypeLowCardinality.h b/src/DataTypes/DataTypeLowCardinality.h index 389e24ef2a9..cd926bb595c 100644 --- a/src/DataTypes/DataTypeLowCardinality.h +++ b/src/DataTypes/DataTypeLowCardinality.h @@ -60,6 +60,8 @@ public: static MutableColumnUniquePtr createColumnUnique(const IDataType & keys_type); static MutableColumnUniquePtr createColumnUnique(const IDataType & keys_type, MutableColumnPtr && keys); + void forEachChild(const ChildCallback & callback) const override; + private: SerializationPtr doGetDefaultSerialization() const override; diff --git a/src/DataTypes/DataTypeMap.cpp b/src/DataTypes/DataTypeMap.cpp index 1f246af74d3..4b85606ff26 100644 --- a/src/DataTypes/DataTypeMap.cpp +++ b/src/DataTypes/DataTypeMap.cpp @@ -143,6 +143,14 @@ DataTypePtr DataTypeMap::getNestedTypeWithUnnamedTuple() const return std::make_shared(std::make_shared(from_tuple.getElements())); } +void DataTypeMap::forEachChild(const DB::IDataType::ChildCallback & callback) const +{ + callback(*key_type); + key_type->forEachChild(callback); + callback(*value_type); + value_type->forEachChild(callback); +} + static DataTypePtr create(const ASTPtr & arguments) { if (!arguments || arguments->children.size() != 2) diff --git a/src/DataTypes/DataTypeMap.h b/src/DataTypes/DataTypeMap.h index 257888a8e44..7281cca1bb1 100644 --- a/src/DataTypes/DataTypeMap.h +++ b/src/DataTypes/DataTypeMap.h @@ -54,6 +54,8 @@ public: static bool checkKeyType(DataTypePtr key_type); + void forEachChild(const ChildCallback & callback) const override; + private: void assertKeyType() const; }; diff --git a/src/DataTypes/DataTypeNullable.cpp b/src/DataTypes/DataTypeNullable.cpp index 484d779551f..16d5d41e5e5 100644 --- a/src/DataTypes/DataTypeNullable.cpp +++ b/src/DataTypes/DataTypeNullable.cpp @@ -61,6 +61,12 @@ SerializationPtr DataTypeNullable::doGetDefaultSerialization() const return std::make_shared(nested_data_type->getDefaultSerialization()); } +void DataTypeNullable::forEachChild(const ChildCallback & callback) const +{ + callback(*nested_data_type); + nested_data_type->forEachChild(callback); +} + static DataTypePtr create(const ASTPtr & arguments) { diff --git a/src/DataTypes/DataTypeNullable.h b/src/DataTypes/DataTypeNullable.h index 7ad0e1ba5f1..b102c767993 100644 --- a/src/DataTypes/DataTypeNullable.h +++ b/src/DataTypes/DataTypeNullable.h @@ -43,6 +43,9 @@ public: bool canBePromoted() const override { return nested_data_type->canBePromoted(); } const DataTypePtr & getNestedType() const { return nested_data_type; } + + void forEachChild(const ChildCallback & callback) const override; + private: SerializationPtr doGetDefaultSerialization() const override; diff --git a/src/DataTypes/DataTypeTuple.cpp b/src/DataTypes/DataTypeTuple.cpp index 5c9d5a3366e..eb218d8efb7 100644 --- a/src/DataTypes/DataTypeTuple.cpp +++ b/src/DataTypes/DataTypeTuple.cpp @@ -376,6 +376,14 @@ SerializationInfoPtr DataTypeTuple::getSerializationInfo(const IColumn & column) return std::make_shared(std::move(infos), names, SerializationInfo::Settings{}); } +void DataTypeTuple::forEachChild(const ChildCallback & callback) const +{ + for (const auto & elem : elems) + { + callback(*elem); + elem->forEachChild(callback); + } +} static DataTypePtr create(const ASTPtr & arguments) { diff --git a/src/DataTypes/DataTypeTuple.h b/src/DataTypes/DataTypeTuple.h index db49b7f22d1..4e5a0c1b33c 100644 --- a/src/DataTypes/DataTypeTuple.h +++ b/src/DataTypes/DataTypeTuple.h @@ -70,6 +70,8 @@ public: String getNameByPosition(size_t i) const; bool haveExplicitNames() const { return have_explicit_names; } + + void forEachChild(const ChildCallback & callback) const override; }; } diff --git a/src/DataTypes/DataTypeVariant.cpp b/src/DataTypes/DataTypeVariant.cpp index 456b4ea03b6..0543507a14d 100644 --- a/src/DataTypes/DataTypeVariant.cpp +++ b/src/DataTypes/DataTypeVariant.cpp @@ -175,6 +175,15 @@ SerializationPtr DataTypeVariant::doGetDefaultSerialization() const return std::make_shared(std::move(serializations), std::move(variant_names), SerializationVariant::getVariantsDeserializeTextOrder(variants), getName()); } +void DataTypeVariant::forEachChild(const DB::IDataType::ChildCallback & callback) const +{ + for (const auto & variant : variants) + { + callback(*variant); + variant->forEachChild(callback); + } +} + static DataTypePtr create(const ASTPtr & arguments) { if (!arguments || arguments->children.empty()) diff --git a/src/DataTypes/DataTypeVariant.h b/src/DataTypes/DataTypeVariant.h index d26ce4ea90f..2a2206f985a 100644 --- a/src/DataTypes/DataTypeVariant.h +++ b/src/DataTypes/DataTypeVariant.h @@ -54,6 +54,8 @@ public: /// Check if Variant has provided type in the list of variants and return its discriminator. std::optional tryGetVariantDiscriminator(const DataTypePtr & type) const; + void forEachChild(const ChildCallback & callback) const override; + private: std::string doGetName() const override; std::string doGetPrettyName(size_t indent) const override; diff --git a/src/DataTypes/IDataType.h b/src/DataTypes/IDataType.h index 48cc127746f..220658afda5 100644 --- a/src/DataTypes/IDataType.h +++ b/src/DataTypes/IDataType.h @@ -111,6 +111,10 @@ public: const SubcolumnCallback & callback, const SubstreamData & data); + /// Call callback for each nested type recursively. + using ChildCallback = std::function; + virtual void forEachChild(const ChildCallback &) const {} + Names getSubcolumnNames() const; virtual MutableSerializationInfoPtr createSerializationInfo(const SerializationInfo::Settings & settings) const; diff --git a/src/Databases/MySQL/MaterializedMySQLSyncThread.cpp b/src/Databases/MySQL/MaterializedMySQLSyncThread.cpp index 20db8036942..2656835f912 100644 --- a/src/Databases/MySQL/MaterializedMySQLSyncThread.cpp +++ b/src/Databases/MySQL/MaterializedMySQLSyncThread.cpp @@ -779,7 +779,7 @@ static void writeFieldsToColumn( casted_int32_column->insertValue(num & 0x800000 ? num | 0xFF000000 : num); } else - throw Exception(ErrorCodes::LOGICAL_ERROR, "MaterializedMySQL is a bug."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: it is a bug."); } } } @@ -844,7 +844,7 @@ static inline bool differenceSortingKeys(const Tuple & row_old_data, const Tuple static inline size_t onUpdateData(const Row & rows_data, Block & buffer, size_t version, const std::vector & sorting_columns_index) { if (rows_data.size() % 2 != 0) - throw Exception(ErrorCodes::LOGICAL_ERROR, "MaterializedMySQL is a bug."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: It is a bug."); size_t prev_bytes = buffer.bytes(); std::vector writeable_rows_mask(rows_data.size()); diff --git a/src/Disks/IO/ThreadPoolReader.h b/src/Disks/IO/ThreadPoolReader.h index 42bc9bf8bb4..b8aff9f22a2 100644 --- a/src/Disks/IO/ThreadPoolReader.h +++ b/src/Disks/IO/ThreadPoolReader.h @@ -3,7 +3,7 @@ #include #include #include -#include +#include namespace DB diff --git a/src/Disks/IO/ThreadPoolRemoteFSReader.h b/src/Disks/IO/ThreadPoolRemoteFSReader.h index cd2bf223f33..abc251b2b10 100644 --- a/src/Disks/IO/ThreadPoolRemoteFSReader.h +++ b/src/Disks/IO/ThreadPoolRemoteFSReader.h @@ -3,7 +3,7 @@ #include #include #include -#include +#include namespace DB { diff --git a/src/Disks/ObjectStorages/IObjectStorage.h b/src/Disks/ObjectStorages/IObjectStorage.h index 049935ad60c..56c269a3fc5 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.h +++ b/src/Disks/ObjectStorages/IObjectStorage.h @@ -22,7 +22,7 @@ #include #include #include -#include +#include #include #include "config.h" diff --git a/src/Disks/ObjectStorages/ObjectStorageIteratorAsync.h b/src/Disks/ObjectStorages/ObjectStorageIteratorAsync.h index a6abe03bac9..5f63e5f6e8a 100644 --- a/src/Disks/ObjectStorages/ObjectStorageIteratorAsync.h +++ b/src/Disks/ObjectStorages/ObjectStorageIteratorAsync.h @@ -2,7 +2,7 @@ #include #include -#include +#include #include #include diff --git a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp index 4cc49288af6..5771eb1ebe0 100644 --- a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp +++ b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp @@ -13,7 +13,7 @@ #include #include #include -#include +#include #include #include diff --git a/src/Formats/FormatFactory.cpp b/src/Formats/FormatFactory.cpp index 8c1f5f40626..0654dd01e49 100644 --- a/src/Formats/FormatFactory.cpp +++ b/src/Formats/FormatFactory.cpp @@ -229,6 +229,7 @@ FormatSettings getFormatSettings(const ContextPtr & context, const Settings & se format_settings.try_infer_integers = settings.input_format_try_infer_integers; format_settings.try_infer_dates = settings.input_format_try_infer_dates; format_settings.try_infer_datetimes = settings.input_format_try_infer_datetimes; + format_settings.try_infer_exponent_floats = settings.input_format_try_infer_exponent_floats; format_settings.markdown.escape_special_characters = settings.output_format_markdown_escape_special_characters; format_settings.bson.output_string_as_string = settings.output_format_bson_string_as_string; format_settings.bson.skip_fields_with_unsupported_types_in_schema_inference = settings.input_format_bson_skip_fields_with_unsupported_types_in_schema_inference; diff --git a/src/Formats/FormatSettings.h b/src/Formats/FormatSettings.h index cda15fd6531..aa37216d381 100644 --- a/src/Formats/FormatSettings.h +++ b/src/Formats/FormatSettings.h @@ -46,6 +46,7 @@ struct FormatSettings bool try_infer_integers = false; bool try_infer_dates = false; bool try_infer_datetimes = false; + bool try_infer_exponent_floats = false; enum class DateTimeInputFormat { diff --git a/src/Formats/SchemaInferenceUtils.cpp b/src/Formats/SchemaInferenceUtils.cpp index 2cfcff75edd..06b52e7a7a2 100644 --- a/src/Formats/SchemaInferenceUtils.cpp +++ b/src/Formats/SchemaInferenceUtils.cpp @@ -17,6 +17,7 @@ #include #include #include +#include #include #include @@ -865,6 +866,13 @@ namespace return std::make_shared(nested_types); } + bool tryReadFloat(Float64 & value, ReadBuffer & buf, const FormatSettings & settings) + { + if (settings.try_infer_exponent_floats) + return tryReadFloatText(value, buf); + return tryReadFloatTextNoExponent(value, buf); + } + DataTypePtr tryInferNumber(ReadBuffer & buf, const FormatSettings & settings) { if (buf.eof()) @@ -903,7 +911,7 @@ namespace buf.position() = number_start; } - if (tryReadFloatText(tmp_float, buf)) + if (tryReadFloat(tmp_float, buf, settings)) { if (read_int && buf.position() == int_end) return std::make_shared(); @@ -937,7 +945,7 @@ namespace peekable_buf.rollbackToCheckpoint(true); } - if (tryReadFloatText(tmp_float, peekable_buf)) + if (tryReadFloat(tmp_float, peekable_buf, settings)) { /// Float parsing reads no fewer bytes than integer parsing, /// so position of the buffer is either the same, or further. @@ -949,7 +957,7 @@ namespace return std::make_shared(); } } - else if (tryReadFloatText(tmp_float, buf)) + else if (tryReadFloat(tmp_float, buf, settings)) { return std::make_shared(); } @@ -1390,7 +1398,7 @@ DataTypePtr tryInferNumberFromString(std::string_view field, const FormatSetting buf.position() = buf.buffer().begin(); Float64 tmp; - if (tryReadFloatText(tmp, buf) && buf.eof()) + if (tryReadFloat(tmp, buf, settings) && buf.eof()) return std::make_shared(); return nullptr; diff --git a/src/Functions/CMakeLists.txt b/src/Functions/CMakeLists.txt index a06e898b7c5..ac3e3671ae0 100644 --- a/src/Functions/CMakeLists.txt +++ b/src/Functions/CMakeLists.txt @@ -9,6 +9,11 @@ extract_into_parent_list(clickhouse_functions_sources dbms_sources FunctionHelpers.cpp extractTimeZoneFromFunctionArguments.cpp FunctionsLogical.cpp + if.cpp + multiIf.cpp + multiMatchAny.cpp + checkHyperscanRegexp.cpp + array/has.cpp CastOverloadResolver.cpp ) extract_into_parent_list(clickhouse_functions_headers dbms_headers diff --git a/src/Functions/EmptyImpl.h b/src/Functions/EmptyImpl.h index d3b2dda024b..52484524e6a 100644 --- a/src/Functions/EmptyImpl.h +++ b/src/Functions/EmptyImpl.h @@ -35,7 +35,7 @@ struct EmptyImpl /// Only make sense if is_fixed_to_constant. static void vectorFixedToConstant(const ColumnString::Chars & /*data*/, size_t /*n*/, UInt8 & /*res*/) { - throw Exception(ErrorCodes::LOGICAL_ERROR, "'vectorFixedToConstant method' is called"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: 'vectorFixedToConstant method' is called"); } static void vectorFixedToVector(const ColumnString::Chars & data, size_t n, PaddedPODArray & res) diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index 1b2519d1ec5..d253095ca01 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -147,11 +147,32 @@ private: /// it's not correct for Decimal public: static constexpr bool allow_decimal = IsOperation::allow_decimal; + using DecimalResultDataType = Switch< + Case, + Case && IsDataTypeDecimal && UseLeftDecimal, LeftDataType>, + Case && IsDataTypeDecimal, RightDataType>, + Case && IsIntegralOrExtended, LeftDataType>, + Case && IsIntegralOrExtended, RightDataType>, + + /// e.g Decimal +-*/ Float, least(Decimal, Float), greatest(Decimal, Float) = Float64 + Case && IsFloatingPoint, DataTypeFloat64>, + Case && IsFloatingPoint, DataTypeFloat64>, + + Case::bit_hamming_distance && IsIntegral && IsIntegral, DataTypeUInt8>, + Case::bit_hamming_distance && IsFixedString && IsFixedString, DataTypeUInt16>, + Case::bit_hamming_distance && IsString && IsString, DataTypeUInt64>, + + /// Decimal Real is not supported (traditional DBs convert Decimal Real to Real) + Case && !IsIntegralOrExtendedOrDecimal, InvalidType>, + Case && !IsIntegralOrExtendedOrDecimal, InvalidType>>; + /// Appropriate result type for binary operator on numeric types. "Date" can also mean /// DateTime, but if both operands are Dates, their type must be the same (e.g. Date - DateTime is invalid). using ResultDataType = Switch< + /// Result must be Integer + Case::div_int || IsOperation::div_int_or_zero, DataTypeFromFieldType>, /// Decimal cases - Case || IsDataTypeDecimal), InvalidType>, + Case || IsDataTypeDecimal, DecimalResultDataType>, Case< IsDataTypeDecimal && IsDataTypeDecimal && UseLeftDecimal, LeftDataType>, @@ -622,7 +643,11 @@ private: if constexpr (op_case == OpCase::RightConstant) { if ((*right_nullmap)[0]) + { + for (size_t i = 0; i < size; ++i) + c[i] = ResultType(); return; + } for (size_t i = 0; i < size; ++i) c[i] = apply_func(undec(a[i]), undec(b)); @@ -1665,7 +1690,9 @@ public: if constexpr (!std::is_same_v) { - if constexpr (IsDataTypeDecimal && IsDataTypeDecimal) + if constexpr (is_div_int || is_div_int_or_zero) + type_res = std::make_shared(); + else if constexpr (IsDataTypeDecimal && IsDataTypeDecimal) { if constexpr (is_division) { @@ -1685,13 +1712,19 @@ public: ResultDataType result_type = decimalResultType(left, right); type_res = std::make_shared(result_type.getPrecision(), result_type.getScale()); } - else if constexpr ((IsDataTypeDecimal && IsFloatingPoint) || - (IsDataTypeDecimal && IsFloatingPoint)) + else if constexpr (((IsDataTypeDecimal && IsFloatingPoint) || + (IsDataTypeDecimal && IsFloatingPoint))) + { type_res = std::make_shared(); + } else if constexpr (IsDataTypeDecimal) + { type_res = std::make_shared(left.getPrecision(), left.getScale()); + } else if constexpr (IsDataTypeDecimal) + { type_res = std::make_shared(right.getPrecision(), right.getScale()); + } else if constexpr (std::is_same_v) { // Special case for DateTime: binary OPS should reuse timezone @@ -2000,6 +2033,7 @@ ColumnPtr executeStringInteger(const ColumnsWithTypeAndName & arguments, const A using LeftDataType = std::decay_t; using RightDataType = std::decay_t; using ResultDataType = typename BinaryOperationTraits::ResultDataType; + using DecimalResultType = typename BinaryOperationTraits::DecimalResultDataType; if constexpr (std::is_same_v) return nullptr; @@ -2051,6 +2085,35 @@ ColumnPtr executeStringInteger(const ColumnsWithTypeAndName & arguments, const A col_left_size, right_nullmap); } + /// Here we check if we have `intDiv` or `intDivOrZero` and at least one of the arguments is decimal, because in this case originally we had result as decimal, so we need to convert result into integer after calculations + else if constexpr (!decimal_with_float && (is_div_int || is_div_int_or_zero) && (IsDataTypeDecimal || IsDataTypeDecimal)) + { + + if constexpr (!std::is_same_v) + { + DataTypePtr type_res; + if constexpr (IsDataTypeDecimal && IsDataTypeDecimal) + { + DecimalResultType result_type = decimalResultType(left, right); + type_res = std::make_shared(result_type.getPrecision(), result_type.getScale()); + } + else if constexpr (IsDataTypeDecimal) + type_res = std::make_shared(left.getPrecision(), left.getScale()); + else + type_res = std::make_shared(right.getPrecision(), right.getScale()); + + auto res = executeNumericWithDecimal( + left, right, + col_left_const, col_right_const, + col_left, col_right, + col_left_size, + right_nullmap); + + auto col = ColumnWithTypeAndName(res, type_res, name); + return castColumn(col, std::make_shared()); + } + return nullptr; + } else // can't avoid else and another indentation level, otherwise the compiler would try to instantiate // ColVecResult for Decimals which would lead to a compile error. { diff --git a/src/Functions/FunctionsComparison.h b/src/Functions/FunctionsComparison.h index 3be675f39b3..d04f76d051a 100644 --- a/src/Functions/FunctionsComparison.h +++ b/src/Functions/FunctionsComparison.h @@ -811,7 +811,7 @@ private: c0_const_size = c0_const_fixed_string->getN(); } else - throw Exception(ErrorCodes::ILLEGAL_COLUMN, "ColumnConst contains not String nor FixedString column"); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Logical error: ColumnConst contains not String nor FixedString column"); } if (c1_const) @@ -830,7 +830,7 @@ private: c1_const_size = c1_const_fixed_string->getN(); } else - throw Exception(ErrorCodes::ILLEGAL_COLUMN, "ColumnConst contains not String nor FixedString column"); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Logical error: ColumnConst contains not String nor FixedString column"); } using StringImpl = StringComparisonImpl>; @@ -1114,7 +1114,7 @@ private: /// This is a paranoid check to protect from a broken query analysis. if (c0->isNullable() != c1->isNullable()) throw Exception(ErrorCodes::LOGICAL_ERROR, - "Columns are assumed to be of identical types, but they are different in Nullable"); + "Logical error: columns are assumed to be of identical types, but they are different in Nullable"); if (c0_const && c1_const) { diff --git a/src/Functions/FunctionsConversion.h b/src/Functions/FunctionsConversion.h index 62148fa8022..4089a5b542b 100644 --- a/src/Functions/FunctionsConversion.h +++ b/src/Functions/FunctionsConversion.h @@ -2560,7 +2560,7 @@ public: if constexpr (std::is_same_v) res = std::make_shared(extractTimeZoneNameFromFunctionArguments(arguments, 1, 0, false)); else if constexpr (std::is_same_v) - throw Exception(ErrorCodes::LOGICAL_ERROR, "MaterializedMySQL is a bug."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: It is a bug."); else if constexpr (to_decimal) { UInt64 scale = extractToDecimalScale(arguments[1]); diff --git a/src/Functions/FunctionsLogical.cpp b/src/Functions/FunctionsLogical.cpp index 2e0f4cd3038..d0795941e1f 100644 --- a/src/Functions/FunctionsLogical.cpp +++ b/src/Functions/FunctionsLogical.cpp @@ -1,5 +1,6 @@ #include #include +#include #include #include @@ -530,7 +531,7 @@ DataTypePtr FunctionAnyArityLogical::getReturnTypeImpl(const DataTyp { has_nullable_arguments = arg_type->isNullable(); if (has_nullable_arguments && !Impl::specialImplementationForNulls()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unexpected type of argument for function \"{}\": " + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: Unexpected type of argument for function \"{}\": " " argument {} is of type {}", getName(), i + 1, arg_type->getName()); } @@ -776,4 +777,21 @@ ColumnPtr FunctionUnaryLogical::executeImpl(const ColumnsWithTypeAnd return res; } +FunctionOverloadResolverPtr createInternalFunctionOrOverloadResolver() +{ + return std::make_unique(std::make_shared()); +} +FunctionOverloadResolverPtr createInternalFunctionAndOverloadResolver() +{ + return std::make_unique(std::make_shared()); +} +FunctionOverloadResolverPtr createInternalFunctionXorOverloadResolver() +{ + return std::make_unique(std::make_shared()); +} +FunctionOverloadResolverPtr createInternalFunctionNotOverloadResolver() +{ + return std::make_unique(std::make_shared()); +} + } diff --git a/src/Functions/IsOperation.h b/src/Functions/IsOperation.h index 8ea53c865ce..b2c7a27d375 100644 --- a/src/Functions/IsOperation.h +++ b/src/Functions/IsOperation.h @@ -61,7 +61,7 @@ struct IsOperation static constexpr bool bit_hamming_distance = IsSameOperation::value; static constexpr bool division = div_floating || div_int || div_int_or_zero || modulo; - + // NOTE: allow_decimal should not fully contain `division` because of divInt static constexpr bool allow_decimal = plus || minus || multiply || division || least || greatest; }; diff --git a/src/Functions/array/arrayDistance.cpp b/src/Functions/array/arrayDistance.cpp index 670442c0c79..71564f6fa93 100644 --- a/src/Functions/array/arrayDistance.cpp +++ b/src/Functions/array/arrayDistance.cpp @@ -90,17 +90,19 @@ struct L2Distance size_t & i_y, State & state) { + static constexpr bool is_float32 = std::is_same_v; + __m512 sums; - if constexpr (std::is_same_v) + if constexpr (is_float32) sums = _mm512_setzero_ps(); else sums = _mm512_setzero_pd(); - const size_t n = (std::is_same_v) ? 16 : 8; + constexpr size_t n = is_float32 ? 16 : 8; for (; i_x + n < i_max; i_x += n, i_y += n) { - if constexpr (std::is_same_v) + if constexpr (is_float32) { __m512 x = _mm512_loadu_ps(data_x + i_x); __m512 y = _mm512_loadu_ps(data_y + i_y); @@ -116,7 +118,7 @@ struct L2Distance } } - if constexpr (std::is_same_v) + if constexpr (is_float32) state.sum = _mm512_reduce_add_ps(sums); else state.sum = _mm512_reduce_add_pd(sums); @@ -247,11 +249,13 @@ struct CosineDistance size_t & i_y, State & state) { + static constexpr bool is_float32 = std::is_same_v; + __m512 dot_products; __m512 x_squareds; __m512 y_squareds; - if constexpr (std::is_same_v) + if constexpr (is_float32) { dot_products = _mm512_setzero_ps(); x_squareds = _mm512_setzero_ps(); @@ -264,11 +268,11 @@ struct CosineDistance y_squareds = _mm512_setzero_pd(); } - const size_t n = (std::is_same_v) ? 16 : 8; + constexpr size_t n = is_float32 ? 16 : 8; for (; i_x + n < i_max; i_x += n, i_y += n) { - if constexpr (std::is_same_v) + if constexpr (is_float32) { __m512 x = _mm512_loadu_ps(data_x + i_x); __m512 y = _mm512_loadu_ps(data_y + i_y); @@ -286,7 +290,7 @@ struct CosineDistance } } - if constexpr (std::is_same_v) + if constexpr (is_float32) { state.dot_prod = _mm512_reduce_add_ps(dot_products); state.x_squared = _mm512_reduce_add_ps(x_squareds); @@ -312,7 +316,11 @@ template class FunctionArrayDistance : public IFunction { public: - String getName() const override { static auto name = String("array") + Kernel::name + "Distance"; return name; } + String getName() const override + { + static auto name = String("array") + Kernel::name + "Distance"; + return name; + } static FunctionPtr create(ContextPtr) { return std::make_shared>(); } size_t getNumberOfArguments() const override { return 2; } ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {}; } diff --git a/src/Functions/array/arrayDotProduct.cpp b/src/Functions/array/arrayDotProduct.cpp index 47e865785d4..6c615a058c3 100644 --- a/src/Functions/array/arrayDotProduct.cpp +++ b/src/Functions/array/arrayDotProduct.cpp @@ -1,44 +1,51 @@ +#include +#include +#include #include -#include -#include -#include -#include -#include -#include #include +#include +#include +#include +#include +#include +#if USE_MULTITARGET_CODE +#include +#endif namespace DB { namespace ErrorCodes { + extern const int BAD_ARGUMENTS; + extern const int ILLEGAL_COLUMN; extern const int ILLEGAL_TYPE_OF_ARGUMENT; + extern const int LOGICAL_ERROR; } -struct NameArrayDotProduct + +struct DotProduct { static constexpr auto name = "arrayDotProduct"; -}; -class ArrayDotProductImpl -{ -public: static DataTypePtr getReturnType(const DataTypePtr & left, const DataTypePtr & right) { using Types = TypeList; + Types types; DataTypePtr result_type; - bool valid = castTypeToEither(Types{}, left.get(), [&](const auto & left_) + bool valid = castTypeToEither(types, left.get(), [&](const auto & left_) { - return castTypeToEither(Types{}, right.get(), [&](const auto & right_) + return castTypeToEither(types, right.get(), [&](const auto & right_) { - using LeftDataType = typename std::decay_t::FieldType; - using RightDataType = typename std::decay_t::FieldType; - using ResultType = typename NumberTraits::ResultOfAdditionMultiplication::Type; - if (std::is_same_v && std::is_same_v) + using LeftType = typename std::decay_t::FieldType; + using RightType = typename std::decay_t::FieldType; + using ResultType = typename NumberTraits::ResultOfAdditionMultiplication::Type; + + if constexpr (std::is_same_v && std::is_same_v) result_type = std::make_shared(); else result_type = std::make_shared>(); @@ -49,26 +56,268 @@ public: if (!valid) throw Exception( ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, - "Arguments of function {} " - "only support: UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64, Float32, Float64.", - std::string(NameArrayDotProduct::name)); + "Arguments of function {} only support: UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64, Float32, Float64.", name); return result_type; } - template - static inline NO_SANITIZE_UNDEFINED ResultType apply( - const T * left, - const U * right, - size_t size) + template + struct State { - ResultType result = 0; - for (size_t i = 0; i < size; ++i) - result += static_cast(left[i]) * static_cast(right[i]); - return result; + Type sum = 0; + }; + + template + static void accumulate(State & state, Type x, Type y) + { + state.sum += x * y; + } + + template + static void combine(State & state, const State & other_state) + { + state.sum += other_state.sum; + } + +#if USE_MULTITARGET_CODE + template + AVX512_FUNCTION_SPECIFIC_ATTRIBUTE static void accumulateCombine( + const Type * __restrict data_x, + const Type * __restrict data_y, + size_t i_max, + size_t & i, + State & state) + { + static constexpr bool is_float32 = std::is_same_v; + + __m512 sums; + if constexpr (is_float32) + sums = _mm512_setzero_ps(); + else + sums = _mm512_setzero_pd(); + + constexpr size_t n = is_float32 ? 16 : 8; + + for (; i + n < i_max; i += n) + { + if constexpr (is_float32) + { + __m512 x = _mm512_loadu_ps(data_x + i); + __m512 y = _mm512_loadu_ps(data_y + i); + sums = _mm512_fmadd_ps(x, y, sums); + } + else + { + __m512 x = _mm512_loadu_pd(data_x + i); + __m512 y = _mm512_loadu_pd(data_y + i); + sums = _mm512_fmadd_pd(x, y, sums); + } + } + + if constexpr (is_float32) + state.sum = _mm512_reduce_add_ps(sums); + else + state.sum = _mm512_reduce_add_pd(sums); + } +#endif + + template + static Type finalize(const State & state) + { + return state.sum; + } + +}; + + +/// The implementation is modeled after the implementation of distance functions arrayL1Distance, arrayL2Distance, etc. +/// The main difference is that arrayDotProduct() interferes the result type differently. +template +class FunctionArrayScalarProduct : public IFunction +{ +public: + static constexpr auto name = Kernel::name; + + String getName() const override { return name; } + static FunctionPtr create(ContextPtr) { return std::make_shared(); } + size_t getNumberOfArguments() const override { return 2; } + bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return true; } + + DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override + { + std::array nested_types; + for (size_t i = 0; i < 2; ++i) + { + const DataTypeArray * array_type = checkAndGetDataType(arguments[i].get()); + if (!array_type) + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Arguments for function {} must be of type Array", getName()); + + const auto & nested_type = array_type->getNestedType(); + if (!isNativeNumber(nested_type)) + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Function {} cannot process values of type {}", getName(), nested_type->getName()); + + nested_types[i] = nested_type; + } + + return Kernel::getReturnType(nested_types[0], nested_types[1]); + } + + ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t /* input_rows_count */) const override + { + switch (result_type->getTypeId()) + { + #define SUPPORTED_TYPE(type) \ + case TypeIndex::type: \ + return executeWithResultType(arguments); \ + break; + + SUPPORTED_TYPE(UInt8) + SUPPORTED_TYPE(UInt16) + SUPPORTED_TYPE(UInt32) + SUPPORTED_TYPE(UInt64) + SUPPORTED_TYPE(Int8) + SUPPORTED_TYPE(Int16) + SUPPORTED_TYPE(Int32) + SUPPORTED_TYPE(Int64) + SUPPORTED_TYPE(Float32) + SUPPORTED_TYPE(Float64) + #undef SUPPORTED_TYPE + + default: + throw Exception(ErrorCodes::LOGICAL_ERROR, "Unexpected result type {}", result_type->getName()); + } + } + +private: + template + ColumnPtr executeWithResultType(const ColumnsWithTypeAndName & arguments) const + { + ColumnPtr res; + if (!((res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)) + || (res = executeWithResultTypeAndLeft(arguments)))) + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Illegal column {} of first argument of function {}", arguments[0].column->getName(), getName()); + + return res; + } + + template + ColumnPtr executeWithResultTypeAndLeft(const ColumnsWithTypeAndName & arguments) const + { + ColumnPtr res; + if ( (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments)) + || (res = executeWithResultTypeAndLeftAndRight(arguments))) + return res; + + return nullptr; + } + + template + ColumnPtr executeWithResultTypeAndLeftAndRight(const ColumnsWithTypeAndName & arguments) const + { + ColumnPtr col_left = arguments[0].column->convertToFullColumnIfConst(); + ColumnPtr col_right = arguments[1].column->convertToFullColumnIfConst(); + if (!col_left || !col_right) + return nullptr; + + const ColumnArray * col_arr_left = checkAndGetColumn(col_left.get()); + const ColumnArray * cokl_arr_right = checkAndGetColumn(col_right.get()); + if (!col_arr_left || !cokl_arr_right) + return nullptr; + + const ColumnVector * col_arr_nested_left = checkAndGetColumn>(col_arr_left->getData()); + const ColumnVector * col_arr_nested_right = checkAndGetColumn>(cokl_arr_right->getData()); + if (!col_arr_nested_left || !col_arr_nested_right) + return nullptr; + + if (!col_arr_left->hasEqualOffsets(*cokl_arr_right)) + throw Exception(ErrorCodes::BAD_ARGUMENTS, "Array arguments for function {} must have equal sizes", getName()); + + auto col_res = ColumnVector::create(); + + vector( + col_arr_nested_left->getData(), + col_arr_nested_right->getData(), + col_arr_left->getOffsets(), + col_res->getData()); + + return col_res; + } + + template + static void vector( + const PaddedPODArray & left, + const PaddedPODArray & right, + const ColumnArray::Offsets & offsets, + PaddedPODArray & result) + { + size_t size = offsets.size(); + result.resize(size); + + ColumnArray::Offset current_offset = 0; + for (size_t row = 0; row < size; ++row) + { + size_t array_size = offsets[row] - current_offset; + + typename Kernel::template State state; + size_t i = 0; + + /// SIMD optimization: process multiple elements in both input arrays at once. + /// To avoid combinatorial explosion of SIMD kernels, focus on + /// - the two most common input/output types (Float32 x Float32) --> Float32 and (Float64 x Float64) --> Float64 instead of 10 x + /// 10 input types x 8 output types, + /// - the most powerful SIMD instruction set (AVX-512F). +#if USE_MULTITARGET_CODE + if constexpr ((std::is_same_v || std::is_same_v) + && std::is_same_v && std::is_same_v) + { + if (isArchSupported(TargetArch::AVX512F)) + Kernel::template accumulateCombine(&left[current_offset], &right[current_offset], array_size, i, state); + } +#else + /// Process chunks in vectorized manner + static constexpr size_t VEC_SIZE = 4; + typename Kernel::template State states[VEC_SIZE]; + for (; i + VEC_SIZE < array_size; i += VEC_SIZE) + { + for (size_t j = 0; j < VEC_SIZE; ++j) + Kernel::template accumulate(states[j], static_cast(left[i + j]), static_cast(right[i + j])); + } + + for (const auto & other_state : states) + Kernel::template combine(state, other_state); +#endif + + /// Process the tail + for (; i < array_size; ++i) + Kernel::template accumulate(state, static_cast(left[i]), static_cast(right[i])); + + /// ResultType res = Kernel::template finalize(state); + result[row] = Kernel::template finalize(state); + + current_offset = offsets[row]; + } } }; -using FunctionArrayDotProduct = FunctionArrayScalarProduct; +using FunctionArrayDotProduct = FunctionArrayScalarProduct; REGISTER_FUNCTION(ArrayDotProduct) { @@ -77,4 +326,5 @@ REGISTER_FUNCTION(ArrayDotProduct) // These functions are used by TupleOrArrayFunction in Function/vectorFunctions.cpp FunctionPtr createFunctionArrayDotProduct(ContextPtr context_) { return FunctionArrayDotProduct::create(context_); } + } diff --git a/src/Functions/array/arrayReduce.cpp b/src/Functions/array/arrayReduce.cpp index 5a6a99ef785..d47d1ae98cc 100644 --- a/src/Functions/array/arrayReduce.cpp +++ b/src/Functions/array/arrayReduce.cpp @@ -1,14 +1,15 @@ -#include -#include -#include -#include -#include -#include -#include #include #include #include #include +#include +#include +#include +#include +#include +#include +#include +#include #include #include @@ -48,6 +49,11 @@ public: bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return true; } bool useDefaultImplementationForConstants() const override { return true; } + /// As we parse the function name and deal with arrays we don't want to default NULL handler, which will hide + /// nullability from us (which also means hidden from the aggregate functions) + bool useDefaultImplementationForNulls() const override { return false; } + /// Same for low cardinality. We want to return exactly what the aggregate function returns, no meddling + bool useDefaultImplementationForLowCardinalityColumns() const override { return false; } ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {0}; } DataTypePtr getReturnTypeImpl(const ColumnsWithTypeAndName & arguments) const override; @@ -115,7 +121,8 @@ ColumnPtr FunctionArrayReduce::executeImpl(const ColumnsWithTypeAndName & argume const IAggregateFunction & agg_func = *aggregate_function; std::unique_ptr arena = std::make_unique(); - /// Aggregate functions do not support constant columns. Therefore, we materialize them. + /// Aggregate functions do not support constant or lowcardinality columns. Therefore, we materialize them and + /// keep a reference so they are alive until we finish using their nested columns (array data/offset) std::vector materialized_columns; const size_t num_arguments_columns = arguments.size() - 1; @@ -126,6 +133,12 @@ ColumnPtr FunctionArrayReduce::executeImpl(const ColumnsWithTypeAndName & argume for (size_t i = 0; i < num_arguments_columns; ++i) { const IColumn * col = arguments[i + 1].column.get(); + auto col_no_lowcardinality = recursiveRemoveLowCardinality(arguments[i + 1].column); + if (col_no_lowcardinality != arguments[i + 1].column) + { + materialized_columns.emplace_back(col_no_lowcardinality); + col = col_no_lowcardinality.get(); + } const ColumnArray::Offsets * offsets_i = nullptr; if (const ColumnArray * arr = checkAndGetColumn(col)) diff --git a/src/Functions/array/arrayScalarProduct.h b/src/Functions/array/arrayScalarProduct.h deleted file mode 100644 index 374a2d8a194..00000000000 --- a/src/Functions/array/arrayScalarProduct.h +++ /dev/null @@ -1,182 +0,0 @@ -#pragma once - -#include -#include -#include -#include -#include -#include -#include - - -namespace DB -{ - -class Context; - -namespace ErrorCodes -{ - extern const int ILLEGAL_COLUMN; - extern const int ILLEGAL_TYPE_OF_ARGUMENT; - extern const int BAD_ARGUMENTS; - extern const int LOGICAL_ERROR; -} - - -template -class FunctionArrayScalarProduct : public IFunction -{ -public: - static constexpr auto name = Name::name; - static FunctionPtr create(ContextPtr) { return std::make_shared(); } - -private: - - template - ColumnPtr executeNumber(const ColumnsWithTypeAndName & arguments) const - { - ColumnPtr res; - if ( (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments)) - || (res = executeNumberNumber(arguments))) - return res; - - return nullptr; - } - - - template - ColumnPtr executeNumberNumber(const ColumnsWithTypeAndName & arguments) const - { - ColumnPtr col1 = arguments[0].column->convertToFullColumnIfConst(); - ColumnPtr col2 = arguments[1].column->convertToFullColumnIfConst(); - if (!col1 || !col2) - return nullptr; - - const ColumnArray * col_array1 = checkAndGetColumn(col1.get()); - const ColumnArray * col_array2 = checkAndGetColumn(col2.get()); - if (!col_array1 || !col_array2) - return nullptr; - - if (!col_array1->hasEqualOffsets(*col_array2)) - throw Exception(ErrorCodes::BAD_ARGUMENTS, "Array arguments for function {} must have equal sizes", getName()); - - const ColumnVector * col_nested1 = checkAndGetColumn>(col_array1->getData()); - const ColumnVector * col_nested2 = checkAndGetColumn>(col_array2->getData()); - if (!col_nested1 || !col_nested2) - return nullptr; - - auto col_res = ColumnVector::create(); - - vector( - col_nested1->getData(), - col_nested2->getData(), - col_array1->getOffsets(), - col_res->getData()); - - return col_res; - } - - template - static NO_INLINE void vector( - const PaddedPODArray & data1, - const PaddedPODArray & data2, - const ColumnArray::Offsets & offsets, - PaddedPODArray & result) - { - size_t size = offsets.size(); - result.resize(size); - - ColumnArray::Offset current_offset = 0; - for (size_t i = 0; i < size; ++i) - { - size_t array_size = offsets[i] - current_offset; - result[i] = Method::template apply(&data1[current_offset], &data2[current_offset], array_size); - current_offset = offsets[i]; - } - } - -public: - String getName() const override { return name; } - size_t getNumberOfArguments() const override { return 2; } - - bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return true; } - - - DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override - { - // Basic type check - std::vector nested_types(2, nullptr); - for (size_t i = 0; i < getNumberOfArguments(); ++i) - { - const DataTypeArray * array_type = checkAndGetDataType(arguments[i].get()); - if (!array_type) - throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "All arguments for function {} must be an array.", getName()); - - const auto & nested_type = array_type->getNestedType(); - if (!isNativeNumber(nested_type) && !isEnum(nested_type)) - throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "{} cannot process values of type {}", - getName(), nested_type->getName()); - nested_types[i] = nested_type; - } - - // Detail type check in Method, then return ReturnType - return Method::getReturnType(nested_types[0], nested_types[1]); - } - - template - ColumnPtr executeWithResultType(const ColumnsWithTypeAndName & arguments) const - { - ColumnPtr res; - if (!