From 566c08086afd037e9b7ce244ac6c222d286af48d Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Fri, 6 May 2022 14:44:00 +0000 Subject: [PATCH 001/239] support Object type inside other types --- src/Columns/ColumnArray.h | 2 + src/Columns/ColumnObject.h | 5 +- src/Columns/ColumnTuple.cpp | 6 + src/Columns/ColumnTuple.h | 1 + src/Columns/IColumn.h | 2 + src/DataTypes/DataTypeArray.h | 1 + src/DataTypes/DataTypeObject.h | 1 + src/DataTypes/DataTypeTuple.cpp | 5 + src/DataTypes/DataTypeTuple.h | 1 + src/DataTypes/IDataType.h | 2 + src/DataTypes/ObjectUtils.cpp | 267 ++++++++++++++---- src/DataTypes/ObjectUtils.h | 39 +-- src/Functions/FunctionsConversion.h | 5 +- src/Interpreters/InterpreterCreateQuery.cpp | 4 +- src/Processors/Formats/IRowInputFormat.cpp | 5 +- .../Formats/Impl/ValuesBlockInputFormat.cpp | 4 +- src/Storages/MergeTree/MergeTask.cpp | 2 +- src/Storages/MergeTree/MergeTreeData.cpp | 16 +- src/Storages/MergeTree/MergeTreeData.h | 6 +- .../MergeTree/MergeTreeDataWriter.cpp | 12 +- src/Storages/MergeTree/MergeTreeDataWriter.h | 2 - src/Storages/MergeTree/MergeTreeSink.cpp | 4 +- .../MergeTree/ReplicatedMergeTreeSink.cpp | 5 +- src/Storages/StorageDistributed.cpp | 6 +- src/Storages/StorageInMemoryMetadata.cpp | 6 +- src/Storages/StorageMemory.cpp | 6 +- src/Storages/StorageSnapshot.cpp | 2 +- src/Storages/getStructureOfRemoteTable.cpp | 2 +- .../01825_type_json_in_array.reference | 17 ++ .../0_stateless/01825_type_json_in_array.sql | 27 ++ 30 files changed, 339 insertions(+), 124 deletions(-) create mode 100644 tests/queries/0_stateless/01825_type_json_in_array.reference create mode 100644 tests/queries/0_stateless/01825_type_json_in_array.sql diff --git a/src/Columns/ColumnArray.h b/src/Columns/ColumnArray.h index 1d88b2e6a26..1657160fbae 100644 --- a/src/Columns/ColumnArray.h +++ b/src/Columns/ColumnArray.h @@ -168,6 +168,8 @@ public: void getIndicesOfNonDefaultRows(Offsets & indices, size_t from, size_t limit) const override; + void finalize() override { data->finalize(); } + bool isCollationSupported() const override { return getData().isCollationSupported(); } size_t getNumberOfDimensions() const; diff --git a/src/Columns/ColumnObject.h b/src/Columns/ColumnObject.h index 9d61b165042..27a00e44698 100644 --- a/src/Columns/ColumnObject.h +++ b/src/Columns/ColumnObject.h @@ -177,8 +177,6 @@ public: Subcolumns & getSubcolumns() { return subcolumns; } PathsInData getKeys() const; - /// Finalizes all subcolumns. - void finalize(); bool isFinalized() const; /// Part of interface @@ -199,6 +197,9 @@ public: Field operator[](size_t n) const override; void get(size_t n, Field & res) const override; + /// Finalizes all subcolumns. + void finalize() override; + /// All other methods throw exception. ColumnPtr decompress() const override { throwMustBeConcrete(); } diff --git a/src/Columns/ColumnTuple.cpp b/src/Columns/ColumnTuple.cpp index 7c8d73edd16..8d6b2dbdbd9 100644 --- a/src/Columns/ColumnTuple.cpp +++ b/src/Columns/ColumnTuple.cpp @@ -572,4 +572,10 @@ SerializationInfoPtr ColumnTuple::getSerializationInfo() const return std::make_shared(std::move(infos), SerializationInfo::Settings{}); } +void ColumnTuple::finalize() +{ + for (auto & column : columns) + column->finalize(); +} + } diff --git a/src/Columns/ColumnTuple.h b/src/Columns/ColumnTuple.h index e6797c2582c..7b554ed2fea 100644 --- a/src/Columns/ColumnTuple.h +++ b/src/Columns/ColumnTuple.h @@ -103,6 +103,7 @@ public: double getRatioOfDefaultRows(double sample_ratio) const override; void getIndicesOfNonDefaultRows(Offsets & indices, size_t from, size_t limit) const override; SerializationInfoPtr getSerializationInfo() const override; + void finalize() override; size_t tupleSize() const { return columns.size(); } diff --git a/src/Columns/IColumn.h b/src/Columns/IColumn.h index f4986799d47..4bf364d9b52 100644 --- a/src/Columns/IColumn.h +++ b/src/Columns/IColumn.h @@ -442,6 +442,8 @@ public: return getPtr(); } + /// Some columns may require finalization before using of other operations. + virtual void finalize() {} [[nodiscard]] static MutablePtr mutate(Ptr ptr) { diff --git a/src/DataTypes/DataTypeArray.h b/src/DataTypes/DataTypeArray.h index 122ac8e03a3..033a657c845 100644 --- a/src/DataTypes/DataTypeArray.h +++ b/src/DataTypes/DataTypeArray.h @@ -48,6 +48,7 @@ public: bool textCanContainOnlyValidUTF8() const override { return nested->textCanContainOnlyValidUTF8(); } bool isComparable() const override { return nested->isComparable(); } bool canBeComparedWithCollation() const override { return nested->canBeComparedWithCollation(); } + bool hasDynamicSubcolumns() const override { return nested->hasDynamicSubcolumns(); } bool isValueUnambiguouslyRepresentedInContiguousMemoryRegion() const override { diff --git a/src/DataTypes/DataTypeObject.h b/src/DataTypes/DataTypeObject.h index 503947c3738..2f6ad961512 100644 --- a/src/DataTypes/DataTypeObject.h +++ b/src/DataTypes/DataTypeObject.h @@ -36,6 +36,7 @@ public: bool haveSubtypes() const override { return false; } bool equals(const IDataType & rhs) const override; bool isParametric() const override { return true; } + bool hasDynamicSubcolumns() const override { return true; } SerializationPtr doGetDefaultSerialization() const override; diff --git a/src/DataTypes/DataTypeTuple.cpp b/src/DataTypes/DataTypeTuple.cpp index 558b13927c1..6b49d0c969c 100644 --- a/src/DataTypes/DataTypeTuple.cpp +++ b/src/DataTypes/DataTypeTuple.cpp @@ -233,6 +233,11 @@ bool DataTypeTuple::haveMaximumSizeOfValue() const return std::all_of(elems.begin(), elems.end(), [](auto && elem) { return elem->haveMaximumSizeOfValue(); }); } +bool DataTypeTuple::hasDynamicSubcolumns() const +{ + return std::any_of(elems.begin(), elems.end(), [](auto && elem) { return elem->hasDynamicSubcolumns(); }); +} + bool DataTypeTuple::isComparable() const { return std::all_of(elems.begin(), elems.end(), [](auto && elem) { return elem->isComparable(); }); diff --git a/src/DataTypes/DataTypeTuple.h b/src/DataTypes/DataTypeTuple.h index 009a2284a0a..0c5c1e82fab 100644 --- a/src/DataTypes/DataTypeTuple.h +++ b/src/DataTypes/DataTypeTuple.h @@ -48,6 +48,7 @@ public: bool isComparable() const override; bool textCanContainOnlyValidUTF8() const override; bool haveMaximumSizeOfValue() const override; + bool hasDynamicSubcolumns() const override; size_t getMaximumSizeOfValueInMemory() const override; size_t getSizeOfValueInMemory() const override; diff --git a/src/DataTypes/IDataType.h b/src/DataTypes/IDataType.h index fc9e50dc55b..6d1d17c9e40 100644 --- a/src/DataTypes/IDataType.h +++ b/src/DataTypes/IDataType.h @@ -287,6 +287,8 @@ public: /// Strings, Numbers, Date, DateTime, Nullable virtual bool canBeInsideLowCardinality() const { return false; } + virtual bool hasDynamicSubcolumns() const { return false; } + /// Updates avg_value_size_hint for newly read column. Uses to optimize deserialization. Zero expected for first column. static void updateAvgValueSizeHint(const IColumn & column, double & avg_value_size_hint); diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index ab4ac847a1c..430bfe2af72 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -8,6 +8,7 @@ #include #include #include +#include #include #include #include @@ -128,44 +129,109 @@ static auto extractVector(const std::vector & vec) return res; } -void convertObjectsToTuples(Block & block, const NamesAndTypesList & extended_storage_columns) +static DataTypePtr recreateTupleWithElements(const DataTypeTuple & type_tuple, const DataTypes & elements) { - std::unordered_map storage_columns_map; - for (const auto & [name, type] : extended_storage_columns) - storage_columns_map[name] = type; + return type_tuple.haveExplicitNames() + ? std::make_shared(elements, type_tuple.getElementNames()) + : std::make_shared(elements); +} - for (auto & column : block) +static std::pair convertObjectColumnToTuple( + const ColumnObject & column_object, const DataTypeObject & type_object) +{ + const auto & subcolumns = column_object.getSubcolumns(); + + if (!column_object.isFinalized()) + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Cannot convert to tuple column of type {}. Column should be finalized first", type_object.getName()); + + PathsInData tuple_paths; + DataTypes tuple_types; + Columns tuple_columns; + + for (const auto & entry : subcolumns) { - if (!isObject(column.type)) - continue; + tuple_paths.emplace_back(entry->path); + tuple_types.emplace_back(entry->data.getLeastCommonType()); + tuple_columns.emplace_back(entry->data.getFinalizedColumnPtr()); + } - const auto & column_object = assert_cast(*column.column); - const auto & subcolumns = column_object.getSubcolumns(); + return unflattenTuple(tuple_paths, tuple_types, tuple_columns); +} - if (!column_object.isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, - "Cannot convert to tuple column '{}' from type {}. Column should be finalized first", - column.name, column.type->getName()); +static std::pair recursivlyConvertDynamicColumnToTuple( + const ColumnPtr & column, const DataTypePtr & type) +{ + if (!type->hasDynamicSubcolumns()) + return {column, type}; - PathsInData tuple_paths; - DataTypes tuple_types; - Columns tuple_columns; + if (const auto * type_object = typeid_cast(type.get())) + { + const auto & column_object = assert_cast(*column); + return convertObjectColumnToTuple(column_object, *type_object); + } - for (const auto & entry : subcolumns) + if (const auto * type_array = typeid_cast(type.get())) + { + const auto & column_array = assert_cast(*column); + auto [new_column, new_type] = recursivlyConvertDynamicColumnToTuple( + column_array.getDataPtr(), type_array->getNestedType()); + + return { - tuple_paths.emplace_back(entry->path); - tuple_types.emplace_back(entry->data.getLeastCommonType()); - tuple_columns.emplace_back(entry->data.getFinalizedColumnPtr()); + ColumnArray::create(std::move(new_column), column_array.getOffsetsPtr()), + std::make_shared(std::move(new_type)), + }; + } + + if (const auto * type_tuple = typeid_cast(type.get())) + { + const auto & column_tuple = assert_cast(*column); + + const auto & tuple_columns = column_tuple.getColumns(); + const auto & tuple_types = type_tuple->getElements(); + assert(tuple_columns.size() == tuple_types.size()); + const size_t tuple_size = tuple_types.size(); + + Columns new_tuple_columns(tuple_size); + DataTypes new_tuple_types(tuple_size); + + for (size_t i = 0; i < tuple_size; ++i) + { + std::tie(new_tuple_columns[i], new_tuple_types[i]) + = recursivlyConvertDynamicColumnToTuple(tuple_columns[i], tuple_types[i]); } - auto it = storage_columns_map.find(column.name); - if (it == storage_columns_map.end()) + return + { + ColumnTuple::create(new_tuple_columns), + recreateTupleWithElements(*type_tuple, new_tuple_types) + }; + } + + throw