((res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)) - || (res = executeNumber(arguments)))) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Illegal column {} of first argument of function {}", arguments[0].column->getName(), getName()); - - return res; - } - - ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t /* input_rows_count */) const override - { - switch (result_type->getTypeId()) - { - #define SUPPORTED_TYPE(type) \ - case TypeIndex::type: \ - return executeWithResultType(arguments); \ - break; - - SUPPORTED_TYPE(UInt8) - SUPPORTED_TYPE(UInt16) - SUPPORTED_TYPE(UInt32) - SUPPORTED_TYPE(UInt64) - SUPPORTED_TYPE(Int8) - SUPPORTED_TYPE(Int16) - SUPPORTED_TYPE(Int32) - SUPPORTED_TYPE(Int64) - SUPPORTED_TYPE(Float32) - SUPPORTED_TYPE(Float64) - #undef SUPPORTED_TYPE - - default: - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unexpected result type {}", result_type->getName()); - } - } -}; - -} - diff --git a/src/Functions/array/has.cpp b/src/Functions/array/has.cpp index f08a4f29d2d..a17dcdcfbf9 100644 --- a/src/Functions/array/has.cpp +++ b/src/Functions/array/has.cpp @@ -9,4 +9,10 @@ struct NameHas { static constexpr auto name = "has"; }; using FunctionHas = FunctionArrayIndex; REGISTER_FUNCTION(Has) { factory.registerFunction(); } + +FunctionOverloadResolverPtr createInternalFunctionHasOverloadResolver() +{ + return std::make_unique(std::make_shared()); +} + } diff --git a/src/Functions/array/has.h b/src/Functions/array/has.h new file mode 100644 index 00000000000..226662d4051 --- /dev/null +++ b/src/Functions/array/has.h @@ -0,0 +1,12 @@ +#pragma once +#include + +namespace DB +{ + +class IFunctionOverloadResolver; +using FunctionOverloadResolverPtr = std::shared_ptr; + +FunctionOverloadResolverPtr createInternalFunctionHasOverloadResolver(); + +} diff --git a/src/Functions/if.cpp b/src/Functions/if.cpp index 7306dc4173e..70aced8842a 100644 --- a/src/Functions/if.cpp +++ b/src/Functions/if.cpp @@ -1413,4 +1413,9 @@ REGISTER_FUNCTION(If) factory.registerFunction({}, FunctionFactory::CaseInsensitive); } +FunctionOverloadResolverPtr createInternalFunctionIfOverloadResolver(bool allow_experimental_variant_type, bool use_variant_as_common_type) +{ + return std::make_unique(std::make_shared(allow_experimental_variant_type && use_variant_as_common_type)); +} + } diff --git a/src/Functions/if.h b/src/Functions/if.h new file mode 100644 index 00000000000..09a7a6a3e78 --- /dev/null +++ b/src/Functions/if.h @@ -0,0 +1,12 @@ +#pragma once +#include + +namespace DB +{ + +class IFunctionOverloadResolver; +using FunctionOverloadResolverPtr = std::shared_ptr; + +FunctionOverloadResolverPtr createInternalFunctionIfOverloadResolver(bool allow_experimental_variant_type, bool use_variant_as_common_type); + +} diff --git a/src/Functions/logical.h b/src/Functions/logical.h new file mode 100644 index 00000000000..d2d07f6cec7 --- /dev/null +++ b/src/Functions/logical.h @@ -0,0 +1,15 @@ +#pragma once +#include + +namespace DB +{ + +class IFunctionOverloadResolver; +using FunctionOverloadResolverPtr = std::shared_ptr; + +FunctionOverloadResolverPtr createInternalFunctionOrOverloadResolver(); +FunctionOverloadResolverPtr createInternalFunctionAndOverloadResolver(); +FunctionOverloadResolverPtr createInternalFunctionXorOverloadResolver(); +FunctionOverloadResolverPtr createInternalFunctionNotOverloadResolver(); + +} diff --git a/src/Functions/multiIf.cpp b/src/Functions/multiIf.cpp index cb946b55c73..af7afb75e1a 100644 --- a/src/Functions/multiIf.cpp +++ b/src/Functions/multiIf.cpp @@ -40,9 +40,17 @@ class FunctionMultiIf final : public FunctionIfBase { public: static constexpr auto name = "multiIf"; - static FunctionPtr create(ContextPtr context_) { return std::make_shared(context_); } + static FunctionPtr create(ContextPtr context_) + { + const auto & settings = context_->getSettingsRef(); + return std::make_shared(settings.allow_execute_multiif_columnar, settings.allow_experimental_variant_type, settings.use_variant_as_common_type); + } - explicit FunctionMultiIf(ContextPtr context_) : context(context_) { } + explicit FunctionMultiIf(bool allow_execute_multiif_columnar_, bool allow_experimental_variant_type_, bool use_variant_as_common_type_) + : allow_execute_multiif_columnar(allow_execute_multiif_columnar_) + , allow_experimental_variant_type(allow_experimental_variant_type_) + , use_variant_as_common_type(use_variant_as_common_type_) + {} String getName() const override { return name; } bool isVariadic() const override { return true; } @@ -118,7 +126,7 @@ public: types_of_branches.emplace_back(arg); }); - if (context->getSettingsRef().allow_experimental_variant_type && context->getSettingsRef().use_variant_as_common_type) + if (allow_experimental_variant_type && use_variant_as_common_type) return getLeastSupertypeOrVariant(types_of_branches); return getLeastSupertype(types_of_branches); @@ -240,10 +248,9 @@ public: } } - const auto & settings = context->getSettingsRef(); const WhichDataType which(removeNullable(result_type)); bool execute_multiif_columnar - = settings.allow_execute_multiif_columnar && !contains_short && (which.isInt() || which.isUInt() || which.isFloat()); + = allow_execute_multiif_columnar && !contains_short && (which.isInt() || which.isUInt() || which.isFloat()); size_t rows = input_rows_count; if (!execute_multiif_columnar) @@ -507,7 +514,9 @@ private: executeColumnIfNeeded(arguments[i], true); } - ContextPtr context; + const bool allow_execute_multiif_columnar; + const bool allow_experimental_variant_type; + const bool use_variant_as_common_type; }; } @@ -521,6 +530,11 @@ REGISTER_FUNCTION(MultiIf) factory.registerFunction("caseWithoutExpression"); } +FunctionOverloadResolverPtr createInternalMultiIfOverloadResolver(bool allow_execute_multiif_columnar, bool allow_experimental_variant_type, bool use_variant_as_common_type) +{ + return std::make_unique(std::make_shared(allow_execute_multiif_columnar, allow_experimental_variant_type, use_variant_as_common_type)); +} + } diff --git a/src/Functions/multiIf.h b/src/Functions/multiIf.h new file mode 100644 index 00000000000..617d63b89bc --- /dev/null +++ b/src/Functions/multiIf.h @@ -0,0 +1,12 @@ +#pragma once +#include + +namespace DB +{ + +class IFunctionOverloadResolver; +using FunctionOverloadResolverPtr = std::shared_ptr; + +FunctionOverloadResolverPtr createInternalMultiIfOverloadResolver(bool allow_execute_multiif_columnar, bool allow_experimental_variant_type, bool use_variant_as_common_type); + +} diff --git a/src/Functions/multiMatchAny.cpp b/src/Functions/multiMatchAny.cpp index 6e6abe61898..054a60fce2d 100644 --- a/src/Functions/multiMatchAny.cpp +++ b/src/Functions/multiMatchAny.cpp @@ -22,4 +22,9 @@ REGISTER_FUNCTION(MultiMatchAny) factory.registerFunction(); } +FunctionOverloadResolverPtr createInternalMultiMatchAnyOverloadResolver(bool allow_hyperscan, size_t max_hyperscan_regexp_length, size_t max_hyperscan_regexp_total_length, bool reject_expensive_hyperscan_regexps) +{ + return std::make_unique(std::make_shared(allow_hyperscan, max_hyperscan_regexp_length, max_hyperscan_regexp_total_length, reject_expensive_hyperscan_regexps)); +} + } diff --git a/src/Functions/multiMatchAny.h b/src/Functions/multiMatchAny.h new file mode 100644 index 00000000000..4548ec1d593 --- /dev/null +++ b/src/Functions/multiMatchAny.h @@ -0,0 +1,12 @@ +#pragma once +#include + +namespace DB +{ + +class IFunctionOverloadResolver; +using FunctionOverloadResolverPtr = std::shared_ptr; + +FunctionOverloadResolverPtr createInternalMultiMatchAnyOverloadResolver(bool allow_hyperscan, size_t max_hyperscan_regexp_length, size_t max_hyperscan_regexp_total_length, bool reject_expensive_hyperscan_regexps); + +} diff --git a/src/Functions/trap.cpp b/src/Functions/trap.cpp index 6ce696fedb5..99430f039a4 100644 --- a/src/Functions/trap.cpp +++ b/src/Functions/trap.cpp @@ -177,7 +177,7 @@ public: } else if (mode == "logical error") { - throw Exception(ErrorCodes::LOGICAL_ERROR, "Trap"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: trap"); } else throw Exception(ErrorCodes::BAD_ARGUMENTS, "Unknown trap mode"); diff --git a/src/Functions/vectorFunctions.cpp b/src/Functions/vectorFunctions.cpp index 33b0e9f6039..de4a6fb0a5c 100644 --- a/src/Functions/vectorFunctions.cpp +++ b/src/Functions/vectorFunctions.cpp @@ -1,9 +1,9 @@ #include #include #include +#include #include #include -#include #include #include #include @@ -1364,11 +1364,11 @@ public: ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t input_rows_count) const override { - if (getReturnTypeImpl(arguments)->isNullable()) - { - return DataTypeNullable(std::make_shared()) - .createColumnConstWithDefaultValue(input_rows_count); - } + /// TODO: cosineDistance does not support nullable arguments + /// https://github.com/ClickHouse/ClickHouse/pull/27933#issuecomment-916670286 + auto return_type = getReturnTypeImpl(arguments); + if (return_type->isNullable()) + return return_type->createColumnConstWithDefaultValue(input_rows_count); FunctionDotProduct dot(context); ColumnWithTypeAndName dot_result{dot.executeImpl(arguments, DataTypePtr(), input_rows_count), diff --git a/src/IO/AzureBlobStorage/copyAzureBlobStorageFile.h b/src/IO/AzureBlobStorage/copyAzureBlobStorageFile.h index 83814f42693..1433f8d18ba 100644 --- a/src/IO/AzureBlobStorage/copyAzureBlobStorageFile.h +++ b/src/IO/AzureBlobStorage/copyAzureBlobStorageFile.h @@ -6,7 +6,7 @@ #include #include -#include +#include #include #include #include diff --git a/src/IO/ParallelReadBuffer.h b/src/IO/ParallelReadBuffer.h index e76b40f77b7..daac1190399 100644 --- a/src/IO/ParallelReadBuffer.h +++ b/src/IO/ParallelReadBuffer.h @@ -3,7 +3,7 @@ #include #include #include -#include +#include #include namespace DB diff --git a/src/IO/S3/Client.cpp b/src/IO/S3/Client.cpp index 7f0ede72740..182e7ad18cd 100644 --- a/src/IO/S3/Client.cpp +++ b/src/IO/S3/Client.cpp @@ -27,7 +27,6 @@ #include -#include namespace ProfileEvents { @@ -48,7 +47,6 @@ namespace ErrorCodes { extern const int LOGICAL_ERROR; extern const int TOO_MANY_REDIRECTS; - extern const int BAD_ARGUMENTS; } namespace S3 @@ -106,19 +104,6 @@ void verifyClientConfiguration(const Aws::Client::ClientConfiguration & client_c assert_cast(*client_config.retryStrategy); } -void validateCredentials(const Aws::Auth::AWSCredentials& auth_credentials) -{ - if (auth_credentials.GetAWSAccessKeyId().empty()) - { - return; - } - /// Follow https://docs.aws.amazon.com/IAM/latest/APIReference/API_AccessKey.html - if (!std::all_of(auth_credentials.GetAWSAccessKeyId().begin(), auth_credentials.GetAWSAccessKeyId().end(), isWordCharASCII)) - { - throw Exception(ErrorCodes::BAD_ARGUMENTS, "Access key id has an invalid character"); - } -} - void addAdditionalAMZHeadersToCanonicalHeadersList( Aws::AmazonWebServiceRequest & request, const HTTPHeaderEntries & extra_headers @@ -144,7 +129,6 @@ std::unique_ptr Client::create( const ClientSettings & client_settings) { verifyClientConfiguration(client_configuration); - validateCredentials(credentials_provider->GetAWSCredentials()); return std::unique_ptr( new Client(max_redirects_, std::move(sse_kms_config_), credentials_provider, client_configuration, sign_payloads, client_settings)); } diff --git a/src/IO/S3/copyS3File.cpp b/src/IO/S3/copyS3File.cpp index 98024e74f8e..51518df268c 100644 --- a/src/IO/S3/copyS3File.cpp +++ b/src/IO/S3/copyS3File.cpp @@ -746,7 +746,12 @@ namespace break; } - if (outcome.GetError().GetExceptionName() == "EntityTooLarge" || outcome.GetError().GetExceptionName() == "InvalidRequest" || outcome.GetError().GetExceptionName() == "InvalidArgument") + if (outcome.GetError().GetExceptionName() == "EntityTooLarge" || + outcome.GetError().GetExceptionName() == "InvalidRequest" || + outcome.GetError().GetExceptionName() == "InvalidArgument" || + (outcome.GetError().GetExceptionName() == "InternalError" && + outcome.GetError().GetResponseCode() == Aws::Http::HttpResponseCode::GATEWAY_TIMEOUT && + outcome.GetError().GetMessage().contains("use the Rewrite method in the JSON API"))) { if (!supports_multipart_copy) { diff --git a/src/IO/S3/copyS3File.h b/src/IO/S3/copyS3File.h index 607be51ed25..093d26ba7bb 100644 --- a/src/IO/S3/copyS3File.h +++ b/src/IO/S3/copyS3File.h @@ -5,7 +5,7 @@ #if USE_AWS_S3 #include -#include +#include #include #include #include diff --git a/src/IO/S3Common.cpp b/src/IO/S3Common.cpp index 5039059f522..56e3e0df21b 100644 --- a/src/IO/S3Common.cpp +++ b/src/IO/S3Common.cpp @@ -1,7 +1,9 @@ #include #include +#include #include + #include "config.h" #if USE_AWS_S3 @@ -124,6 +126,15 @@ AuthSettings AuthSettings::loadFromConfig(const std::string & config_elem, const HTTPHeaderEntries headers = getHTTPHeaders(config_elem, config); ServerSideEncryptionKMSConfig sse_kms_config = getSSEKMSConfig(config_elem, config); + std::unordered_set users; + Poco::Util::AbstractConfiguration::Keys keys; + config.keys(config_elem, keys); + for (const auto & key : keys) + { + if (startsWith(key, "user")) + users.insert(config.getString(config_elem + "." + key)); + } + return AuthSettings { std::move(access_key_id), std::move(secret_access_key), std::move(session_token), @@ -134,10 +145,16 @@ AuthSettings AuthSettings::loadFromConfig(const std::string & config_elem, const use_environment_credentials, use_insecure_imds_request, expiration_window_seconds, - no_sign_request + no_sign_request, + std::move(users) }; } +bool AuthSettings::canBeUsedByUser(const String & user) const +{ + return users.empty() || users.contains(user); +} + bool AuthSettings::hasUpdates(const AuthSettings & other) const { AuthSettings copy = *this; @@ -173,6 +190,8 @@ void AuthSettings::updateFrom(const AuthSettings & from) if (from.no_sign_request.has_value()) no_sign_request = from.no_sign_request; + + users.insert(from.users.begin(), from.users.end()); } } diff --git a/src/IO/S3Common.h b/src/IO/S3Common.h index 6ee8d96ed09..b3e01bd6132 100644 --- a/src/IO/S3Common.h +++ b/src/IO/S3Common.h @@ -6,6 +6,7 @@ #include #include +#include #include "config.h" @@ -92,9 +93,13 @@ struct AuthSettings std::optional expiration_window_seconds; std::optional no_sign_request; + std::unordered_set users; + bool hasUpdates(const AuthSettings & other) const; void updateFrom(const AuthSettings & from); + bool canBeUsedByUser(const String & user) const; + private: bool operator==(const AuthSettings & other) const = default; }; diff --git a/src/IO/WriteBufferFromS3.h b/src/IO/WriteBufferFromS3.h index 230f39b074e..5dc269990a1 100644 --- a/src/IO/WriteBufferFromS3.h +++ b/src/IO/WriteBufferFromS3.h @@ -10,7 +10,7 @@ #include #include #include -#include +#include #include #include diff --git a/src/IO/readFloatText.cpp b/src/IO/readFloatText.cpp index d1143f7c62c..17ccc1b25b7 100644 --- a/src/IO/readFloatText.cpp +++ b/src/IO/readFloatText.cpp @@ -67,4 +67,7 @@ template void readFloatText(Float64 &, ReadBuffer &); template bool tryReadFloatText(Float32 &, ReadBuffer &); template bool tryReadFloatText(Float64 &, ReadBuffer &); +template bool tryReadFloatTextNoExponent(Float32 &, ReadBuffer &); +template bool tryReadFloatTextNoExponent(Float64 &, ReadBuffer &); + } diff --git a/src/IO/readFloatText.h b/src/IO/readFloatText.h index 23e904f305a..51964636389 100644 --- a/src/IO/readFloatText.h +++ b/src/IO/readFloatText.h @@ -324,7 +324,7 @@ static inline void readUIntTextUpToNSignificantDigits(T & x, ReadBuffer & buf) } -template +template ReturnType readFloatTextFastImpl(T & x, ReadBuffer & in) { static_assert(std::is_same_v || std::is_same_v, "Argument for readFloatTextImpl must be float or double"); @@ -395,30 +395,33 @@ ReturnType readFloatTextFastImpl(T & x, ReadBuffer & in) after_point_exponent = (read_digits > significant_digits ? -significant_digits : static_cast(-read_digits)) - after_point_num_leading_zeros; } - if (checkChar('e', in) || checkChar('E', in)) + if constexpr (allow_exponent) { - if (in.eof()) + if (checkChar('e', in) || checkChar('E', in)) { - if constexpr (throw_exception) - throw Exception(ErrorCodes::CANNOT_PARSE_NUMBER, "Cannot read floating point value: nothing after exponent"); - else - return false; - } + if (in.eof()) + { + if constexpr (throw_exception) + throw Exception(ErrorCodes::CANNOT_PARSE_NUMBER, "Cannot read floating point value: nothing after exponent"); + else + return false; + } - bool exponent_negative = false; - if (*in.position() == '-') - { - exponent_negative = true; - ++in.position(); - } - else if (*in.position() == '+') - { - ++in.position(); - } + bool exponent_negative = false; + if (*in.position() == '-') + { + exponent_negative = true; + ++in.position(); + } + else if (*in.position() == '+') + { + ++in.position(); + } - readUIntTextUpToNSignificantDigits<4>(exponent, in); - if (exponent_negative) - exponent = -exponent; + readUIntTextUpToNSignificantDigits<4>(exponent, in); + if (exponent_negative) + exponent = -exponent; + } } if (after_point) @@ -604,4 +607,7 @@ template bool tryReadFloatTextSimple(T & x, ReadBuffer & in) { retu template void readFloatText(T & x, ReadBuffer & in) { readFloatTextFast(x, in); } template bool tryReadFloatText(T & x, ReadBuffer & in) { return tryReadFloatTextFast(x, in); } +/// Don't read exponent part of the number. +template bool tryReadFloatTextNoExponent(T & x, ReadBuffer & in) { return readFloatTextFastImpl(x, in); } + } diff --git a/src/Interpreters/Aggregator.cpp b/src/Interpreters/Aggregator.cpp index 50fab486568..331cd991ea1 100644 --- a/src/Interpreters/Aggregator.cpp +++ b/src/Interpreters/Aggregator.cpp @@ -624,7 +624,7 @@ Aggregator::Aggregator(const Block & header_, const Params & params_) { size_t alignment_of_next_state = params.aggregates[i + 1].function->alignOfData(); if ((alignment_of_next_state & (alignment_of_next_state - 1)) != 0) - throw Exception(ErrorCodes::LOGICAL_ERROR, "`alignOfData` is not 2^N"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: alignOfData is not 2^N"); /// Extend total_size to next alignment requirement /// Add padding by rounding up 'total_size_of_aggregate_states' to be a multiplier of alignment_of_next_state. @@ -857,7 +857,7 @@ AggregatedDataVariants::Type Aggregator::chooseAggregationMethod() return AggregatedDataVariants::Type::low_cardinality_keys128; if (size_of_field == 32) return AggregatedDataVariants::Type::low_cardinality_keys256; - throw Exception(ErrorCodes::LOGICAL_ERROR, "LowCardinality numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: low cardinality numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); } if (size_of_field == 1) @@ -872,7 +872,7 @@ AggregatedDataVariants::Type Aggregator::chooseAggregationMethod() return AggregatedDataVariants::Type::keys128; if (size_of_field == 32) return AggregatedDataVariants::Type::keys256; - throw Exception(ErrorCodes::LOGICAL_ERROR, "Numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); } if (params.keys_size == 1 && isFixedString(types_removed_nullable[0])) diff --git a/src/Interpreters/ArrayJoinedColumnsVisitor.h b/src/Interpreters/ArrayJoinedColumnsVisitor.h index f16751c4561..3bbd6982213 100644 --- a/src/Interpreters/ArrayJoinedColumnsVisitor.h +++ b/src/Interpreters/ArrayJoinedColumnsVisitor.h @@ -62,7 +62,7 @@ private: { auto [array_join_expression_list, _] = node.arrayJoinExpressionList(); if (!array_join_expression_list) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No ARRAY JOIN"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no ARRAY JOIN"); std::vector out; out.reserve(array_join_expression_list->children.size()); diff --git a/src/Interpreters/BackupLog.cpp b/src/Interpreters/BackupLog.cpp index d34e982ffc3..d5b69bc0728 100644 --- a/src/Interpreters/BackupLog.cpp +++ b/src/Interpreters/BackupLog.cpp @@ -28,6 +28,7 @@ ColumnsDescription BackupLogElement::getColumnsDescription() {"id", std::make_shared()}, {"name", std::make_shared()}, {"base_backup_name", std::make_shared()}, + {"query_id", std::make_shared()}, {"status", std::make_shared(getBackupStatusEnumValues())}, {"error", std::make_shared()}, {"start_time", std::make_shared()}, @@ -51,6 +52,7 @@ void BackupLogElement::appendToBlock(MutableColumns & columns) const columns[i++]->insert(info.id); columns[i++]->insert(info.name); columns[i++]->insert(info.base_backup_name); + columns[i++]->insert(info.query_id); columns[i++]->insert(static_cast(info.status)); columns[i++]->insert(info.error_message); columns[i++]->insert(static_cast(std::chrono::system_clock::to_time_t(info.start_time))); diff --git a/src/Interpreters/ClientInfo.cpp b/src/Interpreters/ClientInfo.cpp index e4778edeb9c..347ec115aba 100644 --- a/src/Interpreters/ClientInfo.cpp +++ b/src/Interpreters/ClientInfo.cpp @@ -23,7 +23,7 @@ namespace ErrorCodes void ClientInfo::write(WriteBuffer & out, UInt64 server_protocol_revision) const { if (server_protocol_revision < DBMS_MIN_REVISION_WITH_CLIENT_INFO) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Method ClientInfo::write is called for unsupported server revision"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: method ClientInfo::write is called for unsupported server revision"); writeBinary(static_cast(query_kind), out); if (empty()) @@ -103,7 +103,7 @@ void ClientInfo::write(WriteBuffer & out, UInt64 server_protocol_revision) const void ClientInfo::read(ReadBuffer & in, UInt64 client_protocol_revision) { if (client_protocol_revision < DBMS_MIN_REVISION_WITH_CLIENT_INFO) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Method ClientInfo::read is called for unsupported client revision"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: method ClientInfo::read is called for unsupported client revision"); UInt8 read_query_kind = 0; readBinary(read_query_kind, in); diff --git a/src/Interpreters/CrossToInnerJoinVisitor.cpp b/src/Interpreters/CrossToInnerJoinVisitor.cpp index e3e8b80e437..42af164f4ad 100644 --- a/src/Interpreters/CrossToInnerJoinVisitor.cpp +++ b/src/Interpreters/CrossToInnerJoinVisitor.cpp @@ -173,7 +173,7 @@ std::vector getTables(const ASTSelectQuery & select) { const auto * table_element = child->as(); if (!table_element) - throw Exception(ErrorCodes::LOGICAL_ERROR, "TablesInSelectQueryElement expected"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: TablesInSelectQueryElement expected"); JoinedElement & t = joined_tables.emplace_back(*table_element); t.rewriteCommaToCross(); @@ -224,7 +224,7 @@ void CrossToInnerJoinMatcher::visit(ASTSelectQuery & select, ASTPtr &, Data & da { if (joined_tables.size() != data.tables_with_columns.size()) throw Exception(ErrorCodes::LOGICAL_ERROR, - "Inconsistent number of tables: {} != {}", + "Logical error: inconsistent number of tables: {} != {}", joined_tables.size(), data.tables_with_columns.size()); for (size_t i = 0; i < joined_tables.size(); ++i) diff --git a/src/Interpreters/DatabaseAndTableWithAlias.cpp b/src/Interpreters/DatabaseAndTableWithAlias.cpp index 329391b45d7..db020cb9166 100644 --- a/src/Interpreters/DatabaseAndTableWithAlias.cpp +++ b/src/Interpreters/DatabaseAndTableWithAlias.cpp @@ -71,7 +71,7 @@ DatabaseAndTableWithAlias::DatabaseAndTableWithAlias(const ASTTableExpression & alias = table_expression.subquery->tryGetAlias(); } else - throw Exception(ErrorCodes::LOGICAL_ERROR, "No known elements in ASTTableExpression"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no known elements in ASTTableExpression"); } bool DatabaseAndTableWithAlias::satisfies(const DatabaseAndTableWithAlias & db_table, bool table_may_be_an_alias) const diff --git a/src/Interpreters/HashJoin.cpp b/src/Interpreters/HashJoin.cpp index 73487a0914a..33dc178ca00 100644 --- a/src/Interpreters/HashJoin.cpp +++ b/src/Interpreters/HashJoin.cpp @@ -368,7 +368,7 @@ HashJoin::Type HashJoin::chooseMethod(JoinKind kind, const ColumnRawPtrs & key_c return Type::keys128; if (size_of_field == 32) return Type::keys256; - throw Exception(ErrorCodes::LOGICAL_ERROR, "Numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); } /// If the keys fit in N bits, we will use a hash table for N-bit-packed keys diff --git a/src/Interpreters/InJoinSubqueriesPreprocessor.cpp b/src/Interpreters/InJoinSubqueriesPreprocessor.cpp index ec4241a2740..3858830a43b 100644 --- a/src/Interpreters/InJoinSubqueriesPreprocessor.cpp +++ b/src/Interpreters/InJoinSubqueriesPreprocessor.cpp @@ -103,12 +103,12 @@ private: /// Already processed. } else - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unexpected function name {}", concrete->name); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: unexpected function name {}", concrete->name); } else if (table_join) table_join->locality = JoinLocality::Global; else - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unexpected AST node"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: unexpected AST node"); } else if (distributed_product_mode == DistributedProductMode::DENY) { diff --git a/src/Interpreters/InterpreterCreateQuery.cpp b/src/Interpreters/InterpreterCreateQuery.cpp index a9942ac7cde..08dda0fe811 100644 --- a/src/Interpreters/InterpreterCreateQuery.cpp +++ b/src/Interpreters/InterpreterCreateQuery.cpp @@ -50,6 +50,7 @@ #include #include #include +#include #include @@ -99,7 +100,6 @@ namespace ErrorCodes extern const int DATABASE_ALREADY_EXISTS; extern const int BAD_ARGUMENTS; extern const int BAD_DATABASE_FOR_TEMPORARY_TABLE; - extern const int SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY; extern const int ILLEGAL_SYNTAX_FOR_DATA_TYPE; extern const int ILLEGAL_COLUMN; extern const int ILLEGAL_INDEX; @@ -911,66 +911,13 @@ void InterpreterCreateQuery::validateTableStructure(const ASTCreateQuery & creat const auto & settings = getContext()->getSettingsRef(); - /// Check low cardinality types in creating table if it was not allowed in setting - if (!create.attach && !settings.allow_suspicious_low_cardinality_types && !create.is_materialized_view) + /// If it's not attach and not materialized view to existing table, + /// we need to validate data types (check for experimental or suspicious types). + if (!create.attach && !create.is_materialized_view) { + DataTypeValidationSettings validation_settings(settings); for (const auto & name_and_type_pair : properties.columns.getAllPhysical()) - { - if (const auto * current_type_ptr = typeid_cast(name_and_type_pair.type.get())) - { - if (!isStringOrFixedString(*removeNullable(current_type_ptr->getDictionaryType()))) - throw Exception(ErrorCodes::SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY, - "Creating columns of type {} is prohibited by default " - "due to expected negative impact on performance. " - "It can be enabled with the \"allow_suspicious_low_cardinality_types\" setting.", - current_type_ptr->getName()); - } - } - } - - if (!create.attach && !settings.allow_experimental_object_type) - { - for (const auto & [name, type] : properties.columns.getAllPhysical()) - { - if (type->hasDynamicSubcolumns()) - { - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Cannot create table with column '{}' which type is '{}' " - "because experimental Object type is not allowed. " - "Set setting allow_experimental_object_type = 1 in order to allow it", - name, type->getName()); - } - } - } - if (!create.attach && !settings.allow_suspicious_fixed_string_types) - { - for (const auto & [name, type] : properties.columns.getAllPhysical()) - { - auto basic_type = removeLowCardinalityAndNullable(type); - if (const auto * fixed_string = typeid_cast(basic_type.get())) - { - if (fixed_string->getN() > MAX_FIXEDSTRING_SIZE_WITHOUT_SUSPICIOUS) - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Cannot create table with column '{}' which type is '{}' " - "because fixed string with size > {} is suspicious. " - "Set setting allow_suspicious_fixed_string_types = 1 in order to allow it", - name, type->getName(), MAX_FIXEDSTRING_SIZE_WITHOUT_SUSPICIOUS); - } - } - } - if (!create.attach && !settings.allow_experimental_variant_type) - { - for (const auto & [name, type] : properties.columns.getAllPhysical()) - { - if (isVariant(type)) - { - throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Cannot create table with column '{}' which type is '{}' " - "because experimental Variant type is not allowed. " - "Set setting allow_experimental_variant_type = 1 in order to allow it", - name, type->getName()); - } - } + validateDataType(name_and_type_pair.type, validation_settings); } } diff --git a/src/Interpreters/InterpreterSelectWithUnionQuery.cpp b/src/Interpreters/InterpreterSelectWithUnionQuery.cpp index cc1d7dd6531..16bc4b1fe2e 100644 --- a/src/Interpreters/InterpreterSelectWithUnionQuery.cpp +++ b/src/Interpreters/InterpreterSelectWithUnionQuery.cpp @@ -56,7 +56,7 @@ InterpreterSelectWithUnionQuery::InterpreterSelectWithUnionQuery( size_t num_children = ast->list_of_selects->children.size(); if (!num_children) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No children in ASTSelectWithUnionQuery"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no children in ASTSelectWithUnionQuery"); /// Note that we pass 'required_result_column_names' to first SELECT. /// And for the rest, we pass names at the corresponding positions of 'required_result_column_names' in the result of first SELECT, diff --git a/src/Interpreters/JoinToSubqueryTransformVisitor.cpp b/src/Interpreters/JoinToSubqueryTransformVisitor.cpp index 6251a9604e1..bf2d1eb79cd 100644 --- a/src/Interpreters/JoinToSubqueryTransformVisitor.cpp +++ b/src/Interpreters/JoinToSubqueryTransformVisitor.cpp @@ -168,7 +168,7 @@ private: has_asterisks = true; if (!qualified_asterisk->qualifier) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Qualified asterisk must have a qualifier"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: qualified asterisk must have a qualifier"); auto & identifier = qualified_asterisk->qualifier->as(); @@ -183,7 +183,7 @@ private: transformer->as()) IASTColumnsTransformer::transform(transformer, columns); else - throw Exception(ErrorCodes::LOGICAL_ERROR, "Qualified asterisk must only have children of IASTColumnsTransformer type"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: qualified asterisk must only have children of IASTColumnsTransformer type"); } } } diff --git a/src/Interpreters/MergeJoin.cpp b/src/Interpreters/MergeJoin.cpp index d5fb0208d45..901c82029ee 100644 --- a/src/Interpreters/MergeJoin.cpp +++ b/src/Interpreters/MergeJoin.cpp @@ -239,7 +239,7 @@ public: /// SortCursorImpl can work with permutation, but MergeJoinCursor can't. if (impl.permutation) - throw Exception(ErrorCodes::LOGICAL_ERROR, "MergeJoinCursor doesn't support permutation"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: MergeJoinCursor doesn't support permutation"); } size_t position() const { return impl.getRow(); } diff --git a/src/Interpreters/MySQL/InterpretersMySQLDDLQuery.cpp b/src/Interpreters/MySQL/InterpretersMySQLDDLQuery.cpp index 107b435ded4..0fdc9347ee9 100644 --- a/src/Interpreters/MySQL/InterpretersMySQLDDLQuery.cpp +++ b/src/Interpreters/MySQL/InterpretersMySQLDDLQuery.cpp @@ -337,7 +337,7 @@ static ASTPtr getPartitionPolicy(const NamesAndTypesList & primary_keys) WhichDataType which(type); if (which.isNullable()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "MySQL's primary key must be not null, it is a bug."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: MySQL primary key must be not null, it is a bug."); if (which.isDate() || which.isDate32() || which.isDateTime() || which.isDateTime64()) { diff --git a/src/Interpreters/ProcessList.cpp b/src/Interpreters/ProcessList.cpp index 3bd7b2d4206..5b3b87114ae 100644 --- a/src/Interpreters/ProcessList.cpp +++ b/src/Interpreters/ProcessList.cpp @@ -295,7 +295,7 @@ ProcessListEntry::~ProcessListEntry() auto user_process_list_it = parent.user_to_queries.find(user); if (user_process_list_it == parent.user_to_queries.end()) { - LOG_ERROR(getLogger("ProcessList"), "Cannot find user in ProcessList"); + LOG_ERROR(getLogger("ProcessList"), "Logical error: cannot find user in ProcessList"); std::terminate(); } @@ -323,7 +323,7 @@ ProcessListEntry::~ProcessListEntry() if (!found) { - LOG_ERROR(getLogger("ProcessList"), "Cannot find query by query_id and pointer to ProcessListElement in ProcessListForUser"); + LOG_ERROR(getLogger("ProcessList"), "Logical error: cannot find query by query_id and pointer to ProcessListElement in ProcessListForUser"); std::terminate(); } diff --git a/src/Interpreters/S3QueueLog.cpp b/src/Interpreters/S3QueueLog.cpp index 967becb6e0f..3ed58de0f87 100644 --- a/src/Interpreters/S3QueueLog.cpp +++ b/src/Interpreters/S3QueueLog.cpp @@ -28,7 +28,9 @@ ColumnsDescription S3QueueLogElement::getColumnsDescription() {"hostname", std::make_shared(std::make_shared())}, {"event_date", std::make_shared()}, {"event_time", std::make_shared()}, - {"table_uuid", std::make_shared()}, + {"database", std::make_shared()}, + {"table", std::make_shared()}, + {"uuid", std::make_shared()}, {"file_name", std::make_shared()}, {"rows_processed", std::make_shared()}, {"status", status_datatype}, @@ -45,7 +47,9 @@ void S3QueueLogElement::appendToBlock(MutableColumns & columns) const columns[i++]->insert(getFQDNOrHostName()); columns[i++]->insert(DateLUT::instance().toDayNum(event_time).toUnderType()); columns[i++]->insert(event_time); - columns[i++]->insert(table_uuid); + columns[i++]->insert(database); + columns[i++]->insert(table); + columns[i++]->insert(uuid); columns[i++]->insert(file_name); columns[i++]->insert(rows_processed); columns[i++]->insert(status); diff --git a/src/Interpreters/S3QueueLog.h b/src/Interpreters/S3QueueLog.h index e0362bf9716..b6bc138d42c 100644 --- a/src/Interpreters/S3QueueLog.h +++ b/src/Interpreters/S3QueueLog.h @@ -12,7 +12,11 @@ namespace DB struct S3QueueLogElement { time_t event_time{}; - std::string table_uuid; + + std::string database; + std::string table; + std::string uuid; + std::string file_name; size_t rows_processed = 0; diff --git a/src/Interpreters/Set.cpp b/src/Interpreters/Set.cpp index 8f11754b3be..84260faafd4 100644 --- a/src/Interpreters/Set.cpp +++ b/src/Interpreters/Set.cpp @@ -275,7 +275,7 @@ void Set::appendSetElements(SetKeyColumns & holder) void Set::checkIsCreated() const { if (!is_created.load()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Trying to use set before it has been built."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: Trying to use set before it has been built."); } ColumnPtr Set::execute(const ColumnsWithTypeAndName & columns, bool negative) const @@ -283,7 +283,7 @@ ColumnPtr Set::execute(const ColumnsWithTypeAndName & columns, bool negative) co size_t num_key_columns = columns.size(); if (0 == num_key_columns) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No columns passed to Set::execute method."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no columns passed to Set::execute method."); auto res = ColumnUInt8::create(); ColumnUInt8::Container & vec_res = res->getData(); diff --git a/src/Interpreters/SetVariants.cpp b/src/Interpreters/SetVariants.cpp index 0fb2e5189d4..cd9148a01cf 100644 --- a/src/Interpreters/SetVariants.cpp +++ b/src/Interpreters/SetVariants.cpp @@ -146,7 +146,7 @@ typename SetVariantsTemplate::Type SetVariantsTemplate::choose return Type::keys128; if (size_of_field == 32) return Type::keys256; - throw Exception(ErrorCodes::LOGICAL_ERROR, "Numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: numeric column has sizeOfField not in 1, 2, 4, 8, 16, 32."); } /// If the keys fit in N bits, we will use a hash table for N-bit-packed keys diff --git a/src/Interpreters/TablesStatus.cpp b/src/Interpreters/TablesStatus.cpp index 911a028f813..005a4515c3a 100644 --- a/src/Interpreters/TablesStatus.cpp +++ b/src/Interpreters/TablesStatus.cpp @@ -35,7 +35,7 @@ void TableStatus::read(ReadBuffer & in) void TablesStatusRequest::write(WriteBuffer & out, UInt64 server_protocol_revision) const { if (server_protocol_revision < DBMS_MIN_REVISION_WITH_TABLES_STATUS) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Method TablesStatusRequest::write is called for unsupported server revision"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: method TablesStatusRequest::write is called for unsupported server revision"); writeVarUInt(tables.size(), out); for (const auto & table_name : tables) diff --git a/src/Interpreters/TranslateQualifiedNamesVisitor.cpp b/src/Interpreters/TranslateQualifiedNamesVisitor.cpp index 3de7e217e53..130ce2194fd 100644 --- a/src/Interpreters/TranslateQualifiedNamesVisitor.cpp +++ b/src/Interpreters/TranslateQualifiedNamesVisitor.cpp @@ -158,7 +158,7 @@ void TranslateQualifiedNamesMatcher::visit(ASTFunction & node, const ASTPtr &, D void TranslateQualifiedNamesMatcher::visit(const ASTQualifiedAsterisk & node, const ASTPtr &, Data & data) { if (!node.qualifier) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Qualified asterisk must have a qualifier"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: qualified asterisk must have a qualifier"); /// @note it could contain table alias as table name. DatabaseAndTableWithAlias db_and_table(node.qualifier); diff --git a/src/Interpreters/TreeRewriter.cpp b/src/Interpreters/TreeRewriter.cpp index bb6df2da8d9..d01bdd16c53 100644 --- a/src/Interpreters/TreeRewriter.cpp +++ b/src/Interpreters/TreeRewriter.cpp @@ -73,7 +73,6 @@ namespace ErrorCodes extern const int NOT_IMPLEMENTED; extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH; extern const int UNKNOWN_IDENTIFIER; - extern const int UNEXPECTED_EXPRESSION; } namespace @@ -786,16 +785,6 @@ void expandOrderByAll(ASTSelectQuery * select_query) for (const auto & expr : select_query->select()->children) { - if (auto * identifier = expr->as(); identifier != nullptr) - if (Poco::toUpper(identifier->name()) == "ALL" || Poco::toUpper(identifier->alias) == "ALL") - throw Exception(ErrorCodes::UNEXPECTED_EXPRESSION, - "Cannot use ORDER BY ALL to sort a column with name 'all', please disable setting `enable_order_by_all` and try again"); - - if (auto * function = expr->as(); function != nullptr) - if (Poco::toUpper(function->alias) == "ALL") - throw Exception(ErrorCodes::UNEXPECTED_EXPRESSION, - "Cannot use ORDER BY ALL to sort a column with name 'all', please disable setting `enable_order_by_all` and try again"); - auto elem = std::make_shared(); elem->direction = all_elem->direction; elem->nulls_direction = all_elem->nulls_direction; @@ -1322,8 +1311,8 @@ TreeRewriterResultPtr TreeRewriter::analyzeSelect( if (select_query->group_by_all) expandGroupByAll(select_query); - // expand ORDER BY ALL - if (settings.enable_order_by_all && select_query->order_by_all) + // expand ORDER BY * + if (select_query->order_by_all) expandOrderByAll(select_query); /// Remove unneeded columns according to 'required_result_columns'. diff --git a/src/Interpreters/convertFieldToType.cpp b/src/Interpreters/convertFieldToType.cpp index c3b8405659a..346180c3613 100644 --- a/src/Interpreters/convertFieldToType.cpp +++ b/src/Interpreters/convertFieldToType.cpp @@ -493,10 +493,12 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID { /// Promote data type to avoid overflows. Note that overflows in the largest data type are still possible. /// But don't promote Float32, since we want to keep the exact same value + /// Also don't promote domain types (like bool) because we would otherwise use the serializer of the promoted type (e.g. UInt64 for + /// bool, which does not allow 'true' and 'false' as input values) const IDataType * type_to_parse = &type; DataTypePtr holder; - if (type.canBePromoted() && !which_type.isFloat32()) + if (type.canBePromoted() && !which_type.isFloat32() && !type.getCustomSerialization()) { holder = type.promoteNumericType(); type_to_parse = holder.get(); diff --git a/src/Interpreters/evaluateConstantExpression.cpp b/src/Interpreters/evaluateConstantExpression.cpp index b5c3e00e299..00d36750cc1 100644 --- a/src/Interpreters/evaluateConstantExpression.cpp +++ b/src/Interpreters/evaluateConstantExpression.cpp @@ -106,7 +106,7 @@ std::optional evaluateConstantExpressionImpl(c if (result_column->empty()) throw Exception(ErrorCodes::LOGICAL_ERROR, - "Empty result column after evaluation " + "Logical error: empty result column after evaluation " "of constant expression for IN, VALUES, or LIMIT, or aggregate function parameter, or a table function argument"); /// Expressions like rand() or now() are not constant diff --git a/src/Interpreters/executeQuery.cpp b/src/Interpreters/executeQuery.cpp index 808a2ee2d81..f2aa51bd6de 100644 --- a/src/Interpreters/executeQuery.cpp +++ b/src/Interpreters/executeQuery.cpp @@ -722,7 +722,7 @@ static std::tuple executeQueryImpl( /// TODO: parser should fail early when max_query_size limit is reached. ast = parseQuery(parser, begin, end, "", max_query_size, settings.max_parser_depth); -#ifndef NDEBUG +#if 0 /// Verify that AST formatting is consistent: /// If you format AST, parse it back, and format it again, you get the same string. diff --git a/src/Interpreters/getHeaderForProcessingStage.cpp b/src/Interpreters/getHeaderForProcessingStage.cpp index 67a909ba6b4..d16e01ef2d2 100644 --- a/src/Interpreters/getHeaderForProcessingStage.cpp +++ b/src/Interpreters/getHeaderForProcessingStage.cpp @@ -167,7 +167,8 @@ Block getHeaderForProcessingStage( return result; } } - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown processed stage."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical Error: unknown processed stage."); } } + diff --git a/src/Interpreters/parseColumnsListForTableFunction.cpp b/src/Interpreters/parseColumnsListForTableFunction.cpp index 551a883d093..1499568cec9 100644 --- a/src/Interpreters/parseColumnsListForTableFunction.cpp +++ b/src/Interpreters/parseColumnsListForTableFunction.cpp @@ -8,7 +8,6 @@ #include #include - namespace DB { @@ -20,57 +19,64 @@ namespace ErrorCodes } -void validateDataType(const DataTypePtr & type, const DataTypeValidationSettings & settings) +void validateDataType(const DataTypePtr & type_to_check, const DataTypeValidationSettings & settings) { - if (!settings.allow_suspicious_low_cardinality_types) + auto validate_callback = [&](const IDataType & data_type) { - if (const auto * lc_type = typeid_cast(type.get())) + if (!settings.allow_suspicious_low_cardinality_types) { - if (!isStringOrFixedString(*removeNullable(lc_type->getDictionaryType()))) - throw Exception( - ErrorCodes::SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY, - "Creating columns of type {} is prohibited by default due to expected negative impact on performance. " - "It can be enabled with the \"allow_suspicious_low_cardinality_types\" setting.", - lc_type->getName()); + if (const auto * lc_type = typeid_cast(&data_type)) + { + if (!isStringOrFixedString(*removeNullable(lc_type->getDictionaryType()))) + throw Exception( + ErrorCodes::SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY, + "Creating columns of type {} is prohibited by default due to expected negative impact on performance. " + "It can be enabled with the \"allow_suspicious_low_cardinality_types\" setting.", + lc_type->getName()); + } } - } - if (!settings.allow_experimental_object_type) - { - if (type->hasDynamicSubcolumns()) + if (!settings.allow_experimental_object_type) { - throw Exception( - ErrorCodes::ILLEGAL_COLUMN, - "Cannot create column with type '{}' because experimental Object type is not allowed. " - "Set setting allow_experimental_object_type = 1 in order to allow it", type->getName()); - } - } - - if (!settings.allow_suspicious_fixed_string_types) - { - auto basic_type = removeLowCardinalityAndNullable(type); - if (const auto * fixed_string = typeid_cast(basic_type.get())) - { - if (fixed_string->getN() > MAX_FIXEDSTRING_SIZE_WITHOUT_SUSPICIOUS) + if (data_type.hasDynamicSubcolumns()) + { throw Exception( ErrorCodes::ILLEGAL_COLUMN, - "Cannot create column with type '{}' because fixed string with size > {} is suspicious. " - "Set setting allow_suspicious_fixed_string_types = 1 in order to allow it", - type->getName(), - MAX_FIXEDSTRING_SIZE_WITHOUT_SUSPICIOUS); + "Cannot create column with type '{}' because experimental Object type is not allowed. " + "Set setting allow_experimental_object_type = 1 in order to allow it", + data_type.getName()); + } } - } - if (!settings.allow_experimental_variant_type) - { - if (isVariant(type)) + if (!settings.allow_suspicious_fixed_string_types) { - throw Exception( - ErrorCodes::ILLEGAL_COLUMN, - "Cannot create column with type '{}' because experimental Variant type is not allowed. " - "Set setting allow_experimental_variant_type = 1 in order to allow it", type->getName()); + if (const auto * fixed_string = typeid_cast(&data_type)) + { + if (fixed_string->getN() > MAX_FIXEDSTRING_SIZE_WITHOUT_SUSPICIOUS) + throw Exception( + ErrorCodes::ILLEGAL_COLUMN, + "Cannot create column with type '{}' because fixed string with size > {} is suspicious. " + "Set setting allow_suspicious_fixed_string_types = 1 in order to allow it", + data_type.getName(), + MAX_FIXEDSTRING_SIZE_WITHOUT_SUSPICIOUS); + } } - } + + if (!settings.allow_experimental_variant_type) + { + if (isVariant(data_type)) + { + throw Exception( + ErrorCodes::ILLEGAL_COLUMN, + "Cannot create column with type '{}' because experimental Variant type is not allowed. " + "Set setting allow_experimental_variant_type = 1 in order to allow it", + data_type.getName()); + } + } + }; + + validate_callback(*type_to_check); + type_to_check->forEachChild(validate_callback); } ColumnsDescription parseColumnsListFromString(const std::string & structure, const ContextPtr & context) diff --git a/src/Parsers/ASTFunction.cpp b/src/Parsers/ASTFunction.cpp index ae9b8ddbe85..ba4c7db96e6 100644 --- a/src/Parsers/ASTFunction.cpp +++ b/src/Parsers/ASTFunction.cpp @@ -380,7 +380,7 @@ namespace findMySQLFunctionSecretArguments(); } else if ((engine_name == "S3") || (engine_name == "COSN") || (engine_name == "OSS") || - (engine_name == "DeltaLake") || (engine_name == "Hudi") || (engine_name == "Iceberg")) + (engine_name == "DeltaLake") || (engine_name == "Hudi") || (engine_name == "Iceberg") || (engine_name == "S3Queue")) { /// S3('url', ['aws_access_key_id', 'aws_secret_access_key',] ...) findS3TableEngineSecretArguments(); diff --git a/src/Parsers/ASTSelectQuery.cpp b/src/Parsers/ASTSelectQuery.cpp index 2115de1c124..d38e0933981 100644 --- a/src/Parsers/ASTSelectQuery.cpp +++ b/src/Parsers/ASTSelectQuery.cpp @@ -165,7 +165,7 @@ void ASTSelectQuery::formatImpl(const FormatSettings & s, FormatState & state, F if (order_by_all) { - s.ostr << (s.hilite ? hilite_keyword : "") << s.nl_or_ws << indent_str << "ORDER BY ALL" << (s.hilite ? hilite_none : ""); + s.ostr << (s.hilite ? hilite_keyword : "") << s.nl_or_ws << indent_str << "ORDER BY *" << (s.hilite ? hilite_none : ""); auto * elem = orderBy()->children[0]->as(); s.ostr << (s.hilite ? hilite_keyword : "") diff --git a/src/Parsers/ExpressionElementParsers.cpp b/src/Parsers/ExpressionElementParsers.cpp index 486555ae86d..62c480e0f6b 100644 --- a/src/Parsers/ExpressionElementParsers.cpp +++ b/src/Parsers/ExpressionElementParsers.cpp @@ -934,7 +934,7 @@ bool ParserNumber::parseImpl(Pos & pos, ASTPtr & node, Expected & expected) { if (float_value < 0) throw Exception(ErrorCodes::LOGICAL_ERROR, - "Token number cannot begin with minus, " + "Logical error: token number cannot begin with minus, " "but parsed float number is less than zero."); if (negative) diff --git a/src/Parsers/IParser.h b/src/Parsers/IParser.h index 198ec0346ff..d53b58baa7c 100644 --- a/src/Parsers/IParser.h +++ b/src/Parsers/IParser.h @@ -9,7 +9,6 @@ #include #include #include -#include namespace DB @@ -74,21 +73,6 @@ public: if (unlikely(max_depth > 0 && depth > max_depth)) throw Exception(ErrorCodes::TOO_DEEP_RECURSION, "Maximum parse depth ({}) exceeded. " "Consider rising max_parser_depth parameter.", max_depth); - - /** Sometimes the maximum parser depth can be set to a high value by the user, - * but we still want to avoid stack overflow. - * For this purpose, we can use the checkStackSize function, but it is too heavy. - * The solution is to check not too frequently. - * The frequency is arbitrary, but not too large, not too small, - * and a power of two to simplify the division. - */ -#if defined(USE_MUSL) || defined(SANITIZER) || !defined(NDEBUG) - static constexpr uint32_t check_frequency = 128; -#else - static constexpr uint32_t check_frequency = 8192; -#endif - if (depth % check_frequency == 0) - checkStackSize(); } ALWAYS_INLINE void decreaseDepth() diff --git a/src/Parsers/ParserSelectQuery.cpp b/src/Parsers/ParserSelectQuery.cpp index 641e74b5f18..6397a2a2a55 100644 --- a/src/Parsers/ParserSelectQuery.cpp +++ b/src/Parsers/ParserSelectQuery.cpp @@ -1,21 +1,23 @@ -#include +#include + +#include +#include +#include +#include #include +#include #include -#include #include #include #include -#include +#include #include -#include +#include #include #include -#include -#include -#include -#include #include +#include namespace DB { @@ -290,9 +292,9 @@ bool ParserSelectQuery::parseImpl(Pos & pos, ASTPtr & node, Expected & expected) } else if (order_expression_list->children.size() == 1) { - /// ORDER BY ALL - auto * identifier = order_expression_list->children[0]->as()->children[0]->as(); - if (identifier != nullptr && Poco::toUpper(identifier->name()) == "ALL") + /// ORDER BY * + auto * asterisk = order_expression_list->children[0]->as()->children[0]->as(); + if (asterisk != nullptr) select_query->order_by_all = true; } } diff --git a/src/Planner/PlannerJoinTree.cpp b/src/Planner/PlannerJoinTree.cpp index 227ac86d3a5..e6a459d0e8a 100644 --- a/src/Planner/PlannerJoinTree.cpp +++ b/src/Planner/PlannerJoinTree.cpp @@ -801,14 +801,18 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres table_expression_query_info.prewhere_info->prewhere_actions = filter_info.actions; table_expression_query_info.prewhere_info->prewhere_column_name = filter_info.column_name; table_expression_query_info.prewhere_info->remove_prewhere_column = filter_info.do_remove_column; + table_expression_query_info.prewhere_info->need_filter = true; } - else + else if (!table_expression_query_info.prewhere_info->row_level_filter) { table_expression_query_info.prewhere_info->row_level_filter = filter_info.actions; table_expression_query_info.prewhere_info->row_level_column_name = filter_info.column_name; + table_expression_query_info.prewhere_info->need_filter = true; + } + else + { + where_filters.emplace_back(filter_info, std::move(description)); } - - table_expression_query_info.prewhere_info->need_filter = true; } else { diff --git a/src/Processors/Executors/PipelineExecutor.cpp b/src/Processors/Executors/PipelineExecutor.cpp index 580aaa2b259..a06bacd7d3b 100644 --- a/src/Processors/Executors/PipelineExecutor.cpp +++ b/src/Processors/Executors/PipelineExecutor.cpp @@ -138,8 +138,8 @@ bool PipelineExecutor::executeStep(std::atomic_bool * yield_flag) initializeExecution(1, true); // Acquire slot until we are done - single_thread_slot = slots->tryAcquire(); - chassert(single_thread_slot && "Unable to allocate slot for the first thread, but we just allocated at least one slot"); + single_thread_cpu_slot = cpu_slots->tryAcquire(); + chassert(single_thread_cpu_slot && "Unable to allocate cpu slot for the first thread, but we just allocated at least one slot"); if (yield_flag && *yield_flag) return true; @@ -155,7 +155,7 @@ bool PipelineExecutor::executeStep(std::atomic_bool * yield_flag) if (node->exception) std::rethrow_exception(node->exception); - single_thread_slot.reset(); + single_thread_cpu_slot.reset(); finalizeExecution(); return false; @@ -333,8 +333,8 @@ void PipelineExecutor::initializeExecution(size_t num_threads, bool concurrency_ /// Allocate CPU slots from concurrency control size_t min_threads = concurrency_control ? 1uz : num_threads; - slots = ConcurrencyControl::instance().allocate(min_threads, num_threads); - use_threads = slots->grantedCount(); + cpu_slots = ConcurrencyControl::instance().allocate(min_threads, num_threads); + use_threads = cpu_slots->grantedCount(); Queue queue; graph->initializeExecution(queue); @@ -348,7 +348,7 @@ void PipelineExecutor::initializeExecution(size_t num_threads, bool concurrency_ void PipelineExecutor::spawnThreads() { - while (auto slot = slots->tryAcquire()) + while (auto slot = cpu_slots->tryAcquire()) { size_t thread_num = threads.fetch_add(1); @@ -405,7 +405,7 @@ void PipelineExecutor::executeImpl(size_t num_threads, bool concurrency_control) } else { - auto slot = slots->tryAcquire(); + auto slot = cpu_slots->tryAcquire(); executeSingleThread(0); } diff --git a/src/Processors/Executors/PipelineExecutor.h b/src/Processors/Executors/PipelineExecutor.h index 862a460f0ed..cb74b524163 100644 --- a/src/Processors/Executors/PipelineExecutor.h +++ b/src/Processors/Executors/PipelineExecutor.h @@ -68,8 +68,8 @@ private: ExecutorTasks tasks; /// Concurrency control related - ConcurrencyControl::AllocationPtr slots; - ConcurrencyControl::SlotPtr single_thread_slot; // slot for single-thread mode to work using executeStep() + SlotAllocationPtr cpu_slots; + AcquiredSlotPtr single_thread_cpu_slot; // cpu slot for single-thread mode to work using executeStep() std::unique_ptr pool; std::atomic_size_t threads = 0; diff --git a/src/Processors/Formats/Impl/JSONEachRowRowInputFormat.cpp b/src/Processors/Formats/Impl/JSONEachRowRowInputFormat.cpp index 6fa94356cd3..0ef19a9c14f 100644 --- a/src/Processors/Formats/Impl/JSONEachRowRowInputFormat.cpp +++ b/src/Processors/Formats/Impl/JSONEachRowRowInputFormat.cpp @@ -179,7 +179,7 @@ void JSONEachRowRowInputFormat::readJSONObject(MutableColumns & columns) else if (column_index == NESTED_FIELD) readNestedData(name_ref.toString(), columns); else - throw Exception(ErrorCodes::LOGICAL_ERROR, "Illegal value of column_index"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: illegal value of column_index"); } else { diff --git a/src/Processors/Formats/RowInputFormatWithDiagnosticInfo.cpp b/src/Processors/Formats/RowInputFormatWithDiagnosticInfo.cpp index fcf338577f8..a56c24a740a 100644 --- a/src/Processors/Formats/RowInputFormatWithDiagnosticInfo.cpp +++ b/src/Processors/Formats/RowInputFormatWithDiagnosticInfo.cpp @@ -136,7 +136,7 @@ bool RowInputFormatWithDiagnosticInfo::deserializeFieldAndPrintDiagnosticInfo(co auto * curr_position = in->position(); if (curr_position < prev_position) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Parsing is non-deterministic."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: parsing is non-deterministic."); if (isNativeNumber(type) || isDate(type) || isDateTime(type) || isDateTime64(type)) { diff --git a/src/Processors/Sources/WaitForAsyncInsertSource.h b/src/Processors/Sources/WaitForAsyncInsertSource.h index 78af6294202..1029c164941 100644 --- a/src/Processors/Sources/WaitForAsyncInsertSource.h +++ b/src/Processors/Sources/WaitForAsyncInsertSource.h @@ -33,7 +33,7 @@ protected: { auto status = insert_future.wait_for(std::chrono::milliseconds(timeout_ms)); if (status == std::future_status::deferred) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Got future in deferred state"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: got future in deferred state"); if (status == std::future_status::timeout) throw Exception(ErrorCodes::TIMEOUT_EXCEEDED, "Wait for async insert timeout ({} ms) exceeded)", timeout_ms); diff --git a/src/Processors/Transforms/CreatingSetsTransform.cpp b/src/Processors/Transforms/CreatingSetsTransform.cpp index eeb8f4a6060..cc0b5926e66 100644 --- a/src/Processors/Transforms/CreatingSetsTransform.cpp +++ b/src/Processors/Transforms/CreatingSetsTransform.cpp @@ -163,7 +163,7 @@ void CreatingSetsTransform::startSubquery() done_with_table = !external_table; if ((done_with_set && !set_from_cache) && done_with_table) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Nothing to do with subquery"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: nothing to do with subquery"); if (table_out.initialized()) { diff --git a/src/Processors/Transforms/getSourceFromASTInsertQuery.cpp b/src/Processors/Transforms/getSourceFromASTInsertQuery.cpp index 8a13973b970..6c7c7447070 100644 --- a/src/Processors/Transforms/getSourceFromASTInsertQuery.cpp +++ b/src/Processors/Transforms/getSourceFromASTInsertQuery.cpp @@ -37,7 +37,7 @@ InputFormatPtr getInputFormatFromASTInsertQuery( const auto * ast_insert_query = ast->as(); if (!ast_insert_query) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Query requires data to insert, but it is not INSERT query"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: query requires data to insert, but it is not INSERT query"); if (ast_insert_query->infile && context->getApplicationType() == Context::ApplicationType::SERVER) throw Exception(ErrorCodes::UNKNOWN_TYPE_OF_QUERY, "Query has infile and was send directly to server"); @@ -47,7 +47,7 @@ InputFormatPtr getInputFormatFromASTInsertQuery( if (input_function) throw Exception(ErrorCodes::INVALID_USAGE_OF_INPUT, "FORMAT must be specified for function input()"); else - throw Exception(ErrorCodes::LOGICAL_ERROR, "INSERT query requires format to be set"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: INSERT query requires format to be set"); } /// Data could be in parsed (ast_insert_query.data) and in not parsed yet (input_buffer_tail_part) part of query. @@ -105,7 +105,7 @@ std::unique_ptr getReadBufferFromASTInsertQuery(const ASTPtr & ast) { const auto * insert_query = ast->as(); if (!insert_query) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Query requires data to insert, but it is not INSERT query"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: query requires data to insert, but it is not INSERT query"); if (insert_query->infile) { diff --git a/src/QueryPipeline/ExecutionSpeedLimits.cpp b/src/QueryPipeline/ExecutionSpeedLimits.cpp index 05fd394db77..f8ae4c76d0f 100644 --- a/src/QueryPipeline/ExecutionSpeedLimits.cpp +++ b/src/QueryPipeline/ExecutionSpeedLimits.cpp @@ -113,7 +113,7 @@ static bool handleOverflowMode(OverflowMode mode, int code, FormatStringHelper #include #include +#include #include #include #include diff --git a/src/Server/HTTPHandlerFactory.cpp b/src/Server/HTTPHandlerFactory.cpp index 9e4a440ddb2..06ca1182be5 100644 --- a/src/Server/HTTPHandlerFactory.cpp +++ b/src/Server/HTTPHandlerFactory.cpp @@ -120,7 +120,7 @@ HTTPRequestHandlerFactoryPtr createHandlerFactory(IServer & server, const Poco:: return createPrometheusMainHandlerFactory(server, config, metrics_writer, name); } - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown HTTP handler factory name."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: Unknown HTTP handler factory name."); } diff --git a/src/Server/TCPHandler.cpp b/src/Server/TCPHandler.cpp index 9464ef74586..833f8ecc818 100644 --- a/src/Server/TCPHandler.cpp +++ b/src/Server/TCPHandler.cpp @@ -943,7 +943,7 @@ void TCPHandler::processInsertQuery() auto wait_status = result.future.wait_for(std::chrono::milliseconds(timeout_ms)); if (wait_status == std::future_status::deferred) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Got future in deferred state"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: got future in deferred state"); if (wait_status == std::future_status::timeout) throw Exception(ErrorCodes::TIMEOUT_EXCEEDED, "Wait for async insert timeout ({} ms) exceeded)", timeout_ms); diff --git a/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.h b/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.h index f7d7553851a..a1b436bb9c8 100644 --- a/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.h +++ b/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.h @@ -6,9 +6,7 @@ #include #include #include -#include #include -#include namespace CurrentMetrics { class Increment; } diff --git a/src/Storages/MergeTree/AlterConversions.cpp b/src/Storages/MergeTree/AlterConversions.cpp index a98cd6d99f9..31f8f17e2c1 100644 --- a/src/Storages/MergeTree/AlterConversions.cpp +++ b/src/Storages/MergeTree/AlterConversions.cpp @@ -9,6 +9,11 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } +bool AlterConversions::supportsMutationCommandType(MutationCommand::Type t) +{ + return t == MutationCommand::Type::RENAME_COLUMN; +} + void AlterConversions::addMutationCommand(const MutationCommand & command) { /// Currently only RENAME_COLUMN is applied on-fly. diff --git a/src/Storages/MergeTree/AlterConversions.h b/src/Storages/MergeTree/AlterConversions.h index 4410b9c56e2..0f857d351dd 100644 --- a/src/Storages/MergeTree/AlterConversions.h +++ b/src/Storages/MergeTree/AlterConversions.h @@ -35,6 +35,8 @@ public: /// Get column old name before rename (lookup by key in rename_map) std::string getColumnOldName(const std::string & new_name) const; + static bool supportsMutationCommandType(MutationCommand::Type); + private: /// Rename map new_name -> old_name. std::vector rename_map; diff --git a/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp b/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp index 000d36752cb..e31d991ef09 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDiskBase.cpp @@ -701,7 +701,9 @@ void DataPartStorageOnDiskBase::remove( if (file_name.starts_with(proj_dir_name)) files_not_to_remove_for_projection.emplace(fs::path(file_name).filename()); - LOG_DEBUG(log, "Will not remove files [{}] for projection {}", fmt::join(files_not_to_remove_for_projection, ", "), projection.name); + if (!files_not_to_remove_for_projection.empty()) + LOG_DEBUG( + log, "Will not remove files [{}] for projection {}", fmt::join(files_not_to_remove_for_projection, ", "), projection.name); CanRemoveDescription proj_description { diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index 168c5f729ce..ce70fbe18e5 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -903,7 +903,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDisk( || part_name.empty() || std::string::npos != tmp_prefix.find_first_of("/.") || std::string::npos != part_name.find_first_of("/.")) - throw Exception(ErrorCodes::LOGICAL_ERROR, "`tmp_prefix` and `part_name` cannot be empty or contain '.' or '/' characters."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: tmp_prefix and part_name cannot be empty or contain '.' or '/' characters."); auto part_dir = tmp_prefix + part_name; auto part_relative_path = data.getRelativeDataPath() + String(to_detached ? "detached/" : ""); diff --git a/src/Storages/MergeTree/EphemeralLockInZooKeeper.cpp b/src/Storages/MergeTree/EphemeralLockInZooKeeper.cpp index cbdeabffa97..1ffb5177430 100644 --- a/src/Storages/MergeTree/EphemeralLockInZooKeeper.cpp +++ b/src/Storages/MergeTree/EphemeralLockInZooKeeper.cpp @@ -17,7 +17,7 @@ EphemeralLockInZooKeeper::EphemeralLockInZooKeeper(const String & path_prefix_, : zookeeper(zookeeper_), path_prefix(path_prefix_), path(path_), conflict_path(conflict_path_) { if (conflict_path.empty() && path.size() <= path_prefix.size()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Name of the main node is shorter than prefix."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: name of the main node is shorter than prefix."); } template @@ -179,7 +179,7 @@ EphemeralLocksInAllPartitions::EphemeralLocksInAllPartitions( size_t prefix_size = block_numbers_path.size() + 1 + partitions[i].size() + 1 + path_prefix.size(); const String & path = dynamic_cast(*lock_responses[i]).path_created; if (path.size() <= prefix_size) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Name of the sequential node is shorter than prefix."