Exception(ErrorCodes::LOGICAL_ERROR, "Type {} unexpectedly has dynamic columns", type->getName()); +} + +void convertDynamicColumnsToTuples(Block & block, const StorageSnapshotPtr & storage_snapshot) +{ + for (auto & column : block) + { + if (!column.type->hasDynamicSubcolumns()) + continue; + + std::tie(column.column, column.type) + = recursivlyConvertDynamicColumnToTuple(column.column, column.type); + + GetColumnsOptions options(GetColumnsOptions::AllPhysical); + auto storage_column = storage_snapshot->tryGetColumn(options, column.name); + if (!storage_column) throw Exception(ErrorCodes::LOGICAL_ERROR, "Column '{}' not found in storage", column.name); - std::tie(column.column, column.type) = unflattenTuple(tuple_paths, tuple_types, tuple_columns); + auto storage_column_concrete = storage_snapshot->getColumn(options.withExtendedObjects(), column.name); /// Check that constructed Tuple type and type in storage are compatible. - getLeastCommonTypeForObject({column.type, it->second}, true); + getLeastCommonTypeForDynamicColumns( + storage_column->type, {column.type, storage_column_concrete.type}, true); } } @@ -196,24 +262,8 @@ void checkObjectHasNoAmbiguosPaths(const PathsInData & paths) } } -DataTypePtr getLeastCommonTypeForObject(const DataTypes & types, bool check_ambiguos_paths) +static DataTypePtr getLeastCommonTypeForObject(const DataTypes & types, bool check_ambiguos_paths) { - if (types.empty()) - return nullptr; - - bool all_equal = true; - for (size_t i = 1; i < types.size(); ++i) - { - if (!types[i]->equals(*types[0])) - { - all_equal = false; - break; - } - } - - if (all_equal) - return types[0]; - /// Types of subcolumns by path from all tuples. std::unordered_map subcolumns_types; @@ -266,19 +316,123 @@ DataTypePtr getLeastCommonTypeForObject(const DataTypes & types, bool check_ambi return unflattenTuple(tuple_paths, tuple_types); } -NameSet getNamesOfObjectColumns(const NamesAndTypesList & columns_list) +DataTypePtr getLeastCommonTypeForDynamicColumnsImpl( + const DataTypePtr & type_in_storage, const DataTypes & concrete_types, bool check_ambiguos_paths) { - NameSet res; - for (const auto & [name, type] : columns_list) - if (isObject(type)) - res.insert(name); + if (!type_in_storage->hasDynamicSubcolumns()) + return type_in_storage; - return res; + if (isObject(type_in_storage)) + return getLeastCommonTypeForObject(concrete_types, check_ambiguos_paths); + + if (const auto * type_array = typeid_cast(type_in_storage.get())) + { + DataTypes nested_types; + nested_types.reserve(concrete_types.size()); + + for (const auto & type : concrete_types) + { + const auto * type_array_conctete = typeid_cast(type.get()); + if (!type_array_conctete) + throw Exception(ErrorCodes::TYPE_MISMATCH, "Expected Array type, got {}", type->getName()); + + nested_types.push_back(type_array_conctete->getNestedType()); + } + + return std::make_shared( + getLeastCommonTypeForDynamicColumnsImpl( + type_array->getNestedType(), nested_types, check_ambiguos_paths)); + } + + if (const auto * type_tuple = typeid_cast(type_in_storage.get())) + { + const auto & element_types = type_tuple->getElements(); + DataTypes new_element_types(element_types.size()); + + for (size_t i = 0; i < element_types.size(); ++i) + { + DataTypes concrete_element_types; + concrete_element_types.reserve(concrete_types.size()); + + for (const auto & type : concrete_types) + { + const auto * type_tuple_conctete = typeid_cast(type.get()); + if (!type_tuple_conctete) + throw Exception(ErrorCodes::TYPE_MISMATCH, "Expected Tuple type, got {}", type->getName()); + + concrete_element_types.push_back(type_tuple_conctete->getElement(i)); + } + + new_element_types[i] = getLeastCommonTypeForDynamicColumnsImpl( + element_types[i], concrete_element_types, check_ambiguos_paths); + } + + return recreateTupleWithElements(*type_tuple, new_element_types); + } + + throw Exception(ErrorCodes::LOGICAL_ERROR, "Type {} unexpectedly has dynamic columns", type_in_storage->getName()); } -bool hasObjectColumns(const ColumnsDescription & columns) +DataTypePtr getLeastCommonTypeForDynamicColumns( + const DataTypePtr & type_in_storage, const DataTypes & concrete_types, bool check_ambiguos_paths) { - return std::any_of(columns.begin(), columns.end(), [](const auto & column) { return isObject(column.type); }); + if (concrete_types.empty()) + return nullptr; + + bool all_equal = true; + for (size_t i = 1; i < concrete_types.size(); ++i) + { + if (!concrete_types[i]->equals(*concrete_types[0])) + { + all_equal = false; + break; + } + } + + if (all_equal) + return concrete_types[0]; + + return getLeastCommonTypeForDynamicColumnsImpl(type_in_storage, concrete_types, check_ambiguos_paths); +} + +DataTypePtr createConcreteEmptyDynamicColumn(const DataTypePtr & type_in_storage) +{ + if (!type_in_storage->hasDynamicSubcolumns()) + return type_in_storage; + + if (isObject(type_in_storage)) + { + return std::make_shared( + DataTypes{std::make_shared()}, + Names{ColumnObject::COLUMN_NAME_DUMMY}); + } + + if (const auto * type_array = typeid_cast(type_in_storage.get())) + return std::make_shared( + createConcreteEmptyDynamicColumn(type_array->getNestedType())); + + if (const auto * type_tuple = typeid_cast(type_in_storage.get())) + { + const auto & elements = type_tuple->getElements(); + DataTypes new_elements; + new_elements.reserve(elements.size()); + + for (const auto & element : elements) + new_elements.push_back(createConcreteEmptyDynamicColumn(element)); + + return recreateTupleWithElements(*type_tuple, new_elements); + } + + throw Exception(ErrorCodes::LOGICAL_ERROR, "Type {} unexpectedly has dynamic columns", type_in_storage->getName()); +} + +bool hasDynamicSubcolumns(const ColumnsDescription & columns) +{ + return std::any_of(columns.begin(), columns.end(), + [](const auto & column) + { + return column.type->hasDynamicSubcolumns(); + }); } void extendObjectColumns(NamesAndTypesList & columns_list, const ColumnsDescription & object_columns, bool with_subcolumns) @@ -299,16 +453,20 @@ void extendObjectColumns(NamesAndTypesList & columns_list, const ColumnsDescript columns_list.splice(columns_list.end(), std::move(subcolumns_list)); } -void updateObjectColumns(ColumnsDescription & object_columns, const NamesAndTypesList & new_columns) +void updateObjectColumns( + ColumnsDescription & object_columns, + const ColumnsDescription & storage_columns, + const NamesAndTypesList & new_columns) { for (const auto & new_column : new_columns) { auto object_column = object_columns.tryGetColumn(GetColumnsOptions::All, new_column.name); if (object_column && !object_column->type->equals(*new_column.type)) { + auto storage_column = storage_columns.getColumn(GetColumnsOptions::All, new_column.name); object_columns.modify(new_column.name, [&](auto & column) { - column.type = getLeastCommonTypeForObject({object_column->type, new_column.type}); + column.type = getLeastCommonTypeForDynamicColumns(storage_column.type, {object_column->type, new_column.type}); }); } } @@ -686,13 +844,6 @@ void replaceMissedSubcolumnsByConstants( addConstantToWithClause(query, name, type); } -void finalizeObjectColumns(MutableColumns & columns) -{ - for (auto & column : columns) - if (auto * column_object = typeid_cast(column.get())) - column_object->finalize(); -} - Field FieldVisitorReplaceScalars::operator()(const Array & x) const { if (num_dimensions_to_keep == 0) diff --git a/src/DataTypes/ObjectUtils.h b/src/DataTypes/ObjectUtils.h index 8dc46ceecf5..815e6654229 100644 --- a/src/DataTypes/ObjectUtils.h +++ b/src/DataTypes/ObjectUtils.h @@ -11,6 +11,9 @@ namespace DB { +struct StorageSnapshot; +using StorageSnapshotPtr = std::shared_ptr; + /// Returns number of dimensions in Array type. 0 if type is not array. size_t getNumberOfDimensions(const IDataType & type); @@ -37,25 +40,28 @@ DataTypePtr getDataTypeByColumn(const IColumn & column); /// Converts Object types and columns to Tuples in @columns_list and @block /// and checks that types are consistent with types in @extended_storage_columns. -void convertObjectsToTuples(Block & block, const NamesAndTypesList & extended_storage_columns); +void convertDynamicColumnsToTuples(Block & block, const StorageSnapshotPtr & storage_snapshot); /// Checks that each path is not the prefix of any other path. void checkObjectHasNoAmbiguosPaths(const PathsInData & paths); /// Receives several Tuple types and deduces the least common type among them. -DataTypePtr getLeastCommonTypeForObject(const DataTypes & types, bool check_ambiguos_paths = false); +DataTypePtr getLeastCommonTypeForDynamicColumns( + const DataTypePtr & type_in_storage, const DataTypes & types, bool check_ambiguos_paths = false); + +DataTypePtr createConcreteEmptyDynamicColumn(const DataTypePtr & type_in_storage); /// Converts types of object columns to tuples in @columns_list /// according to @object_columns and adds all tuple's subcolumns if needed. void extendObjectColumns(NamesAndTypesList & columns_list, const ColumnsDescription & object_columns, bool with_subcolumns); - -NameSet getNamesOfObjectColumns(const NamesAndTypesList & columns_list); -bool hasObjectColumns(const ColumnsDescription & columns); -void finalizeObjectColumns(MutableColumns & columns); +bool hasDynamicSubcolumns(const ColumnsDescription & columns); /// Updates types of objects in @object_columns inplace /// according to types in new_columns. -void updateObjectColumns(ColumnsDescription & object_columns, const NamesAndTypesList & new_columns); +void updateObjectColumns( + ColumnsDescription & object_columns, + const ColumnsDescription & storage_columns, + const NamesAndTypesList & new_columns); using DataTypeTuplePtr = std::shared_ptr; @@ -124,7 +130,7 @@ public: /// columns-like objects from entry to which Iterator points. /// columns-like object should have fields "name" and "type". template -ColumnsDescription getObjectColumns( +ColumnsDescription getConcreteObjectColumns( Iterator begin, Iterator end, const ColumnsDescription & storage_columns, EntryColumnsGetter && entry_columns_getter) @@ -135,14 +141,8 @@ ColumnsDescription getObjectColumns( { for (const auto & column : storage_columns) { - if (isObject(column.type)) - { - auto tuple_type = std::make_shared( - DataTypes{std::make_shared()}, - Names{ColumnObject::COLUMN_NAME_DUMMY}); - - res.add({column.name, std::move(tuple_type)}); - } + if (column.type->hasDynamicSubcolumns()) + res.add({column.name, createConcreteEmptyDynamicColumn(column.type)}); } return res; @@ -156,13 +156,16 @@ ColumnsDescription getObjectColumns( for (const auto & column : entry_columns) { auto storage_column = storage_columns.tryGetPhysical(column.name); - if (storage_column && isObject(storage_column->type)) + if (storage_column && storage_column->type->hasDynamicSubcolumns()) types_in_entries[column.name].push_back(column.type); } } for (const auto & [name, types] : types_in_entries) - res.add({name, getLeastCommonTypeForObject(types)}); + { + auto storage_column = storage_columns.getPhysical(name); + res.add({name, getLeastCommonTypeForDynamicColumns(storage_column.type, types)}); + } return res; } diff --git a/src/Functions/FunctionsConversion.h b/src/Functions/FunctionsConversion.h index 6926899eb63..54f302772de 100644 --- a/src/Functions/FunctionsConversion.h +++ b/src/Functions/FunctionsConversion.h @@ -3277,9 +3277,8 @@ private: { return [] (ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, const ColumnNullable * nullable_source, size_t input_rows_count) { - auto res = ConvertImplGenericFromString::execute(arguments, result_type, nullable_source, input_rows_count); - auto & res_object = assert_cast(res->assumeMutableRef()); - res_object.finalize(); + auto res = ConvertImplGenericFromString::execute(arguments, result_type, nullable_source, input_rows_count)->assumeMutable(); + res->finalize(); return res; }; } diff --git a/src/Interpreters/InterpreterCreateQuery.cpp b/src/Interpreters/InterpreterCreateQuery.cpp index d65f4086e2d..bb7bc62f677 100644 --- a/src/Interpreters/InterpreterCreateQuery.cpp +++ b/src/Interpreters/InterpreterCreateQuery.cpp @@ -766,7 +766,7 @@ void InterpreterCreateQuery::validateTableStructure(const ASTCreateQuery & creat { for (const auto & [name, type] : properties.columns.getAllPhysical()) { - if (isObject(type)) + if (type->hasDynamicSubcolumns()) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot create table with column '{}' which type is '{}' " @@ -1299,7 +1299,7 @@ bool InterpreterCreateQuery::doCreateTable(ASTCreateQuery & create, /// we can safely destroy the object without a call to "shutdown", because there is guarantee /// that no background threads/similar resources remain after exception from "startup". - if (!res->supportsDynamicSubcolumns() && hasObjectColumns(res->getInMemoryMetadataPtr()->getColumns())) + if (!res->supportsDynamicSubcolumns() && hasDynamicSubcolumns(res->getInMemoryMetadataPtr()->getColumns())) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot create table with column of type Object, " diff --git a/src/Processors/Formats/IRowInputFormat.cpp b/src/Processors/Formats/IRowInputFormat.cpp index 72864d83b86..6bed40b0166 100644 --- a/src/Processors/Formats/IRowInputFormat.cpp +++ b/src/Processors/Formats/IRowInputFormat.cpp @@ -211,9 +211,10 @@ Chunk IRowInputFormat::generate() return {}; } - finalizeObjectColumns(columns); + for (const auto & column : columns) + column->finalize(); + Chunk chunk(std::move(columns), num_rows); - //chunk.setChunkInfo(std::move(chunk_missing_values)); return chunk; } diff --git a/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp b/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp index 956e640e202..6802262ca85 100644 --- a/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp +++ b/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp @@ -107,7 +107,9 @@ Chunk ValuesBlockInputFormat::generate() return {}; } - finalizeObjectColumns(columns); + for (const auto & column : columns) + column->finalize(); + size_t rows_in_block = columns[0]->size(); return Chunk{std::move(columns), rows_in_block}; } diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp index cb01fe3fcfc..bd1c78e8e04 100644 --- a/src/Storages/MergeTree/MergeTask.cpp +++ b/src/Storages/MergeTree/MergeTask.cpp @@ -135,7 +135,7 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() global_ctx->all_column_names = global_ctx->metadata_snapshot->getColumns().getNamesOfPhysical(); global_ctx->storage_columns = global_ctx->metadata_snapshot->getColumns().getAllPhysical(); - auto object_columns = MergeTreeData::getObjectColumns(global_ctx->future_part->parts, global_ctx->metadata_snapshot->getColumns()); + auto object_columns = MergeTreeData::getConcreteObjectColumns(global_ctx->future_part->parts, global_ctx->metadata_snapshot->getColumns()); global_ctx->storage_snapshot = std::make_shared(*global_ctx->data, global_ctx->metadata_snapshot, object_columns); extendObjectColumns(global_ctx->storage_columns, object_columns, false); diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index d84fb9d30d3..28f19819c19 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -6643,18 +6643,18 @@ ReservationPtr MergeTreeData::balancedReservation( return reserved_space; } -ColumnsDescription MergeTreeData::getObjectColumns( +ColumnsDescription MergeTreeData::getConcreteObjectColumns( const DataPartsVector & parts, const ColumnsDescription & storage_columns) { - return DB::getObjectColumns( + return DB::getConcreteObjectColumns( parts.begin(), parts.end(), storage_columns, [](const auto & part) -> const auto & { return part->getColumns(); }); } -ColumnsDescription MergeTreeData::getObjectColumns( +ColumnsDescription MergeTreeData::getConcreteObjectColumns( boost::iterator_range range, const ColumnsDescription & storage_columns) { - return DB::getObjectColumns( + return DB::getConcreteObjectColumns( range.begin(), range.end(), storage_columns, [](const auto & part) -> const auto & { return part->getColumns(); }); } @@ -6663,21 +6663,21 @@ void MergeTreeData::resetObjectColumnsFromActiveParts(const DataPartsLock & /*lo { auto metadata_snapshot = getInMemoryMetadataPtr(); const auto & columns = metadata_snapshot->getColumns(); - if (!hasObjectColumns(columns)) + if (!hasDynamicSubcolumns(columns)) return; auto range = getDataPartsStateRange(DataPartState::Active); - object_columns = getObjectColumns(range, columns); + object_columns = getConcreteObjectColumns(range, columns); } void MergeTreeData::updateObjectColumns(const DataPartPtr & part, const DataPartsLock & /*lock*/) { auto metadata_snapshot = getInMemoryMetadataPtr(); const auto & columns = metadata_snapshot->getColumns(); - if (!hasObjectColumns(columns)) + if (!hasDynamicSubcolumns(columns)) return; - DB::updateObjectColumns(object_columns, part->getColumns()); + DB::updateObjectColumns(object_columns, columns, part->getColumns()); } StorageSnapshotPtr MergeTreeData::getStorageSnapshot(const StorageMetadataPtr & metadata_snapshot, ContextPtr query_context) const diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 8515c8583b7..29f3ef3d40a 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -746,10 +746,10 @@ public: return column_sizes; } - const ColumnsDescription & getObjectColumns() const { return object_columns; } + const ColumnsDescription & getConcreteObjectColumns() const { return object_columns; } /// Creates desciprion of columns of data type Object from the range of data parts. - static ColumnsDescription getObjectColumns( + static ColumnsDescription getConcreteObjectColumns( const DataPartsVector & parts, const ColumnsDescription & storage_columns); IndexSizeByName getSecondaryIndexSizes() const override @@ -1101,7 +1101,7 @@ protected: } /// Creates desciprion of columns of data type Object from the range of data parts. - static ColumnsDescription getObjectColumns( + static ColumnsDescription getConcreteObjectColumns( boost::iterator_range range, const ColumnsDescription & storage_columns); std::optional totalRowsByPartitionPredicateImpl( diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 24ca7cc2f5a..2451af7594f 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -287,7 +287,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( auto columns = metadata_snapshot->getColumns().getAllPhysical().filter(block.getNames()); for (auto & column : columns) - if (isObject(column.type)) + if (column.type->hasDynamicSubcolumns()) column.type = block.getByName(column.name).type; static const String TMP_PREFIX = "tmp_insert_"; @@ -469,16 +469,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( return temp_part; } -void MergeTreeDataWriter::deduceTypesOfObjectColumns(const StorageSnapshotPtr & storage_snapshot, Block & block) -{ - if (!storage_snapshot->object_columns.empty()) - { - auto options = GetColumnsOptions(GetColumnsOptions::AllPhysical).withExtendedObjects(); - auto storage_columns = storage_snapshot->getColumns(options); - convertObjectsToTuples(block, storage_columns); - } -} - MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( const String & part_name, MergeTreeDataPartType part_type, diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.h b/src/Storages/MergeTree/MergeTreeDataWriter.h index 7b6bf8fb1db..86909ab2062 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.h +++ b/src/Storages/MergeTree/MergeTreeDataWriter.h @@ -42,8 +42,6 @@ public: */ static BlocksWithPartition splitBlockIntoParts(const Block & block, size_t max_parts, const StorageMetadataPtr & metadata_snapshot, ContextPtr context); - static void deduceTypesOfObjectColumns(const StorageSnapshotPtr & storage_snapshot, Block & block); - /// This structure contains not completely written temporary part. /// Some writes may happen asynchronously, e.g. for blob storages. /// You should call finalize() to wait until all data is written. diff --git a/src/Storages/MergeTree/MergeTreeSink.cpp b/src/Storages/MergeTree/MergeTreeSink.cpp index 93b9f356595..4b0998edf69 100644 --- a/src/Storages/MergeTree/MergeTreeSink.cpp +++ b/src/Storages/MergeTree/MergeTreeSink.cpp @@ -2,6 +2,7 @@ #include #include #include +#include namespace DB @@ -51,8 +52,9 @@ void MergeTreeSink::consume(Chunk chunk) { auto block = getHeader().cloneWithColumns(chunk.detachColumns()); auto storage_snapshot = storage.getStorageSnapshot(metadata_snapshot, context); + if (!storage_snapshot->object_columns.empty()) + convertDynamicColumnsToTuples(block, storage_snapshot); - storage.writer.deduceTypesOfObjectColumns(storage_snapshot, block); auto part_blocks = storage.writer.splitBlockIntoParts(block, max_parts_per_block, metadata_snapshot, context); using DelayedPartitions = std::vector; diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp index 126d34bcc1d..28eeb97bb55 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp +++ b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp @@ -4,6 +4,7 @@ #include #include #include +#include #include #include @@ -151,7 +152,9 @@ void ReplicatedMergeTreeSink::consume(Chunk chunk) checkQuorumPrecondition(zookeeper); auto storage_snapshot = storage.getStorageSnapshot(metadata_snapshot, context); - storage.writer.deduceTypesOfObjectColumns(storage_snapshot, block); + if (!storage_snapshot->object_columns.empty()) + convertDynamicColumnsToTuples(block, storage_snapshot); + auto part_blocks = storage.writer.splitBlockIntoParts(block, max_parts_per_block, metadata_snapshot, context); using DelayedPartitions = std::vector; diff --git a/src/Storages/StorageDistributed.cpp b/src/Storages/StorageDistributed.cpp index ae67a67af7f..66f8b7f8a7f 100644 --- a/src/Storages/StorageDistributed.cpp +++ b/src/Storages/StorageDistributed.cpp @@ -595,7 +595,7 @@ std::optional StorageDistributed::getOptimizedQueryP static bool requiresObjectColumns(const ColumnsDescription & all_columns, ASTPtr query) { - if (!hasObjectColumns(all_columns)) + if (!hasDynamicSubcolumns(all_columns)) return false; if (!query) @@ -610,7 +610,7 @@ static bool requiresObjectColumns(const ColumnsDescription & all_columns, ASTPtr auto name_in_storage = Nested::splitName(required_column).first; auto column_in_storage = all_columns.tryGetPhysical(name_in_storage); - if (column_in_storage && isObject(column_in_storage->type)) + if (column_in_storage && column_in_storage->type->hasDynamicSubcolumns()) return true; } @@ -637,7 +637,7 @@ StorageSnapshotPtr StorageDistributed::getStorageSnapshotForQuery( metadata_snapshot->getColumns(), getContext()); - auto object_columns = DB::getObjectColumns( + auto object_columns = DB::getConcreteObjectColumns( snapshot_data->objects_by_shard.begin(), snapshot_data->objects_by_shard.end(), metadata_snapshot->getColumns(), diff --git a/src/Storages/StorageInMemoryMetadata.cpp b/src/Storages/StorageInMemoryMetadata.cpp index 66dcc938aef..a80f21834db 100644 --- a/src/Storages/StorageInMemoryMetadata.cpp +++ b/src/Storages/StorageInMemoryMetadata.cpp @@ -526,7 +526,7 @@ void StorageInMemoryMetadata::check(const NamesAndTypesList & provided_columns) const auto * available_type = it->getMapped(); - if (!isObject(*available_type) + if (!available_type->hasDynamicSubcolumns() && !column.type->equals(*available_type) && !isCompatibleEnumTypes(available_type, column.type.get())) throw Exception( @@ -575,7 +575,7 @@ void StorageInMemoryMetadata::check(const NamesAndTypesList & provided_columns, const auto * provided_column_type = it->getMapped(); const auto * available_column_type = jt->getMapped(); - if (!isObject(*provided_column_type) + if (!provided_column_type->hasDynamicSubcolumns() && !provided_column_type->equals(*available_column_type) && !isCompatibleEnumTypes(available_column_type, provided_column_type)) throw Exception( @@ -619,7 +619,7 @@ void StorageInMemoryMetadata::check(const Block & block, bool need_all) const listOfColumns(available_columns)); const auto * available_type = it->getMapped(); - if (!isObject(*available_type) + if (!available_type->hasDynamicSubcolumns() && !column.type->equals(*available_type) && !isCompatibleEnumTypes(available_type, column.type.get())) throw Exception( diff --git a/src/Storages/StorageMemory.cpp b/src/Storages/StorageMemory.cpp index c660195c368..9392aec482d 100644 --- a/src/Storages/StorageMemory.cpp +++ b/src/Storages/StorageMemory.cpp @@ -141,7 +141,7 @@ public: auto extended_storage_columns = storage_snapshot->getColumns( GetColumnsOptions(GetColumnsOptions::AllPhysical).withExtendedObjects()); - convertObjectsToTuples(block, extended_storage_columns); + convertDynamicColumnsToTuples(block, storage_snapshot); } if (storage.compress) @@ -207,10 +207,10 @@ StorageSnapshotPtr StorageMemory::getStorageSnapshot(const StorageMetadataPtr & auto snapshot_data = std::make_unique(); snapshot_data->blocks = data.get(); - if (!hasObjectColumns(metadata_snapshot->getColumns())) + if (!hasDynamicSubcolumns(metadata_snapshot->getColumns())) return std::make_shared(*this, metadata_snapshot, ColumnsDescription{}, std::move(snapshot_data)); - auto object_columns = getObjectColumns( + auto object_columns = getConcreteObjectColumns( snapshot_data->blocks->begin(), snapshot_data->blocks->end(), metadata_snapshot->getColumns(), diff --git a/src/Storages/StorageSnapshot.cpp b/src/Storages/StorageSnapshot.cpp index a4b64c798f3..6745b8fb3f5 100644 --- a/src/Storages/StorageSnapshot.cpp +++ b/src/Storages/StorageSnapshot.cpp @@ -60,7 +60,7 @@ std::optional StorageSnapshot::tryGetColumn(const GetColumnsOpt { const auto & columns = getMetadataForQuery()->getColumns(); auto column = columns.tryGetColumn(options, column_name); - if (column && (!isObject(column->type) || !options.with_extended_objects)) + if (column && (!column->type->hasDynamicSubcolumns() || !options.with_extended_objects)) return column; if (options.with_extended_objects) diff --git a/src/Storages/getStructureOfRemoteTable.cpp b/src/Storages/getStructureOfRemoteTable.cpp index 8fa4d02e8e1..d022287a671 100644 --- a/src/Storages/getStructureOfRemoteTable.cpp +++ b/src/Storages/getStructureOfRemoteTable.cpp @@ -189,7 +189,7 @@ ColumnsDescriptionByShardNum getExtendedObjectsOfRemoteTables( auto type_name = get(type_col[i]); auto storage_column = storage_columns.tryGetPhysical(name); - if (storage_column && isObject(storage_column->type)) + if (storage_column && storage_column->type->hasDynamicSubcolumns()) res.add(ColumnDescription(std::move(name), DataTypeFactory::instance().get(type_name))); } } diff --git a/tests/queries/0_stateless/01825_type_json_in_array.reference b/tests/queries/0_stateless/01825_type_json_in_array.reference new file mode 100644 index 00000000000..1b6ca177b57 --- /dev/null +++ b/tests/queries/0_stateless/01825_type_json_in_array.reference @@ -0,0 +1,17 @@ +{"id":1,"arr":[{"k1":1,"k2":{"k3":2,"k4":3,"k5":""}},{"k1":2,"k2":{"k3":0,"k4":0,"k5":"foo"}}]} +{"id":2,"arr":[{"k1":3,"k2":{"k3":4,"k4":5,"k5":""}}]} +1 [1,2] [2,0] [3,0] ['','foo'] +2 [3] [4] [5] [''] +{"arr":{"k1":1,"k2":{"k3":2,"k4":3,"k5":""}}} +{"arr":{"k1":2,"k2":{"k3":0,"k4":0,"k5":"foo"}}} +{"arr":{"k1":3,"k2":{"k3":4,"k4":5,"k5":""}}} +Array(Tuple(k1 Int8, k2 Tuple(k3 Int8, k4 Int8, k5 String))) +{"id":1,"arr":[{"k1":[{"k2":"aaa","k3":"bbb","k4":0},{"k2":"ccc","k3":"","k4":0}],"k5":{"k6":""}}]} +{"id":2,"arr":[{"k1":[{"k2":"","k3":"ddd","k4":10},{"k2":"","k3":"","k4":20}],"k5":{"k6":"foo"}}]} +1 [['aaa','ccc']] [['bbb','']] [[]] [''] +2 [[]] [['ddd','']] [[10,20]] ['foo'] +{"k1":{"k2":"","k3":"","k4":20}} +{"k1":{"k2":"","k3":"ddd","k4":10}} +{"k1":{"k2":"aaa","k3":"bbb","k4":0}} +{"k1":{"k2":"ccc","k3":"","k4":0}} +Tuple(k2 String, k3 String, k4 Int8) diff --git a/tests/queries/0_stateless/01825_type_json_in_array.sql b/tests/queries/0_stateless/01825_type_json_in_array.sql new file mode 100644 index 00000000000..abb117fbfa4 --- /dev/null +++ b/tests/queries/0_stateless/01825_type_json_in_array.sql @@ -0,0 +1,27 @@ +-- Tags: no-fasttest + +SET allow_experimental_object_type = 1; +DROP TABLE IF EXISTS t_json_array; + +CREATE TABLE t_json_array (id UInt32, arr Array(JSON)) ENGINE = MergeTree ORDER BY id; + +INSERT INTO t_json_array FORMAT JSONEachRow {"id": 1, "arr": [{"k1": 1, "k2": {"k3": 2, "k4": 3}}, {"k1": 2, "k2": {"k5": "foo"}}]} +INSERT INTO t_json_array FORMAT JSONEachRow {"id": 2, "arr": [{"k1": 3, "k2": {"k3": 4, "k4": 5}}]} + +SET output_format_json_named_tuples_as_objects = 1; + +SELECT * FROM t_json_array ORDER BY id FORMAT JSONEachRow; +SELECT id, arr.k1, arr.k2.k3, arr.k2.k4, arr.k2.k5 FROM t_json_array ORDER BY id; +SELECT arr FROM t_json_array ARRAY JOIN arr ORDER BY arr.k1 FORMAT JSONEachRow; +SELECT toTypeName(arr) FROM t_json_array LIMIT 1; + +TRUNCATE TABLE t_json_array; + +INSERT INTO t_json_array FORMAT JSONEachRow {"id": 1, "arr": [{"k1": [{"k2": "aaa", "k3": "bbb"}, {"k2": "ccc"}]}]} +INSERT INTO t_json_array FORMAT JSONEachRow {"id": 2, "arr": [{"k1": [{"k3": "ddd", "k4": 10}, {"k4": 20}], "k5": {"k6": "foo"}}]} + +SELECT * FROM t_json_array ORDER BY id FORMAT JSONEachRow; +SELECT id, arr.k1.k2, arr.k1.k3, arr.k1.k4, arr.k5.k6 FROM t_json_array ORDER BY id; + +SELECT arrayJoin(arrayJoin(arr.k1)) AS k1 FROM t_json_array ORDER BY k1 FORMAT JSONEachRow; +SELECT toTypeName(arrayJoin(arrayJoin(arr.k1))) AS arr FROM t_json_array LIMIT 1; From 90c60741a172e641d70e2d3766d0b702c209d880 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Fri, 6 May 2022 17:13:26 +0000 Subject: [PATCH 002/239] more safe operations with ColumnObject --- src/Columns/ColumnObject.cpp | 102 ++++++++++-------- src/Columns/ColumnObject.h | 4 +- .../Serializations/SerializationObject.cpp | 9 +- src/Interpreters/convertFieldToType.cpp | 3 + .../01825_type_json_in_array.reference | 4 + .../0_stateless/01825_type_json_in_array.sql | 7 ++ 6 files changed, 83 insertions(+), 46 deletions(-) diff --git a/src/Columns/ColumnObject.cpp b/src/Columns/ColumnObject.cpp index 0913f6b8da9..8413a709e05 100644 --- a/src/Columns/ColumnObject.cpp +++ b/src/Columns/ColumnObject.cpp @@ -553,11 +553,9 @@ size_t ColumnObject::allocatedBytes() const void ColumnObject::forEachSubcolumn(ColumnCallback callback) { - if (!isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot iterate over non-finalized ColumnObject"); - for (auto & entry : subcolumns) - callback(entry->data.data.back()); + for (auto & part : entry->data.data) + callback(part); } void ColumnObject::insert(const Field & field) @@ -594,26 +592,43 @@ void ColumnObject::insertDefault() Field ColumnObject::operator[](size_t n) const { - if (!isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot get Field from non-finalized ColumnObject"); - - Object object; - for (const auto & entry : subcolumns) - object[entry->path.getPath()] = (*entry->data.data.back())[n]; - + Field object; + get(n, object); return object; } void ColumnObject::get(size_t n, Field & res) const { - if (!isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot get Field from non-finalized ColumnObject"); - + assert(n < size()); + res = Object(); auto & object = res.get(); - for (const auto & entry : subcolumns) + + if (isFinalized()) { - auto it = object.try_emplace(entry->path.getPath()).first; - entry->data.data.back()->get(n, it->second); + for (const auto & entry : subcolumns) + { + auto it = object.try_emplace(entry->path.getPath()).first; + entry->data.data.back()->get(n, it->second); + } + } + else + { + for (const auto & entry : subcolumns) + { + size_t ind = n; + for (const auto & part : entry->data.data) + { + if (ind < part->size()) + { + auto it = object.try_emplace(entry->path.getPath()).first; + part->get(ind, it->second); + it->second = convertFieldToTypeOrThrow(it->second, *entry->data.getLeastCommonType()); + break; + } + + ind -= part->size(); + } + } } } @@ -625,19 +640,25 @@ void ColumnObject::insertFrom(const IColumn & src, size_t n) void ColumnObject::insertRangeFrom(const IColumn & src, size_t start, size_t length) { - const auto & src_object = assert_cast(src); - if (!src_object.isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot insertRangeFrom non-finalized ColumnObject"); + const auto * src_object = &assert_cast(src); + MutableColumnPtr column_holder; + + if (!src_object->isFinalized()) + { + column_holder = IColumn::mutate(src_object->getPtr()); + column_holder->finalize(); + src_object = &assert_cast(*column_holder); + } for (auto & entry : subcolumns) { - if (src_object.hasSubcolumn(entry->path)) - entry->data.insertRangeFrom(src_object.getSubcolumn(entry->path), start, length); + if (src_object->hasSubcolumn(entry->path)) + entry->data.insertRangeFrom(src_object->getSubcolumn(entry->path), start, length); else entry->data.insertManyDefaults(length); } - for (const auto & entry : src_object.subcolumns) + for (const auto & entry : src_object->subcolumns) { if (!hasSubcolumn(entry->path)) { @@ -668,21 +689,6 @@ void ColumnObject::insertRangeFrom(const IColumn & src, size_t start, size_t len finalize(); } -ColumnPtr ColumnObject::replicate(const Offsets & offsets) const -{ - if (!isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot replicate non-finalized ColumnObject"); - - auto res_column = ColumnObject::create(is_nullable); - for (const auto & entry : subcolumns) - { - auto replicated_data = entry->data.data.back()->replicate(offsets)->assumeMutable(); - res_column->addSubcolumn(entry->path, std::move(replicated_data)); - } - - return res_column; -} - void ColumnObject::popBack(size_t length) { for (auto & entry : subcolumns) @@ -692,10 +698,15 @@ void ColumnObject::popBack(size_t length) } template -ColumnPtr ColumnObject::applyForSubcolumns(Func && func, std::string_view func_name) const +ColumnPtr ColumnObject::applyForSubcolumns(Func && func) const { if (!isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot {} non-finalized ColumnObject", func_name); + { + auto finalized = IColumn::mutate(getPtr()); + auto & finalized_object = assert_cast(*finalized); + finalized_object.finalize(); + return finalized_object.applyForSubcolumns(std::move(func)); + } auto res = ColumnObject::create(is_nullable); for (const auto & subcolumn : subcolumns) @@ -708,17 +719,22 @@ ColumnPtr ColumnObject::applyForSubcolumns(Func && func, std::string_view func_n ColumnPtr ColumnObject::permute(const Permutation & perm, size_t limit) const { - return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.permute(perm, limit); }, "permute"); + return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.permute(perm, limit); }); } ColumnPtr ColumnObject::filter(const Filter & filter, ssize_t result_size_hint) const { - return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.filter(filter, result_size_hint); }, "filter"); + return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.filter(filter, result_size_hint); }); } ColumnPtr ColumnObject::index(const IColumn & indexes, size_t limit) const { - return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.index(indexes, limit); }, "index"); + return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.index(indexes, limit); }); +} + +ColumnPtr ColumnObject::replicate(const Offsets & offsets) const +{ + return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.replicate(offsets); }); } const ColumnObject::Subcolumn & ColumnObject::getSubcolumn(const PathInData & key) const diff --git a/src/Columns/ColumnObject.h b/src/Columns/ColumnObject.h index 6c69b0247b7..154122f6743 100644 --- a/src/Columns/ColumnObject.h +++ b/src/Columns/ColumnObject.h @@ -195,13 +195,13 @@ public: void insertDefault() override; void insertFrom(const IColumn & src, size_t n) override; void insertRangeFrom(const IColumn & src, size_t start, size_t length) override; - ColumnPtr replicate(const Offsets & offsets) const override; void popBack(size_t length) override; Field operator[](size_t n) const override; void get(size_t n, Field & res) const override; ColumnPtr permute(const Permutation & perm, size_t limit) const override; ColumnPtr filter(const Filter & filter, ssize_t result_size_hint) const override; ColumnPtr index(const IColumn & indexes, size_t limit) const override; + ColumnPtr replicate(const Offsets & offsets) const override; /// Finalizes all subcolumns. void finalize() override; @@ -237,7 +237,7 @@ private: } template - ColumnPtr applyForSubcolumns(Func && func, std::string_view func_name) const; + ColumnPtr applyForSubcolumns(Func && func) const; }; } diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index 2d6555dcb43..0229754f677 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -241,7 +241,14 @@ void SerializationObject::serializeBinaryBulkWithMultipleStreams( const auto & column_object = assert_cast(column); if (!column_object.isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot write non-finalized ColumnObject"); + { + auto finalized = IColumn::mutate(column_object.getPtr()); + auto & finalized_object = assert_cast(*finalized); + finalized_object.finalize(); + serializeBinaryBulkWithMultipleStreams( + finalized_object, offset, limit, settings, state); + return; + } settings.path.push_back(Substream::ObjectStructure); if (auto * stream = settings.getter(settings.path)) diff --git a/src/Interpreters/convertFieldToType.cpp b/src/Interpreters/convertFieldToType.cpp index dd23ad69ae2..36970895b65 100644 --- a/src/Interpreters/convertFieldToType.cpp +++ b/src/Interpreters/convertFieldToType.cpp @@ -370,6 +370,9 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID } else if (isObject(type)) { + if (src.getType() == Field::Types::Object) + return src; /// Already in needed type. + const auto * from_type_tuple = typeid_cast(from_type_hint); if (src.getType() == Field::Types::Tuple && from_type_tuple && from_type_tuple->haveExplicitNames()) { diff --git a/tests/queries/0_stateless/01825_type_json_in_array.reference b/tests/queries/0_stateless/01825_type_json_in_array.reference index 1b6ca177b57..0bdb9917622 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.reference +++ b/tests/queries/0_stateless/01825_type_json_in_array.reference @@ -15,3 +15,7 @@ Array(Tuple(k1 Int8, k2 Tuple(k3 Int8, k4 Int8, k5 String))) {"k1":{"k2":"aaa","k3":"bbb","k4":0}} {"k1":{"k2":"ccc","k3":"","k4":0}} Tuple(k2 String, k3 String, k4 Int8) +{"arr":[{"x":1}]} +{"arr":{"x":{"y":1},"t":{"y":2}}} +{"arr":[1,{"y":1}]} +{"arr":[{"x":"aaa","y":[1,2,3]}]} diff --git a/tests/queries/0_stateless/01825_type_json_in_array.sql b/tests/queries/0_stateless/01825_type_json_in_array.sql index abb117fbfa4..b02202ec175 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.sql +++ b/tests/queries/0_stateless/01825_type_json_in_array.sql @@ -25,3 +25,10 @@ SELECT id, arr.k1.k2, arr.k1.k3, arr.k1.k4, arr.k5.k6 FROM t_json_array ORDER BY SELECT arrayJoin(arrayJoin(arr.k1)) AS k1 FROM t_json_array ORDER BY k1 FORMAT JSONEachRow; SELECT toTypeName(arrayJoin(arrayJoin(arr.k1))) AS arr FROM t_json_array LIMIT 1; + +DROP TABLE t_json_array; + +SELECT * FROM values('arr Array(JSON)', '[\'{"x" : 1}\']') FORMAT JSONEachRow; +SELECT * FROM values('arr Map(String, JSON)', '{\'x\' : \'{"y" : 1}\', \'t\' : \'{"y" : 2}\'}') FORMAT JSONEachRow; +SELECT * FROM values('arr Tuple(Int32, JSON)', '(1, \'{"y" : 1}\')') FORMAT JSONEachRow; +SELECT * FROM format(JSONEachRow, '{"arr" : [{"x" : "aaa", "y" : [1,2,3]}]}') FORMAT JSONEachRow; From b3e262f605e7f451471a0e93ddcf121ea578cd0c Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Fri, 6 May 2022 18:52:53 +0000 Subject: [PATCH 003/239] better cloneResized in ColumnObject --- src/Columns/ColumnObject.cpp | 20 +++++++++----------- src/Columns/ColumnObject.h | 5 +++-- 2 files changed, 12 insertions(+), 13 deletions(-) diff --git a/src/Columns/ColumnObject.cpp b/src/Columns/ColumnObject.cpp index 8413a709e05..64f92156175 100644 --- a/src/Columns/ColumnObject.cpp +++ b/src/Columns/ColumnObject.cpp @@ -525,16 +525,6 @@ size_t ColumnObject::size() const return num_rows; } -MutableColumnPtr ColumnObject::cloneResized(size_t new_size) const -{ - /// cloneResized with new_size == 0 is used for cloneEmpty(). - if (new_size != 0) - throw Exception(ErrorCodes::NOT_IMPLEMENTED, - "ColumnObject doesn't support resize to non-zero length"); - - return ColumnObject::create(is_nullable); -} - size_t ColumnObject::byteSize() const { size_t res = 0; @@ -698,7 +688,7 @@ void ColumnObject::popBack(size_t length) } template -ColumnPtr ColumnObject::applyForSubcolumns(Func && func) const +MutableColumnPtr ColumnObject::applyForSubcolumns(Func && func) const { if (!isFinalized()) { @@ -737,6 +727,14 @@ ColumnPtr ColumnObject::replicate(const Offsets & offsets) const return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.replicate(offsets); }); } +MutableColumnPtr ColumnObject::cloneResized(size_t new_size) const +{ + if (new_size == 0) + return ColumnObject::create(is_nullable); + + return applyForSubcolumns([&](const auto & subcolumn) { return subcolumn.cloneResized(new_size); }); +} + const ColumnObject::Subcolumn & ColumnObject::getSubcolumn(const PathInData & key) const { if (const auto * node = subcolumns.findLeaf(key)) diff --git a/src/Columns/ColumnObject.h b/src/Columns/ColumnObject.h index 154122f6743..8bed90bd0ce 100644 --- a/src/Columns/ColumnObject.h +++ b/src/Columns/ColumnObject.h @@ -187,7 +187,6 @@ public: TypeIndex getDataType() const override { return TypeIndex::Object; } size_t size() const override; - MutableColumnPtr cloneResized(size_t new_size) const override; size_t byteSize() const override; size_t allocatedBytes() const override; void forEachSubcolumn(ColumnCallback callback) override; @@ -198,10 +197,12 @@ public: void popBack(size_t length) override; Field operator[](size_t n) const override; void get(size_t n, Field & res) const override; + ColumnPtr permute(const Permutation & perm, size_t limit) const override; ColumnPtr filter(const Filter & filter, ssize_t result_size_hint) const override; ColumnPtr index(const IColumn & indexes, size_t limit) const override; ColumnPtr replicate(const Offsets & offsets) const override; + MutableColumnPtr cloneResized(size_t new_size) const override; /// Finalizes all subcolumns. void finalize() override; @@ -237,7 +238,7 @@ private: } template - ColumnPtr applyForSubcolumns(Func && func) const; + MutableColumnPtr applyForSubcolumns(Func && func) const; }; } From e1911a29b81bff19573549cbc5d7960ab89476df Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 9 May 2022 16:12:04 +0000 Subject: [PATCH 004/239] better text serialization of type Object --- src/Columns/ColumnArray.h | 1 + src/Columns/ColumnObject.cpp | 81 +++++++++++-------- src/Columns/ColumnObject.h | 7 +- src/Columns/ColumnTuple.cpp | 5 ++ src/Columns/ColumnTuple.h | 1 + src/Columns/IColumn.h | 8 ++ src/DataTypes/ObjectUtils.cpp | 5 +- .../Serializations/SerializationObject.cpp | 70 ++++++++++++---- .../Serializations/SerializationObject.h | 2 + 9 files changed, 127 insertions(+), 53 deletions(-) diff --git a/src/Columns/ColumnArray.h b/src/Columns/ColumnArray.h index 1657160fbae..3e3462c7cdf 100644 --- a/src/Columns/ColumnArray.h +++ b/src/Columns/ColumnArray.h @@ -169,6 +169,7 @@ public: void getIndicesOfNonDefaultRows(Offsets & indices, size_t from, size_t limit) const override; void finalize() override { data->finalize(); } + bool isFinalized() const override { return data->isFinalized(); } bool isCollationSupported() const