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: name of the sequential node is shorter than prefix."); UInt64 number = parse(path.c_str() + prefix_size, path.size() - prefix_size); locks.push_back(LockInfo{path, partitions[i], number}); diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index 2826c3e23f1..11ede661f78 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -347,7 +347,7 @@ const IMergeTreeDataPart::Index & IMergeTreeDataPart::getIndex() const { std::scoped_lock lock(index_mutex); if (!index_loaded) - loadIndex(lock); + loadIndex(); index_loaded = true; return index; } @@ -569,6 +569,7 @@ void IMergeTreeDataPart::removeIfNeeded() UInt64 IMergeTreeDataPart::getIndexSizeInBytes() const { + std::scoped_lock lock(index_mutex); UInt64 res = 0; for (const ColumnPtr & column : index) res += column->byteSize(); @@ -577,6 +578,7 @@ UInt64 IMergeTreeDataPart::getIndexSizeInBytes() const UInt64 IMergeTreeDataPart::getIndexSizeInAllocatedBytes() const { + std::scoped_lock lock(index_mutex); UInt64 res = 0; for (const ColumnPtr & column : index) res += column->allocatedBytes(); @@ -828,7 +830,7 @@ void IMergeTreeDataPart::appendFilesOfIndexGranularity(Strings & /* files */) co { } -void IMergeTreeDataPart::loadIndex(std::scoped_lock &) const +void IMergeTreeDataPart::loadIndex() const { /// Memory for index must not be accounted as memory usage for query, because it belongs to a table. MemoryTrackerBlockerInThread temporarily_disable_memory_tracker; diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.h b/src/Storages/MergeTree/IMergeTreeDataPart.h index c9dea1afcc5..0d7acfab891 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.h +++ b/src/Storages/MergeTree/IMergeTreeDataPart.h @@ -3,6 +3,7 @@ #include #include #include +#include #include #include #include @@ -565,8 +566,8 @@ protected: /// Lazily loaded in RAM. Contains each index_granularity-th value of primary key tuple. /// Note that marks (also correspond to primary key) are not always in RAM, but cached. See MarkCache.h. mutable std::mutex index_mutex; - mutable Index index; - mutable bool index_loaded = false; + mutable Index index TSA_GUARDED_BY(index_mutex); + mutable bool index_loaded TSA_GUARDED_BY(index_mutex) = false; /// Total size of all columns, calculated once in calcuateColumnSizesOnDisk ColumnSize total_columns_size; @@ -664,7 +665,7 @@ private: virtual void appendFilesOfIndexGranularity(Strings & files) const; /// Loads the index file. - void loadIndex(std::scoped_lock &) const; + void loadIndex() const TSA_REQUIRES(index_mutex); void appendFilesOfIndex(Strings & files) const; diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 5b297de3fda..2e63701dbdb 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -869,7 +869,7 @@ void MergeTreeData::MergingParams::check(const StorageInMemoryMetadata & metadat if (is_optional) return; - throw Exception(ErrorCodes::LOGICAL_ERROR, "Sign column for storage {} is empty", storage); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: Sign column for storage {} is empty", storage); } bool miss_column = true; @@ -896,7 +896,7 @@ void MergeTreeData::MergingParams::check(const StorageInMemoryMetadata & metadat if (is_optional) return; - throw Exception(ErrorCodes::LOGICAL_ERROR, "Version column for storage {} is empty", storage); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: Version column for storage {} is empty", storage); } bool miss_column = true; @@ -925,12 +925,12 @@ void MergeTreeData::MergingParams::check(const StorageInMemoryMetadata & metadat if (is_optional) return; - throw Exception(ErrorCodes::LOGICAL_ERROR, "`is_deleted` ({}) column for storage {} is empty", is_deleted_column, storage); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: is_deleted ({}) column for storage {} is empty", is_deleted_column, storage); } else { if (version_column.empty() && !is_optional) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Version column ({}) for storage {} is empty while is_deleted ({}) is not.", + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: Version column ({}) for storage {} is empty while is_deleted ({}) is not.", version_column, storage, is_deleted_column); bool miss_is_deleted_column = true; @@ -7950,12 +7950,11 @@ bool MergeTreeData::canUsePolymorphicParts(const MergeTreeSettings & settings, S AlterConversionsPtr MergeTreeData::getAlterConversionsForPart(MergeTreeDataPartPtr part) const { - auto commands_map = getAlterMutationCommandsForPart(part); + auto commands = getAlterMutationCommandsForPart(part); auto result = std::make_shared(); - for (const auto & [_, commands] : commands_map) - for (const auto & command : commands) - result->addMutationCommand(command); + for (const auto & command : commands | std::views::reverse) + result->addMutationCommand(command); return result; } diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 4ad440dae00..b06b3018938 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -35,7 +35,7 @@ #include #include #include -#include +#include #include @@ -1356,11 +1356,12 @@ protected: /// mechanisms for parts locking virtual bool partIsAssignedToBackgroundOperation(const DataPartPtr & part) const = 0; - /// Return most recent mutations commands for part which weren't applied - /// Used to receive AlterConversions for part and apply them on fly. This - /// method has different implementations for replicated and non replicated - /// MergeTree because they store mutations in different way. - virtual std::map getAlterMutationCommandsForPart(const DataPartPtr & part) const = 0; + /// Return pending mutations that weren't applied to `part` yet and should be applied on the fly + /// (i.e. when reading from the part). Mutations not supported by AlterConversions + /// (supportsMutationCommandType()) can be omitted. + /// + /// @return list of mutations, in *reverse* order (newest to oldest) + virtual MutationCommands getAlterMutationCommandsForPart(const DataPartPtr & part) const = 0; struct PartBackupEntries { diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index 1bf1d4a3c29..58fddde7b54 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -85,7 +85,7 @@ UInt64 MergeTreeDataMergerMutator::getMaxSourcePartsSizeForMerge(size_t max_coun if (scheduled_tasks_count > max_count) { throw Exception(ErrorCodes::LOGICAL_ERROR, - "Invalid argument passed to getMaxSourcePartsSize: scheduled_tasks_count = {} > max_count = {}", + "Logical error: invalid argument passed to getMaxSourcePartsSize: scheduled_tasks_count = {} > max_count = {}", scheduled_tasks_count, max_count); } @@ -511,7 +511,7 @@ SelectPartsDecision MergeTreeDataMergerMutator::selectPartsToMergeFromRanges( /// Do not allow to "merge" part with itself for regular merges, unless it is a TTL-merge where it is ok to remove some values with expired ttl if (parts_to_merge.size() == 1) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Merge selector returned only one part to merge"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: merge selector returned only one part to merge"); if (parts_to_merge.empty()) { diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index ebf887f5e9e..c9c16b59f9e 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -384,13 +384,13 @@ Block MergeTreeDataWriter::mergeBlock( /// Check that after first merge merging_algorithm is waiting for data from input 0. if (status.required_source != 0) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Required source after the first merge is not 0. Chunk rows: {}, is_finished: {}, required_source: {}, algorithm: {}", status.chunk.getNumRows(), status.is_finished, status.required_source, merging_algorithm->getName()); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: required source after the first merge is not 0. Chunk rows: {}, is_finished: {}, required_source: {}, algorithm: {}", status.chunk.getNumRows(), status.is_finished, status.required_source, merging_algorithm->getName()); status = merging_algorithm->merge(); /// Check that merge is finished. if (!status.is_finished) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Merge is not finished after the second merge."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: merge is not finished after the second merge."); /// Merged Block is sorted and we don't need to use permutation anymore permutation = nullptr; @@ -439,7 +439,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPartImpl( auto max_month = date_lut.toNumYYYYMM(max_date); if (min_month != max_month) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Part spans more than one month."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: part spans more than one month."); part_name = new_part_info.getPartNameV0(min_date, max_date); } diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp index f506230b5ea..da49814b83a 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp @@ -59,7 +59,7 @@ bool maybeTrueOnBloomFilter(const IColumn * hash_column, const BloomFilterPtr & const auto * non_const_column = typeid_cast(hash_column); if (!const_column && !non_const_column) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Hash column must be Const or UInt64."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: hash column must be Const Column or UInt64 Column."); if (const_column) { diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h index 8029d6d405b..db85c804d8d 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h @@ -53,7 +53,7 @@ public: if (const auto & bf_granule = typeid_cast(granule.get())) return mayBeTrueOnGranule(bf_granule); - throw Exception(ErrorCodes::LOGICAL_ERROR, "Requires bloom filter index granule."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "LOGICAL ERROR: require bloom filter index granule."); } private: diff --git a/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp b/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp index da89d52a9ff..4e339964de3 100644 --- a/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp @@ -54,9 +54,9 @@ MarkType::MarkType(bool adaptive_, bool compressed_, MergeTreeDataPartType::Valu : adaptive(adaptive_), compressed(compressed_), part_type(part_type_) { if (!adaptive && part_type != MergeTreeDataPartType::Wide) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Non-Wide data part type with non-adaptive granularity"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: non-Wide data part type with non-adaptive granularity"); if (part_type == MergeTreeDataPartType::Unknown) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown data part type"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: unknown data part type"); } bool MarkType::isMarkFileExtension(std::string_view extension) @@ -71,7 +71,7 @@ std::string MarkType::getFileExtension() const if (!adaptive) { if (part_type != MergeTreeDataPartType::Wide) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Non-Wide data part type with non-adaptive granularity"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: non-Wide data part type with non-adaptive granularity"); return res; } @@ -84,7 +84,7 @@ std::string MarkType::getFileExtension() const case MergeTreeDataPartType::InMemory: return ""; case MergeTreeDataPartType::Unknown: - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown data part type"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: unknown data part type"); } } diff --git a/src/Storages/MergeTree/MergeTreeMarksLoader.cpp b/src/Storages/MergeTree/MergeTreeMarksLoader.cpp index eb7b60b0727..8250050412f 100644 --- a/src/Storages/MergeTree/MergeTreeMarksLoader.cpp +++ b/src/Storages/MergeTree/MergeTreeMarksLoader.cpp @@ -1,6 +1,6 @@ #include #include -#include +#include #include #include #include diff --git a/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp b/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp index 47c2fe07bb4..8d8b0f1cc79 100644 --- a/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp +++ b/src/Storages/MergeTree/MergeTreePrefetchedReadPool.cpp @@ -1,6 +1,6 @@ #include #include -#include +#include #include #include #include diff --git a/src/Storages/MergeTree/MergeTreeSource.cpp b/src/Storages/MergeTree/MergeTreeSource.cpp index a450505f7a8..e1d1d0951e4 100644 --- a/src/Storages/MergeTree/MergeTreeSource.cpp +++ b/src/Storages/MergeTree/MergeTreeSource.cpp @@ -1,6 +1,6 @@ #include #include -#include +#include #include #include diff --git a/src/Storages/MergeTree/MutateTask.cpp b/src/Storages/MergeTree/MutateTask.cpp index 6882963fd24..6bacce9e2c5 100644 --- a/src/Storages/MergeTree/MutateTask.cpp +++ b/src/Storages/MergeTree/MutateTask.cpp @@ -73,6 +73,7 @@ static void splitAndModifyMutationCommands( LoggerPtr log) { auto part_columns = part->getColumnsDescription(); + const auto & table_columns = metadata_snapshot->getColumns(); if (!isWidePart(part) || !isFullPartStorage(part->getDataPartStorage())) { @@ -81,9 +82,19 @@ static void splitAndModifyMutationCommands( for (const auto & command : commands) { + if (command.type == MutationCommand::Type::MATERIALIZE_COLUMN) + { + /// For ordinary column with default or materialized expression, MATERIALIZE COLUMN should not override past values + /// So we only mutate column if `command.column_name` is a default/materialized column or if the part does not have physical column file + auto column_ordinary = table_columns.getOrdinary().tryGetByName(command.column_name); + if (!column_ordinary || !part->tryGetColumn(command.column_name) || !part->hasColumnFiles(*column_ordinary)) + { + for_interpreter.push_back(command); + mutated_columns.emplace(command.column_name); + } + } if (command.type == MutationCommand::Type::MATERIALIZE_INDEX || command.type == MutationCommand::Type::MATERIALIZE_STATISTIC - || command.type == MutationCommand::Type::MATERIALIZE_COLUMN || command.type == MutationCommand::Type::MATERIALIZE_PROJECTION || command.type == MutationCommand::Type::MATERIALIZE_TTL || command.type == MutationCommand::Type::DELETE @@ -93,9 +104,6 @@ static void splitAndModifyMutationCommands( for_interpreter.push_back(command); for (const auto & [column_name, expr] : command.column_to_update_expression) mutated_columns.emplace(column_name); - - if (command.type == MutationCommand::Type::MATERIALIZE_COLUMN) - mutated_columns.emplace(command.column_name); } else if (command.type == MutationCommand::Type::DROP_INDEX || command.type == MutationCommand::Type::DROP_PROJECTION @@ -205,8 +213,15 @@ static void splitAndModifyMutationCommands( { for (const auto & command : commands) { - if (command.type == MutationCommand::Type::MATERIALIZE_INDEX - || command.type == MutationCommand::Type::MATERIALIZE_COLUMN + if (command.type == MutationCommand::Type::MATERIALIZE_COLUMN) + { + /// For ordinary column with default or materialized expression, MATERIALIZE COLUMN should not override past values + /// So we only mutate column if `command.column_name` is a default/materialized column or if the part does not have physical column file + auto column_ordinary = table_columns.getOrdinary().tryGetByName(command.column_name); + if (!column_ordinary || !part->tryGetColumn(command.column_name) || !part->hasColumnFiles(*column_ordinary)) + for_interpreter.push_back(command); + } + else if (command.type == MutationCommand::Type::MATERIALIZE_INDEX || command.type == MutationCommand::Type::MATERIALIZE_STATISTIC || command.type == MutationCommand::Type::MATERIALIZE_PROJECTION || command.type == MutationCommand::Type::MATERIALIZE_TTL diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeQueue.cpp b/src/Storages/MergeTree/ReplicatedMergeTreeQueue.cpp index 2d4b4e0edba..ab256f07c89 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeQueue.cpp +++ b/src/Storages/MergeTree/ReplicatedMergeTreeQueue.cpp @@ -1789,7 +1789,7 @@ ReplicatedMergeTreeMergePredicate ReplicatedMergeTreeQueue::getMergePredicate(zk } -std::map ReplicatedMergeTreeQueue::getAlterMutationCommandsForPart(const MergeTreeData::DataPartPtr & part) const +MutationCommands ReplicatedMergeTreeQueue::getAlterMutationCommandsForPart(const MergeTreeData::DataPartPtr & part) const { std::unique_lock lock(state_mutex); @@ -1799,9 +1799,8 @@ std::map ReplicatedMergeTreeQueue::getAlterMutationCo Int64 part_data_version = part->info.getDataVersion(); Int64 part_metadata_version = part->getMetadataVersion(); - LOG_TEST(log, "Looking for mutations for part {} (part data version {}, part metadata version {})", part->name, part_data_version, part_metadata_version); - std::map result; + MutationCommands result; bool seen_all_data_mutations = false; bool seen_all_metadata_mutations = false; @@ -1814,7 +1813,15 @@ std::map ReplicatedMergeTreeQueue::getAlterMutationCo if (seen_all_data_mutations && seen_all_metadata_mutations) break; - auto alter_version = mutation_status->entry->alter_version; + auto & entry = mutation_status->entry; + + auto add_to_result = [&] { + for (const auto & command : entry->commands | std::views::reverse) + if (AlterConversions::supportsMutationCommandType(command.type)) + result.emplace_back(command); + }; + + auto alter_version = entry->alter_version; if (alter_version != -1) { if (alter_version > storage.getInMemoryMetadataPtr()->getMetadataVersion()) @@ -1822,22 +1829,19 @@ std::map ReplicatedMergeTreeQueue::getAlterMutationCo /// We take commands with bigger metadata version if (alter_version > part_metadata_version) - result[mutation_version] = mutation_status->entry->commands; + add_to_result(); else seen_all_metadata_mutations = true; } else { if (mutation_version > part_data_version) - result[mutation_version] = mutation_status->entry->commands; + add_to_result(); else seen_all_data_mutations = true; } } - LOG_TEST(log, "Got {} commands for part {} (part data version {}, part metadata version {})", - result.size(), part->name, part_data_version, part_metadata_version); - return result; } diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeQueue.h b/src/Storages/MergeTree/ReplicatedMergeTreeQueue.h index 84106565dff..743ca7fc258 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeQueue.h +++ b/src/Storages/MergeTree/ReplicatedMergeTreeQueue.h @@ -401,7 +401,7 @@ public: /// Return mutation commands for part which could be not applied to /// it according to part mutation version. Used when we apply alter commands on fly, /// without actual data modification on disk. - std::map getAlterMutationCommandsForPart(const MergeTreeData::DataPartPtr & part) const; + MutationCommands getAlterMutationCommandsForPart(const MergeTreeData::DataPartPtr & part) const; /// Mark finished mutations as done. If the function needs to be called again at some later time /// (because some mutations are probably done but we are not sure yet), returns true. diff --git a/src/Storages/RabbitMQ/RabbitMQConsumer.cpp b/src/Storages/RabbitMQ/RabbitMQConsumer.cpp index 1843bebe3c7..28dc239ae37 100644 --- a/src/Storages/RabbitMQ/RabbitMQConsumer.cpp +++ b/src/Storages/RabbitMQ/RabbitMQConsumer.cpp @@ -128,6 +128,32 @@ bool RabbitMQConsumer::ackMessages(const CommitInfo & commit_info) return false; } +bool RabbitMQConsumer::nackMessages(const CommitInfo & commit_info) +{ + if (state != State::OK) + return false; + + /// Nothing to nack. + if (!commit_info.delivery_tag || commit_info.delivery_tag <= last_commited_delivery_tag) + return false; + + if (consumer_channel->reject(commit_info.delivery_tag, AMQP::multiple)) + { + LOG_TRACE( + log, "Consumer rejected messages with deliveryTags from {} to {} on channel {}", + last_commited_delivery_tag, commit_info.delivery_tag, channel_id); + + return true; + } + + LOG_ERROR( + log, + "Failed to reject messages for {}:{}, (current commit point {}:{})", + commit_info.channel_id, commit_info.delivery_tag, + channel_id, last_commited_delivery_tag); + + return false; +} void RabbitMQConsumer::updateChannel(RabbitMQConnection & connection) { @@ -161,7 +187,7 @@ void RabbitMQConsumer::updateChannel(RabbitMQConnection & connection) consumer_channel->onError([&](const char * message) { - LOG_ERROR(log, "Channel {} in an error state: {}", channel_id, message); + LOG_ERROR(log, "Channel {} in in error state: {}", channel_id, message); state = State::ERROR; }); } diff --git a/src/Storages/RabbitMQ/RabbitMQConsumer.h b/src/Storages/RabbitMQ/RabbitMQConsumer.h index c78b33bfc7c..9dad175dda3 100644 --- a/src/Storages/RabbitMQ/RabbitMQConsumer.h +++ b/src/Storages/RabbitMQ/RabbitMQConsumer.h @@ -50,7 +50,9 @@ public: UInt64 delivery_tag = 0; String channel_id; }; + const MessageData & currentMessage() { return current; } + const String & getChannelID() const { return channel_id; } /// Return read buffer containing next available message /// or nullptr if there are no messages to process. @@ -63,6 +65,7 @@ public: bool isConsumerStopped() const { return stopped.load(); } bool ackMessages(const CommitInfo & commit_info); + bool nackMessages(const CommitInfo & commit_info); bool hasPendingMessages() { return !received.empty(); } diff --git a/src/Storages/RabbitMQ/RabbitMQSource.cpp b/src/Storages/RabbitMQ/RabbitMQSource.cpp index 3cec448fc11..72196e7dd3c 100644 --- a/src/Storages/RabbitMQ/RabbitMQSource.cpp +++ b/src/Storages/RabbitMQ/RabbitMQSource.cpp @@ -123,7 +123,11 @@ Chunk RabbitMQSource::generateImpl() } if (is_finished || !consumer || consumer->isConsumerStopped()) + { + LOG_TRACE(log, "RabbitMQSource is stopped (is_finished: {}, consumer_stopped: {})", + is_finished, consumer ? toString(consumer->isConsumerStopped()) : "No consumer"); return {}; + } /// Currently it is one time usage source: to make sure data is flushed /// strictly by timeout or by block size. @@ -254,13 +258,12 @@ Chunk RabbitMQSource::generateImpl() bool RabbitMQSource::sendAck() { - if (!consumer) - return false; + return consumer && consumer->ackMessages(commit_info); +} - if (!consumer->ackMessages(commit_info)) - return false; - - return true; +bool RabbitMQSource::sendNack() +{ + return consumer && consumer->nackMessages(commit_info); } } diff --git a/src/Storages/RabbitMQ/RabbitMQSource.h b/src/Storages/RabbitMQ/RabbitMQSource.h index 21d059bfae2..0d6fad97054 100644 --- a/src/Storages/RabbitMQ/RabbitMQSource.h +++ b/src/Storages/RabbitMQ/RabbitMQSource.h @@ -33,6 +33,7 @@ public: bool needChannelUpdate(); void updateChannel(); bool sendAck(); + bool sendNack(); private: StorageRabbitMQ & storage; diff --git a/src/Storages/RabbitMQ/StorageRabbitMQ.cpp b/src/Storages/RabbitMQ/StorageRabbitMQ.cpp index 868f48d0b7d..ec2048cca70 100644 --- a/src/Storages/RabbitMQ/StorageRabbitMQ.cpp +++ b/src/Storages/RabbitMQ/StorageRabbitMQ.cpp @@ -1061,7 +1061,8 @@ bool StorageRabbitMQ::tryStreamToViews() for (size_t i = 0; i < num_created_consumers; ++i) { auto source = std::make_shared( - *this, storage_snapshot, rabbitmq_context, column_names, block_size, max_execution_time_ms, rabbitmq_settings->rabbitmq_handle_error_mode, false); + *this, storage_snapshot, rabbitmq_context, column_names, block_size, + max_execution_time_ms, rabbitmq_settings->rabbitmq_handle_error_mode, false); sources.emplace_back(source); pipes.emplace_back(source); @@ -1069,13 +1070,25 @@ bool StorageRabbitMQ::tryStreamToViews() block_io.pipeline.complete(Pipe::unitePipes(std::move(pipes))); + std::atomic_size_t rows = 0; + block_io.pipeline.setProgressCallback([&](const Progress & progress) { rows += progress.read_rows.load(); }); + if (!connection->getHandler().loopRunning()) startLoop(); + bool write_failed = false; + try { CompletedPipelineExecutor executor(block_io.pipeline); executor.execute(); } + catch (...) + { + LOG_ERROR(log, "Failed to push to views. Error: {}", getCurrentExceptionMessage(true)); + write_failed = true; + } + + LOG_TRACE(log, "Processed {} rows", rows); /* Note: sending ack() with loop running in another thread will lead to a lot of data races inside the library, but only in case * error occurs or connection is lost while ack is being sent @@ -1083,13 +1096,6 @@ bool StorageRabbitMQ::tryStreamToViews() deactivateTask(looping_task, false, true); size_t queue_empty = 0; - if (!hasDependencies(getStorageID())) - { - /// Do not commit to rabbitmq if the dependency was removed. - LOG_TRACE(log, "No dependencies, reschedule"); - return false; - } - if (!connection->isConnected()) { if (shutdown_called) @@ -1130,7 +1136,7 @@ bool StorageRabbitMQ::tryStreamToViews() * the same channel will also commit all previously not-committed messages. Anyway I do not think that for ack frame this * will ever happen. */ - if (!source->sendAck()) + if (write_failed ? source->sendNack() : source->sendAck()) { /// Iterate loop to activate error callbacks if they happened connection->getHandler().iterateLoop(); @@ -1142,6 +1148,19 @@ bool StorageRabbitMQ::tryStreamToViews() } } + if (write_failed) + { + LOG_TRACE(log, "Write failed, reschedule"); + return false; + } + + if (!hasDependencies(getStorageID())) + { + /// Do not commit to rabbitmq if the dependency was removed. + LOG_TRACE(log, "No dependencies, reschedule"); + return false; + } + if ((queue_empty == num_created_consumers) && (++read_attempts == MAX_FAILED_READ_ATTEMPTS)) { connection->heartbeat(); diff --git a/src/Storages/S3Queue/S3QueueSource.cpp b/src/Storages/S3Queue/S3QueueSource.cpp index b4f5f957f76..933238d8614 100644 --- a/src/Storages/S3Queue/S3QueueSource.cpp +++ b/src/Storages/S3Queue/S3QueueSource.cpp @@ -352,7 +352,11 @@ void StorageS3QueueSource::applyActionAfterProcessing(const String & path) } } -void StorageS3QueueSource::appendLogElement(const std::string & filename, S3QueueFilesMetadata::FileStatus & file_status_, size_t processed_rows, bool processed) +void StorageS3QueueSource::appendLogElement( + const std::string & filename, + S3QueueFilesMetadata::FileStatus & file_status_, + size_t processed_rows, + bool processed) { if (!s3_queue_log) return; @@ -363,6 +367,9 @@ void StorageS3QueueSource::appendLogElement(const std::string & filename, S3Queu elem = S3QueueLogElement { .event_time = std::chrono::system_clock::to_time_t(std::chrono::system_clock::now()), + .database = storage_id.database_name, + .table = storage_id.table_name, + .uuid = toString(storage_id.uuid), .file_name = filename, .rows_processed = processed_rows, .status = processed ? S3QueueLogElement::S3QueueStatus::Processed : S3QueueLogElement::S3QueueStatus::Failed, diff --git a/src/Storages/StorageBuffer.cpp b/src/Storages/StorageBuffer.cpp index d5c135bb81d..2925038ec8e 100644 --- a/src/Storages/StorageBuffer.cpp +++ b/src/Storages/StorageBuffer.cpp @@ -1,40 +1,41 @@ -#include -#include #include +#include +#include #include #include -#include #include -#include -#include -#include -#include -#include -#include -#include #include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include +#include +#include +#include +#include #include +#include +#include +#include #include #include #include -#include -#include -#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include namespace ProfileEvents @@ -56,6 +57,9 @@ namespace CurrentMetrics { extern const Metric StorageBufferRows; extern const Metric StorageBufferBytes; + extern const Metric StorageBufferFlushThreads; + extern const Metric StorageBufferFlushThreadsActive; + extern const Metric StorageBufferFlushThreadsScheduled; } @@ -153,6 +157,12 @@ StorageBuffer::StorageBuffer( storage_metadata.setComment(comment); setInMemoryMetadata(storage_metadata); + if (num_shards > 1) + { + flush_pool = std::make_unique( + CurrentMetrics::StorageBufferFlushThreads, CurrentMetrics::StorageBufferFlushThreadsActive, CurrentMetrics::StorageBufferFlushThreadsScheduled, + num_shards, 0, num_shards); + } flush_handle = bg_pool.createTask(log->name() + "/Bg", [this]{ backgroundFlush(); }); } @@ -802,7 +812,22 @@ bool StorageBuffer::checkThresholdsImpl(bool direct, size_t rows, size_t bytes, void StorageBuffer::flushAllBuffers(bool check_thresholds) { for (auto & buf : buffers) - flushBuffer(buf, check_thresholds, false); + { + if (flush_pool) + { + scheduleFromThreadPool([&] () + { + flushBuffer(buf, check_thresholds, false); + }, *flush_pool, "BufferFlush"); + } + else + { + flushBuffer(buf, check_thresholds, false); + } + } + + if (flush_pool) + flush_pool->wait(); } diff --git a/src/Storages/StorageBuffer.h b/src/Storages/StorageBuffer.h index 47f6239b173..6c15c7e0238 100644 --- a/src/Storages/StorageBuffer.h +++ b/src/Storages/StorageBuffer.h @@ -3,6 +3,7 @@ #include #include #include +#include #include @@ -149,6 +150,7 @@ private: /// There are `num_shards` of independent buffers. const size_t num_shards; + std::unique_ptr flush_pool; std::vector buffers; const Thresholds min_thresholds; diff --git a/src/Storages/StorageJoin.cpp b/src/Storages/StorageJoin.cpp index b122674466f..b9e082c0b22 100644 --- a/src/Storages/StorageJoin.cpp +++ b/src/Storages/StorageJoin.cpp @@ -500,7 +500,7 @@ protected: Chunk chunk; if (!joinDispatch(join->kind, join->strictness, join->data->maps.front(), [&](auto kind, auto strictness, auto & map) { chunk = createChunk(map); })) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown JOIN strictness"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: unknown JOIN strictness"); return chunk; } diff --git a/src/Storages/StorageLog.cpp b/src/Storages/StorageLog.cpp index 99192fe1e50..c7b0a9d0644 100644 --- a/src/Storages/StorageLog.cpp +++ b/src/Storages/StorageLog.cpp @@ -241,7 +241,7 @@ void LogSource::readData(const NameAndTypePair & name_and_type, ColumnPtr & colu const auto & data_file_it = storage.data_files_by_names.find(data_file_name); if (data_file_it == storage.data_files_by_names.end()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No information about file {} in StorageLog", data_file_name); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no information about file {} in StorageLog", data_file_name); const auto & data_file = *data_file_it->second; size_t offset = stream_for_prefix ? 0 : offsets[data_file.index]; @@ -448,7 +448,7 @@ ISerialization::OutputStreamGetter LogSink::createStreamGetter(const NameAndType String data_file_name = ISerialization::getFileNameForStream(name_and_type, path); auto it = streams.find(data_file_name); if (it == streams.end()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Stream was not created when writing data in LogSink"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: stream was not created when writing data in LogSink"); Stream & stream = it->second; if (stream.written) @@ -473,7 +473,7 @@ void LogSink::writeData(const NameAndTypePair & name_and_type, const IColumn & c { const auto & data_file_it = storage.data_files_by_names.find(data_file_name); if (data_file_it == storage.data_files_by_names.end()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No information about file {} in StorageLog", data_file_name); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no information about file {} in StorageLog", data_file_name); const auto & data_file = *data_file_it->second; const auto & columns = metadata_snapshot->getColumns(); diff --git a/src/Storages/StorageMergeTree.cpp b/src/Storages/StorageMergeTree.cpp index 678535da732..3458bd18ed3 100644 --- a/src/Storages/StorageMergeTree.cpp +++ b/src/Storages/StorageMergeTree.cpp @@ -2394,19 +2394,21 @@ void StorageMergeTree::attachRestoredParts(MutableDataPartsVector && parts) } -std::map StorageMergeTree::getAlterMutationCommandsForPart(const DataPartPtr & part) const +MutationCommands StorageMergeTree::getAlterMutationCommandsForPart(const DataPartPtr & part) const { std::lock_guard lock(currently_processing_in_background_mutex); UInt64 part_data_version = part->info.getDataVersion(); - std::map result; + MutationCommands result; for (const auto & [mutation_version, entry] : current_mutations_by_version | std::views::reverse) { - if (mutation_version > part_data_version) - result[mutation_version] = entry.commands; - else + if (mutation_version <= part_data_version) break; + + for (const auto & command : entry.commands | std::views::reverse) + if (AlterConversions::supportsMutationCommandType(command.type)) + result.emplace_back(command); } return result; diff --git a/src/Storages/StorageMergeTree.h b/src/Storages/StorageMergeTree.h index 359fa1d262d..8c41664b23c 100644 --- a/src/Storages/StorageMergeTree.h +++ b/src/Storages/StorageMergeTree.h @@ -308,7 +308,7 @@ private: }; protected: - std::map getAlterMutationCommandsForPart(const DataPartPtr & part) const override; + MutationCommands getAlterMutationCommandsForPart(const DataPartPtr & part) const override; }; } diff --git a/src/Storages/StorageReplicatedMergeTree.cpp b/src/Storages/StorageReplicatedMergeTree.cpp index 5870671736d..320fb09fd02 100644 --- a/src/Storages/StorageReplicatedMergeTree.cpp +++ b/src/Storages/StorageReplicatedMergeTree.cpp @@ -513,8 +513,15 @@ StorageReplicatedMergeTree::StorageReplicatedMergeTree( if (same_structure) { Coordination::Stat metadata_stat; - current_zookeeper->get(zookeeper_path + "/metadata", &metadata_stat); + current_zookeeper->get(fs::path(zookeeper_path) / "metadata", &metadata_stat); + + /** We change metadata_snapshot so that `createReplica` method will create `metadata_version` node in ZooKeeper + * with version of table '/metadata' node in Zookeeper. + * + * Otherwise `metadata_version` for not first replica will be initialized with 0 by default. + */ setInMemoryMetadata(metadata_snapshot->withMetadataVersion(metadata_stat.version)); + metadata_snapshot = getInMemoryMetadataPtr(); } } catch (Coordination::Exception & e) @@ -2043,7 +2050,7 @@ bool StorageReplicatedMergeTree::executeFetch(LogEntry & entry, bool need_to_che if (entry.quorum) { if (entry.type != LogEntry::GET_PART) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Log entry with quorum but type is not GET_PART"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: log entry with quorum but type is not GET_PART"); LOG_DEBUG(log, "No active replica has part {} which needs to be written with quorum. Will try to mark that quorum as failed.", entry.new_part_name); @@ -2106,7 +2113,7 @@ bool StorageReplicatedMergeTree::executeFetch(LogEntry & entry, bool need_to_che auto part_info = MergeTreePartInfo::fromPartName(entry.new_part_name, format_version); if (part_info.min_block != part_info.max_block) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Log entry with quorum for part covering more than one block number"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: log entry with quorum for part covering more than one block number"); ops.emplace_back(zkutil::makeCreateRequest( fs::path(zookeeper_path) / "quorum" / "failed_parts" / entry.new_part_name, @@ -5817,6 +5824,7 @@ bool StorageReplicatedMergeTree::executeMetadataAlter(const StorageReplicatedMer Coordination::Requests requests; requests.emplace_back(zkutil::makeSetRequest(fs::path(replica_path) / "columns", entry.columns_str, -1)); requests.emplace_back(zkutil::makeSetRequest(fs::path(replica_path) / "metadata", entry.metadata_str, -1)); + requests.emplace_back(zkutil::makeSetRequest(fs::path(replica_path) / "metadata_version", std::to_string(entry.alter_version), -1)); auto table_id = getStorageID(); auto alter_context = getContext(); @@ -5863,10 +5871,6 @@ bool StorageReplicatedMergeTree::executeMetadataAlter(const StorageReplicatedMer resetObjectColumnsFromActiveParts(parts_lock); } - /// This transaction may not happen, but it's OK, because on the next retry we will eventually create/update this node - /// TODO Maybe do in in one transaction for Replicated database? - zookeeper->createOrUpdate(fs::path(replica_path) / "metadata_version", std::to_string(current_metadata->getMetadataVersion()), zkutil::CreateMode::Persistent); - return true; } @@ -6796,7 +6800,7 @@ bool StorageReplicatedMergeTree::tryWaitForReplicaToProcessLogEntry( } else { - throw Exception(ErrorCodes::LOGICAL_ERROR, "Unexpected name of log node: {}", entry.znode_name); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: unexpected name of log node: {}", entry.znode_name); } /** Second - find the corresponding entry in the queue of the specified replica. @@ -7172,7 +7176,7 @@ void StorageReplicatedMergeTree::fetchPartition( } if (best_replica.empty()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot choose best replica."); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: cannot choose best replica."); LOG_INFO(log, "Found {} replicas, {} of them are active. Selected {} to fetch from.", replicas.size(), active_replicas.size(), best_replica); @@ -8953,7 +8957,7 @@ bool StorageReplicatedMergeTree::canUseAdaptiveGranularity() const } -std::map StorageReplicatedMergeTree::getAlterMutationCommandsForPart(const DataPartPtr & part) const +MutationCommands StorageReplicatedMergeTree::getAlterMutationCommandsForPart(const DataPartPtr & part) const { return queue.getAlterMutationCommandsForPart(part); } diff --git a/src/Storages/StorageReplicatedMergeTree.h b/src/Storages/StorageReplicatedMergeTree.h index 79d6d1dce3d..1c2cdb3ec07 100644 --- a/src/Storages/StorageReplicatedMergeTree.h +++ b/src/Storages/StorageReplicatedMergeTree.h @@ -938,7 +938,7 @@ private: void waitMutationToFinishOnReplicas( const Strings & replicas, const String & mutation_id) const; - std::map getAlterMutationCommandsForPart(const DataPartPtr & part) const override; + MutationCommands getAlterMutationCommandsForPart(const DataPartPtr & part) const override; void startBackgroundMovesIfNeeded() override; diff --git a/src/Storages/StorageS3.cpp b/src/Storages/StorageS3.cpp index e29fdd0d4a0..2d8ef3df1c8 100644 --- a/src/Storages/StorageS3.cpp +++ b/src/Storages/StorageS3.cpp @@ -1385,7 +1385,7 @@ const StorageS3::Configuration & StorageS3::getConfiguration() bool StorageS3::Configuration::update(const ContextPtr & context) { - auto s3_settings = context->getStorageS3Settings().getSettings(url.uri.toString()); + auto s3_settings = context->getStorageS3Settings().getSettings(url.uri.toString(), context->getUserName()); request_settings = s3_settings.request_settings; request_settings.updateFromSettings(context->getSettings()); diff --git a/src/Storages/StorageS3.h b/src/Storages/StorageS3.h index 73559ef8571..587145cd1a7 100644 --- a/src/Storages/StorageS3.h +++ b/src/Storages/StorageS3.h @@ -19,7 +19,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/src/Storages/StorageS3Settings.cpp b/src/Storages/StorageS3Settings.cpp index b0c1160429a..2a0d15a2bab 100644 --- a/src/Storages/StorageS3Settings.cpp +++ b/src/Storages/StorageS3Settings.cpp @@ -293,7 +293,7 @@ void StorageS3Settings::loadFromConfig(const String & config_elem, const Poco::U } } -S3Settings StorageS3Settings::getSettings(const String & endpoint) const +S3Settings StorageS3Settings::getSettings(const String & endpoint, const String & user) const { std::lock_guard lock(mutex); auto next_prefix_setting = s3_settings.upper_bound(endpoint); @@ -302,7 +302,8 @@ S3Settings StorageS3Settings::getSettings(const String & endpoint) const for (auto possible_prefix_setting = next_prefix_setting; possible_prefix_setting != s3_settings.begin();) { std::advance(possible_prefix_setting, -1); - if (boost::algorithm::starts_with(endpoint, possible_prefix_setting->first)) + const auto & [endpoint_prefix, settings] = *possible_prefix_setting; + if (boost::algorithm::starts_with(endpoint, endpoint_prefix) && settings.auth_settings.canBeUsedByUser(user)) return possible_prefix_setting->second; } diff --git a/src/Storages/StorageS3Settings.h b/src/Storages/StorageS3Settings.h index 0e152bb2d31..21b6264717e 100644 --- a/src/Storages/StorageS3Settings.h +++ b/src/Storages/StorageS3Settings.h @@ -112,7 +112,7 @@ class StorageS3Settings public: void loadFromConfig(const String & config_elem, const Poco::Util::AbstractConfiguration & config, const Settings & settings); - S3Settings getSettings(const String & endpoint) const; + S3Settings getSettings(const String & endpoint, const String & user) const; private: mutable std::mutex mutex; diff --git a/src/Storages/StorageURL.cpp b/src/Storages/StorageURL.cpp index 6f3599630d3..608e44c3cd0 100644 --- a/src/Storages/StorageURL.cpp +++ b/src/Storages/StorageURL.cpp @@ -5,7 +5,7 @@ #include #include -#include +#include #include #include #include diff --git a/src/Storages/StorageView.cpp b/src/Storages/StorageView.cpp index 5679effbcb2..181fd0ac61c 100644 --- a/src/Storages/StorageView.cpp +++ b/src/Storages/StorageView.cpp @@ -207,12 +207,12 @@ void StorageView::read( static ASTTableExpression * getFirstTableExpression(ASTSelectQuery & select_query) { if (!select_query.tables() || select_query.tables()->children.empty()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "No table expression in view select AST"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: no table expression in view select AST"); auto * select_element = select_query.tables()->children[0]->as(); if (!select_element->table_expression) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Incorrect table expression"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: incorrect table expression"); return select_element->table_expression->as(); } @@ -243,7 +243,7 @@ void StorageView::replaceWithSubquery(ASTSelectQuery & outer_query, ASTPtr view_ } if (!table_expression->database_and_table_name) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Incorrect table expression"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: incorrect table expression"); } DatabaseAndTableWithAlias db_table(table_expression->database_and_table_name); @@ -270,7 +270,7 @@ ASTPtr StorageView::restoreViewName(ASTSelectQuery & select_query, const ASTPtr ASTTableExpression * table_expression = getFirstTableExpression(select_query); if (!table_expression->subquery) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Incorrect table expression"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: incorrect table expression"); ASTPtr subquery = table_expression->subquery; table_expression->subquery = {}; diff --git a/src/Storages/System/StorageSystemBackups.cpp b/src/Storages/System/StorageSystemBackups.cpp index 17fb56e0a92..0063d9e308f 100644 --- a/src/Storages/System/StorageSystemBackups.cpp +++ b/src/Storages/System/StorageSystemBackups.cpp @@ -22,6 +22,7 @@ ColumnsDescription StorageSystemBackups::getColumnsDescription() {"id", std::make_shared(), "Operation ID, can be either passed via SETTINGS id=... or be randomly generated UUID."}, {"name", std::make_shared(), "Operation name, a string like `Disk('backups', 'my_backup')`"}, {"base_backup_name", std::make_shared(), "Base Backup Operation name, a string like `Disk('backups', 'my_base_backup')`"}, + {"query_id", std::make_shared(), "Query ID of a query that started backup."}, {"status", std::make_shared(getBackupStatusEnumValues()), "Status of backup or restore operation."}, {"error", std::make_shared(), "The error message if any."}, {"start_time", std::make_shared(), "The time when operation started."}, @@ -44,6 +45,7 @@ void StorageSystemBackups::fillData(MutableColumns & res_columns, ContextPtr con auto & column_id = assert_cast(*res_columns[column_index++]); auto & column_name = assert_cast(*res_columns[column_index++]); auto & column_base_backup_name = assert_cast(*res_columns[column_index++]); + auto & column_query_id = assert_cast(*res_columns[column_index++]); auto & column_status = assert_cast(*res_columns[column_index++]); auto & column_error = assert_cast(*res_columns[column_index++]); auto & column_start_time = assert_cast(*res_columns[column_index++]); @@ -62,6 +64,7 @@ void StorageSystemBackups::fillData(MutableColumns & res_columns, ContextPtr con column_id.insertData(info.id.data(), info.id.size()); column_name.insertData(info.name.data(), info.name.size()); column_base_backup_name.insertData(info.base_backup_name.data(), info.base_backup_name.size()); + column_query_id.insertData(info.query_id.data(), info.query_id.size()); column_status.insertValue(static_cast(info.status)); column_error.insertData(info.error_message.data(), info.error_message.size()); column_start_time.insertValue(static_cast(std::chrono::system_clock::to_time_t(info.start_time))); diff --git a/src/Storages/System/StorageSystemDetachedParts.cpp b/src/Storages/System/StorageSystemDetachedParts.cpp index a9cd5f2610a..3dae43976f7 100644 --- a/src/Storages/System/StorageSystemDetachedParts.cpp +++ b/src/Storages/System/StorageSystemDetachedParts.cpp @@ -11,7 +11,7 @@ #include #include #include -#include +#include #include diff --git a/src/Storages/System/StorageSystemStackTrace.cpp b/src/Storages/System/StorageSystemStackTrace.cpp index 90eb0ad89ec..82a5fd4e33f 100644 --- a/src/Storages/System/StorageSystemStackTrace.cpp +++ b/src/Storages/System/StorageSystemStackTrace.cpp @@ -168,7 +168,7 @@ bool wait(int timeout_ms) continue; /// Drain delayed notifications. } - throw Exception(ErrorCodes::LOGICAL_ERROR, "Read wrong number of bytes from pipe"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: read wrong number of bytes from pipe"); } } diff --git a/src/Storages/transformQueryForExternalDatabase.cpp b/src/Storages/transformQueryForExternalDatabase.cpp index afc458ea612..4526a38a1c3 100644 --- a/src/Storages/transformQueryForExternalDatabase.cpp +++ b/src/Storages/transformQueryForExternalDatabase.cpp @@ -145,7 +145,7 @@ bool isCompatible(ASTPtr & node) return false; if (!function->arguments) - throw Exception(ErrorCodes::LOGICAL_ERROR, "function->arguments is not set"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Logical error: function->arguments is not set"); String name = function->name; diff --git a/tests/analyzer_tech_debt.txt b/tests/analyzer_tech_debt.txt index 0566dca8f5c..180456fffe2 100644 --- a/tests/analyzer_tech_debt.txt +++ b/tests/analyzer_tech_debt.txt @@ -7,7 +7,6 @@ 01244_optimize_distributed_group_by_sharding_key 01584_distributed_buffer_cannot_find_column 01624_soft_constraints -01656_test_query_log_factories_info 01747_join_view_filter_dictionary 01761_cast_to_enum_nullable 01925_join_materialized_columns @@ -16,7 +15,6 @@ 02354_annoy 02493_inconsistent_hex_and_binary_number 02725_agg_projection_resprect_PK -02763_row_policy_storage_merge_alias # Check after constants refactoring 02901_parallel_replicas_rollup # Flaky. Please don't delete them without fixing them: diff --git a/tests/ci/ci.py b/tests/ci/ci.py index 819152fadc3..320a0ef42d5 100644 --- a/tests/ci/ci.py +++ b/tests/ci/ci.py @@ -6,12 +6,13 @@ from enum import Enum import json import logging import os +import random import re import subprocess import sys import time from pathlib import Path -from typing import Any, Dict, List, Optional, Sequence, Union +from typing import Any, Dict, List, Optional, Sequence, Set, Union import docker_images_helper import upload_result_helper @@ -1107,6 +1108,7 @@ def _configure_jobs( ci_cache.print_status() jobs_to_wait: Dict[str, Dict[str, Any]] = {} + randomization_buckets = {} # type: Dict[str, Set[str]] for job in digests: digest = digests[job] @@ -1115,11 +1117,18 @@ def _configure_jobs( batches_to_do: List[int] = [] add_to_skip = False + if job_config.pr_only and pr_info.is_release_branch(): + continue + if job_config.release_only and not pr_info.is_release_branch(): + continue + + # fill job randomization buckets (for jobs with configured @random_bucket property)) + if job_config.random_bucket: + if not job_config.random_bucket in randomization_buckets: + randomization_buckets[job_config.random_bucket] = set() + randomization_buckets[job_config.random_bucket].add(job) + for batch in range(num_batches): # type: ignore - if job_config.pr_only and pr_info.is_release_branch(): - continue - if job_config.release_only and not pr_info.is_release_branch(): - continue if job_config.run_by_label: # this job controlled by label, add to todo if its label is set in pr if job_config.run_by_label in pr_info.labels: @@ -1167,6 +1176,24 @@ def _configure_jobs( "num_batches": num_batches, } + if not pr_info.is_release_branch(): + # randomization bucket filtering (pick one random job from each bucket, for jobs with configured random_bucket property) + for _, jobs in randomization_buckets.items(): + jobs_to_remove_randomization = set() + bucket_ = list(jobs) + random.shuffle(bucket_) + while len(bucket_) > 1: + random_job = bucket_.pop() + if random_job in jobs_to_do: + jobs_to_remove_randomization.add(random_job) + if jobs_to_remove_randomization: + print( + f"Following jobs will be removed due to randomization bucket: [{jobs_to_remove_randomization}]" + ) + jobs_to_do = [ + job for job in jobs_to_do if job not in jobs_to_remove_randomization + ] + ## c. check CI controlling labels and commit messages if pr_info.labels: jobs_requested_by_label = [] # type: List[str] @@ -1642,13 +1669,7 @@ def main() -> int: if not args.skip_jobs: ci_cache = CiCache(s3, jobs_data["digests"]) - if ( - pr_info.is_release_branch() - or pr_info.event.get("pull_request", {}) - .get("user", {}) - .get("login", "not_maxknv") - == "maxknv" - ): + if pr_info.is_release_branch(): # wait for pending jobs to be finished, await_jobs is a long blocking call # wait pending jobs (for now only on release/master branches) ready_jobs_batches_dict = ci_cache.await_jobs( @@ -1838,7 +1859,7 @@ def main() -> int: pr_info.sha, job_report.test_results, job_report.additional_files, - job_report.check_name or args.job_name, + job_report.check_name or _get_ext_check_name(args.job_name), additional_urls=additional_urls or None, ) commit = get_commit( @@ -1849,7 +1870,7 @@ def main() -> int: job_report.status, check_url, format_description(job_report.description), - job_report.check_name or args.job_name, + job_report.check_name or _get_ext_check_name(args.job_name), pr_info, dump_to_file=True, ) @@ -1867,7 +1888,7 @@ def main() -> int: job_report.duration, job_report.start_time, check_url or "", - job_report.check_name or args.job_name, + job_report.check_name or _get_ext_check_name(args.job_name), ) ch_helper.insert_events_into( db="default", table="checks", events=prepared_events diff --git a/tests/ci/ci_config.py b/tests/ci/ci_config.py index 3ebcbb7ed59..7972a47ee58 100644 --- a/tests/ci/ci_config.py +++ b/tests/ci/ci_config.py @@ -198,6 +198,8 @@ class JobConfig: pr_only: bool = False # job is for release/master branches only release_only: bool = False + # to randomly pick and run one job among jobs in the same @random_bucket. Applied in PR branches only. + random_bucket: str = "" @dataclass @@ -637,16 +639,8 @@ CI_CONFIG = CIConfig( Labels.CI_SET_INTEGRATION: LabelConfig( run_jobs=[ JobNames.STYLE_CHECK, - Build.PACKAGE_ASAN, Build.PACKAGE_RELEASE, - Build.PACKAGE_TSAN, - Build.PACKAGE_AARCH64, - JobNames.INTEGRATION_TEST_ASAN, - JobNames.INTEGRATION_TEST_ARM, JobNames.INTEGRATION_TEST, - JobNames.INTEGRATION_TEST_ASAN_ANALYZER, - JobNames.INTEGRATION_TEST_TSAN, - JobNames.INTEGRATION_TEST_FLAKY, ] ), Labels.CI_SET_REDUCED: LabelConfig( @@ -932,16 +926,16 @@ CI_CONFIG = CIConfig( Build.PACKAGE_DEBUG, job_config=JobConfig(**stateful_test_common_params) # type: ignore ), JobNames.STATEFUL_TEST_PARALLEL_REPL_ASAN: TestConfig( - Build.PACKAGE_ASAN, job_config=JobConfig(**stateful_test_common_params) # type: ignore + Build.PACKAGE_ASAN, job_config=JobConfig(random_bucket="parrepl_with_sanitizer", **stateful_test_common_params) # type: ignore ), JobNames.STATEFUL_TEST_PARALLEL_REPL_MSAN: TestConfig( - Build.PACKAGE_MSAN, job_config=JobConfig(**stateful_test_common_params) # type: ignore + Build.PACKAGE_MSAN, job_config=JobConfig(random_bucket="parrepl_with_sanitizer", **stateful_test_common_params) # type: ignore ), JobNames.STATEFUL_TEST_PARALLEL_REPL_UBSAN: TestConfig( - Build.PACKAGE_UBSAN, job_config=JobConfig(**stateful_test_common_params) # type: ignore + Build.PACKAGE_UBSAN, job_config=JobConfig(random_bucket="parrepl_with_sanitizer", **stateful_test_common_params) # type: ignore ), JobNames.STATEFUL_TEST_PARALLEL_REPL_TSAN: TestConfig( - Build.PACKAGE_TSAN, job_config=JobConfig(**stateful_test_common_params) # type: ignore + Build.PACKAGE_TSAN, job_config=JobConfig(random_bucket="parrepl_with_sanitizer", **stateful_test_common_params) # type: ignore ), # End stateful tests for parallel replicas JobNames.STATELESS_TEST_ASAN: TestConfig( @@ -993,29 +987,29 @@ CI_CONFIG = CIConfig( Build.PACKAGE_TSAN, job_config=JobConfig(num_batches=5, **statless_test_common_params), # type: ignore ), - JobNames.STRESS_TEST_ASAN: TestConfig( - Build.PACKAGE_ASAN, job_config=JobConfig(**stress_test_common_params) # type: ignore + JobNames.STRESS_TEST_DEBUG: TestConfig( + Build.PACKAGE_DEBUG, job_config=JobConfig(**stress_test_common_params) # type: ignore ), JobNames.STRESS_TEST_TSAN: TestConfig( Build.PACKAGE_TSAN, job_config=JobConfig(**stress_test_common_params) # type: ignore ), + JobNames.STRESS_TEST_ASAN: TestConfig( + Build.PACKAGE_ASAN, job_config=JobConfig(random_bucket="stress_with_sanitizer", **stress_test_common_params) # type: ignore + ), JobNames.STRESS_TEST_UBSAN: TestConfig( - Build.PACKAGE_UBSAN, job_config=JobConfig(**stress_test_common_params) # type: ignore + Build.PACKAGE_UBSAN, job_config=JobConfig(random_bucket="stress_with_sanitizer", **stress_test_common_params) # type: ignore ), JobNames.STRESS_TEST_MSAN: TestConfig( - Build.PACKAGE_MSAN, job_config=JobConfig(**stress_test_common_params) # type: ignore - ), - JobNames.STRESS_TEST_DEBUG: TestConfig( - Build.PACKAGE_DEBUG, job_config=JobConfig(**stress_test_common_params) # type: ignore + Build.PACKAGE_MSAN, job_config=JobConfig(random_bucket="stress_with_sanitizer", **stress_test_common_params) # type: ignore ), JobNames.UPGRADE_TEST_ASAN: TestConfig( - Build.PACKAGE_ASAN, job_config=JobConfig(pr_only=True, **upgrade_test_common_params) # type: ignore + Build.PACKAGE_ASAN, job_config=JobConfig(pr_only=True, random_bucket="upgrade_with_sanitizer", **upgrade_test_common_params) # type: ignore ), JobNames.UPGRADE_TEST_TSAN: TestConfig( - Build.PACKAGE_TSAN, job_config=JobConfig(pr_only=True, **upgrade_test_common_params) # type: ignore + Build.PACKAGE_TSAN, job_config=JobConfig(pr_only=True, random_bucket="upgrade_with_sanitizer", **upgrade_test_common_params) # type: ignore ), JobNames.UPGRADE_TEST_MSAN: TestConfig( - Build.PACKAGE_MSAN, job_config=JobConfig(pr_only=True, **upgrade_test_common_params) # type: ignore + Build.PACKAGE_MSAN, job_config=JobConfig(pr_only=True, random_bucket="upgrade_with_sanitizer", **upgrade_test_common_params) # type: ignore ), JobNames.UPGRADE_TEST_DEBUG: TestConfig( Build.PACKAGE_DEBUG, job_config=JobConfig(pr_only=True, **upgrade_test_common_params) # type: ignore @@ -1137,6 +1131,7 @@ CI_CONFIG.validate() # checks required by Mergeable Check REQUIRED_CHECKS = [ "PR Check", + "A Sync", # Cloud sync JobNames.BUILD_CHECK, JobNames.BUILD_CHECK_SPECIAL, JobNames.DOCS_CHECK, diff --git a/tests/config/config.d/keeper_port.xml b/tests/config/config.d/keeper_port.xml index b724d5dd87e..2b04d843a3b 100644 --- a/tests/config/config.d/keeper_port.xml +++ b/tests/config/config.d/keeper_port.xml @@ -3,7 +3,7 @@ 9181 1 - 1 + 0 1 @@ -24,6 +24,9 @@ 0 1 + + 1073741824 + 524288000 diff --git a/tests/config/install.sh b/tests/config/install.sh index cfe810cda84..9873af2f6cd 100755 --- a/tests/config/install.sh +++ b/tests/config/install.sh @@ -134,6 +134,12 @@ fi value=$(($RANDOM % 2)) sed --follow-symlinks -i "s|[01]|$value|" $DEST_SERVER_PATH/config.d/keeper_port.xml +value=$((($RANDOM + 100) * 2048)) +sed --follow-symlinks -i "s|[[:digit:]]\+|$value|" $DEST_SERVER_PATH/config.d/keeper_port.xml + +value=$((($RANDOM + 100) * 2048)) +sed --follow-symlinks -i "s|[[:digit:]]\+|$value|" $DEST_SERVER_PATH/config.d/keeper_port.xml + if [[ -n "$USE_POLYMORPHIC_PARTS" ]] && [[ "$USE_POLYMORPHIC_PARTS" -eq 1 ]]; then ln -sf $SRC_PATH/config.d/polymorphic_parts.xml $DEST_SERVER_PATH/config.d/ fi diff --git a/tests/integration/helpers/cluster.py b/tests/integration/helpers/cluster.py index 1d96563251b..542f757ddd4 100644 --- a/tests/integration/helpers/cluster.py +++ b/tests/integration/helpers/cluster.py @@ -465,7 +465,7 @@ class ClickHouseCluster: self.base_cmd += ["--project-name", self.project_name] self.base_zookeeper_cmd = None - self.base_mysql_cmd = [] + self.base_mysql57_cmd = [] self.base_kafka_cmd = [] self.base_kerberized_kafka_cmd = [] self.base_kerberos_kdc_cmd = [] @@ -479,7 +479,7 @@ class ClickHouseCluster: self.with_zookeeper = False self.with_zookeeper_secure = False self.with_mysql_client = False - self.with_mysql = False + self.with_mysql57 = False self.with_mysql8 = False self.with_mysql_cluster = False self.with_postgres = False @@ -644,12 +644,19 @@ class ClickHouseCluster: self.mysql_client_host = "mysql_client" self.mysql_client_container = None - # available when with_mysql == True - self.mysql_host = "mysql57" - self.mysql_port = 3306 - self.mysql_ip = None - self.mysql_dir = p.abspath(p.join(self.instances_dir, "mysql")) - self.mysql_logs_dir = os.path.join(self.mysql_dir, "logs") + # available when with_mysql57 == True + self.mysql57_host = "mysql57" + self.mysql57_port = 3306 + self.mysql57_ip = None + self.mysql57_dir = p.abspath(p.join(self.instances_dir, "mysql")) + self.mysql57_logs_dir = os.path.join(self.mysql57_dir, "logs") + + # available when with_mysql8 == True + self.mysql8_host = "mysql80" + self.mysql8_port = 3306 + self.mysql8_ip = None + self.mysql8_dir = p.abspath(p.join(self.instances_dir, "mysql8")) + self.mysql8_logs_dir = os.path.join(self.mysql8_dir, "logs") # available when with_mysql_cluster == True self.mysql2_host = "mysql2" @@ -659,14 +666,7 @@ class ClickHouseCluster: self.mysql3_ip = None self.mysql4_ip = None self.mysql_cluster_dir = p.abspath(p.join(self.instances_dir, "mysql")) - self.mysql_cluster_logs_dir = os.path.join(self.mysql_dir, "logs") - - # available when with_mysql8 == True - self.mysql8_host = "mysql80" - self.mysql8_port = 3306 - self.mysql8_ip = None - self.mysql8_dir = p.abspath(p.join(self.instances_dir, "mysql8")) - self.mysql8_logs_dir = os.path.join(self.mysql8_dir, "logs") + self.mysql_cluster_logs_dir = os.path.join(self.mysql8_dir, "logs") # available when with_zookeper_secure == True self.zookeeper_secure_port = 2281 @@ -1045,17 +1045,17 @@ class ClickHouseCluster: return self.base_mysql_client_cmd - def setup_mysql_cmd(self, instance, env_variables, docker_compose_yml_dir): - self.with_mysql = True - env_variables["MYSQL_HOST"] = self.mysql_host - env_variables["MYSQL_PORT"] = str(self.mysql_port) + def setup_mysql57_cmd(self, instance, env_variables, docker_compose_yml_dir): + self.with_mysql57 = True + env_variables["MYSQL_HOST"] = self.mysql57_host + env_variables["MYSQL_PORT"] = str(self.mysql57_port) env_variables["MYSQL_ROOT_HOST"] = "%" - env_variables["MYSQL_LOGS"] = self.mysql_logs_dir + env_variables["MYSQL_LOGS"] = self.mysql57_logs_dir env_variables["MYSQL_LOGS_FS"] = "bind" self.base_cmd.extend( ["--file", p.join(docker_compose_yml_dir, "docker_compose_mysql.yml")] ) - self.base_mysql_cmd = [ + self.base_mysql57_cmd = [ "docker-compose", "--env-file", instance.env_file, @@ -1065,7 +1065,7 @@ class ClickHouseCluster: p.join(docker_compose_yml_dir, "docker_compose_mysql.yml"), ] - return self.base_mysql_cmd + return self.base_mysql57_cmd def setup_mysql8_cmd(self, instance, env_variables, docker_compose_yml_dir): self.with_mysql8 = True @@ -1091,7 +1091,7 @@ class ClickHouseCluster: def setup_mysql_cluster_cmd(self, instance, env_variables, docker_compose_yml_dir): self.with_mysql_cluster = True - env_variables["MYSQL_CLUSTER_PORT"] = str(self.mysql_port) + env_variables["MYSQL_CLUSTER_PORT"] = str(self.mysql8_port) env_variables["MYSQL_CLUSTER_ROOT_HOST"] = "%" env_variables["MYSQL_CLUSTER_LOGS"] = self.mysql_cluster_logs_dir env_variables["MYSQL_CLUSTER_LOGS_FS"] = "bind" @@ -1572,7 +1572,7 @@ class ClickHouseCluster: with_zookeeper=False, with_zookeeper_secure=False, with_mysql_client=False, - with_mysql=False, + with_mysql57=False, with_mysql8=False, with_mysql_cluster=False, with_kafka=False, @@ -1676,7 +1676,7 @@ class ClickHouseCluster: with_zookeeper=with_zookeeper, zookeeper_config_path=self.zookeeper_config_path, with_mysql_client=with_mysql_client, - with_mysql=with_mysql, + with_mysql57=with_mysql57, with_mysql8=with_mysql8, with_mysql_cluster=with_mysql_cluster, with_kafka=with_kafka, @@ -1767,9 +1767,9 @@ class ClickHouseCluster: ) ) - if with_mysql and not self.with_mysql: + if with_mysql57 and not self.with_mysql57: cmds.append( - self.setup_mysql_cmd(instance, env_variables, docker_compose_yml_dir) + self.setup_mysql57_cmd(instance, env_variables, docker_compose_yml_dir) ) if with_mysql8 and not self.with_mysql8: @@ -1805,9 +1805,9 @@ class ClickHouseCluster: if with_odbc_drivers and not self.with_odbc_drivers: self.with_odbc_drivers = True - if not self.with_mysql: + if not self.with_mysql8: cmds.append( - self.setup_mysql_cmd( + self.setup_mysql8_cmd( instance, env_variables, docker_compose_yml_dir ) ) @@ -2148,8 +2148,8 @@ class ClickHouseCluster: logging.error("Can't connect to MySQL Client:{}".format(errors)) raise Exception("Cannot wait MySQL Client container") - def wait_mysql_to_start(self, timeout=180): - self.mysql_ip = self.get_instance_ip("mysql57") + def wait_mysql57_to_start(self, timeout=180): + self.mysql57_ip = self.get_instance_ip("mysql57") start = time.time() errors = [] while time.time() - start < timeout: @@ -2157,8 +2157,8 @@ class ClickHouseCluster: conn = pymysql.connect( user=mysql_user, password=mysql_pass, - host=self.mysql_ip, - port=self.mysql_port, + host=self.mysql57_ip, + port=self.mysql57_port, ) conn.close() logging.debug("Mysql Started") @@ -2205,7 +2205,7 @@ class ClickHouseCluster: user=mysql_user, password=mysql_pass, host=ip, - port=self.mysql_port, + port=self.mysql8_port, ) conn.close() logging.debug(f"Mysql Started {ip}") @@ -2752,15 +2752,15 @@ class ClickHouseCluster: subprocess_check_call(self.base_mysql_client_cmd + common_opts) self.wait_mysql_client_to_start() - if self.with_mysql and self.base_mysql_cmd: + if self.with_mysql57 and self.base_mysql57_cmd: logging.debug("Setup MySQL") - if os.path.exists(self.mysql_dir): - shutil.rmtree(self.mysql_dir) - os.makedirs(self.mysql_logs_dir) - os.chmod(self.mysql_logs_dir, stat.S_IRWXU | stat.S_IRWXO) - subprocess_check_call(self.base_mysql_cmd + common_opts) + if os.path.exists(self.mysql57_dir): + shutil.rmtree(self.mysql57_dir) + os.makedirs(self.mysql57_logs_dir) + os.chmod(self.mysql57_logs_dir, stat.S_IRWXU | stat.S_IRWXO) + subprocess_check_call(self.base_mysql57_cmd + common_opts) self.up_called = True - self.wait_mysql_to_start() + self.wait_mysql57_to_start() if self.with_mysql8 and self.base_mysql8_cmd: logging.debug("Setup MySQL 8") @@ -2775,7 +2775,7 @@ class ClickHouseCluster: print("Setup MySQL") if os.path.exists(self.mysql_cluster_dir): shutil.rmtree(self.mysql_cluster_dir) - os.makedirs(self.mysql_cluster_logs_dir) + os.makedirs(self.mysql_cluster_logs_dir, exist_ok=True) os.chmod(self.mysql_cluster_logs_dir, stat.S_IRWXU | stat.S_IRWXO) subprocess_check_call(self.base_mysql_cluster_cmd + common_opts) @@ -3239,7 +3239,7 @@ class ClickHouseInstance: with_zookeeper, zookeeper_config_path, with_mysql_client, - with_mysql, + with_mysql57, with_mysql8, with_mysql_cluster, with_kafka, @@ -3324,7 +3324,7 @@ class ClickHouseInstance: self.library_bridge_bin_path = library_bridge_bin_path self.with_mysql_client = with_mysql_client - self.with_mysql = with_mysql + self.with_mysql57 = with_mysql57 self.with_mysql8 = with_mysql8 self.with_mysql_cluster = with_mysql_cluster self.with_postgres = with_postgres @@ -3368,7 +3368,7 @@ class ClickHouseInstance: self.env_file = self.cluster.env_file if with_odbc_drivers: self.odbc_ini_path = self.path + "/odbc.ini:/etc/odbc.ini" - self.with_mysql = True + self.with_mysql8 = True else: self.odbc_ini_path = "" @@ -4294,7 +4294,7 @@ class ClickHouseInstance: "Database": odbc_mysql_db, "Uid": odbc_mysql_uid, "Pwd": odbc_mysql_pass, - "Server": self.cluster.mysql_host, + "Server": self.cluster.mysql8_host, }, "PostgreSQL": { "DSN": "postgresql_odbc", @@ -4482,14 +4482,14 @@ class ClickHouseInstance: if self.with_mysql_client: depends_on.append(self.cluster.mysql_client_host) - if self.with_mysql: + if self.with_mysql57: depends_on.append("mysql57") if self.with_mysql8: depends_on.append("mysql80") if self.with_mysql_cluster: - depends_on.append("mysql57") + depends_on.append("mysql80") depends_on.append("mysql2") depends_on.append("mysql3") depends_on.append("mysql4") diff --git a/tests/integration/helpers/external_sources.py b/tests/integration/helpers/external_sources.py index cccf151e73e..033a2f84fa2 100644 --- a/tests/integration/helpers/external_sources.py +++ b/tests/integration/helpers/external_sources.py @@ -119,7 +119,7 @@ class SourceMySQL(ExternalSource): def prepare(self, structure, table_name, cluster): if self.internal_hostname is None: - self.internal_hostname = cluster.mysql_ip + self.internal_hostname = cluster.mysql8_ip self.create_mysql_conn() self.execute_mysql_query( "create database if not exists test default character set 'utf8'" diff --git a/tests/integration/test_backup_restore_s3/configs/s3_settings.xml b/tests/integration/test_backup_restore_s3/configs/s3_settings.xml index 981cf67bbe9..61ef7759b57 100644 --- a/tests/integration/test_backup_restore_s3/configs/s3_settings.xml +++ b/tests/integration/test_backup_restore_s3/configs/s3_settings.xml @@ -1,5 +1,6 @@ + 0 http://minio1:9001/root/data/backups/multipart/ + + + + + Float32 + Float64 + + + + + + CREATE TABLE vecs_{element_type} ( + v Array({element_type}) + ) ENGINE=Memory; + + + + + + + INSERT INTO vecs_{element_type} + SELECT v FROM ( + SELECT + number AS n, + [ + rand(n*10), rand(n*10+1), rand(n*10+2), rand(n*10+3), rand(n*10+4), rand(n*10+5), rand(n*10+6), rand(n*10+7), rand(n*10+8), rand(n*10+9), + rand(n*10+10), rand(n*10+11), rand(n*10+12), rand(n*10+13), rand(n*10+14), rand(n*10+15), rand(n*10+16), rand(n*10+17), rand(n*10+18), rand(n*10+19), + rand(n*10+20), rand(n*10+21), rand(n*10+22), rand(n*10+23), rand(n*10+24), rand(n*10+25), rand(n*10+26), rand(n*10+27), rand(n*10+28), rand(n*10+29), + rand(n*10+30), rand(n*10+31), rand(n*10+32), rand(n*10+33), rand(n*10+34), rand(n*10+35), rand(n*10+36), rand(n*10+37), rand(n*10+38), rand(n*10+39), + rand(n*10+40), rand(n*10+41), rand(n*10+42), rand(n*10+43), rand(n*10+44), rand(n*10+45), rand(n*10+46), rand(n*10+47), rand(n*10+48), rand(n*10+49), + rand(n*10+50), rand(n*10+51), rand(n*10+52), rand(n*10+53), rand(n*10+54), rand(n*10+55), rand(n*10+56), rand(n*10+57), rand(n*10+58), rand(n*10+59), + rand(n*10+60), rand(n*10+61), rand(n*10+62), rand(n*10+63), rand(n*10+64), rand(n*10+65), rand(n*10+66), rand(n*10+67), rand(n*10+68), rand(n*10+69), + rand(n*10+70), rand(n*10+71), rand(n*10+72), rand(n*10+73), rand(n*10+74), rand(n*10+75), rand(n*10+76), rand(n*10+77), rand(n*10+78), rand(n*10+79), + rand(n*10+80), rand(n*10+81), rand(n*10+82), rand(n*10+83), rand(n*10+84), rand(n*10+85), rand(n*10+86), rand(n*10+87), rand(n*10+88), rand(n*10+89), + rand(n*10+90), rand(n*10+91), rand(n*10+92), rand(n*10+93), rand(n*10+94), rand(n*10+95), rand(n*10+96), rand(n*10+97), rand(n*10+98), rand(n*10+99), + rand(n*10+100), rand(n*10+101), rand(n*10+102), rand(n*10+103), rand(n*10+104), rand(n*10+105), rand(n*10+106), rand(n*10+107), rand(n*10+108), rand(n*10+109), + rand(n*10+110), rand(n*10+111), rand(n*10+112), rand(n*10+113), rand(n*10+114), rand(n*10+115), rand(n*10+116), rand(n*10+117), rand(n*10+118), rand(n*10+119), + rand(n*10+120), rand(n*10+121), rand(n*10+122), rand(n*10+123), rand(n*10+124), rand(n*10+125), rand(n*10+126), rand(n*10+127), rand(n*10+128), rand(n*10+129), + rand(n*10+130), rand(n*10+131), rand(n*10+132), rand(n*10+133), rand(n*10+134), rand(n*10+135), rand(n*10+136), rand(n*10+137), rand(n*10+138), rand(n*10+139), + rand(n*10+140), rand(n*10+141), rand(n*10+142), rand(n*10+143), rand(n*10+144), rand(n*10+145), rand(n*10+146), rand(n*10+147), rand(n*10+148), rand(n*10+149) + ] AS v + FROM system.numbers + LIMIT 5000000 + ); + + + + 1 + + + SELECT sum(dp) FROM (SELECT dotProduct(v, v) AS dp FROM vecs_{element_type}) + + DROP TABLE vecs_{element_type} + + diff --git a/tests/performance/norm_distance.xml b/tests/performance/norm_distance.xml index 1e879607dac..69ed71d026f 100644 --- a/tests/performance/norm_distance.xml +++ b/tests/performance/norm_distance.xml @@ -4,11 +4,11 @@ element_type - + - Int32 - Int64 + + Float32 Float64 diff --git a/tests/performance/scripts/compare.sh b/tests/performance/scripts/compare.sh index 39c6854fbf9..9a0fb5b335c 100755 --- a/tests/performance/scripts/compare.sh +++ b/tests/performance/scripts/compare.sh @@ -11,8 +11,14 @@ script_dir="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )" # upstream/master LEFT_SERVER_PORT=9001 +LEFT_SERVER_KEEPER_PORT=9181 +LEFT_SERVER_KEEPER_RAFT_PORT=9234 +LEFT_SERVER_INTERSERVER_PORT=9009 # patched version -RIGHT_SERVER_PORT=9002 +RIGHT_SERVER_PORT=19001 +RIGHT_SERVER_KEEPER_PORT=19181 +RIGHT_SERVER_KEEPER_RAFT_PORT=19234 +RIGHT_SERVER_INTERSERVER_PORT=19009 # abort_conf -- abort if some options is not recognized # abort -- abort if something is not right in the env (i.e. per-cpu arenas does not work) @@ -127,6 +133,10 @@ function restart --user_files_path left/db/user_files --top_level_domains_path "$(left_or_right left top_level_domains)" --tcp_port $LEFT_SERVER_PORT + --keeper_server.tcp_port $LEFT_SERVER_KEEPER_PORT + --keeper_server.raft_configuration.server.port $LEFT_SERVER_KEEPER_RAFT_PORT + --zookeeper.node.port $LEFT_SERVER_KEEPER_PORT + --interserver_http_port $LEFT_SERVER_INTERSERVER_PORT ) left/clickhouse-server "${left_server_opts[@]}" &>> left-server-log.log & left_pid=$! @@ -142,6 +152,10 @@ function restart --user_files_path right/db/user_files --top_level_domains_path "$(left_or_right right top_level_domains)" --tcp_port $RIGHT_SERVER_PORT + --keeper_server.tcp_port $RIGHT_SERVER_KEEPER_PORT + --keeper_server.raft_configuration.server.port $RIGHT_SERVER_KEEPER_RAFT_PORT + --zookeeper.node.port $RIGHT_SERVER_KEEPER_PORT + --interserver_http_port $RIGHT_SERVER_INTERSERVER_PORT ) right/clickhouse-server "${right_server_opts[@]}" &>> right-server-log.log & right_pid=$! diff --git a/tests/performance/scripts/config/config.d/zzz-perf-comparison-tweaks-config.xml b/tests/performance/scripts/config/config.d/zzz-perf-comparison-tweaks-config.xml index 292665c4f68..c2bef2b479a 100644 --- a/tests/performance/scripts/config/config.d/zzz-perf-comparison-tweaks-config.xml +++ b/tests/performance/scripts/config/config.d/zzz-perf-comparison-tweaks-config.xml @@ -2,10 +2,7 @@ - - - :: diff --git a/tests/performance/sum.xml b/tests/performance/sum.xml index 57b879a360d..36b898436bf 100644 --- a/tests/performance/sum.xml +++ b/tests/performance/sum.xml @@ -17,6 +17,13 @@ SELECT sumKahan(toNullable(toFloat32(number))) FROM numbers(100000000) SELECT sumKahan(toNullable(toFloat64(number))) FROM numbers(100000000) + select sumIf(number::Decimal128(3), rand32() % 2 = 0) from numbers(100000000) + select sumIf(number::Decimal256(3), rand32() % 2 = 0) from numbers(100000000) + select sumIf(number::Int128, rand32() % 2 = 0) from numbers(100000000) + select sumIf(number::UInt128, rand32() % 2 = 0) from numbers(100000000) + select sumIf(number::Int256, rand32() % 2 = 0) from numbers(100000000) + select sumIf(number::UInt256, rand32() % 2 = 0) from numbers(100000000) + CREATE TABLE nullfloat32 (x Nullable(Float32)) ENGINE = Memory INSERT INTO nullfloat32 diff --git a/tests/queries/0_stateless/00700_decimal_arithm.reference b/tests/queries/0_stateless/00700_decimal_arithm.reference index 811946c87e0..109c0632fb1 100644 --- a/tests/queries/0_stateless/00700_decimal_arithm.reference +++ b/tests/queries/0_stateless/00700_decimal_arithm.reference @@ -10,18 +10,18 @@ 63 21 -42 882 -882 2 0 2 0 63 21 -42 882 -882 2 0 2 0 1.00305798474369219219752355409390731264 -0.16305798474369219219752355409390731264 1.490591730234615865843651857942052864 -1.38847100762815390390123822295304634368 1.38847100762815390390123822295304634368 0.02 0.005 -63.42 21.42 -41.58 890.82 -890.82 2.02 0.505 2.02 0.505 -63.42 21.42 -41.58 890.82 -890.82 2.02 0.505 2.02 0.505 -63.42 21.42 -41.58 890.82 -890.82 2.02 0.505 2.02 0.505 -63.42 21.42 -41.58 890.82 -890.82 2.02 0.5 2.02 0.5 +63.42 21.42 -41.58 890.82 -890.82 2.02 0.505 2 0 +63.42 21.42 -41.58 890.82 -890.82 2.02 0.505 2 0 +63.42 21.42 -41.58 890.82 -890.82 2.02 0.505 2 0 +63.42 21.42 -41.58 890.82 -890.82 2.02 0.5 2 0 63 -21 42 882 -882 0 2 0 2 63 -21 42 882 -882 0 2 0 2 63 -21 42 882 -882 0 2 0 2 1.00305798474369219219752355409390731264 0.16305798474369219219752355409390731264 -1.490591730234615865843651857942052864 -1.38847100762815390390123822295304634368 1.38847100762815390390123822295304634368 -0.00000000000000000000000000000000000001 0.00000000000000000000000000000000000001 -63.42 -21.42 41.58 890.82 -890.82 0.495 1.98 0.495 1.98 +63.42 -21.42 41.58 890.82 -890.82 0.495 1.98 0 1 63.42 -21.42 41.58 890.82 -890.82 -63.42 -21.42 41.58 890.82 -890.82 0.495049504950495049 1.980198019801980198 0.495049504950495049 1.980198019801980198 -63.42 -21.42 41.58 890.82 -890.82 0.49 1.98 0.49 1.98 +63.42 -21.42 41.58 890.82 -890.82 0.495049504950495049 1.980198019801980198 0 1 +63.42 -21.42 41.58 890.82 -890.82 0.49 1.98 0 1 -42 42 42 42 0.42 0.42 0.42 42.42 42.42 42.42 0 0 0 0 0 0 0 0 0 0 42 -42 -42 -42 -0.42 -0.42 -0.42 -42.42 -42.42 -42.42 diff --git a/tests/queries/0_stateless/00752_low_cardinality_lambda_argument.sql b/tests/queries/0_stateless/00752_low_cardinality_lambda_argument.sql index a4bdbd5653c..998ff2f54d3 100644 --- a/tests/queries/0_stateless/00752_low_cardinality_lambda_argument.sql +++ b/tests/queries/0_stateless/00752_low_cardinality_lambda_argument.sql @@ -1,3 +1,4 @@ +set allow_suspicious_low_cardinality_types=1; drop table if exists lc_lambda; create table lc_lambda (arr Array(LowCardinality(UInt64))) engine = Memory; insert into lc_lambda select range(number) from system.numbers limit 10; diff --git a/tests/queries/0_stateless/00752_low_cardinality_left_array_join.sql b/tests/queries/0_stateless/00752_low_cardinality_left_array_join.sql index 1c19700e34d..2d65f01a1b9 100644 --- a/tests/queries/0_stateless/00752_low_cardinality_left_array_join.sql +++ b/tests/queries/0_stateless/00752_low_cardinality_left_array_join.sql @@ -1,3 +1,4 @@ +set allow_suspicious_low_cardinality_types=1; drop table if exists lc_left_aj; CREATE TABLE lc_left_aj ( diff --git a/tests/queries/0_stateless/00873_t64_codec_date.reference b/tests/queries/0_stateless/00873_t64_codec_date.reference new file mode 100644 index 00000000000..9353696610c --- /dev/null +++ b/tests/queries/0_stateless/00873_t64_codec_date.reference @@ -0,0 +1,4 @@ +1970-01-01 1970-01-01 1950-01-01 1950-01-01 +1970-01-01 1970-01-01 1970-01-01 1970-01-01 +2149-06-06 2149-06-06 2149-06-06 2149-06-06 +2149-06-06 2149-06-06 2149-06-08 2149-06-08 diff --git a/tests/queries/0_stateless/00873_t64_codec_date.sql b/tests/queries/0_stateless/00873_t64_codec_date.sql new file mode 100644 index 00000000000..c6e21baba12 --- /dev/null +++ b/tests/queries/0_stateless/00873_t64_codec_date.sql @@ -0,0 +1,26 @@ +DROP TABLE IF EXISTS t64; + +CREATE TABLE t64 +( + date16 Date, + t_date16 Date Codec(T64, ZSTD), + date_32 Date32, + t_date32 Date32 Codec(T64, ZSTD) +) ENGINE MergeTree() ORDER BY tuple(); + +INSERT INTO t64 values ('1970-01-01', '1970-01-01', '1970-01-01', '1970-01-01'); +INSERT INTO t64 values ('2149-06-06', '2149-06-06', '2149-06-06', '2149-06-06'); +INSERT INTO t64 values ('2149-06-08', '2149-06-08', '2149-06-08', '2149-06-08'); +INSERT INTO t64 values ('1950-01-01', '1950-01-01', '1950-01-01', '1950-01-01'); + +SELECT * FROM t64 ORDER BY date_32; + +SELECT * FROM t64 WHERE date16 != t_date16; +SELECT * FROM t64 WHERE date_32 != t_date32; + +OPTIMIZE TABLE t64 FINAL; + +SELECT * FROM t64 WHERE date16 != t_date16; +SELECT * FROM t64 WHERE date_32 != t_date32; + +DROP TABLE t64; diff --git a/tests/queries/0_stateless/00945_bloom_filter_index.sql b/tests/queries/0_stateless/00945_bloom_filter_index.sql index dc47e858c4d..faa7feda04d 100644 --- a/tests/queries/0_stateless/00945_bloom_filter_index.sql +++ b/tests/queries/0_stateless/00945_bloom_filter_index.sql @@ -1,3 +1,4 @@ +SET allow_suspicious_low_cardinality_types=1; DROP TABLE IF EXISTS single_column_bloom_filter; diff --git a/tests/queries/0_stateless/01414_low_cardinality_nullable.sql b/tests/queries/0_stateless/01414_low_cardinality_nullable.sql index 2d3d31e9b5c..cd5111faf45 100644 --- a/tests/queries/0_stateless/01414_low_cardinality_nullable.sql +++ b/tests/queries/0_stateless/01414_low_cardinality_nullable.sql @@ -1,3 +1,5 @@ +SET allow_suspicious_low_cardinality_types=1; + DROP TABLE IF EXISTS lc_nullable; CREATE TABLE lc_nullable ( diff --git a/tests/queries/0_stateless/01441_low_cardinality_array_index.sql b/tests/queries/0_stateless/01441_low_cardinality_array_index.sql index 4b31a86edfb..b5e14c957c6 100644 --- a/tests/queries/0_stateless/01441_low_cardinality_array_index.sql +++ b/tests/queries/0_stateless/01441_low_cardinality_array_index.sql @@ -1,3 +1,5 @@ +SET allow_suspicious_low_cardinality_types=1; + DROP TABLE IF EXISTS t_01411; CREATE TABLE t_01411( diff --git a/tests/queries/0_stateless/01651_lc_insert_tiny_log.sql b/tests/queries/0_stateless/01651_lc_insert_tiny_log.sql index 22532529812..d405bb01fd9 100644 --- a/tests/queries/0_stateless/01651_lc_insert_tiny_log.sql +++ b/tests/queries/0_stateless/01651_lc_insert_tiny_log.sql @@ -1,3 +1,4 @@ +set allow_suspicious_low_cardinality_types=1; drop table if exists perf_lc_num; CREATE TABLE perf_lc_num(　 num UInt8,　 arr Array(LowCardinality(Int64)) default [num]　 ) ENGINE = TinyLog; diff --git a/tests/queries/0_stateless/01656_test_query_log_factories_info.sql b/tests/queries/0_stateless/01656_test_query_log_factories_info.sql index 020d7cc5e72..8a6b604b053 100644 --- a/tests/queries/0_stateless/01656_test_query_log_factories_info.sql +++ b/tests/queries/0_stateless/01656_test_query_log_factories_info.sql @@ -41,7 +41,9 @@ FROM system.query_log WHERE current_database = currentDatabase() AND type = 'Que ORDER BY query_start_time DESC LIMIT 1 FORMAT TabSeparatedWithNames; SELECT ''; -SELECT arraySort(used_functions) +-- 1. analyzer includes arrayJoin into functions list +-- 2. for crc32 (CaseInsensitive function) we use lower case now +SELECT arraySort(arrayMap(x -> x == 'crc32' ? 'CRC32' : x, arrayFilter(x-> x != 'arrayJoin', used_functions))) as `arraySort(used_functions)` FROM system.query_log WHERE current_database = currentDatabase() AND type = 'QueryFinish' AND (query LIKE '%toDate(\'2000-12-05\')%') ORDER BY query_start_time DESC LIMIT 1 FORMAT TabSeparatedWithNames; SELECT ''; diff --git a/tests/queries/0_stateless/01717_int_div_float_too_large_ubsan.sql b/tests/queries/0_stateless/01717_int_div_float_too_large_ubsan.sql index c4f26a079f0..dc1e5b37050 100644 --- a/tests/queries/0_stateless/01717_int_div_float_too_large_ubsan.sql +++ b/tests/queries/0_stateless/01717_int_div_float_too_large_ubsan.sql @@ -1,2 +1,2 @@ -SELECT intDiv(9223372036854775807, 0.9998999834060669); -- { serverError 153 } -SELECT intDiv(9223372036854775807, 1.); -- { serverError 153 } +SELECT intDiv(18446744073709551615, 0.9998999834060669); -- { serverError 153 } +SELECT intDiv(18446744073709551615, 1.); -- { serverError 153 } diff --git a/tests/queries/0_stateless/02008_materialize_column.sql b/tests/queries/0_stateless/02008_materialize_column.sql index a78920d2525..cc7d3096402 100644 --- a/tests/queries/0_stateless/02008_materialize_column.sql +++ b/tests/queries/0_stateless/02008_materialize_column.sql @@ -17,6 +17,7 @@ ALTER TABLE tmp MATERIALIZE COLUMN s; ALTER TABLE tmp MODIFY COLUMN s String DEFAULT toString(x+2); SELECT arraySort(groupArray(x)), groupArray(s) FROM tmp; +ALTER TABLE tmp CLEAR COLUMN s; -- Need to clear because MATERIALIZE COLUMN won't override past values; ALTER TABLE tmp MATERIALIZE COLUMN s; ALTER TABLE tmp MODIFY COLUMN s String DEFAULT toString(x+3); SELECT arraySort(groupArray(x)), groupArray(s) FROM tmp; diff --git a/tests/queries/0_stateless/02010_array_index_bad_cast.sql b/tests/queries/0_stateless/02010_array_index_bad_cast.sql index 19c58bb28a7..42a6556fc77 100644 --- a/tests/queries/0_stateless/02010_array_index_bad_cast.sql +++ b/tests/queries/0_stateless/02010_array_index_bad_cast.sql @@ -1,2 +1,3 @@ -- This query throws exception about uncomparable data types (but at least it does not introduce bad cast in code). +SET allow_suspicious_low_cardinality_types=1; SELECT has(materialize(CAST(['2021-07-14'] AS Array(LowCardinality(Nullable(DateTime))))), materialize('2021-07-14'::DateTime64(7))); -- { serverError 44 } diff --git a/tests/queries/0_stateless/02131_row_policies_combination.reference b/tests/queries/0_stateless/02131_row_policies_combination.reference index b76028d5077..5015cb14456 100644 --- a/tests/queries/0_stateless/02131_row_policies_combination.reference +++ b/tests/queries/0_stateless/02131_row_policies_combination.reference @@ -12,6 +12,15 @@ R1, R2, R3: (x == 1) OR (x == 2) OR (x == 3) 1 2 3 +R1, R2, R3 + additional_table_filters and PREWHERE: (x == 1) OR (x == 2) OR (x == 3) AND (x < 3) AND (x > 1) +2 +3 +R1, R2, R3 + additional_result_filter and PREWHERE: (x == 1) OR (x == 2) OR (x == 3) AND (x < 3) AND (x > 1) +2 +3 +R1, R2, R3 + additional_table_filters and WHERE: (x == 1) OR (x == 2) OR (x == 3) AND (x < 3) AND (x > 1) +2 +3 R1, R2, R3, R4: ((x == 1) OR (x == 2) OR (x == 3)) AND (x <= 2) 1 2 diff --git a/tests/queries/0_stateless/02131_row_policies_combination.sql b/tests/queries/0_stateless/02131_row_policies_combination.sql index b5be672bb1b..02f2365eed8 100644 --- a/tests/queries/0_stateless/02131_row_policies_combination.sql +++ b/tests/queries/0_stateless/02131_row_policies_combination.sql @@ -23,6 +23,24 @@ CREATE ROW POLICY 02131_filter_3 ON 02131_rptable USING x=3 AS permissive TO ALL SELECT 'R1, R2, R3: (x == 1) OR (x == 2) OR (x == 3)'; SELECT * FROM 02131_rptable; +SELECT 'R1, R2, R3 + additional_table_filters and PREWHERE: (x == 1) OR (x == 2) OR (x == 3) AND (x < 3) AND (x > 1)'; +SELECT * FROM 02131_rptable +PREWHERE x >= 2 +SETTINGS additional_table_filters = {'02131_rptable': 'x > 1'} +; + +SELECT 'R1, R2, R3 + additional_result_filter and PREWHERE: (x == 1) OR (x == 2) OR (x == 3) AND (x < 3) AND (x > 1)'; +SELECT * FROM 02131_rptable +PREWHERE x >= 2 +SETTINGS additional_result_filter = 'x > 1' +; + +SELECT 'R1, R2, R3 + additional_table_filters and WHERE: (x == 1) OR (x == 2) OR (x == 3) AND (x < 3) AND (x > 1)'; +SELECT * FROM 02131_rptable +WHERE x >= 2 +SETTINGS additional_table_filters = {'02131_rptable': 'x > 1'} +; + CREATE ROW POLICY 02131_filter_4 ON 02131_rptable USING x<=2 AS restrictive TO ALL; SELECT 'R1, R2, R3, R4: ((x == 1) OR (x == 2) OR (x == 3)) AND (x <= 2)'; SELECT * FROM 02131_rptable; diff --git a/tests/queries/0_stateless/02184_nested_tuple.sql b/tests/queries/0_stateless/02184_nested_tuple.sql index 67a20e3dce1..09ed8eb7200 100644 --- a/tests/queries/0_stateless/02184_nested_tuple.sql +++ b/tests/queries/0_stateless/02184_nested_tuple.sql @@ -1,3 +1,4 @@ +SET allow_suspicious_low_cardinality_types=1; DROP TABLE IF EXISTS t_nested_tuple; CREATE TABLE t_nested_tuple diff --git a/tests/queries/0_stateless/02235_remote_fs_cache_stress.sh b/tests/queries/0_stateless/02235_remote_fs_cache_stress.sh index bc1a4cbfdd1..0b6b9f461b0 100755 --- a/tests/queries/0_stateless/02235_remote_fs_cache_stress.sh +++ b/tests/queries/0_stateless/02235_remote_fs_cache_stress.sh @@ -6,7 +6,7 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) . "$CUR_DIR"/../shell_config.sh -${CLICKHOUSE_CLIENT} --multiquery --multiline --query=""" +${CLICKHOUSE_CLIENT} --allow_suspicious_low_cardinality_types=1 --multiquery --multiline --query=""" DROP TABLE IF EXISTS t_01411; DROP TABLE IF EXISTS t_01411_num; diff --git a/tests/queries/0_stateless/02327_try_infer_integers_schema_inference.sql b/tests/queries/0_stateless/02327_try_infer_integers_schema_inference.sql index 0ceed178865..a4a69f4fa40 100644 --- a/tests/queries/0_stateless/02327_try_infer_integers_schema_inference.sql +++ b/tests/queries/0_stateless/02327_try_infer_integers_schema_inference.sql @@ -1,6 +1,7 @@ -- Tags: no-fasttest set input_format_try_infer_integers=1; +set input_format_try_infer_exponent_floats=1; select 'JSONEachRow'; desc format(JSONEachRow, '{"x" : 123}'); diff --git a/tests/queries/0_stateless/02500_numbers_inference.sh b/tests/queries/0_stateless/02500_numbers_inference.sh index ce9cd5bdc9f..5d863bd616f 100755 --- a/tests/queries/0_stateless/02500_numbers_inference.sh +++ b/tests/queries/0_stateless/02500_numbers_inference.sh @@ -8,10 +8,10 @@ $CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : 1.2}')"; echo '{"x" : 1.2}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test"; $CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : 1}')"; echo '{"x" : 1}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test"; -$CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : 1e10}')"; -echo '{"x" : 1e10}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test"; -$CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : [1, 42.42, 1, 1e10]}')"; -echo '{"x" : [1, 42.42, 1, 1e10]}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test"; +$CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : 1e10}')" --input_format_try_infer_exponent_floats=1; +echo '{"x" : 1e10}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test" --input_format_try_infer_exponent_floats=1; +$CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : [1, 42.42, 1, 1e10]}')" --input_format_try_infer_exponent_floats=1; +echo '{"x" : [1, 42.42, 1, 1e10]}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test" --input_format_try_infer_exponent_floats=1; $CLICKHOUSE_LOCAL -q "desc format(JSONEachRow, '{\"x\" : [1, 42.42, false]}')"; echo '{"x" : [1, 42.42, false]}' | $CLICKHOUSE_LOCAL --input-format='JSONEachRow' --table='test' -q "desc test"; @@ -19,10 +19,10 @@ $CLICKHOUSE_LOCAL -q "desc format(TSV, '1.2')"; echo '1.2' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test"; $CLICKHOUSE_LOCAL -q "desc format(TSV, '1')"; echo '1' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test"; -$CLICKHOUSE_LOCAL -q "desc format(TSV, '1e10')"; -echo '1e10' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test"; -$CLICKHOUSE_LOCAL -q "desc format(TSV, '[1, 42.42, 1, 1e10]')"; -echo '[1, 42.42, 1, 1e10]' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test"; +$CLICKHOUSE_LOCAL -q "desc format(TSV, '1e10')" --input_format_try_infer_exponent_floats=1; +echo '1e10' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test" --input_format_try_infer_exponent_floats=1; +$CLICKHOUSE_LOCAL -q "desc format(TSV, '[1, 42.42, 1, 1e10]')" --input_format_try_infer_exponent_floats=1; +echo '[1, 42.42, 1, 1e10]' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test" --input_format_try_infer_exponent_floats=1; $CLICKHOUSE_LOCAL -q "desc format(TSV, '[1, 42.42, false]')"; echo '[1, 42.42, false]' | $CLICKHOUSE_LOCAL --input-format='TSV' --table='test' -q "desc test"; diff --git a/tests/queries/0_stateless/02567_and_consistency.sql b/tests/queries/0_stateless/02567_and_consistency.sql index 0eeab99e539..b1fa526e33f 100644 --- a/tests/queries/0_stateless/02567_and_consistency.sql +++ b/tests/queries/0_stateless/02567_and_consistency.sql @@ -5,7 +5,7 @@ FROM ) GROUP BY number HAVING 1 AND sin(sum(number)) -ORDER BY ALL +ORDER BY * SETTINGS enable_optimize_predicate_expression = 0; SELECT '====='; @@ -17,7 +17,7 @@ FROM ) GROUP BY number HAVING 1 AND sin(1) -ORDER BY ALL +ORDER BY * SETTINGS enable_optimize_predicate_expression = 0; SELECT '====='; @@ -29,7 +29,7 @@ FROM ) GROUP BY number HAVING x AND sin(sum(number)) -ORDER BY ALL +ORDER BY * SETTINGS enable_optimize_predicate_expression = 1; SELECT '====='; @@ -41,7 +41,7 @@ FROM ) GROUP BY number HAVING 1 AND sin(sum(number)) -ORDER BY ALL +ORDER BY * SETTINGS enable_optimize_predicate_expression = 0; SELECT '====='; @@ -61,7 +61,7 @@ FROM ) GROUP BY number HAVING 1 AND sin(sum(number)) -ORDER BY ALL +ORDER BY * SETTINGS enable_optimize_predicate_expression = 1; select '#45440'; diff --git a/tests/queries/0_stateless/02708_dotProduct.reference b/tests/queries/0_stateless/02708_dotProduct.reference new file mode 100644 index 00000000000..5cc9a9f0502 --- /dev/null +++ b/tests/queries/0_stateless/02708_dotProduct.reference @@ -0,0 +1,34 @@ +-- Negative tests +-- Tests + -- Array +[1,2,3] [4,5,6] 32 UInt16 +[1,2,3] [4,5,6] 32 UInt32 +[1,2,3] [4,5,6] 32 UInt64 +[1,2,3] [4,5,6] 32 UInt64 +[-1,-2,-3] [4,5,6] -32 Int16 +[-1,-2,-3] [4,5,6] -32 Int32 +[-1,-2,-3] [4,5,6] -32 Int64 +[-1,-2,-3] [4,5,6] -32 Int64 +[1,2,3] [4,5,6] 32 Float32 +[1,2,3] [4,5,6] 32 Float64 + -- Tuple +(1,2,3) (4,5,6) 32 UInt64 +(1,2,3) (4,5,6) 32 UInt64 +(1,2,3) (4,5,6) 32 UInt64 +(1,2,3) (4,5,6) 32 UInt64 +(-1,-2,-3) (4,5,6) -32 Int64 +(-1,-2,-3) (4,5,6) -32 Int64 +(-1,-2,-3) (4,5,6) -32 Int64 +(-1,-2,-3) (4,5,6) -32 Int64 +(1,2,3) (4,5,6) 32 Float64 +(1,2,3) (4,5,6) 32 Float64 +-- Non-const argument +[1,2,3] [4,5,6] 32 UInt16 + -- Array with mixed element arguments types (result type is the supertype) +[1,2,3] [4,5,6] 32 Float32 + -- Tuple with mixed element arguments types +(1,2,3) (4,5,6) 32 Float64 +-- Aliases +32 +32 +32 diff --git a/tests/queries/0_stateless/02708_dotProduct.sql b/tests/queries/0_stateless/02708_dotProduct.sql new file mode 100644 index 00000000000..6ad615664e8 --- /dev/null +++ b/tests/queries/0_stateless/02708_dotProduct.sql @@ -0,0 +1,47 @@ +SELECT '-- Negative tests'; + +SELECT arrayDotProduct([1, 2]); -- { serverError NUMBER_OF_ARGUMENTS_DOESNT_MATCH } +SELECT arrayDotProduct([1, 2], 'abc'); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT arrayDotProduct('abc', [1, 2]); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT arrayDotProduct([1, 2], ['abc', 'def']); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT arrayDotProduct([1, 2], [3, 4, 5]); -- { serverError BAD_ARGUMENTS } +SELECT dotProduct([1, 2], (3, 4, 5)); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } + +SELECT '-- Tests'; +SELECT ' -- Array'; +SELECT [1, 2, 3]::Array(UInt8) AS x, [4, 5, 6]::Array(UInt8) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [1, 2, 3]::Array(UInt16) AS x, [4, 5, 6]::Array(UInt16) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [1, 2, 3]::Array(UInt32) AS x, [4, 5, 6]::Array(UInt32) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [1, 2, 3]::Array(UInt64) AS x, [4, 5, 6]::Array(UInt64) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [-1, -2, -3]::Array(Int8) AS x, [4, 5, 6]::Array(Int8) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [-1, -2, -3]::Array(Int16) AS x, [4, 5, 6]::Array(Int16) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [-1, -2, -3]::Array(Int32) AS x, [4, 5, 6]::Array(Int32) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [-1, -2, -3]::Array(Int64) AS x, [4, 5, 6]::Array(Int64) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [1, 2, 3]::Array(Float32) AS x, [4, 5, 6]::Array(Float32) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT [1, 2, 3]::Array(Float64) AS x, [4, 5, 6]::Array(Float64) AS y, dotProduct(x, y) AS res, toTypeName(res); + +SELECT ' -- Tuple'; +SELECT (1::UInt8, 2::UInt8, 3::UInt8) AS x, (4::UInt8, 5::UInt8, 6::UInt8) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (1::UInt16, 2::UInt16, 3::UInt16) AS x, (4::UInt16, 5::UInt16, 6::UInt16) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (1::UInt32, 2::UInt32, 3::UInt32) AS x, (4::UInt32, 5::UInt32, 6::UInt32) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (1::UInt64, 2::UInt64, 3::UInt64) AS x, (4::UInt64, 5::UInt64, 6::UInt64) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (-1::Int8, -2::Int8, -3::Int8) AS x, (4::Int8, 5::Int8, 6::Int8) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (-1::Int16, -2::Int16, -3::Int16) AS x, (4::Int16, 5::Int16, 6::Int16) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (-1::Int32, -2::Int32, -3::Int32) AS x, (4::Int32, 5::Int32, 6::Int32) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (-1::Int64, -2::Int64, -3::Int64) AS x, (4::Int64, 5::Int64, 6::Int64) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (1::Float32, 2::Float32, 3::Float32) AS x, (4::Float32, 5::Float32, 6::Float32) AS y, dotProduct(x, y) AS res, toTypeName(res); +SELECT (1::Float64, 2::Float64, 3::Float64) AS x, (4::Float64, 5::Float64, 6::Float64) AS y, dotProduct(x, y) AS res, toTypeName(res); + +SELECT '-- Non-const argument'; +SELECT materialize([1::UInt8, 2::UInt8, 3::UInt8]) AS x, [4::UInt8, 5::UInt8, 6::UInt8] AS y, dotProduct(x, y) AS res, toTypeName(res); + +SELECT ' -- Array with mixed element arguments types (result type is the supertype)'; +SELECT [1::UInt16, 2::UInt8, 3::Float32] AS x, [4::Int16, 5::Float32, 6::UInt8] AS y, dotProduct(x, y) AS res, toTypeName(res); + +SELECT ' -- Tuple with mixed element arguments types'; +SELECT (1::UInt16, 2::UInt8, 3::Float32) AS x, (4::Int16, 5::Float32, 6::UInt8) AS y, dotProduct(x, y) AS res, toTypeName(res); + +SELECT '-- Aliases'; +SELECT scalarProduct([1, 2, 3], [4, 5, 6]); +SELECT scalarProduct((1, 2, 3), (4, 5, 6)); +SELECT arrayDotProduct([1, 2, 3], [4, 5, 6]); -- actually no alias but the internal function for arrays diff --git a/tests/queries/0_stateless/02708_dot_product.reference b/tests/queries/0_stateless/02708_dot_product.reference deleted file mode 100644 index 45e53871aa2..00000000000 --- a/tests/queries/0_stateless/02708_dot_product.reference +++ /dev/null @@ -1,14 +0,0 @@ -3881.304 -3881.304 -3881.304 -376.5 -230 -0 -0 -Float64 -Float32 -Float64 -Float64 -UInt16 -UInt64 -Int64 diff --git a/tests/queries/0_stateless/02708_dot_product.sql b/tests/queries/0_stateless/02708_dot_product.sql deleted file mode 100644 index e94cb577bf4..00000000000 --- a/tests/queries/0_stateless/02708_dot_product.sql +++ /dev/null @@ -1,55 +0,0 @@ -SELECT dotProduct([12, 2.22, 302], [1.32, 231.2, 11.1]); - -SELECT scalarProduct([12, 2.22, 302], [1.32, 231.2, 11.1]); - -SELECT arrayDotProduct([12, 2.22, 302], [1.32, 231.2, 11.1]); - -SELECT dotProduct([1.3, 2, 3, 4, 5], [222, 12, 5.3, 2, 8]); - -SELECT dotProduct([1, 1, 1, 1, 1], [222, 12, 0, -12, 8]); - -SELECT round(dotProduct([12345678901234567], [1]) - dotProduct(tuple(12345678901234567), tuple(1)), 2); - -SELECT round(dotProduct([-1, 2, 3.002], [2, 3.4, 4]) - dotProduct((-1, 2, 3.002), (2, 3.4, 4)), 2); - -DROP TABLE IF EXISTS product_fp64_fp64; -CREATE TABLE product_fp64_fp64 (x Array(Float64), y Array(Float64)) engine = MergeTree() order by x; -INSERT INTO TABLE product_fp64_fp64 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_fp64_fp64; -DROP TABLE product_fp64_fp64; - -DROP TABLE IF EXISTS product_fp32_fp32; -CREATE TABLE product_fp32_fp32 (x Array(Float32), y Array(Float32)) engine = MergeTree() order by x; -INSERT INTO TABLE product_fp32_fp32 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_fp32_fp32; -DROP TABLE product_fp32_fp32; - -DROP TABLE IF EXISTS product_fp32_fp64; -CREATE TABLE product_fp32_fp64 (x Array(Float32), y Array(Float64)) engine = MergeTree() order by x; -INSERT INTO TABLE product_fp32_fp64 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_fp32_fp64; -DROP TABLE product_fp32_fp64; - -DROP TABLE IF EXISTS product_uint8_fp64; -CREATE TABLE product_uint8_fp64 (x Array(UInt8), y Array(Float64)) engine = MergeTree() order by x; -INSERT INTO TABLE product_uint8_fp64 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_uint8_fp64; -DROP TABLE product_uint8_fp64; - -DROP TABLE IF EXISTS product_uint8_uint8; -CREATE TABLE product_uint8_uint8 (x Array(UInt8), y Array(UInt8)) engine = MergeTree() order by x; -INSERT INTO TABLE product_uint8_uint8 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_uint8_uint8; -DROP TABLE product_uint8_uint8; - -DROP TABLE IF EXISTS product_uint64_uint64; -CREATE TABLE product_uint64_uint64 (x Array(UInt64), y Array(UInt64)) engine = MergeTree() order by x; -INSERT INTO TABLE product_uint64_uint64 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_uint64_uint64; -DROP TABLE product_uint64_uint64; - -DROP TABLE IF EXISTS product_int32_uint64; -CREATE TABLE product_int32_uint64 (x Array(Int32), y Array(UInt64)) engine = MergeTree() order by x; -INSERT INTO TABLE product_int32_uint64 (x, y) values ([1, 2], [3, 4]); -SELECT toTypeName(dotProduct(x, y)) from product_int32_uint64; -DROP TABLE product_int32_uint64; diff --git a/tests/queries/0_stateless/02732_transform_fuzz.sql b/tests/queries/0_stateless/02732_transform_fuzz.sql index c2918d4da81..872cf3a6599 100644 --- a/tests/queries/0_stateless/02732_transform_fuzz.sql +++ b/tests/queries/0_stateless/02732_transform_fuzz.sql @@ -1 +1 @@ -SELECT caseWithExpr(arrayReduce(NULL, []), []); -- { serverError BAD_ARGUMENTS } +SELECT caseWithExpr(arrayReduce(NULL, []), []); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } diff --git a/tests/queries/0_stateless/02735_parquet_encoder.sql b/tests/queries/0_stateless/02735_parquet_encoder.sql index 19125abf8da..fe45a2a317d 100644 --- a/tests/queries/0_stateless/02735_parquet_encoder.sql +++ b/tests/queries/0_stateless/02735_parquet_encoder.sql @@ -6,6 +6,7 @@ set output_format_parquet_data_page_size = 800; set output_format_parquet_batch_size = 100; set output_format_parquet_row_group_size_bytes = 1000000000; set engine_file_truncate_on_insert=1; +set allow_suspicious_low_cardinality_types=1; -- Write random data to parquet file, then read from it and check that it matches what we wrote. -- Do this for all kinds of data types: primitive, Nullable(primitive), Array(primitive), diff --git a/tests/queries/0_stateless/02763_row_policy_storage_merge_alias.sql.j2 b/tests/queries/0_stateless/02763_row_policy_storage_merge_alias.sql.j2 index bdd456951dd..99ac89c4eb4 100644 --- a/tests/queries/0_stateless/02763_row_policy_storage_merge_alias.sql.j2 +++ b/tests/queries/0_stateless/02763_row_policy_storage_merge_alias.sql.j2 @@ -12,8 +12,6 @@ CREATE TABLE 02763_a_merge (x UInt8, y UInt64, z UInt64) ENGINE = Merge(currentD {% for prew in [0 , 1] -%} - - SELECT 'x, y, z FROM 02763_a_merge'; SELECT x, y, z FROM 02763_a_merge ORDER BY x SETTINGS optimize_move_to_prewhere= {{prew}}; SELECT '* FROM 02763_a_merge'; diff --git a/tests/queries/0_stateless/02797_join_nested_lowcardinality_convert.sql.j2 b/tests/queries/0_stateless/02797_join_nested_lowcardinality_convert.sql.j2 index 79a7c654f10..95bac76c591 100644 --- a/tests/queries/0_stateless/02797_join_nested_lowcardinality_convert.sql.j2 +++ b/tests/queries/0_stateless/02797_join_nested_lowcardinality_convert.sql.j2 @@ -1,4 +1,4 @@ - +SET allow_suspicious_low_cardinality_types=1; DROP TABLE IF EXISTS test1__fuzz_36; DROP TABLE IF EXISTS test1__fuzz_38; diff --git a/tests/queries/0_stateless/02884_string_distance_function.sql b/tests/queries/0_stateless/02884_string_distance_function.sql index fddbf41f0e5..95604c6f401 100644 --- a/tests/queries/0_stateless/02884_string_distance_function.sql +++ b/tests/queries/0_stateless/02884_string_distance_function.sql @@ -29,13 +29,13 @@ CREATE TABLE t INSERT INTO t VALUES ('', '') ('abc', '') ('', 'abc') ('abc', 'abc') ('abc', 'ab') ('abc', 'bc') ('clickhouse', 'mouse'); SELECT '-- non-const arguments'; -SELECT 'byteHammingDistance', s1, s2, byteHammingDistance(s1, s2) FROM t ORDER BY ALL; -SELECT 'editDistance', s1, s2, editDistance(s1, s2) FROM t ORDER BY ALL; -SELECT 'damerauLevenshteinDistance', s1, s2, damerauLevenshteinDistance(s1, s2) FROM t ORDER BY ALL; -SELECT 'stringJaccardIndex', s1, s2, stringJaccardIndex(s1, s2) FROM t ORDER BY ALL; -SELECT 'stringJaccardIndexUTF8', s1, s2, stringJaccardIndexUTF8(s1, s2) FROM t ORDER BY ALL; -SELECT 'jaroSimilarity', s1, s2, jaroSimilarity(s1, s2) FROM t ORDER BY ALL; -SELECT 'jaroWinklerSimilarity', s1, s2, jaroWinklerSimilarity(s1, s2) FROM t ORDER BY ALL; +SELECT 'byteHammingDistance', s1, s2, byteHammingDistance(s1, s2) FROM t ORDER BY *; +SELECT 'editDistance', s1, s2, editDistance(s1, s2) FROM t ORDER BY *; +SELECT 'damerauLevenshteinDistance', s1, s2, damerauLevenshteinDistance(s1, s2) FROM t ORDER BY *; +SELECT 'stringJaccardIndex', s1, s2, stringJaccardIndex(s1, s2) FROM t ORDER BY *; +SELECT 'stringJaccardIndexUTF8', s1, s2, stringJaccardIndexUTF8(s1, s2) FROM t ORDER BY *; +SELECT 'jaroSimilarity', s1, s2, jaroSimilarity(s1, s2) FROM t ORDER BY *; +SELECT 'jaroWinklerSimilarity', s1, s2, jaroWinklerSimilarity(s1, s2) FROM t ORDER BY *; SELECT '-- Special UTF-8 tests'; -- We do not perform full UTF8 validation, so sometimes it just returns some result diff --git a/tests/queries/0_stateless/02933_compare_with_bool_as_string.reference b/tests/queries/0_stateless/02933_compare_with_bool_as_string.reference new file mode 100644 index 00000000000..d00491fd7e5 --- /dev/null +++ b/tests/queries/0_stateless/02933_compare_with_bool_as_string.reference @@ -0,0 +1 @@ +1 diff --git a/tests/queries/0_stateless/02933_compare_with_bool_as_string.sql b/tests/queries/0_stateless/02933_compare_with_bool_as_string.sql new file mode 100644 index 00000000000..5dbacd5fbbf --- /dev/null +++ b/tests/queries/0_stateless/02933_compare_with_bool_as_string.sql @@ -0,0 +1 @@ +select true = 'true'; diff --git a/tests/queries/0_stateless/02943_order_by_all.reference b/tests/queries/0_stateless/02943_order_by_all.reference index 6eed33cc68d..ef399fe8e2d 100644 --- a/tests/queries/0_stateless/02943_order_by_all.reference +++ b/tests/queries/0_stateless/02943_order_by_all.reference @@ -49,45 +49,9 @@ A 2 2 A 3 B \N C --- what happens if some column "all" already exists? -B 3 10 -D 1 20 -A 2 30 -C \N 40 -B 3 10 -D 1 20 -A 2 30 -C \N 40 -D 1 +-- Special case: all columns in SELECT clause, ORDER BY * A 2 B 3 C \N D 1 -A 2 -B 3 -C \N -A 2 -B 3 -D 1 -\N -A 2 -B 3 -D 1 -\N -B 3 10 -D 1 20 -A 2 30 -C \N 40 -B 3 10 -D 1 20 -A 2 30 -C \N 40 --- test SELECT * ORDER BY ALL with no "all" column in the SELECT clause -A 2 30 -B 3 10 -C \N 40 -D 1 20 -A 2 30 -B 3 10 -C \N 40 -D 1 20 +-- "*" must appear stand-alone in ORDER BY diff --git a/tests/queries/0_stateless/02943_order_by_all.sql b/tests/queries/0_stateless/02943_order_by_all.sql index 0960d75ad96..2fe628e9b95 100644 --- a/tests/queries/0_stateless/02943_order_by_all.sql +++ b/tests/queries/0_stateless/02943_order_by_all.sql @@ -1,4 +1,4 @@ --- Tests that sort expression ORDER BY ALL +-- Tests that sort expression ORDER BY * DROP TABLE IF EXISTS order_by_all; @@ -6,104 +6,48 @@ CREATE TABLE order_by_all ( a String, b Nullable(Int32), - all UInt64, ) ENGINE = Memory; -INSERT INTO order_by_all VALUES ('B', 3, 10), ('C', NULL, 40), ('D', 1, 20), ('A', 2, 30); +INSERT INTO order_by_all VALUES ('B', 3), ('C', NULL), ('D', 1), ('A', 2); SELECT '-- no modifiers'; SET allow_experimental_analyzer = 0; -SELECT a, b FROM order_by_all ORDER BY ALL; -SELECT b, a FROM order_by_all ORDER BY ALL; +SELECT a, b FROM order_by_all ORDER BY *; +SELECT b, a FROM order_by_all ORDER BY *; SET allow_experimental_analyzer = 1; -SELECT a, b FROM order_by_all ORDER BY ALL; -SELECT b, a FROM order_by_all ORDER BY ALL; +SELECT a, b FROM order_by_all ORDER BY *; +SELECT b, a FROM order_by_all ORDER BY *; SELECT '-- with ASC/DESC modifiers'; SET allow_experimental_analyzer = 0; -SELECT a, b FROM order_by_all ORDER BY ALL ASC; -SELECT a, b FROM order_by_all ORDER BY ALL DESC; +SELECT a, b FROM order_by_all ORDER BY * ASC; +SELECT a, b FROM order_by_all ORDER BY * DESC; SET allow_experimental_analyzer = 1; -SELECT a, b FROM order_by_all ORDER BY ALL ASC; -SELECT a, b FROM order_by_all ORDER BY ALL DESC; +SELECT a, b FROM order_by_all ORDER BY * ASC; +SELECT a, b FROM order_by_all ORDER BY * DESC; SELECT '-- with NULLS FIRST/LAST modifiers'; SET allow_experimental_analyzer = 0; -SELECT b, a FROM order_by_all ORDER BY ALL NULLS FIRST; -SELECT b, a FROM order_by_all ORDER BY ALL NULLS LAST; +SELECT b, a FROM order_by_all ORDER BY * NULLS FIRST; +SELECT b, a FROM order_by_all ORDER BY * NULLS LAST; SET allow_experimental_analyzer = 1; -SELECT b, a FROM order_by_all ORDER BY ALL NULLS FIRST; -SELECT b, a FROM order_by_all ORDER BY ALL NULLS LAST; +SELECT b, a FROM order_by_all ORDER BY * NULLS FIRST; +SELECT b, a FROM order_by_all ORDER BY * NULLS LAST; -SELECT '-- what happens if some column "all" already exists?'; +SELECT '-- Special case: all columns in SELECT clause, ORDER BY *'; +SELECT * FROM order_by_all ORDER BY * NULLS LAST; --- columns +SELECT '-- "*" must appear stand-alone in ORDER BY'; SET allow_experimental_analyzer = 0; -SELECT a, b, all FROM order_by_all ORDER BY all; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b, all