override { return getData().isCollationSupported(); } diff --git a/src/Columns/ColumnObject.cpp b/src/Columns/ColumnObject.cpp index 64f92156175..d0c5997c071 100644 --- a/src/Columns/ColumnObject.cpp +++ b/src/Columns/ColumnObject.cpp @@ -24,6 +24,7 @@ namespace ErrorCodes extern const int NUMBER_OF_DIMENSIONS_MISMATHED; extern const int NOT_IMPLEMENTED; extern const int SIZES_OF_COLUMNS_DOESNT_MATCH; + extern const int ARGUMENT_OUT_OF_BOUND; } namespace @@ -179,7 +180,7 @@ ColumnObject::Subcolumn::Subcolumn( { } -size_t ColumnObject::Subcolumn::Subcolumn::size() const +size_t ColumnObject::Subcolumn::size() const { size_t res = num_of_defaults_in_prefix; for (const auto & part : data) @@ -187,7 +188,7 @@ size_t ColumnObject::Subcolumn::Subcolumn::size() const return res; } -size_t ColumnObject::Subcolumn::Subcolumn::byteSize() const +size_t ColumnObject::Subcolumn::byteSize() const { size_t res = 0; for (const auto & part : data) @@ -195,7 +196,7 @@ size_t ColumnObject::Subcolumn::Subcolumn::byteSize() const return res; } -size_t ColumnObject::Subcolumn::Subcolumn::allocatedBytes() const +size_t ColumnObject::Subcolumn::allocatedBytes() const { size_t res = 0; for (const auto & part : data) @@ -203,6 +204,37 @@ size_t ColumnObject::Subcolumn::Subcolumn::allocatedBytes() const return res; } +void ColumnObject::Subcolumn::get(size_t n, Field & res) const +{ + if (isFinalized()) + { + getFinalizedColumn().get(n, res); + return; + } + + size_t ind = n; + if (ind < num_of_defaults_in_prefix) + { + res = least_common_type.get()->getDefault(); + return; + } + + ind -= num_of_defaults_in_prefix; + for (const auto & part : data) + { + if (ind < part->size()) + { + part->get(ind, res); + res = convertFieldToTypeOrThrow(res, *least_common_type.get()); + return; + } + + ind -= part->size(); + } + + throw Exception(ErrorCodes::ARGUMENT_OUT_OF_BOUND, "Index ({}) for getting field is out of range", n); +} + void ColumnObject::Subcolumn::checkTypes() const { DataTypes prefix_types; @@ -593,39 +625,16 @@ void ColumnObject::get(size_t n, Field & res) const res = Object(); auto & object = res.get(); - if (isFinalized()) + for (const auto & entry : subcolumns) { - for (const auto & entry : subcolumns) - { - auto it = object.try_emplace(entry->path.getPath()).first; - entry->data.data.back()->get(n, it->second); - } - } - else - { - for (const auto & entry : subcolumns) - { - size_t ind = n; - for (const auto & part : entry->data.data) - { - if (ind < part->size()) - { - auto it = object.try_emplace(entry->path.getPath()).first; - part->get(ind, it->second); - it->second = convertFieldToTypeOrThrow(it->second, *entry->data.getLeastCommonType()); - break; - } - - ind -= part->size(); - } - } + auto it = object.try_emplace(entry->path.getPath()).first; + entry->data.get(n, it->second); } } void ColumnObject::insertFrom(const IColumn & src, size_t n) { insert(src[n]); - finalize(); } void ColumnObject::insertRangeFrom(const IColumn & src, size_t start, size_t length) @@ -635,8 +644,7 @@ void ColumnObject::insertRangeFrom(const IColumn & src, size_t start, size_t len if (!src_object->isFinalized()) { - column_holder = IColumn::mutate(src_object->getPtr()); - column_holder->finalize(); + column_holder = src.cloneFinalized(); src_object = &assert_cast(*column_holder); } @@ -692,10 +700,9 @@ MutableColumnPtr ColumnObject::applyForSubcolumns(Func && func) const { if (!isFinalized()) { - auto finalized = IColumn::mutate(getPtr()); + auto finalized = cloneFinalized(); auto & finalized_object = assert_cast(*finalized); - finalized_object.finalize(); - return finalized_object.applyForSubcolumns(std::move(func)); + return finalized_object.applyForSubcolumns(std::forward(func)); } auto res = ColumnObject::create(is_nullable); @@ -704,6 +711,7 @@ MutableColumnPtr ColumnObject::applyForSubcolumns(Func && func) const auto new_subcolumn = func(subcolumn->data.getFinalizedColumn()); res->addSubcolumn(subcolumn->path, new_subcolumn->assumeMutable()); } + return res; } @@ -843,13 +851,16 @@ bool ColumnObject::isFinalized() const void ColumnObject::finalize() { + if (isFinalized()) + return; + size_t old_size = size(); Subcolumns new_subcolumns; for (auto && entry : subcolumns) { const auto & least_common_type = entry->data.getLeastCommonType(); - /// Do not add subcolumns, which consists only from NULLs. + /// Do not add subcolumns, which consist only from NULLs. if (isNothing(getBaseTypeOfArray(least_common_type))) continue; diff --git a/src/Columns/ColumnObject.h b/src/Columns/ColumnObject.h index 8bed90bd0ce..762049afd38 100644 --- a/src/Columns/ColumnObject.h +++ b/src/Columns/ColumnObject.h @@ -65,6 +65,7 @@ public: size_t size() const; size_t byteSize() const; size_t allocatedBytes() const; + void get(size_t n, Field & res) const; bool isFinalized() const; const DataTypePtr & getLeastCommonType() const { return least_common_type.get(); } @@ -101,6 +102,9 @@ public: const IColumn & getFinalizedColumn() const; const ColumnPtr & getFinalizedColumnPtr() const; + const std::vector & getData() const { return data; } + size_t getNumberOfDefaultsInPrefix() const { return num_of_defaults_in_prefix; } + friend class ColumnObject; private: @@ -179,8 +183,6 @@ public: Subcolumns & getSubcolumns() { return subcolumns; } PathsInData getKeys() const; - bool isFinalized() const; - /// Part of interface const char * getFamilyName() const override { return "Object"; } @@ -206,6 +208,7 @@ public: /// Finalizes all subcolumns. void finalize() override; + bool isFinalized() const override; /// All other methods throw exception. diff --git a/src/Columns/ColumnTuple.cpp b/src/Columns/ColumnTuple.cpp index 8d6b2dbdbd9..c260e48dde5 100644 --- a/src/Columns/ColumnTuple.cpp +++ b/src/Columns/ColumnTuple.cpp @@ -578,4 +578,9 @@ void ColumnTuple::finalize() column->finalize(); } +bool ColumnTuple::isFinalized() const +{ + return std::all_of(columns.begin(), columns.end(), [](const auto & column) { return column->isFinalized(); }); +} + } diff --git a/src/Columns/ColumnTuple.h b/src/Columns/ColumnTuple.h index 7b554ed2fea..93b937251c0 100644 --- a/src/Columns/ColumnTuple.h +++ b/src/Columns/ColumnTuple.h @@ -104,6 +104,7 @@ public: void getIndicesOfNonDefaultRows(Offsets & indices, size_t from, size_t limit) const override; SerializationInfoPtr getSerializationInfo() const override; void finalize() override; + bool isFinalized() const override; size_t tupleSize() const { return columns.size(); } diff --git a/src/Columns/IColumn.h b/src/Columns/IColumn.h index 4bf364d9b52..1b67a3d65ad 100644 --- a/src/Columns/IColumn.h +++ b/src/Columns/IColumn.h @@ -444,6 +444,14 @@ public: /// Some columns may require finalization before using of other operations. virtual void finalize() {} + virtual bool isFinalized() const { return true; } + + MutablePtr cloneFinalized() const + { + auto finalized = IColumn::mutate(getPtr()); + finalized->finalize(); + return finalized; + } [[nodiscard]] static MutablePtr mutate(Ptr ptr) { diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index 112998e176d..dea3dfe9e33 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -105,10 +105,11 @@ Array createEmptyArrayField(size_t num_dimensions) DataTypePtr getDataTypeByColumn(const IColumn & column) { auto idx = column.getDataType(); - if (WhichDataType(idx).isSimple()) + WhichDataType which(idx); + if (which.isSimple()) return DataTypeFactory::instance().get(String(magic_enum::enum_name(idx))); - if (WhichDataType(idx).isNothing()) + if (which.isNothing()) return std::make_shared(); if (const auto * column_array = checkAndGetColumn(&column)) diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index 0229754f677..56d098b47b3 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -10,8 +10,6 @@ #include #include -#include - #include #include #include @@ -242,11 +240,8 @@ void SerializationObject::serializeBinaryBulkWithMultipleStreams( if (!column_object.isFinalized()) { - auto finalized = IColumn::mutate(column_object.getPtr()); - auto & finalized_object = assert_cast(*finalized); - finalized_object.finalize(); - serializeBinaryBulkWithMultipleStreams( - finalized_object, offset, limit, settings, state); + auto finalized = column_object.cloneFinalized(); + serializeBinaryBulkWithMultipleStreams(*finalized, offset, limit, settings, state); return; } @@ -374,21 +369,68 @@ void SerializationObject::serializeTextImpl(const IColumn & column, size const auto & column_object = assert_cast(column); const auto & subcolumns = column_object.getSubcolumns(); + bool first = true; writeChar('{', ostr); - for (auto it = subcolumns.begin(); it != subcolumns.end(); ++it) - { - if (it != subcolumns.begin()) - writeCString(",", ostr); - writeDoubleQuoted((*it)->path.getPath(), ostr); + for (const auto & entry : subcolumns) + { + WriteBufferFromOwnString value_buf; + bool have_value = serializeTextFromSubcolumn(entry->data, row_num, value_buf, settings); + if (!have_value) + continue; + + if (!first) + writeCString(",", ostr); + else + first = false; + + writeDoubleQuoted(entry->path.getPath(), ostr); writeChar(':', ostr); - auto serialization = (*it)->data.getLeastCommonType()->getDefaultSerialization(); - serialization->serializeTextJSON((*it)->data.getFinalizedColumn(), row_num, ostr, settings); + auto value = value_buf.stringRef(); + ostr.write(value.data, value.size); } writeChar('}', ostr); } +template +bool SerializationObject::serializeTextFromSubcolumn( + const ColumnObject::Subcolumn & subcolumn, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const +{ + if (subcolumn.isFinalized()) + { + const auto & finalized_column = subcolumn.getFinalizedColumn(); + if (finalized_column.isDefaultAt(row_num)) + return false; + + auto serialization = subcolumn.getLeastCommonType()->getSerialization(*finalized_column.getSerializationInfo()); + serialization->serializeTextJSON(finalized_column, row_num, ostr, settings); + return true; + } + + size_t ind = row_num; + if (ind < subcolumn.getNumberOfDefaultsInPrefix()) + return false; + + ind -= subcolumn.getNumberOfDefaultsInPrefix(); + for (const auto & part : subcolumn.getData()) + { + if (ind < part->size()) + { + if (part->isDefaultAt(ind)) + return false; + + auto serialization = getDataTypeByColumn(*part)->getSerialization(*part->getSerializationInfo()); + serialization->serializeTextJSON(*part, ind, ostr, settings); + return true; + } + + ind -= part->size(); + } + + return false; +} + template void SerializationObject::serializeText(const IColumn & column, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const { diff --git a/src/DataTypes/Serializations/SerializationObject.h b/src/DataTypes/Serializations/SerializationObject.h index 549c8735aee..74ba2e28bdb 100644 --- a/src/DataTypes/Serializations/SerializationObject.h +++ b/src/DataTypes/Serializations/SerializationObject.h @@ -1,5 +1,6 @@ #pragma once +#include "Columns/ColumnObject.h" #include #include @@ -65,6 +66,7 @@ private: void deserializeTextImpl(IColumn & column, Reader && reader) const; void serializeTextImpl(const IColumn & column, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const; + bool serializeTextFromSubcolumn(const ColumnObject::Subcolumn & subcolumn, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const; /// Pool of parser objects to make SerializationObject thread safe. mutable SimpleObjectPool parsers_pool; From 9d2a4d970dcaf0adca714a88d715a21ee6753989 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 9 May 2022 18:05:55 +0000 Subject: [PATCH 005/239] better text serialization of type Object --- src/Columns/ColumnObject.cpp | 4 +- src/DataTypes/ObjectUtils.cpp | 11 +++-- .../Serializations/SerializationObject.cpp | 42 ++++++++----------- .../Serializations/SerializationObject.h | 2 +- .../01825_type_json_in_array.reference | 1 + .../0_stateless/01825_type_json_in_array.sql | 2 +- 6 files changed, 29 insertions(+), 33 deletions(-) diff --git a/src/Columns/ColumnObject.cpp b/src/Columns/ColumnObject.cpp index d0c5997c071..85a80c29756 100644 --- a/src/Columns/ColumnObject.cpp +++ b/src/Columns/ColumnObject.cpp @@ -364,8 +364,8 @@ void ColumnObject::Subcolumn::insertRangeFrom(const Subcolumn & src, size_t star bool ColumnObject::Subcolumn::isFinalized() const { - return data.empty() || - (data.size() == 1 && !data[0]->isSparse() && num_of_defaults_in_prefix == 0); + return num_of_defaults_in_prefix == 0 && + (data.empty() || (data.size() == 1 && !data[0]->isSparse())); } void ColumnObject::Subcolumn::finalize() diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index dea3dfe9e33..4117bcc5c08 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -143,11 +143,14 @@ static DataTypePtr recreateTupleWithElements(const DataTypeTuple & type_tuple, c static std::pair convertObjectColumnToTuple( const ColumnObject & column_object, const DataTypeObject & type_object) { - const auto & subcolumns = column_object.getSubcolumns(); - if (!column_object.isFinalized()) - throw Exception(ErrorCodes::LOGICAL_ERROR, - "Cannot convert to tuple column of type {}. Column should be finalized first", type_object.getName()); + { + auto finalized = column_object.cloneFinalized(); + const auto & finalized_object = assert_cast(*finalized); + return convertObjectColumnToTuple(finalized_object, type_object); + } + + const auto & subcolumns = column_object.getSubcolumns(); PathsInData tuple_paths; DataTypes tuple_types; diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index 56d098b47b3..593a9a1bb02 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -23,6 +23,7 @@ namespace ErrorCodes extern const int INCORRECT_DATA; extern const int CANNOT_READ_ALL_DATA; extern const int LOGICAL_ERROR; + extern const int ARGUMENT_OUT_OF_BOUND; } namespace @@ -369,66 +370,57 @@ void SerializationObject::serializeTextImpl(const IColumn & column, size const auto & column_object = assert_cast(column); const auto & subcolumns = column_object.getSubcolumns(); - bool first = true; writeChar('{', ostr); - - for (const auto & entry : subcolumns) + for (auto it = subcolumns.begin(); it != subcolumns.end(); ++it) { - WriteBufferFromOwnString value_buf; - bool have_value = serializeTextFromSubcolumn(entry->data, row_num, value_buf, settings); - if (!have_value) - continue; - - if (!first) + const auto & entry = *it; + if (it != subcolumns.begin()) writeCString(",", ostr); - else - first = false; writeDoubleQuoted(entry->path.getPath(), ostr); writeChar(':', ostr); - - auto value = value_buf.stringRef(); - ostr.write(value.data, value.size); + serializeTextFromSubcolumn(entry->data, row_num, ostr, settings); } writeChar('}', ostr); } template -bool SerializationObject::serializeTextFromSubcolumn( +void SerializationObject::serializeTextFromSubcolumn( const ColumnObject::Subcolumn & subcolumn, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const { if (subcolumn.isFinalized()) { const auto & finalized_column = subcolumn.getFinalizedColumn(); - if (finalized_column.isDefaultAt(row_num)) - return false; - auto serialization = subcolumn.getLeastCommonType()->getSerialization(*finalized_column.getSerializationInfo()); serialization->serializeTextJSON(finalized_column, row_num, ostr, settings); - return true; + return; } size_t ind = row_num; if (ind < subcolumn.getNumberOfDefaultsInPrefix()) - return false; + { + /// Suboptimal, but it should happen rarely. + auto tmp_column = subcolumn.getLeastCommonType()->createColumn(); + tmp_column->insertDefault(); + auto serialization = subcolumn.getLeastCommonType()->getSerialization(*tmp_column->getSerializationInfo()); + serialization->serializeTextJSON(*tmp_column, 0, ostr, settings); + return; + } ind -= subcolumn.getNumberOfDefaultsInPrefix(); for (const auto & part : subcolumn.getData()) { if (ind < part->size()) { - if (part->isDefaultAt(ind)) - return false; - auto serialization = getDataTypeByColumn(*part)->getSerialization(*part->getSerializationInfo()); serialization->serializeTextJSON(*part, ind, ostr, settings); - return true; + return; } ind -= part->size(); } - return false; + throw Exception(ErrorCodes::ARGUMENT_OUT_OF_BOUND, "Index ({}) for text serialization is out of range", row_num); } template diff --git a/src/DataTypes/Serializations/SerializationObject.h b/src/DataTypes/Serializations/SerializationObject.h index 74ba2e28bdb..0abb6a98d63 100644 --- a/src/DataTypes/Serializations/SerializationObject.h +++ b/src/DataTypes/Serializations/SerializationObject.h @@ -66,7 +66,7 @@ private: void deserializeTextImpl(IColumn & column, Reader && reader) const; void serializeTextImpl(const IColumn & column, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const; - bool serializeTextFromSubcolumn(const ColumnObject::Subcolumn & subcolumn, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const; + void serializeTextFromSubcolumn(const ColumnObject::Subcolumn & subcolumn, size_t row_num, WriteBuffer & ostr, const FormatSettings & settings) const; /// Pool of parser objects to make SerializationObject thread safe. mutable SimpleObjectPool parsers_pool; diff --git a/tests/queries/0_stateless/01825_type_json_in_array.reference b/tests/queries/0_stateless/01825_type_json_in_array.reference index 0bdb9917622..3f6dc6ccd88 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.reference +++ b/tests/queries/0_stateless/01825_type_json_in_array.reference @@ -18,4 +18,5 @@ Tuple(k2 String, k3 String, k4 Int8) {"arr":[{"x":1}]} {"arr":{"x":{"y":1},"t":{"y":2}}} {"arr":[1,{"y":1}]} +{"arr":[2,{"y":2}]} {"arr":[{"x":"aaa","y":[1,2,3]}]} diff --git a/tests/queries/0_stateless/01825_type_json_in_array.sql b/tests/queries/0_stateless/01825_type_json_in_array.sql index b02202ec175..935969272fa 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.sql +++ b/tests/queries/0_stateless/01825_type_json_in_array.sql @@ -30,5 +30,5 @@ DROP TABLE t_json_array; SELECT * FROM values('arr Array(JSON)', '[\'{"x" : 1}\']') FORMAT JSONEachRow; SELECT * FROM values('arr Map(String, JSON)', '{\'x\' : \'{"y" : 1}\', \'t\' : \'{"y" : 2}\'}') FORMAT JSONEachRow; -SELECT * FROM values('arr Tuple(Int32, JSON)', '(1, \'{"y" : 1}\')') FORMAT JSONEachRow; +SELECT * FROM values('arr Tuple(Int32, JSON)', '(1, \'{"y" : 1}\')', '(2, \'{"y" : 2}\')') FORMAT JSONEachRow; SELECT * FROM format(JSONEachRow, '{"arr" : [{"x" : "aaa", "y" : [1,2,3]}]}') FORMAT JSONEachRow; From 1f91655cf95b257f8dd5ffec33250bf945cc2f6f Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 10 May 2022 15:32:17 +0000 Subject: [PATCH 006/239] support Object inside Map --- src/Columns/ColumnMap.h | 2 + src/Columns/ColumnObject.cpp | 13 +- src/DataTypes/DataTypeMap.cpp | 21 +++ src/DataTypes/DataTypeMap.h | 2 + src/DataTypes/IDataType.h | 1 + src/DataTypes/ObjectUtils.cpp | 121 +++++++++++------- .../Serializations/SerializationObject.h | 4 +- .../01825_type_json_in_other_types.reference | 17 +++ .../01825_type_json_in_other_types.sh | 91 +++++++++++++ 9 files changed, 219 insertions(+), 53 deletions(-) create mode 100644 tests/queries/0_stateless/01825_type_json_in_other_types.reference create mode 100755 tests/queries/0_stateless/01825_type_json_in_other_types.sh diff --git a/src/Columns/ColumnMap.h b/src/Columns/ColumnMap.h index 95838e70d10..3b9ff07c09e 100644 --- a/src/Columns/ColumnMap.h +++ b/src/Columns/ColumnMap.h @@ -92,6 +92,8 @@ public: bool structureEquals(const IColumn & rhs) const override; double getRatioOfDefaultRows(double sample_ratio) const override; void getIndicesOfNonDefaultRows(Offsets & indices, size_t from, size_t limit) const override; + void finalize() override { nested->finalize(); } + bool isFinalized() const override { return nested->isFinalized(); } const ColumnArray & getNestedColumn() const { return assert_cast(*nested); } ColumnArray & getNestedColumn() { return assert_cast(*nested); } diff --git a/src/Columns/ColumnObject.cpp b/src/Columns/ColumnObject.cpp index 85a80c29756..34919ad2398 100644 --- a/src/Columns/ColumnObject.cpp +++ b/src/Columns/ColumnObject.cpp @@ -276,8 +276,8 @@ static bool isConversionRequiredBetweenIntegers(const IDataType & lhs, const IDa bool is_native_int = which_lhs.isNativeInt() && which_rhs.isNativeInt(); bool is_native_uint = which_lhs.isNativeUInt() && which_rhs.isNativeUInt(); - return (is_native_int || is_native_uint) - && lhs.getSizeOfValueInMemory() <= rhs.getSizeOfValueInMemory(); + return (!is_native_int && !is_native_uint) + || lhs.getSizeOfValueInMemory() > rhs.getSizeOfValueInMemory(); } void ColumnObject::Subcolumn::insert(Field field, FieldInfo info) @@ -320,7 +320,7 @@ void ColumnObject::Subcolumn::insert(Field field, FieldInfo info) } else if (!least_common_base_type->equals(*base_type) && !isNothing(base_type)) { - if (!isConversionRequiredBetweenIntegers(*base_type, *least_common_base_type)) + if (isConversionRequiredBetweenIntegers(*base_type, *least_common_base_type)) { base_type = getLeastSupertype(DataTypes{std::move(base_type), least_common_base_type}, true); type_changed = true; @@ -832,6 +832,10 @@ void ColumnObject::addNestedSubcolumn(const PathInData & key, const FieldInfo & if (num_rows == 0) num_rows = new_size; + else if (new_size != num_rows) + throw Exception(ErrorCodes::SIZES_OF_COLUMNS_DOESNT_MATCH, + "Required size of subcolumn {} ({}) is inconsistent with column size ({})", + key.getPath(), new_size, num_rows); } PathsInData ColumnObject::getKeys() const @@ -851,9 +855,6 @@ bool ColumnObject::isFinalized() const void ColumnObject::finalize() { - if (isFinalized()) - return; - size_t old_size = size(); Subcolumns new_subcolumns; for (auto && entry : subcolumns) diff --git a/src/DataTypes/DataTypeMap.cpp