FROM order_by_all ORDER BY ALL; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b, all FROM order_by_all ORDER BY all SETTINGS enable_order_by_all = false; +SELECT a, b FROM order_by_all ORDER BY *, a; -- { serverError UNKNOWN_IDENTIFIER } SET allow_experimental_analyzer = 1; -SELECT a, b, all FROM order_by_all ORDER BY all; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b, all FROM order_by_all ORDER BY ALL; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b, all FROM order_by_all ORDER BY all SETTINGS enable_order_by_all = false; - --- column aliases - -SET allow_experimental_analyzer = 0; -SELECT a, b AS all FROM order_by_all ORDER BY all; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b AS all FROM order_by_all ORDER BY ALL; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b AS all FROM order_by_all ORDER BY all SETTINGS enable_order_by_all = false; - -SET allow_experimental_analyzer = 1; -SELECT a, b AS all FROM order_by_all ORDER BY all; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b AS all FROM order_by_all ORDER BY ALL; -- { serverError UNEXPECTED_EXPRESSION } -SELECT a, b AS all FROM order_by_all ORDER BY all SETTINGS enable_order_by_all = false; - --- expressions - -SET allow_experimental_analyzer = 0; -SELECT format('{} {}', a, b) AS all FROM order_by_all ORDER BY all; -- { serverError UNEXPECTED_EXPRESSION } -SELECT format('{} {}', a, b) AS all FROM order_by_all ORDER BY ALL; -- { serverError UNEXPECTED_EXPRESSION } -SELECT format('{} {}', a, b) AS all FROM order_by_all ORDER BY all SETTINGS enable_order_by_all = false; - -SET allow_experimental_analyzer = 1; -SELECT format('{} {}', a, b) AS all FROM order_by_all ORDER BY all; -- { serverError UNEXPECTED_EXPRESSION } -SELECT format('{} {}', a, b) AS all FROM order_by_all ORDER BY ALL; -- { serverError UNEXPECTED_EXPRESSION } -SELECT format('{} {}', a, b) AS all FROM order_by_all ORDER BY all SETTINGS enable_order_by_all = false; - -SET allow_experimental_analyzer = 0; -SELECT a, b, all FROM order_by_all ORDER BY all, a; - -SET allow_experimental_analyzer = 1; -SELECT a, b, all FROM order_by_all ORDER BY all, a; - -DROP TABLE order_by_all; - -SELECT '-- test SELECT * ORDER BY ALL with no "all" column in the SELECT clause'; - -CREATE TABLE order_by_all -( - a String, - b Nullable(Int32), - c UInt64, -) - ENGINE = Memory; - -INSERT INTO order_by_all VALUES ('B', 3, 10), ('C', NULL, 40), ('D', 1, 20), ('A', 2, 30); - -SET allow_experimental_analyzer = 0; -SELECT * FROM order_by_all ORDER BY ALL; - -SET allow_experimental_analyzer = 1; -SELECT * FROM order_by_all ORDER BY ALL; - -DROP TABLE order_by_all; +SELECT a, b FROM order_by_all ORDER BY *, a; -- { serverError UNSUPPORTED_METHOD } diff --git a/tests/queries/0_stateless/02946_materialize_column_must_not_override_past_values.reference b/tests/queries/0_stateless/02946_materialize_column_must_not_override_past_values.reference new file mode 100644 index 00000000000..461075e9607 --- /dev/null +++ b/tests/queries/0_stateless/02946_materialize_column_must_not_override_past_values.reference @@ -0,0 +1,45 @@ +DEFAULT expressions +-- Compact parts +Before materialize +1 1 +2 54321 +After materialize +1 1 +2 54321 +-- Wide parts +Before materialize +1 1 +2 54321 +After materialize +1 1 +2 54321 +-- Nullable column != physically absent +Before materialize +1 1 +2 \N +3 54321 +After materialize +1 1 +2 \N +3 54321 +-- Parts with renamed column +Before materialize +1 1 +2 54321 +After rename +1 1 +2 54321 +After materialize +1 1 +2 54321 +MATERIALIZED expressions +-- Compact parts +Before materialize +1 54321 +After materialize +1 65432 +-- Compact parts +Before materialize +1 54321 +After materialize +1 65432 diff --git a/tests/queries/0_stateless/02946_materialize_column_must_not_override_past_values.sql b/tests/queries/0_stateless/02946_materialize_column_must_not_override_past_values.sql new file mode 100644 index 00000000000..cfdde287712 --- /dev/null +++ b/tests/queries/0_stateless/02946_materialize_column_must_not_override_past_values.sql @@ -0,0 +1,85 @@ +SET mutations_sync = 2; + +DROP TABLE IF EXISTS tab; + +-- Tests that existing parts which contain a non-default value in columns with DEFAULT expression remain unchanged by MATERIALIZE COLUMN> +SELECT 'DEFAULT expressions'; + +SELECT '-- Compact parts'; + +CREATE TABLE tab (id Int64, dflt Int64 DEFAULT 54321) ENGINE MergeTree ORDER BY id; +INSERT INTO tab (id, dflt) VALUES (1, 1); +INSERT INTO tab (id) VALUES (2); +SELECT 'Before materialize'; +SELECT * FROM tab ORDER BY id; +ALTER TABLE tab MATERIALIZE COLUMN dflt; +SELECT 'After materialize'; +SELECT * FROM tab ORDER BY id; +DROP TABLE tab; + +SELECT '-- Wide parts'; + +CREATE TABLE tab (id Int64, dflt Int64 DEFAULT 54321) ENGINE MergeTree ORDER BY id SETTINGS min_bytes_for_wide_part = 1; +INSERT INTO tab (id, dflt) VALUES (1, 1); +INSERT INTO tab (id) VALUES (2); +SELECT 'Before materialize'; +SELECT * FROM tab ORDER BY id; +ALTER TABLE tab MATERIALIZE COLUMN dflt; +SELECT 'After materialize'; +SELECT * FROM tab ORDER BY id; +DROP TABLE tab; + +SELECT '-- Nullable column != physically absent'; + +CREATE TABLE tab (id Int64, dflt Nullable(Int64) DEFAULT 54321) ENGINE MergeTree ORDER BY id SETTINGS min_bytes_for_wide_part = 1; +INSERT INTO tab (id, dflt) VALUES (1, 1); +INSERT INTO tab (id, dflt) VALUES (2, NULL); +INSERT INTO tab (id) VALUES (3); +SELECT 'Before materialize'; +SELECT * FROM tab ORDER BY id; +ALTER TABLE tab MATERIALIZE COLUMN dflt; +SELECT 'After materialize'; +SELECT * FROM tab ORDER BY id; +DROP TABLE tab; + +SELECT '-- Parts with renamed column'; + +CREATE TABLE tab (id Int64, dflt Int64 DEFAULT 54321) ENGINE MergeTree ORDER BY id; +INSERT INTO tab (id, dflt) VALUES (1, 1); +INSERT INTO tab (id) VALUES (2); +SELECT 'Before materialize'; +SELECT * FROM tab ORDER BY id; +ALTER TABLE tab RENAME COLUMN dflt TO dflt2; +SELECT 'After rename'; +SELECT * FROM tab ORDER BY id; +ALTER TABLE tab MATERIALIZE COLUMN dflt2; +SELECT 'After materialize'; +SELECT * FROM tab ORDER BY id; +DROP TABLE tab; + +-- But for columns with MATERIALIZED expression, all existing parts should be rewritten in case a new expression was set in the meantime. +SELECT 'MATERIALIZED expressions'; + +SELECT '-- Compact parts'; + +CREATE TABLE tab (id Int64, mtrl Int64 MATERIALIZED 54321) ENGINE MergeTree ORDER BY id; +INSERT INTO tab (id) VALUES (1); +SELECT 'Before materialize'; +SELECT id, mtrl FROM tab ORDER BY id; +ALTER TABLE tab MODIFY COLUMN mtrl Int64 MATERIALIZED 65432; +ALTER TABLE tab MATERIALIZE COLUMN mtrl; +SELECT 'After materialize'; +SELECT id, mtrl FROM tab ORDER BY id; +DROP TABLE tab; + +SELECT '-- Compact parts'; + +CREATE TABLE tab (id Int64, mtrl Int64 MATERIALIZED 54321) ENGINE MergeTree ORDER BY id SETTINGS min_bytes_for_wide_part = 1; +INSERT INTO tab (id) VALUES (1); +SELECT 'Before materialize'; +SELECT id, mtrl FROM tab ORDER BY id; +ALTER TABLE tab MODIFY COLUMN mtrl Int64 MATERIALIZED 65432; +ALTER TABLE tab MATERIALIZE COLUMN mtrl; +SELECT 'After materialize'; +SELECT id, mtrl FROM tab ORDER BY id; +DROP TABLE tab; diff --git a/tests/queries/0_stateless/02962_join_using_bug_57894.sql b/tests/queries/0_stateless/02962_join_using_bug_57894.sql index 87aef8b1a71..c9570be7053 100644 --- a/tests/queries/0_stateless/02962_join_using_bug_57894.sql +++ b/tests/queries/0_stateless/02962_join_using_bug_57894.sql @@ -11,23 +11,23 @@ INSERT INTO r VALUES (NULL, NULL); SET allow_experimental_analyzer = 0; -SELECT x FROM t FULL JOIN r USING (x) ORDER BY ALL +SELECT x FROM t FULL JOIN r USING (x) ORDER BY * ; -SELECT x FROM t FULL JOIN r USING (x) ORDER BY ALL +SELECT x FROM t FULL JOIN r USING (x) ORDER BY * SETTINGS join_algorithm = 'partial_merge'; -SELECT x FROM t FULL JOIN r USING (x) ORDER BY ALL +SELECT x FROM t FULL JOIN r USING (x) ORDER BY * SETTINGS join_algorithm = 'full_sorting_merge'; SET allow_experimental_analyzer = 1; -SELECT x FROM t FULL JOIN r USING (x) ORDER BY ALL +SELECT x FROM t FULL JOIN r USING (x) ORDER BY * ; -SELECT x FROM t FULL JOIN r USING (x) ORDER BY ALL +SELECT x FROM t FULL JOIN r USING (x) ORDER BY * SETTINGS join_algorithm = 'partial_merge'; -SELECT x FROM t FULL JOIN r USING (x) ORDER BY ALL +SELECT x FROM t FULL JOIN r USING (x) ORDER BY * SETTINGS join_algorithm = 'full_sorting_merge'; diff --git a/tests/queries/0_stateless/02966_s3_access_key_id_restriction.sql b/tests/queries/0_stateless/02966_s3_access_key_id_restriction.sql deleted file mode 100644 index c1ca0b4bcd5..00000000000 --- a/tests/queries/0_stateless/02966_s3_access_key_id_restriction.sql +++ /dev/null @@ -1,6 +0,0 @@ --- Tags: no-fasttest - -select * from s3('http://localhost:11111/test/a.tsv', '\ninjection\n', 'admin'); -- { serverError 36 } -select * from deltaLake('http://localhost:11111/test/a.tsv', '\ninjection\n', 'admin'); -- { serverError 36 } -select * from hudi('http://localhost:11111/test/a.tsv', '\ninjection\n', 'admin'); -- { serverError 36 } -select * from iceberg('http://localhost:11111/test/a.tsv', '\ninjection\n', 'admin'); -- { serverError 36 } diff --git a/tests/queries/0_stateless/02975_intdiv_with_decimal.reference b/tests/queries/0_stateless/02975_intdiv_with_decimal.reference new file mode 100644 index 00000000000..5540734ae4c --- /dev/null +++ b/tests/queries/0_stateless/02975_intdiv_with_decimal.reference @@ -0,0 +1,68 @@ +2 +2 +1 +2 +2 +2 +2 +2 +2 +2 +2 +1 +2 +1 +2 +1 +2 +2 +2 +2 +2 +2 +2 +2 +2 +2 +1 +1 +1 +1 +2 +2 +2 +2 +2 +2 +1 +2 +2 +2 +2 +2 +2 +2 +2 +1 +2 +1 +2 +1 +2 +2 +2 +2 +2 +2 +2 +2 +2 +2 +1 +1 +1 +1 +2 +2 +2 +2 diff --git a/tests/queries/0_stateless/02975_intdiv_with_decimal.sql b/tests/queries/0_stateless/02975_intdiv_with_decimal.sql new file mode 100644 index 00000000000..0911a481251 --- /dev/null +++ b/tests/queries/0_stateless/02975_intdiv_with_decimal.sql @@ -0,0 +1,70 @@ +--intDiv-- +SELECT intDiv(4,2); +SELECT intDiv(toDecimal32(4.4, 2), 2); +SELECT intDiv(4, toDecimal32(2.2, 2)); +SELECT intDiv(toDecimal32(4.4, 2), 2); +SELECT intDiv(toDecimal32(4.4, 2), toDecimal32(2.2, 2)); +SELECT intDiv(toDecimal64(4.4, 3), 2); +SELECT intDiv(toDecimal64(4.4, 3), toDecimal32(2.2, 2)); +SELECT intDiv(toDecimal128(4.4, 4), 2); +SELECT intDiv(toDecimal128(4.4, 4), toDecimal32(2.2, 2)); +SELECT intDiv(toDecimal256(4.4, 5), 2); +SELECT intDiv(toDecimal256(4.4, 5), toDecimal32(2.2, 2)); +SELECT intDiv(4, toDecimal64(2.2, 2)); +SELECT intDiv(toDecimal32(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDiv(4, toDecimal128(2.2, 3)); +SELECT intDiv(toDecimal32(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDiv(4, toDecimal256(2.2, 4)); +SELECT intDiv(toDecimal32(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDiv(toDecimal64(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDiv(toDecimal128(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDiv(toDecimal256(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDiv(toDecimal64(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDiv(toDecimal128(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDiv(toDecimal256(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDiv(toDecimal64(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDiv(toDecimal128(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDiv(toDecimal256(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDiv(4.2, toDecimal32(2.2, 2)); +SELECT intDiv(4.2, toDecimal64(2.2, 2)); +SELECT intDiv(4.2, toDecimal128(2.2, 2)); +SELECT intDiv(4.2, toDecimal256(2.2, 2)); +SELECT intDiv(toDecimal32(4.4, 2), 2.2); +SELECT intDiv(toDecimal64(4.4, 2), 2.2); +SELECT intDiv(toDecimal128(4.4, 2), 2.2); +SELECT intDiv(toDecimal256(4.4, 2), 2.2); +--intDivOrZero-- +SELECT intDivOrZero(4,2); +SELECT intDivOrZero(toDecimal32(4.4, 2), 2); +SELECT intDivOrZero(4, toDecimal32(2.2, 2)); +SELECT intDivOrZero(toDecimal32(4.4, 2), 2); +SELECT intDivOrZero(toDecimal32(4.4, 2), toDecimal32(2.2, 2)); +SELECT intDivOrZero(toDecimal64(4.4, 3), 2); +SELECT intDivOrZero(toDecimal64(4.4, 3), toDecimal32(2.2, 2)); +SELECT intDivOrZero(toDecimal128(4.4, 4), 2); +SELECT intDivOrZero(toDecimal128(4.4, 4), toDecimal32(2.2, 2)); +SELECT intDivOrZero(toDecimal256(4.4, 5), 2); +SELECT intDivOrZero(toDecimal256(4.4, 5), toDecimal32(2.2, 2)); +SELECT intDivOrZero(4, toDecimal64(2.2, 2)); +SELECT intDivOrZero(toDecimal32(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDivOrZero(4, toDecimal128(2.2, 3)); +SELECT intDivOrZero(toDecimal32(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDivOrZero(4, toDecimal256(2.2, 4)); +SELECT intDivOrZero(toDecimal32(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDivOrZero(toDecimal64(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDivOrZero(toDecimal128(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDivOrZero(toDecimal256(4.4, 2), toDecimal64(2.2, 2)); +SELECT intDivOrZero(toDecimal64(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDivOrZero(toDecimal128(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDivOrZero(toDecimal256(4.4, 2), toDecimal128(2.2, 2)); +SELECT intDivOrZero(toDecimal64(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDivOrZero(toDecimal128(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDivOrZero(toDecimal256(4.4, 2), toDecimal256(2.2, 2)); +SELECT intDivOrZero(4.2, toDecimal32(2.2, 2)); +SELECT intDivOrZero(4.2, toDecimal64(2.2, 2)); +SELECT intDivOrZero(4.2, toDecimal128(2.2, 2)); +SELECT intDivOrZero(4.2, toDecimal256(2.2, 2)); +SELECT intDivOrZero(toDecimal32(4.4, 2), 2.2); +SELECT intDivOrZero(toDecimal64(4.4, 2), 2.2); +SELECT intDivOrZero(toDecimal128(4.4, 2), 2.2); +SELECT intDivOrZero(toDecimal256(4.4, 2), 2.2); diff --git a/tests/queries/0_stateless/02966_s3_access_key_id_restriction.reference b/tests/queries/0_stateless/02981_nested_bad_types.reference similarity index 100% rename from tests/queries/0_stateless/02966_s3_access_key_id_restriction.reference rename to tests/queries/0_stateless/02981_nested_bad_types.reference diff --git a/tests/queries/0_stateless/02981_nested_bad_types.sql b/tests/queries/0_stateless/02981_nested_bad_types.sql new file mode 100644 index 00000000000..87bc80693c8 --- /dev/null +++ b/tests/queries/0_stateless/02981_nested_bad_types.sql @@ -0,0 +1,58 @@ +set allow_suspicious_low_cardinality_types=0; +set allow_suspicious_fixed_string_types=0; +set allow_experimental_variant_type=0; + +select [42]::Array(LowCardinality(UInt64)); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +select [[[42]]]::Array(Array(Array(LowCardinality(UInt64)))); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +select map('a', 42)::Map(String, LowCardinality(UInt64)); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +select map('a', map('b', [42]))::Map(String, Map(String, Array(LowCardinality(UInt64)))); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +select tuple('a', 42)::Tuple(String, LowCardinality(UInt64)); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +select tuple('a', [map('b', 42)])::Tuple(String, Array(Map(String, LowCardinality(UInt64)))); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} + +create table test (x Array(LowCardinality(UInt64))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Array(Array(LowCardinality(UInt64)))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Map(String, LowCardinality(UInt64))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Map(String, Map(String, LowCardinality(UInt64)))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Tuple(String, LowCardinality(UInt64))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Tuple(String, Array(Map(String, LowCardinality(UInt64))))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} + + +select ['42']::Array(FixedString(1000000)); -- {serverError ILLEGAL_COLUMN} +select ['42']::Array(FixedString(1000000)); -- {serverError ILLEGAL_COLUMN} +select [[['42']]]::Array(Array(Array(FixedString(1000000)))); -- {serverError ILLEGAL_COLUMN} +select map('a', '42')::Map(String, FixedString(1000000)); -- {serverError ILLEGAL_COLUMN} +select map('a', map('b', ['42']))::Map(String, Map(String, Array(FixedString(1000000)))); -- {serverError ILLEGAL_COLUMN} +select tuple('a', '42')::Tuple(String, FixedString(1000000)); -- {serverError ILLEGAL_COLUMN} +select tuple('a', [map('b', '42')])::Tuple(String, Array(Map(String, FixedString(1000000)))); -- {serverError ILLEGAL_COLUMN} + +create table test (x Array(FixedString(1000000))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Array(Array(FixedString(1000000)))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Map(String, FixedString(1000000))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Map(String, Map(String, FixedString(1000000)))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Tuple(String, FixedString(1000000))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Tuple(String, Array(Map(String, FixedString(1000000))))) engine=Memory; -- {serverError ILLEGAL_COLUMN} + +select [42]::Array(Variant(String, UInt64)); -- {serverError ILLEGAL_COLUMN} +select [[[42]]]::Array(Array(Array(Variant(String, UInt64)))); -- {serverError ILLEGAL_COLUMN} +select map('a', 42)::Map(String, Variant(String, UInt64)); -- {serverError ILLEGAL_COLUMN} +select map('a', map('b', [42]))::Map(String, Map(String, Array(Variant(String, UInt64)))); -- {serverError ILLEGAL_COLUMN} +select tuple('a', 42)::Tuple(String, Variant(String, UInt64)); -- {serverError ILLEGAL_COLUMN} +select tuple('a', [map('b', 42)])::Tuple(String, Array(Map(String, Variant(String, UInt64)))); -- {serverError ILLEGAL_COLUMN} + +create table test (x Array(Variant(String, UInt64))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Array(Array(Variant(String, UInt64)))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Map(String, Variant(String, UInt64))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Map(String, Map(String, Variant(String, UInt64)))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Tuple(String, Variant(String, UInt64))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Tuple(String, Array(Map(String, Variant(String, UInt64))))) engine=Memory; -- {serverError ILLEGAL_COLUMN} + +set allow_experimental_variant_type=1; +select 42::Variant(String, LowCardinality(UInt64)) settings allow_experimental_variant_type=1; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +select tuple('a', [map('b', 42)])::Tuple(String, Array(Map(String, Variant(LowCardinality(UInt64), UInt8)))); -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Variant(LowCardinality(UInt64), UInt8)) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} +create table test (x Tuple(String, Array(Map(String, Variant(LowCardinality(UInt64), UInt8))))) engine=Memory; -- {serverError SUSPICIOUS_TYPE_FOR_LOW_CARDINALITY} + +select '42'::Variant(UInt64, FixedString(1000000)); -- {serverError ILLEGAL_COLUMN} +select tuple('a', [map('b', '42')])::Tuple(String, Array(Map(String, Variant(UInt32, FixedString(1000000))))); -- {serverError ILLEGAL_COLUMN} +create table test (x Variant(UInt64, FixedString(1000000))) engine=Memory; -- {serverError ILLEGAL_COLUMN} +create table test (x Tuple(String, Array(Map(String, FixedString(1000000))))) engine=Memory; -- {serverError ILLEGAL_COLUMN} diff --git a/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference new file mode 100644 index 00000000000..b6d1ff865e5 --- /dev/null +++ b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.reference @@ -0,0 +1,2 @@ +c1 Nullable(String) +c1 Nullable(Float64) diff --git a/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql new file mode 100644 index 00000000000..2a281e898f1 --- /dev/null +++ b/tests/queries/0_stateless/02982_dont_infer_exponent_floats.sql @@ -0,0 +1,2 @@ +DESC format(CSV, '1E20\n1.1E20') settings input_format_try_infer_exponent_floats = 0; +DESC format(CSV, '1E20\n1.1E20') settings input_format_try_infer_exponent_floats = 1; diff --git a/tests/queries/0_stateless/02985_if_over_big_int_decimal.reference b/tests/queries/0_stateless/02985_if_over_big_int_decimal.reference new file mode 100644 index 00000000000..1dfad945ee2 --- /dev/null +++ b/tests/queries/0_stateless/02985_if_over_big_int_decimal.reference @@ -0,0 +1,12 @@ +49500 +49500 +49500 +49500 +49500 +49500 +450000 +450000 +450000 +450000 +450000 +450000 diff --git a/tests/queries/0_stateless/02985_if_over_big_int_decimal.sql b/tests/queries/0_stateless/02985_if_over_big_int_decimal.sql new file mode 100644 index 00000000000..0295a64a092 --- /dev/null +++ b/tests/queries/0_stateless/02985_if_over_big_int_decimal.sql @@ -0,0 +1,14 @@ +select sumIf(number::Int128, number % 10 == 0) from numbers(1000); +select sumIf(number::UInt128, number % 10 == 0) from numbers(1000); +select sumIf(number::Int256, number % 10 == 0) from numbers(1000); +select sumIf(number::UInt256, number % 10 == 0) from numbers(1000); +select sumIf(number::Decimal128(3), number % 10 == 0) from numbers(1000); +select sumIf(number::Decimal256(3), number % 10 == 0) from numbers(1000); + +-- Test when the condition is neither 0 nor 1 +select sumIf(number::Int128, number % 10) from numbers(1000); +select sumIf(number::UInt128, number % 10) from numbers(1000); +select sumIf(number::Int256, number % 10) from numbers(1000); +select sumIf(number::UInt256, number % 10) from numbers(1000); +select sumIf(number::Decimal128(3), number % 10) from numbers(1000); +select sumIf(number::Decimal256(3), number % 10) from numbers(1000); diff --git a/tests/queries/0_stateless/02985_parser_check_stack_size.reference b/tests/queries/0_stateless/02985_parser_check_stack_size.reference deleted file mode 100644 index f83e0818db2..00000000000 --- a/tests/queries/0_stateless/02985_parser_check_stack_size.reference +++ /dev/null @@ -1 +0,0 @@ -TOO_DEEP diff --git a/tests/queries/0_stateless/02985_parser_check_stack_size.sh b/tests/queries/0_stateless/02985_parser_check_stack_size.sh deleted file mode 100755 index c91a0a3eacc..00000000000 --- a/tests/queries/0_stateless/02985_parser_check_stack_size.sh +++ /dev/null @@ -1,7 +0,0 @@ -#!/usr/bin/env bash - -CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) -# shellcheck source=../shell_config.sh -. "$CURDIR"/../shell_config.sh - -$CLICKHOUSE_CLIENT --query "select 'create table test (x ' || repeat('Array(', 10000) || 'UInt64' || repeat(')', 10000) || ') engine=Memory' format TSVRaw" | $CLICKHOUSE_CURL "${CLICKHOUSE_URL}&max_parser_depth=100000" --data-binary @- | grep -o -F 'TOO_DEEP' diff --git a/tests/queries/0_stateless/02989_replicated_merge_tree_invalid_metadata_version.reference b/tests/queries/0_stateless/02989_replicated_merge_tree_invalid_metadata_version.reference new file mode 100644 index 00000000000..128e3adcc0a --- /dev/null +++ b/tests/queries/0_stateless/02989_replicated_merge_tree_invalid_metadata_version.reference @@ -0,0 +1,14 @@ +Row 1: +────── +name: metadata +version: 1 +-- +Row 1: +────── +name: metadata_version +value: 1 +-- +id UInt64 +value String +insert_time DateTime +insert_time_updated DateTime diff --git a/tests/queries/0_stateless/02989_replicated_merge_tree_invalid_metadata_version.sql b/tests/queries/0_stateless/02989_replicated_merge_tree_invalid_metadata_version.sql new file mode 100644 index 00000000000..15633586aa8 --- /dev/null +++ b/tests/queries/0_stateless/02989_replicated_merge_tree_invalid_metadata_version.sql @@ -0,0 +1,40 @@ +-- Tags: zookeeper + +DROP TABLE IF EXISTS test_table_replicated; +CREATE TABLE test_table_replicated +( + id UInt64, + value String +) ENGINE=ReplicatedMergeTree('/clickhouse/tables/{database}/test_table_replicated', '1_replica') ORDER BY id; + +ALTER TABLE test_table_replicated ADD COLUMN insert_time DateTime; + +SELECT name, version FROM system.zookeeper +WHERE path = (SELECT zookeeper_path FROM system.replicas WHERE database = currentDatabase() AND table = 'test_table_replicated') +AND name = 'metadata' FORMAT Vertical; + +DROP TABLE IF EXISTS test_table_replicated_second; +CREATE TABLE test_table_replicated_second +( + id UInt64, + value String, + insert_time DateTime +) ENGINE=ReplicatedMergeTree('/clickhouse/tables/{database}/test_table_replicated', '2_replica') ORDER BY id; + +DROP TABLE test_table_replicated; + +SELECT '--'; + +SELECT name, value FROM system.zookeeper +WHERE path = (SELECT replica_path FROM system.replicas WHERE database = currentDatabase() AND table = 'test_table_replicated_second') +AND name = 'metadata_version' FORMAT Vertical; + +SYSTEM RESTART REPLICA test_table_replicated_second; + +ALTER TABLE test_table_replicated_second ADD COLUMN insert_time_updated DateTime; + +SELECT '--'; + +DESCRIBE test_table_replicated_second; + +DROP TABLE test_table_replicated_second; diff --git a/tests/queries/0_stateless/02994_cosineDistanceNullable.reference b/tests/queries/0_stateless/02994_cosineDistanceNullable.reference new file mode 100644 index 00000000000..e4fe1f97e7e --- /dev/null +++ b/tests/queries/0_stateless/02994_cosineDistanceNullable.reference @@ -0,0 +1,11 @@ +\N +\N +\N +\N +\N +\N +\N +\N +\N +\N +\N diff --git a/tests/queries/0_stateless/02994_cosineDistanceNullable.sql b/tests/queries/0_stateless/02994_cosineDistanceNullable.sql new file mode 100644 index 00000000000..a62216982f3 --- /dev/null +++ b/tests/queries/0_stateless/02994_cosineDistanceNullable.sql @@ -0,0 +1,3 @@ +-- https://github.com/ClickHouse/ClickHouse/issues/59596 +SELECT cosineDistance((1, 1), (toNullable(0.5), 0.1)); +SELECT cosineDistance((1, 1), (toNullable(0.5), 0.1)) from numbers(10); diff --git a/tests/queries/0_stateless/02995_baseline_23_12_1.tsv b/tests/queries/0_stateless/02995_baseline_23_12_1.tsv new file mode 100644 index 00000000000..4c0c9125b46 --- /dev/null +++ b/tests/queries/0_stateless/02995_baseline_23_12_1.tsv @@ -0,0 +1,940 @@ +add_http_cors_header 0 +additional_result_filter +additional_table_filters {} +aggregate_functions_null_for_empty 0 +aggregation_in_order_max_block_bytes 50000000 +aggregation_memory_efficient_merge_threads 0 +allow_aggregate_partitions_independently 0 +allow_asynchronous_read_from_io_pool_for_merge_tree 0 +allow_changing_replica_until_first_data_packet 0 +allow_create_index_without_type 0 +allow_custom_error_code_in_throwif 0 +allow_ddl 1 +allow_deprecated_database_ordinary 0 +allow_deprecated_syntax_for_merge_tree 0 +allow_distributed_ddl 1 +allow_drop_detached 0 +allow_execute_multiif_columnar 1 +allow_experimental_alter_materialized_view_structure 1 +allow_experimental_analyzer 0 +allow_experimental_annoy_index 0 +allow_experimental_bigint_types 1 +allow_experimental_codecs 0 +allow_experimental_database_atomic 1 +allow_experimental_database_materialized_mysql 0 +allow_experimental_database_materialized_postgresql 0 +allow_experimental_database_replicated 0 +allow_experimental_funnel_functions 0 +allow_experimental_geo_types 1 +allow_experimental_hash_functions 0 +allow_experimental_inverted_index 0 +allow_experimental_lightweight_delete 1 +allow_experimental_live_view 0 +allow_experimental_map_type 1 +allow_experimental_materialized_postgresql_table 0 +allow_experimental_nlp_functions 0 +allow_experimental_object_type 0 +allow_experimental_parallel_reading_from_replicas 0 +allow_experimental_projection_optimization 1 +allow_experimental_query_cache 1 +allow_experimental_query_deduplication 0 +allow_experimental_refreshable_materialized_view 0 +allow_experimental_s3queue 1 +allow_experimental_shared_merge_tree 0 +allow_experimental_statistic 0 +allow_experimental_undrop_table_query 1 +allow_experimental_usearch_index 0 +allow_experimental_window_functions 1 +allow_experimental_window_view 0 +allow_hyperscan 1 +allow_introspection_functions 0 +allow_named_collection_override_by_default 1 +allow_non_metadata_alters 1 +allow_nonconst_timezone_arguments 0 +allow_nondeterministic_mutations 0 +allow_nondeterministic_optimize_skip_unused_shards 0 +allow_prefetched_read_pool_for_local_filesystem 0 +allow_prefetched_read_pool_for_remote_filesystem 1 +allow_push_predicate_when_subquery_contains_with 1 +allow_settings_after_format_in_insert 0 +allow_simdjson 1 +allow_statistic_optimize 0 +allow_suspicious_codecs 0 +allow_suspicious_fixed_string_types 0 +allow_suspicious_indices 0 +allow_suspicious_low_cardinality_types 0 +allow_suspicious_ttl_expressions 0 +allow_unrestricted_reads_from_keeper 0 +alter_move_to_space_execute_async 0 +alter_partition_verbose_result 0 +alter_sync 1 +analyze_index_with_space_filling_curves 1 +annoy_index_search_k_nodes -1 +any_join_distinct_right_table_keys 0 +apply_deleted_mask 1 +apply_mutations_on_fly 0 +asterisk_include_alias_columns 0 +asterisk_include_materialized_columns 0 +async_insert 0 +async_insert_busy_timeout_ms 200 +async_insert_cleanup_timeout_ms 1000 +async_insert_deduplicate 0 +async_insert_max_data_size 1000000 +async_insert_max_query_number 450 +async_insert_stale_timeout_ms 0 +async_insert_threads 16 +async_query_sending_for_remote 1 +async_socket_for_remote 1 +azure_create_new_file_on_insert 0 +azure_list_object_keys_size 1000 +azure_max_single_part_upload_size 104857600 +azure_max_single_read_retries 4 +azure_truncate_on_insert 0 +background_buffer_flush_schedule_pool_size 16 +background_common_pool_size 8 +background_distributed_schedule_pool_size 16 +background_fetches_pool_size 8 +background_merges_mutations_concurrency_ratio 2 +background_message_broker_schedule_pool_size 16 +background_move_pool_size 8 +background_pool_size 16 +background_schedule_pool_size 128 +backup_restore_batch_size_for_keeper_multi 1000 +backup_restore_batch_size_for_keeper_multiread 10000 +backup_restore_keeper_fault_injection_probability 0 +backup_restore_keeper_fault_injection_seed 0 +backup_restore_keeper_max_retries 20 +backup_restore_keeper_retry_initial_backoff_ms 100 +backup_restore_keeper_retry_max_backoff_ms 5000 +backup_restore_keeper_value_max_size 1048576 +backup_threads 16 +bool_false_representation false +bool_true_representation true +cache_warmer_threads 4 +calculate_text_stack_trace 1 +cancel_http_readonly_queries_on_client_close 0 +cast_ipv4_ipv6_default_on_conversion_error 0 +cast_keep_nullable 0 +check_query_single_value_result 1 +check_referential_table_dependencies 0 +check_table_dependencies 1 +checksum_on_read 1 +cloud_mode 0 +cloud_mode_engine 1 +cluster_for_parallel_replicas +collect_hash_table_stats_during_aggregation 1 +column_names_for_schema_inference +compatibility +compatibility_ignore_auto_increment_in_create_table 0 +compatibility_ignore_collation_in_create_table 1 +compile_aggregate_expressions 1 +compile_expressions 0 +compile_sort_description 1 +connect_timeout 10 +connect_timeout_with_failover_ms 1000 +connect_timeout_with_failover_secure_ms 1000 +connection_pool_max_wait_ms 0 +connections_with_failover_max_tries 3 +convert_query_to_cnf 0 +count_distinct_implementation uniqExact +count_distinct_optimization 0 +create_index_ignore_unique 0 +create_replicated_merge_tree_fault_injection_probability 0 +create_table_empty_primary_key_by_default 0 +cross_to_inner_join_rewrite 1 +data_type_default_nullable 0 +database_atomic_wait_for_drop_and_detach_synchronously 0 +database_replicated_allow_only_replicated_engine 0 +database_replicated_allow_replicated_engine_arguments 1 +database_replicated_always_detach_permanently 0 +database_replicated_ddl_output 1 +database_replicated_enforce_synchronous_settings 0 +database_replicated_initial_query_timeout_sec 300 +date_time_input_format basic +date_time_output_format simple +date_time_overflow_behavior ignore +decimal_check_overflow 1 +deduplicate_blocks_in_dependent_materialized_views 0 +default_database_engine Atomic +default_max_bytes_in_join 1000000000 +default_table_engine None +default_temporary_table_engine Memory +describe_compact_output 0 +describe_extend_object_types 0 +describe_include_subcolumns 0 +describe_include_virtual_columns 0 +dialect clickhouse +dictionary_use_async_executor 0 +distinct_overflow_mode throw +distributed_aggregation_memory_efficient 1 +distributed_background_insert_batch 0 +distributed_background_insert_max_sleep_time_ms 30000 +distributed_background_insert_sleep_time_ms 100 +distributed_background_insert_split_batch_on_failure 0 +distributed_background_insert_timeout 0 +distributed_connections_pool_size 1024 +distributed_ddl_entry_format_version 5 +distributed_ddl_output_mode throw +distributed_ddl_task_timeout 180 +distributed_directory_monitor_batch_inserts 0 +distributed_directory_monitor_max_sleep_time_ms 30000 +distributed_directory_monitor_sleep_time_ms 100 +distributed_directory_monitor_split_batch_on_failure 0 +distributed_foreground_insert 0 +distributed_group_by_no_merge 0 +distributed_product_mode deny +distributed_push_down_limit 1 +distributed_replica_error_cap 1000 +distributed_replica_error_half_life 60 +distributed_replica_max_ignored_errors 0 +do_not_merge_across_partitions_select_final 0 +drain_timeout 3 +empty_result_for_aggregation_by_constant_keys_on_empty_set 1 +empty_result_for_aggregation_by_empty_set 0 +enable_debug_queries 0 +enable_deflate_qpl_codec 0 +enable_early_constant_folding 1 +enable_extended_results_for_datetime_functions 0 +enable_filesystem_cache 1 +enable_filesystem_cache_log 0 +enable_filesystem_cache_on_write_operations 0 +enable_filesystem_read_prefetches_log 0 +enable_global_with_statement 1 +enable_http_compression 0 +enable_job_stack_trace 0 +enable_lightweight_delete 1 +enable_memory_bound_merging_of_aggregation_results 1 +enable_multiple_prewhere_read_steps 1 +enable_optimize_predicate_expression 1 +enable_optimize_predicate_expression_to_final_subquery 1 +enable_order_by_all 1 +enable_positional_arguments 1 +enable_reads_from_query_cache 1 +enable_s3_requests_logging 0 +enable_scalar_subquery_optimization 1 +enable_sharing_sets_for_mutations 1 +enable_software_prefetch_in_aggregation 1 +enable_unaligned_array_join 0 +enable_url_encoding 1 +enable_writes_to_query_cache 1 +engine_file_allow_create_multiple_files 0 +engine_file_empty_if_not_exists 0 +engine_file_skip_empty_files 0 +engine_file_truncate_on_insert 0 +engine_url_skip_empty_files 0 +errors_output_format CSV +exact_rows_before_limit 0 +except_default_mode ALL +external_storage_connect_timeout_sec 10 +external_storage_max_read_bytes 0 +external_storage_max_read_rows 0 +external_storage_rw_timeout_sec 300 +external_table_functions_use_nulls 1 +external_table_strict_query 0 +extract_kvp_max_pairs_per_row 1000 +extremes 0 +fallback_to_stale_replicas_for_distributed_queries 1 +filesystem_cache_max_download_size 137438953472 +filesystem_cache_segments_batch_size 20 +filesystem_prefetch_max_memory_usage 1073741824 +filesystem_prefetch_min_bytes_for_single_read_task 2097152 +filesystem_prefetch_step_bytes 0 +filesystem_prefetch_step_marks 0 +filesystem_prefetches_limit 200 +final 0 +flatten_nested 1 +force_aggregate_partitions_independently 0 +force_aggregation_in_order 0 +force_data_skipping_indices +force_grouping_standard_compatibility 1 +force_index_by_date 0 +force_optimize_projection 0 +force_optimize_projection_name +force_optimize_skip_unused_shards 0 +force_optimize_skip_unused_shards_nesting 0 +force_primary_key 0 +force_remove_data_recursively_on_drop 0 +format_avro_schema_registry_url +format_binary_max_array_size 1073741824 +format_binary_max_string_size 1073741824 +format_capn_proto_enum_comparising_mode by_values +format_capn_proto_use_autogenerated_schema 1 +format_csv_allow_double_quotes 1 +format_csv_allow_single_quotes 0 +format_csv_delimiter , +format_csv_null_representation \\N +format_custom_escaping_rule Escaped +format_custom_field_delimiter \t +format_custom_result_after_delimiter +format_custom_result_before_delimiter +format_custom_row_after_delimiter \n +format_custom_row_before_delimiter +format_custom_row_between_delimiter +format_display_secrets_in_show_and_select 0 +format_json_object_each_row_column_for_object_name +format_protobuf_use_autogenerated_schema 1 +format_regexp +format_regexp_escaping_rule Raw +format_regexp_skip_unmatched 0 +format_schema +format_template_resultset +format_template_row +format_template_rows_between_delimiter \n +format_tsv_null_representation \\N +formatdatetime_f_prints_single_zero 0 +formatdatetime_format_without_leading_zeros 0 +formatdatetime_parsedatetime_m_is_month_name 1 +fsync_metadata 1 +function_implementation +function_json_value_return_type_allow_complex 0 +function_json_value_return_type_allow_nullable 0 +function_range_max_elements_in_block 500000000 +function_sleep_max_microseconds_per_block 3000000 +glob_expansion_max_elements 1000 +grace_hash_join_initial_buckets 1 +grace_hash_join_max_buckets 1024 +group_by_overflow_mode throw +group_by_two_level_threshold 100000 +group_by_two_level_threshold_bytes 50000000 +group_by_use_nulls 0 +handle_kafka_error_mode default +handshake_timeout_ms 10000 +hdfs_create_new_file_on_insert 0 +hdfs_replication 0 +hdfs_skip_empty_files 0 +hdfs_truncate_on_insert 0 +hedged_connection_timeout_ms 50 +hsts_max_age 0 +http_connection_timeout 1 +http_headers_progress_interval_ms 100 +http_make_head_request 1 +http_max_chunk_size 107374182400 +http_max_field_name_size 131072 +http_max_field_value_size 131072 +http_max_fields 1000000 +http_max_multipart_form_data_size 1073741824 +http_max_request_param_data_size 10485760 +http_max_tries 10 +http_max_uri_size 1048576 +http_native_compression_disable_checksumming_on_decompress 0 +http_receive_timeout 30 +http_response_buffer_size 0 +http_retry_initial_backoff_ms 100 +http_retry_max_backoff_ms 10000 +http_send_timeout 30 +http_skip_not_found_url_for_globs 1 +http_wait_end_of_query 0 +http_write_exception_in_output_format 1 +http_zlib_compression_level 3 +idle_connection_timeout 3600 +ignore_cold_parts_seconds 0 +ignore_data_skipping_indices +ignore_on_cluster_for_replicated_access_entities_queries 0 +ignore_on_cluster_for_replicated_udf_queries 0 +implicit_transaction 0 +input_format_allow_errors_num 0 +input_format_allow_errors_ratio 0 +input_format_allow_seeks 1 +input_format_arrow_allow_missing_columns 1 +input_format_arrow_case_insensitive_column_matching 0 +input_format_arrow_import_nested 0 +input_format_arrow_skip_columns_with_unsupported_types_in_schema_inference 0 +input_format_avro_allow_missing_fields 0 +input_format_avro_null_as_default 0 +input_format_bson_skip_fields_with_unsupported_types_in_schema_inference 0 +input_format_capn_proto_skip_fields_with_unsupported_types_in_schema_inference 0 +input_format_csv_allow_cr_end_of_line 0 +input_format_csv_allow_variable_number_of_columns 0 +input_format_csv_allow_whitespace_or_tab_as_delimiter 0 +input_format_csv_arrays_as_nested_csv 0 +input_format_csv_detect_header 1 +input_format_csv_empty_as_default 1 +input_format_csv_enum_as_number 0 +input_format_csv_skip_first_lines 0 +input_format_csv_skip_trailing_empty_lines 0 +input_format_csv_trim_whitespaces 1 +input_format_csv_try_infer_numbers_from_strings 0 +input_format_csv_use_best_effort_in_schema_inference 1 +input_format_csv_use_default_on_bad_values 0 +input_format_custom_allow_variable_number_of_columns 0 +input_format_custom_detect_header 1 +input_format_custom_skip_trailing_empty_lines 0 +input_format_defaults_for_omitted_fields 1 +input_format_hive_text_collection_items_delimiter  +input_format_hive_text_fields_delimiter  +input_format_hive_text_map_keys_delimiter  +input_format_import_nested_json 0 +input_format_ipv4_default_on_conversion_error 0 +input_format_ipv6_default_on_conversion_error 0 +input_format_json_compact_allow_variable_number_of_columns 0 +input_format_json_defaults_for_missing_elements_in_named_tuple 1 +input_format_json_ignore_unknown_keys_in_named_tuple 1 +input_format_json_infer_incomplete_types_as_strings 1 +input_format_json_named_tuples_as_objects 1 +input_format_json_read_arrays_as_strings 1 +input_format_json_read_bools_as_numbers 1 +input_format_json_read_numbers_as_strings 1 +input_format_json_read_objects_as_strings 1 +input_format_json_try_infer_named_tuples_from_objects 1 +input_format_json_try_infer_numbers_from_strings 0 +input_format_json_validate_types_from_metadata 1 +input_format_max_bytes_to_read_for_schema_inference 33554432 +input_format_max_rows_to_read_for_schema_inference 25000 +input_format_msgpack_number_of_columns 0 +input_format_mysql_dump_map_column_names 1 +input_format_mysql_dump_table_name +input_format_native_allow_types_conversion 1 +input_format_null_as_default 1 +input_format_orc_allow_missing_columns 1 +input_format_orc_case_insensitive_column_matching 0 +input_format_orc_filter_push_down 1 +input_format_orc_import_nested 0 +input_format_orc_row_batch_size 100000 +input_format_orc_skip_columns_with_unsupported_types_in_schema_inference 0 +input_format_orc_use_fast_decoder 1 +input_format_parallel_parsing 1 +input_format_parquet_allow_missing_columns 1 +input_format_parquet_case_insensitive_column_matching 0 +input_format_parquet_filter_push_down 1 +input_format_parquet_import_nested 0 +input_format_parquet_local_file_min_bytes_for_seek 8192 +input_format_parquet_max_block_size 8192 +input_format_parquet_preserve_order 0 +input_format_parquet_skip_columns_with_unsupported_types_in_schema_inference 0 +input_format_protobuf_flatten_google_wrappers 0 +input_format_protobuf_skip_fields_with_unsupported_types_in_schema_inference 0 +input_format_record_errors_file_path +input_format_skip_unknown_fields 1 +input_format_try_infer_dates 1 +input_format_try_infer_datetimes 1 +input_format_try_infer_integers 1 +input_format_tsv_allow_variable_number_of_columns 0 +input_format_tsv_detect_header 1 +input_format_tsv_empty_as_default 0 +input_format_tsv_enum_as_number 0 +input_format_tsv_skip_first_lines 0 +input_format_tsv_skip_trailing_empty_lines 0 +input_format_tsv_use_best_effort_in_schema_inference 1 +input_format_values_accurate_types_of_literals 1 +input_format_values_allow_data_after_semicolon 0 +input_format_values_deduce_templates_of_expressions 1 +input_format_values_interpret_expressions 1 +input_format_with_names_use_header 1 +input_format_with_types_use_header 1 +insert_allow_materialized_columns 0 +insert_deduplicate 1 +insert_deduplication_token +insert_distributed_one_random_shard 0 +insert_distributed_sync 0 +insert_distributed_timeout 0 +insert_keeper_fault_injection_probability 0 +insert_keeper_fault_injection_seed 0 +insert_keeper_max_retries 20 +insert_keeper_retry_initial_backoff_ms 100 +insert_keeper_retry_max_backoff_ms 10000 +insert_null_as_default 1 +insert_quorum 0 +insert_quorum_parallel 1 +insert_quorum_timeout 600000 +insert_shard_id 0 +interactive_delay 100000 +intersect_default_mode ALL +interval_output_format numeric +join_algorithm default +join_any_take_last_row 0 +join_default_strictness ALL +join_on_disk_max_files_to_merge 64 +join_overflow_mode throw +join_use_nulls 0 +joined_subquery_requires_alias 1 +kafka_disable_num_consumers_limit 0 +kafka_max_wait_ms 5000 +keeper_map_strict_mode 0 +legacy_column_name_of_tuple_literal 0 +limit 0 +live_view_heartbeat_interval 15 +load_balancing random +load_balancing_first_offset 0 +load_marks_asynchronously 0 +local_filesystem_read_method pread_threadpool +local_filesystem_read_prefetch 0 +lock_acquire_timeout 120 +log_comment +log_formatted_queries 0 +log_processors_profiles 0 +log_profile_events 1 +log_queries 1 +log_queries_cut_to_length 100000 +log_queries_min_query_duration_ms 0 +log_queries_min_type QUERY_START +log_queries_probability 1 +log_query_settings 1 +log_query_threads 0 +log_query_views 1 +low_cardinality_allow_in_native_format 1 +low_cardinality_max_dictionary_size 8192 +low_cardinality_use_single_dictionary_for_part 0 +materialize_ttl_after_modify 1 +materialized_views_ignore_errors 0 +max_alter_threads \'auto(16)\' +max_analyze_depth 5000 +max_ast_depth 1000 +max_ast_elements 50000 +max_backup_bandwidth 0 +max_block_size 65409 +max_bytes_before_external_group_by 0 +max_bytes_before_external_sort 0 +max_bytes_before_remerge_sort 1000000000 +max_bytes_in_distinct 0 +max_bytes_in_join 0 +max_bytes_in_set 0 +max_bytes_to_read 0 +max_bytes_to_read_leaf 0 +max_bytes_to_sort 0 +max_bytes_to_transfer 0 +max_columns_to_read 0 +max_compress_block_size 1048576 +max_concurrent_queries_for_all_users 0 +max_concurrent_queries_for_user 0 +max_distributed_connections 1024 +max_distributed_depth 5 +max_download_buffer_size 10485760 +max_download_threads 4 +max_entries_for_hash_table_stats 10000 +max_execution_speed 0 +max_execution_speed_bytes 0 +max_execution_time 0 +max_execution_time_leaf 0 +max_expanded_ast_elements 500000 +max_fetch_partition_retries_count 5 +max_final_threads \'auto(16)\' +max_http_get_redirects 0 +max_hyperscan_regexp_length 0 +max_hyperscan_regexp_total_length 0 +max_insert_block_size 1048449 +max_insert_delayed_streams_for_parallel_write 0 +max_insert_threads 0 +max_joined_block_size_rows 65409 +max_limit_for_ann_queries 1000000 +max_live_view_insert_blocks_before_refresh 64 +max_local_read_bandwidth 0 +max_local_write_bandwidth 0 +max_memory_usage 0 +max_memory_usage_for_all_queries 0 +max_memory_usage_for_user 0 +max_network_bandwidth 0 +max_network_bandwidth_for_all_users 0 +max_network_bandwidth_for_user 0 +max_network_bytes 0 +max_number_of_partitions_for_independent_aggregation 128 +max_parallel_replicas 1 +max_parser_depth 1000 +max_partition_size_to_drop 50000000000 +max_partitions_per_insert_block 100 +max_partitions_to_read -1 +max_pipeline_depth 0 +max_query_size 262144 +max_read_buffer_size 1048576 +max_read_buffer_size_local_fs 131072 +max_read_buffer_size_remote_fs 0 +max_remote_read_network_bandwidth 0 +max_remote_read_network_bandwidth_for_server 0 +max_remote_write_network_bandwidth 0 +max_remote_write_network_bandwidth_for_server 0 +max_replica_delay_for_distributed_queries 300 +max_replicated_fetches_network_bandwidth_for_server 0 +max_replicated_sends_network_bandwidth_for_server 0 +max_result_bytes 0 +max_result_rows 0 +max_rows_in_distinct 0 +max_rows_in_join 0 +max_rows_in_set 0 +max_rows_in_set_to_optimize_join 100000 +max_rows_to_group_by 0 +max_rows_to_read 0 +max_rows_to_read_leaf 0 +max_rows_to_sort 0 +max_rows_to_transfer 0 +max_sessions_for_user 0 +max_size_to_preallocate_for_aggregation 100000000 +max_streams_for_merge_tree_reading 0 +max_streams_multiplier_for_merge_tables 5 +max_streams_to_max_threads_ratio 1 +max_subquery_depth 100 +max_table_size_to_drop 50000000000 +max_temporary_columns 0 +max_temporary_data_on_disk_size_for_query 0 +max_temporary_data_on_disk_size_for_user 0 +max_temporary_non_const_columns 0 +max_threads \'auto(16)\' +max_threads_for_annoy_index_creation 4 +max_threads_for_indexes 0 +max_untracked_memory 4194304 +memory_overcommit_ratio_denominator 1073741824 +memory_overcommit_ratio_denominator_for_user 1073741824 +memory_profiler_sample_max_allocation_size 0 +memory_profiler_sample_min_allocation_size 0 +memory_profiler_sample_probability 0 +memory_profiler_step 4194304 +memory_tracker_fault_probability 0 +memory_usage_overcommit_max_wait_microseconds 5000000 +merge_tree_clear_old_parts_interval_seconds 1 +merge_tree_clear_old_temporary_directories_interval_seconds 60 +merge_tree_coarse_index_granularity 8 +merge_tree_compact_parts_min_granules_to_multibuffer_read 16 +merge_tree_determine_task_size_by_prewhere_columns 1 +merge_tree_max_bytes_to_use_cache 2013265920 +merge_tree_max_rows_to_use_cache 1048576 +merge_tree_min_bytes_for_concurrent_read 251658240 +merge_tree_min_bytes_for_concurrent_read_for_remote_filesystem 251658240 +merge_tree_min_bytes_for_seek 0 +merge_tree_min_bytes_per_task_for_remote_reading 4194304 +merge_tree_min_rows_for_concurrent_read 163840 +merge_tree_min_rows_for_concurrent_read_for_remote_filesystem 163840 +merge_tree_min_rows_for_seek 0 +merge_tree_use_const_size_tasks_for_remote_reading 1 +metrics_perf_events_enabled 0 +metrics_perf_events_list +min_bytes_to_use_direct_io 0 +min_bytes_to_use_mmap_io 0 +min_chunk_bytes_for_parallel_parsing 10485760 +min_compress_block_size 65536 +min_count_to_compile_aggregate_expression 3 +min_count_to_compile_expression 3 +min_count_to_compile_sort_description 3 +min_execution_speed 0 +min_execution_speed_bytes 0 +min_free_disk_space_for_temporary_data 0 +min_hit_rate_to_use_consecutive_keys_optimization 0.5 +min_insert_block_size_bytes 268402944 +min_insert_block_size_bytes_for_materialized_views 0 +min_insert_block_size_rows 1048449 +min_insert_block_size_rows_for_materialized_views 0 +move_all_conditions_to_prewhere 1 +move_primary_key_columns_to_end_of_prewhere 1 +multiple_joins_rewriter_version 0 +multiple_joins_try_to_keep_original_names 0 +mutations_execute_nondeterministic_on_initiator 0 +mutations_execute_subqueries_on_initiator 0 +mutations_max_literal_size_to_replace 16384 +mutations_sync 0 +mysql_datatypes_support_level +mysql_map_fixed_string_to_text_in_show_columns 0 +mysql_map_string_to_text_in_show_columns 0 +mysql_max_rows_to_insert 65536 +network_compression_method LZ4 +network_zstd_compression_level 1 +normalize_function_names 1 +number_of_mutations_to_delay 0 +number_of_mutations_to_throw 0 +odbc_bridge_connection_pool_size 16 +odbc_bridge_use_connection_pooling 1 +odbc_max_field_size 0 +offset 0 +opentelemetry_start_trace_probability 0 +opentelemetry_trace_processors 0 +optimize_aggregation_in_order 0 +optimize_aggregators_of_group_by_keys 1 +optimize_append_index 0 +optimize_arithmetic_operations_in_aggregate_functions 1 +optimize_count_from_files 1 +optimize_distinct_in_order 1 +optimize_distributed_group_by_sharding_key 1 +optimize_duplicate_order_by_and_distinct 0 +optimize_functions_to_subcolumns 0 +optimize_fuse_sum_count_avg 0 +optimize_group_by_constant_keys 1 +optimize_group_by_function_keys 1 +optimize_if_chain_to_multiif 0 +optimize_if_transform_strings_to_enum 0 +optimize_injective_functions_inside_uniq 1 +optimize_min_equality_disjunction_chain_length 3 +optimize_min_inequality_conjunction_chain_length 3 +optimize_monotonous_functions_in_order_by 0 +optimize_move_functions_out_of_any 0 +optimize_move_to_prewhere 1 +optimize_move_to_prewhere_if_final 0 +optimize_multiif_to_if 1 +optimize_normalize_count_variants 1 +optimize_on_insert 1 +optimize_or_like_chain 0 +optimize_read_in_order 1 +optimize_read_in_window_order 1 +optimize_redundant_functions_in_order_by 1 +optimize_respect_aliases 1 +optimize_rewrite_aggregate_function_with_if 1 +optimize_rewrite_array_exists_to_has 0 +optimize_rewrite_sum_if_to_count_if 0 +optimize_skip_merged_partitions 0 +optimize_skip_unused_shards 0 +optimize_skip_unused_shards_limit 1000 +optimize_skip_unused_shards_nesting 0 +optimize_skip_unused_shards_rewrite_in 1 +optimize_sorting_by_input_stream_properties 1 +optimize_substitute_columns 0 +optimize_syntax_fuse_functions 0 +optimize_throw_if_noop 0 +optimize_trivial_approximate_count_query 0 +optimize_trivial_count_query 1 +optimize_trivial_insert_select 1 +optimize_uniq_to_count 1 +optimize_use_implicit_projections 1 +optimize_use_projections 1 +optimize_using_constraints 0 +os_thread_priority 0 +output_format_arrow_compression_method lz4_frame +output_format_arrow_fixed_string_as_fixed_byte_array 1 +output_format_arrow_low_cardinality_as_dictionary 0 +output_format_arrow_string_as_string 0 +output_format_avro_codec +output_format_avro_rows_in_file 1 +output_format_avro_string_column_pattern +output_format_avro_sync_interval 16384 +output_format_bson_string_as_string 0 +output_format_csv_crlf_end_of_line 0 +output_format_decimal_trailing_zeros 0 +output_format_enable_streaming 0 +output_format_json_array_of_rows 0 +output_format_json_escape_forward_slashes 1 +output_format_json_named_tuples_as_objects 1 +output_format_json_quote_64bit_floats 0 +output_format_json_quote_64bit_integers 1 +output_format_json_quote_decimals 0 +output_format_json_quote_denormals 0 +output_format_json_skip_null_value_in_named_tuples 0 +output_format_json_validate_utf8 0 +output_format_markdown_escape_special_characters 0 +output_format_msgpack_uuid_representation ext +output_format_orc_compression_method lz4 +output_format_orc_row_index_stride 10000 +output_format_orc_string_as_string 0 +output_format_parallel_formatting 1 +output_format_parquet_batch_size 1024 +output_format_parquet_compliant_nested_types 1 +output_format_parquet_compression_method lz4 +output_format_parquet_data_page_size 1048576 +output_format_parquet_fixed_string_as_fixed_byte_array 1 +output_format_parquet_parallel_encoding 1 +output_format_parquet_row_group_size 1000000 +output_format_parquet_row_group_size_bytes 536870912 +output_format_parquet_string_as_string 0 +output_format_parquet_use_custom_encoder 0 +output_format_parquet_version 2.latest +output_format_pretty_color 1 +output_format_pretty_grid_charset UTF-8 +output_format_pretty_max_column_pad_width 250 +output_format_pretty_max_rows 10000 +output_format_pretty_max_value_width 10000 +output_format_pretty_row_numbers 0 +output_format_protobuf_nullables_with_google_wrappers 0 +output_format_schema +output_format_sql_insert_include_column_names 1 +output_format_sql_insert_max_batch_size 65409 +output_format_sql_insert_quote_names 1 +output_format_sql_insert_table_name table +output_format_sql_insert_use_replace 0 +output_format_tsv_crlf_end_of_line 0 +output_format_write_statistics 1 +parallel_distributed_insert_select 0 +parallel_replica_offset 0 +parallel_replicas_count 0 +parallel_replicas_custom_key +parallel_replicas_custom_key_filter_type default +parallel_replicas_for_non_replicated_merge_tree 0 +parallel_replicas_min_number_of_granules_to_enable 0 +parallel_replicas_min_number_of_rows_per_replica 0 +parallel_replicas_single_task_marks_count_multiplier 2 +parallel_view_processing 0 +parallelize_output_from_storages 1 +parsedatetime_parse_without_leading_zeros 1 +partial_merge_join_left_table_buffer_bytes 0 +partial_merge_join_optimizations 0 +partial_merge_join_rows_in_right_blocks 65536 +partial_result_on_first_cancel 0 +parts_to_delay_insert 0 +parts_to_throw_insert 0 +periodic_live_view_refresh 60 +poll_interval 10 +postgresql_connection_pool_auto_close_connection 0 +postgresql_connection_pool_size 16 +postgresql_connection_pool_wait_timeout 5000 +precise_float_parsing 0 +prefer_column_name_to_alias 0 +prefer_global_in_and_join 0 +prefer_localhost_replica 1 +prefer_warmed_unmerged_parts_seconds 0 +preferred_block_size_bytes 1000000 +preferred_max_column_in_block_size_bytes 0 +preferred_optimize_projection_name +prefetch_buffer_size 1048576 +print_pretty_type_names 0 +priority 0 +query_cache_compress_entries 1 +query_cache_max_entries 0 +query_cache_max_size_in_bytes 0 +query_cache_min_query_duration 0 +query_cache_min_query_runs 0 +query_cache_nondeterministic_function_handling throw +query_cache_share_between_users 0 +query_cache_squash_partial_results 1 +query_cache_store_results_of_queries_with_nondeterministic_functions 0 +query_cache_ttl 60 +query_plan_aggregation_in_order 1 +query_plan_enable_multithreading_after_window_functions 1 +query_plan_enable_optimizations 1 +query_plan_execute_functions_after_sorting 1 +query_plan_filter_push_down 1 +query_plan_lift_up_array_join 1 +query_plan_lift_up_union 1 +query_plan_max_optimizations_to_apply 10000 +query_plan_merge_expressions 1 +query_plan_optimize_primary_key 1 +query_plan_optimize_projection 1 +query_plan_push_down_limit 1 +query_plan_read_in_order 1 +query_plan_remove_redundant_distinct 1 +query_plan_remove_redundant_sorting 1 +query_plan_reuse_storage_ordering_for_window_functions 1 +query_plan_split_filter 1 +query_profiler_cpu_time_period_ns 1000000000 +query_profiler_real_time_period_ns 1000000000 +queue_max_wait_ms 0 +rabbitmq_max_wait_ms 5000 +read_backoff_max_throughput 1048576 +read_backoff_min_concurrency 1 +read_backoff_min_events 2 +read_backoff_min_interval_between_events_ms 1000 +read_backoff_min_latency_ms 1000 +read_from_filesystem_cache_if_exists_otherwise_bypass_cache 0 +read_in_order_two_level_merge_threshold 100 +read_overflow_mode throw +read_overflow_mode_leaf throw +read_priority 0 +readonly 0 +receive_data_timeout_ms 2000 +receive_timeout 300 +regexp_dict_allow_hyperscan 1 +regexp_dict_flag_case_insensitive 0 +regexp_dict_flag_dotall 0 +regexp_max_matches_per_row 1000 +reject_expensive_hyperscan_regexps 1 +remerge_sort_lowered_memory_bytes_ratio 2 +remote_filesystem_read_method threadpool +remote_filesystem_read_prefetch 1 +remote_fs_read_backoff_max_tries 5 +remote_fs_read_max_backoff_ms 10000 +remote_read_min_bytes_for_seek 4194304 +rename_files_after_processing +replace_running_query 0 +replace_running_query_max_wait_ms 5000 +replication_alter_columns_timeout 60 +replication_alter_partitions_sync 1 +replication_wait_for_inactive_replica_timeout 120 +restore_threads 16 +result_overflow_mode throw +rewrite_count_distinct_if_with_count_distinct_implementation 0 +s3_allow_parallel_part_upload 1 +s3_check_objects_after_upload 0 +s3_create_new_file_on_insert 0 +s3_disable_checksum 0 +s3_http_connection_pool_size 1000 +s3_list_object_keys_size 1000 +s3_max_connections 1024 +s3_max_get_burst 0 +s3_max_get_rps 0 +s3_max_inflight_parts_for_one_file 20 +s3_max_put_burst 0 +s3_max_put_rps 0 +s3_max_redirects 10 +s3_max_single_part_upload_size 33554432 +s3_max_single_read_retries 4 +s3_max_unexpected_write_error_retries 4 +s3_max_upload_part_size 5368709120 +s3_min_upload_part_size 16777216 +s3_request_timeout_ms 30000 +s3_retry_attempts 100 +s3_skip_empty_files 0 +s3_strict_upload_part_size 0 +s3_throw_on_zero_files_match 0 +s3_truncate_on_insert 0 +s3_upload_part_size_multiply_factor 2 +s3_upload_part_size_multiply_parts_count_threshold 500 +s3_use_adaptive_timeouts 1 +s3queue_default_zookeeper_path /clickhouse/s3queue/ +s3queue_enable_logging_to_s3queue_log 0 +schema_inference_cache_require_modification_time_for_url 1 +schema_inference_hints +schema_inference_make_columns_nullable 1 +schema_inference_mode default +schema_inference_use_cache_for_azure 1 +schema_inference_use_cache_for_file 1 +schema_inference_use_cache_for_hdfs 1 +schema_inference_use_cache_for_s3 1 +schema_inference_use_cache_for_url 1 +select_sequential_consistency 0 +send_logs_level fatal +send_logs_source_regexp +send_progress_in_http_headers 0 +send_timeout 300 +session_timezone +set_overflow_mode throw +short_circuit_function_evaluation enable +show_table_uuid_in_table_create_query_if_not_nil 0 +single_join_prefer_left_table 1 +skip_download_if_exceeds_query_cache 1 +skip_unavailable_shards 0 +sleep_after_receiving_query_ms 0 +sleep_in_send_data_ms 0 +sleep_in_send_tables_status_ms 0 +sort_overflow_mode throw +splitby_max_substrings_includes_remaining_string 0 +stop_refreshable_materialized_views_on_startup 0 +storage_file_read_method pread +storage_system_stack_trace_pipe_read_timeout_ms 100 +stream_flush_interval_ms 7500 +stream_like_engine_allow_direct_select 0 +stream_like_engine_insert_queue +stream_poll_timeout_ms 500 +system_events_show_zero_values 0 +table_function_remote_max_addresses 1000 +tcp_keep_alive_timeout 290 +temporary_files_codec LZ4 +temporary_live_view_timeout 1 +throw_if_no_data_to_insert 1 +throw_on_error_from_cache_on_write_operations 0 +throw_on_max_partitions_per_insert_block 1 +throw_on_unsupported_query_inside_transaction 1 +timeout_before_checking_execution_speed 10 +timeout_overflow_mode throw +timeout_overflow_mode_leaf throw +totals_auto_threshold 0.5 +totals_mode after_having_exclusive +trace_profile_events 0 +transfer_overflow_mode throw +transform_null_in 0 +union_default_mode +unknown_packet_in_send_data 0 +use_cache_for_count_from_files 1 +use_client_time_zone 0 +use_compact_format_in_distributed_parts_names 1 +use_concurrency_control 1 +use_hedged_requests 1 +use_index_for_in_with_subqueries 1 +use_index_for_in_with_subqueries_max_values 0 +use_local_cache_for_remote_storage 1 +use_mysql_types_in_show_columns 0 +use_query_cache 0 +use_skip_indexes 1 +use_skip_indexes_if_final 0 +use_structure_from_insertion_table_in_table_functions 2 +use_uncompressed_cache 0 +use_with_fill_by_sorting_prefix 1 +validate_polygons 1 +wait_changes_become_visible_after_commit_mode wait_unknown +wait_for_async_insert 1 +wait_for_async_insert_timeout 120 +wait_for_window_view_fire_signal_timeout 10 +window_view_clean_interval 60 +window_view_heartbeat_interval 15 +workload default +zstd_window_log_max 0 diff --git a/tests/queries/0_stateless/02995_new_settings_history.reference b/tests/queries/0_stateless/02995_new_settings_history.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02995_new_settings_history.sh b/tests/queries/0_stateless/02995_new_settings_history.sh new file mode 100755 index 00000000000..8de98c55b6a --- /dev/null +++ b/tests/queries/0_stateless/02995_new_settings_history.sh @@ -0,0 +1,46 @@ +#!/usr/bin/env bash +# Tags: no-tsan, no-asan, no-ubsan, no-msan, no-cpu-aarch64, no-random-settings +# Some settings can be different for builds with sanitizers or aarch64 + +CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +# shellcheck source=../shell_config.sh +. "$CUR_DIR"/../shell_config.sh + +# Note that this is a broad check. A per version check is done in the upgrade test +# Baseline generated with 23.12.1 +# clickhouse local --query "select name, default from system.settings order by name format TSV" > 02995_baseline_23_12_1.tsv +$CLICKHOUSE_LOCAL --query " + WITH old_settings AS + ( + SELECT * FROM file('${CUR_DIR}/02995_baseline_23_12_1.tsv', 'TSV', 'name String, default String') + ), + new_settings AS + ( + -- Ignore settings that depend on the machine config (max_threads and similar) + SELECT name, default FROM system.settings WHERE default NOT LIKE '%auto(%' + ) + SELECT * FROM + ( + SELECT 'PLEASE ADD THE NEW SETTING TO SettingsChangesHistory.h: ' || name || ' WAS ADDED', + FROM new_settings + WHERE (name NOT IN ( + SELECT name + FROM old_settings + )) AND (name NOT IN ( + SELECT arrayJoin(tupleElement(changes, 'name')) + FROM system.settings_changes + WHERE splitByChar('.', version())[1] >= '24' + )) + UNION ALL + ( + SELECT 'PLEASE ADD THE SETTING VALUE CHANGE TO SettingsChangesHistory.h: ' || name || ' WAS CHANGED FROM ' || old_settings.default || ' TO ' || new_settings.default, + FROM new_settings + LEFT JOIN old_settings ON new_settings.name = old_settings.name + WHERE (new_settings.default != old_settings.default) AND (name NOT IN ( + SELECT arrayJoin(tupleElement(changes, 'name')) + FROM system.settings_changes + WHERE splitByChar('.', version())[1] >= '24' + )) + ) + ) +" diff --git a/tests/queries/0_stateless/02996_analyzer_prewhere_projection.reference b/tests/queries/0_stateless/02996_analyzer_prewhere_projection.reference new file mode 100644 index 00000000000..72749c905a3 --- /dev/null +++ b/tests/queries/0_stateless/02996_analyzer_prewhere_projection.reference @@ -0,0 +1 @@ +1 1 1 diff --git a/tests/queries/0_stateless/02996_analyzer_prewhere_projection.sql b/tests/queries/0_stateless/02996_analyzer_prewhere_projection.sql new file mode 100644 index 00000000000..9d676001010 --- /dev/null +++ b/tests/queries/0_stateless/02996_analyzer_prewhere_projection.sql @@ -0,0 +1,7 @@ +SET allow_suspicious_low_cardinality_types=1; + +CREATE TABLE t__fuzz_0 (`i` LowCardinality(Int32), `j` Int32, `k` Int32, PROJECTION p (SELECT * ORDER BY j)) ENGINE = MergeTree ORDER BY i SETTINGS index_granularity = 1; +INSERT INTO t__fuzz_0 Select number, number, number FROM numbers(100); + +SELECT * FROM t__fuzz_0 PREWHERE 7 AND (i < 2147483647) AND (j IN (2147483646, -2, 1)) +SETTINGS allow_experimental_analyzer = true; diff --git a/tests/queries/0_stateless/02996_nullable_arrayReduce.reference b/tests/queries/0_stateless/02996_nullable_arrayReduce.reference new file mode 100644 index 00000000000..96afb8546ef --- /dev/null +++ b/tests/queries/0_stateless/02996_nullable_arrayReduce.reference @@ -0,0 +1,15 @@ +-- { echoOn } +SELECT arrayReduce('sum', []::Array(UInt8)) as a, toTypeName(a); +0 UInt64 +SELECT arrayReduce('sumOrNull', []::Array(UInt8)) as a, toTypeName(a); +\N Nullable(UInt64) +SELECT arrayReduce('sum', [NULL]::Array(Nullable(UInt8))) as a, toTypeName(a); +\N Nullable(UInt64) +SELECT arrayReduce('sum', [NULL, 10]::Array(Nullable(UInt8))) as a, toTypeName(a); +10 Nullable(UInt64) +SELECT arrayReduce('any_respect_nulls', [NULL, 10]::Array(Nullable(UInt8))) as a, toTypeName(a); +\N Nullable(UInt8) +SELECT arrayReduce('any_respect_nulls', [10, NULL]::Array(Nullable(UInt8))) as a, toTypeName(a); +10 Nullable(UInt8) +SELECT arrayReduce('median', [toLowCardinality(toNullable(8))]) as t, toTypeName(t); +8 Nullable(Float64) diff --git a/tests/queries/0_stateless/02996_nullable_arrayReduce.sql b/tests/queries/0_stateless/02996_nullable_arrayReduce.sql new file mode 100644 index 00000000000..8f69296dbe5 --- /dev/null +++ b/tests/queries/0_stateless/02996_nullable_arrayReduce.sql @@ -0,0 +1,17 @@ +-- https://github.com/ClickHouse/ClickHouse/issues/59600 +SELECT arrayReduce(toNullable('stddevSampOrNull'), [1]); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT arrayReduce(toNullable('median'), [toDecimal32OrNull(toFixedString('1', 1), 2), 8]); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT toFixedString('--- Int Empty ---', toLowCardinality(17)), arrayReduce(toNullable('avgOrNull'), [1]); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT arrayReduce('any', toNullable(3)); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } +SELECT arrayReduce(toLowCardinality('median'), [toLowCardinality(toNullable(8))]); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT } + +-- { echoOn } +SELECT arrayReduce('sum', []::Array(UInt8)) as a, toTypeName(a); +SELECT arrayReduce('sumOrNull', []::Array(UInt8)) as a, toTypeName(a); +SELECT arrayReduce('sum', [NULL]::Array(Nullable(UInt8))) as a, toTypeName(a); +SELECT arrayReduce('sum', [NULL, 10]::Array(Nullable(UInt8))) as a, toTypeName(a); +SELECT arrayReduce('any_respect_nulls', [NULL, 10]::Array(Nullable(UInt8))) as a, toTypeName(a); +SELECT arrayReduce('any_respect_nulls', [10, NULL]::Array(Nullable(UInt8))) as a, toTypeName(a); + +SELECT arrayReduce('median', [toLowCardinality(toNullable(8))]) as t, toTypeName(t); +-- { echoOff } diff --git a/tests/queries/0_stateless/02997_projections_formatting.reference b/tests/queries/0_stateless/02997_projections_formatting.reference new file mode 100644 index 00000000000..6a60da1089a --- /dev/null +++ b/tests/queries/0_stateless/02997_projections_formatting.reference @@ -0,0 +1,26 @@ +CREATE TEMPORARY TABLE t_proj +( + `t` DateTime, + `id` UInt64, + PROJECTION p + ( + SELECT + id, + t + ORDER BY toStartOfDay(t) + ) +) +ENGINE = MergeTree +ORDER BY id +CREATE TEMPORARY TABLE t_proj2 +( + `a` UInt32, + `b` UInt32, + PROJECTION p + ( + SELECT a + ORDER BY b * 2 + ) +) +ENGINE = MergeTree +ORDER BY a diff --git a/tests/queries/0_stateless/02997_projections_formatting.sql b/tests/queries/0_stateless/02997_projections_formatting.sql new file mode 100644 index 00000000000..b593c2576b1 --- /dev/null +++ b/tests/queries/0_stateless/02997_projections_formatting.sql @@ -0,0 +1,5 @@ +CREATE TEMPORARY TABLE t_proj (t DateTime, id UInt64, PROJECTION p (SELECT id, t ORDER BY toStartOfDay(t))) ENGINE = MergeTree ORDER BY id; +SHOW CREATE TEMPORARY TABLE t_proj FORMAT TSVRaw; + +CREATE TEMPORARY TABLE t_proj2 (a UInt32, b UInt32, PROJECTION p (SELECT a ORDER BY b * 2)) ENGINE = MergeTree ORDER BY a; +SHOW CREATE TEMPORARY TABLE t_proj2 FORMAT TSVRaw; diff --git a/utils/check-style/check-style b/utils/check-style/check-style index a71dac91683..6c12970c4bb 100755 --- a/utils/check-style/check-style +++ b/utils/check-style/check-style @@ -448,8 +448,3 @@ find $ROOT_PATH/{src,base,programs,utils} -name '*.h' -or -name '*.cpp' | grep -vP $EXCLUDE_DIRS | xargs grep -P 'Sql|Html|Xml|Cpu|Tcp|Udp|Http|Db|Json|Yaml' | grep -v -P 'RabbitMQ|Azure|Aws|aws|Avro|IO/S3' && echo "Abbreviations such as SQL, XML, HTTP, should be in all caps. For example, SQL is right, Sql is wrong. XMLHttpRequest is very wrong." - -find $ROOT_PATH/{src,base,programs,utils} -name '*.h' -or -name '*.cpp' | - grep -vP $EXCLUDE_DIRS | - xargs grep -F -i 'ErrorCodes::LOGICAL_ERROR, "Logical error:' && - echo "If an exception has LOGICAL_ERROR code, there is no need to include the text 'Logical error' in the exception message, because then the phrase 'Logical error' will be printed twice." diff --git a/utils/keeper-data-dumper/main.cpp b/utils/keeper-data-dumper/main.cpp index 351a4ab90bc..21626665a42 100644 --- a/utils/keeper-data-dumper/main.cpp +++ b/utils/keeper-data-dumper/main.cpp @@ -63,11 +63,11 @@ int main(int argc, char *argv[]) ResponsesQueue queue(std::numeric_limits::max()); SnapshotsQueue snapshots_queue{1}; CoordinationSettingsPtr settings = std::make_shared(); - KeeperContextPtr keeper_context = std::make_shared(true); + KeeperContextPtr keeper_context = std::make_shared(true, settings); keeper_context->setLogDisk(std::make_shared("LogDisk", argv[2])); keeper_context->setSnapshotDisk(std::make_shared("SnapshotDisk", argv[1])); - auto state_machine = std::make_shared(queue, snapshots_queue, settings, keeper_context, nullptr); + auto state_machine = std::make_shared(queue, snapshots_queue, keeper_context, nullptr); state_machine->init(); size_t last_commited_index = state_machine->last_commit_index();