b/src/DataTypes/DataTypeMap.cpp index 42ec739c33b..d49c205fc59 100644 --- a/src/DataTypes/DataTypeMap.cpp +++ b/src/DataTypes/DataTypeMap.cpp @@ -22,6 +22,27 @@ namespace ErrorCodes extern const int BAD_ARGUMENTS; } +DataTypeMap::DataTypeMap(const DataTypePtr & nested_) + : nested(nested_) +{ + const auto * type_array = typeid_cast(nested.get()); + if (!type_array) + throw Exception(ErrorCodes::BAD_ARGUMENTS, + "Expected Array(Tuple(key, value)) type, got {}", nested->getName()); + + const auto * type_tuple = typeid_cast(type_array->getNestedType().get()); + if (!type_tuple) + throw Exception(ErrorCodes::BAD_ARGUMENTS, + "Expected Array(Tuple(key, value)) type, got {}", nested->getName()); + + if (type_tuple->getElements().size() != 2) + throw Exception(ErrorCodes::BAD_ARGUMENTS, + "Expected Array(Tuple(key, value)) type, got {}", nested->getName()); + + key_type = type_tuple->getElement(0); + value_type = type_tuple->getElement(1); + assertKeyType(); +} DataTypeMap::DataTypeMap(const DataTypes & elems_) { diff --git a/src/DataTypes/DataTypeMap.h b/src/DataTypes/DataTypeMap.h index 479008031fe..2ab5c602a25 100644 --- a/src/DataTypes/DataTypeMap.h +++ b/src/DataTypes/DataTypeMap.h @@ -23,6 +23,7 @@ private: public: static constexpr bool is_parametric = true; + explicit DataTypeMap(const DataTypePtr & nested_); explicit DataTypeMap(const DataTypes & elems); DataTypeMap(const DataTypePtr & key_type_, const DataTypePtr & value_type_); @@ -40,6 +41,7 @@ public: bool isComparable() const override { return key_type->isComparable() && value_type->isComparable(); } bool isParametric() const override { return true; } bool haveSubtypes() const override { return true; } + bool hasDynamicSubcolumns() const override { return nested->hasDynamicSubcolumns(); } const DataTypePtr & getKeyType() const { return key_type; } const DataTypePtr & getValueType() const { return value_type; } diff --git a/src/DataTypes/IDataType.h b/src/DataTypes/IDataType.h index 6d1d17c9e40..1fcac18b124 100644 --- a/src/DataTypes/IDataType.h +++ b/src/DataTypes/IDataType.h @@ -287,6 +287,7 @@ public: /// Strings, Numbers, Date, DateTime, Nullable virtual bool canBeInsideLowCardinality() const { return false; } + /// Object, Array(Object), Tuple(..., Object, ...) virtual bool hasDynamicSubcolumns() const { return false; } /// Updates avg_value_size_hint for newly read column. Uses to optimize deserialization. Zero expected for first column. diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index 4117bcc5c08..4aea9a5f7e3 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -2,6 +2,7 @@ #include #include #include +#include #include #include #include @@ -12,6 +13,7 @@ #include #include #include +#include #include #include #include @@ -191,6 +193,19 @@ static std::pair recursivlyConvertDynamicColumnToTuple( }; } + if (const auto * type_map = typeid_cast(type.get())) + { + const auto & column_map = assert_cast(*column); + auto [new_column, new_type] = recursivlyConvertDynamicColumnToTuple( + column_map.getNestedColumnPtr(), type_map->getNestedType()); + + return + { + ColumnMap::create(std::move(new_column)), + std::make_shared(std::move(new_type)), + }; + } + if (const auto * type_tuple = typeid_cast(type.get())) { const auto & column_tuple = assert_cast(*column); @@ -343,7 +358,58 @@ static DataTypePtr getLeastCommonTypeForObject(const DataTypes & types, bool che return unflattenTuple(tuple_paths, tuple_types); } -DataTypePtr getLeastCommonTypeForDynamicColumnsImpl( +static DataTypePtr getLeastCommonTypeForDynamicColumnsImpl( + const DataTypePtr & type_in_storage, const DataTypes & concrete_types, bool check_ambiguos_paths); + +template +static DataTypePtr getLeastCommonTypeForColumnWithNestedType( + const Type & type, const DataTypes & concrete_types, bool check_ambiguos_paths) +{ + DataTypes nested_types; + nested_types.reserve(concrete_types.size()); + + for (const auto & concrete_type : concrete_types) + { + const auto * type_with_nested_conctete = typeid_cast(concrete_type.get()); + if (!type_with_nested_conctete) + throw Exception(ErrorCodes::TYPE_MISMATCH, "Expected {} type, got {}", demangle(typeid(Type).name()), concrete_type->getName()); + + nested_types.push_back(type_with_nested_conctete->getNestedType()); + } + + return std::make_shared( + getLeastCommonTypeForDynamicColumnsImpl( + type.getNestedType(), nested_types, check_ambiguos_paths)); +} + +static DataTypePtr getLeastCommonTypeForTuple( + const DataTypeTuple & type, const DataTypes & concrete_types, bool check_ambiguos_paths) +{ + const auto & element_types = type.getElements(); + DataTypes new_element_types(element_types.size()); + + for (size_t i = 0; i < element_types.size(); ++i) + { + DataTypes concrete_element_types; + concrete_element_types.reserve(concrete_types.size()); + + for (const auto & type_concrete : concrete_types) + { + const auto * type_tuple_conctete = typeid_cast(type_concrete.get()); + if (!type_tuple_conctete) + throw Exception(ErrorCodes::TYPE_MISMATCH, "Expected Tuple type, got {}", type_concrete->getName()); + + concrete_element_types.push_back(type_tuple_conctete->getElement(i)); + } + + new_element_types[i] = getLeastCommonTypeForDynamicColumnsImpl( + element_types[i], concrete_element_types, check_ambiguos_paths); + } + + return recreateTupleWithElements(type, new_element_types); +} + +static DataTypePtr getLeastCommonTypeForDynamicColumnsImpl( const DataTypePtr & type_in_storage, const DataTypes & concrete_types, bool check_ambiguos_paths) { if (!type_in_storage->hasDynamicSubcolumns()) @@ -353,49 +419,13 @@ DataTypePtr getLeastCommonTypeForDynamicColumnsImpl( return getLeastCommonTypeForObject(concrete_types, check_ambiguos_paths); if (const auto * type_array = typeid_cast(type_in_storage.get())) - { - DataTypes nested_types; - nested_types.reserve(concrete_types.size()); + return getLeastCommonTypeForColumnWithNestedType(*type_array, concrete_types, check_ambiguos_paths); - for (const auto & type : concrete_types) - { - const auto * type_array_conctete = typeid_cast(type.get()); - if (!type_array_conctete) - throw Exception(ErrorCodes::TYPE_MISMATCH, "Expected Array type, got {}", type->getName()); - - nested_types.push_back(type_array_conctete->getNestedType()); - } - - return std::make_shared( - getLeastCommonTypeForDynamicColumnsImpl( - type_array->getNestedType(), nested_types, check_ambiguos_paths)); - } + if (const auto * type_map = typeid_cast(type_in_storage.get())) + return getLeastCommonTypeForColumnWithNestedType(*type_map, concrete_types, check_ambiguos_paths); if (const auto * type_tuple = typeid_cast(type_in_storage.get())) - { - const auto & element_types = type_tuple->getElements(); - DataTypes new_element_types(element_types.size()); - - for (size_t i = 0; i < element_types.size(); ++i) - { - DataTypes concrete_element_types; - concrete_element_types.reserve(concrete_types.size()); - - for (const auto & type : concrete_types) - { - const auto * type_tuple_conctete = typeid_cast(type.get()); - if (!type_tuple_conctete) - throw Exception(ErrorCodes::TYPE_MISMATCH, "Expected Tuple type, got {}", type->getName()); - - concrete_element_types.push_back(type_tuple_conctete->getElement(i)); - } - - new_element_types[i] = getLeastCommonTypeForDynamicColumnsImpl( - element_types[i], concrete_element_types, check_ambiguos_paths); - } - - return recreateTupleWithElements(*type_tuple, new_element_types); - } + return getLeastCommonTypeForTuple(*type_tuple, concrete_types, check_ambiguos_paths); throw Exception(ErrorCodes::LOGICAL_ERROR, "Type {} unexpectedly has dynamic columns", type_in_storage->getName()); } @@ -428,16 +458,17 @@ DataTypePtr createConcreteEmptyDynamicColumn(const DataTypePtr & type_in_storage return type_in_storage; if (isObject(type_in_storage)) - { return std::make_shared( - DataTypes{std::make_shared()}, - Names{ColumnObject::COLUMN_NAME_DUMMY}); - } + DataTypes{std::make_shared()}, Names{ColumnObject::COLUMN_NAME_DUMMY}); if (const auto * type_array = typeid_cast(type_in_storage.get())) return std::make_shared( createConcreteEmptyDynamicColumn(type_array->getNestedType())); + if (const auto * type_map = typeid_cast(type_in_storage.get())) + return std::make_shared( + createConcreteEmptyDynamicColumn(type_map->getNestedType())); + if (const auto * type_tuple = typeid_cast(type_in_storage.get())) { const auto & elements = type_tuple->getElements(); diff --git a/src/DataTypes/Serializations/SerializationObject.h b/src/DataTypes/Serializations/SerializationObject.h index 0abb6a98d63..fc34508f1bb 100644 --- a/src/DataTypes/Serializations/SerializationObject.h +++ b/src/DataTypes/Serializations/SerializationObject.h @@ -1,6 +1,6 @@ #pragma once -#include "Columns/ColumnObject.h" +#include #include #include @@ -8,7 +8,7 @@ namespace DB { /// Serialization for data type Object. -/// Supported only test serialization/deserialization. +/// Supported only text serialization/deserialization. /// and binary bulk serialization/deserialization without position independent /// encoding, i.e. serialization/deserialization into Native format. template diff --git a/tests/queries/0_stateless/01825_type_json_in_other_types.reference b/tests/queries/0_stateless/01825_type_json_in_other_types.reference new file mode 100644 index 00000000000..b94885a65ab --- /dev/null +++ b/tests/queries/0_stateless/01825_type_json_in_other_types.reference @@ -0,0 +1,17 @@ +Tuple(String, Map(String, Array(Tuple(k1 Nested(k2 Int8, k3 Int8, k5 String), k4 String))), Tuple(k1 String, k2 Tuple(k3 String, k4 String))) +============= +{"id":1,"data":["foo",{"aa":[{"k1":[{"k2":1,"k3":2,"k5":""},{"k2":0,"k3":3,"k5":""}],"k4":""},{"k1":[{"k2":4,"k3":0,"k5":""},{"k2":0,"k3":5,"k5":""},{"k2":6,"k3":0,"k5":""}],"k4":"qqq"}],"bb":[{"k1":[],"k4":"www"},{"k1":[{"k2":7,"k3":8,"k5":""},{"k2":9,"k3":10,"k5":""},{"k2":11,"k3":12,"k5":""}],"k4":""}]},{"k1":"aa","k2":{"k3":"bb","k4":"c"}}]} +{"id":2,"data":["bar",{"aa":[{"k1":[{"k2":13,"k3":14,"k5":""},{"k2":15,"k3":16,"k5":""}],"k4":"www"}]},{"k1":"","k2":{"k3":"","k4":""}}]} +{"id":3,"data":["some",{"aa":[{"k1":[{"k2":0,"k3":20,"k5":"some"}],"k4":""}]},{"k1":"eee","k2":{"k3":"","k4":""}}]} +============= +{"aa":[{"k1":[{"k2":1,"k3":2,"k5":""},{"k2":0,"k3":3,"k5":""}],"k4":""},{"k1":[{"k2":4,"k3":0,"k5":""},{"k2":0,"k3":5,"k5":""},{"k2":6,"k3":0,"k5":""}],"k4":"qqq"}],"bb":[{"k1":[],"k4":"www"},{"k1":[{"k2":7,"k3":8,"k5":""},{"k2":9,"k3":10,"k5":""},{"k2":11,"k3":12,"k5":""}],"k4":""}]} +{"aa":[{"k1":[{"k2":13,"k3":14,"k5":""},{"k2":15,"k3":16,"k5":""}],"k4":"www"}],"bb":[]} +{"aa":[{"k1":[{"k2":0,"k3":20,"k5":"some"}],"k4":""}],"bb":[]} +============= +{"k1":[[{"k2":1,"k3":2,"k5":""},{"k2":0,"k3":3,"k5":""}],[{"k2":4,"k3":0,"k5":""},{"k2":0,"k3":5,"k5":""},{"k2":6,"k3":0,"k5":""}]],"k4":["","qqq"]} +{"k1":[[{"k2":13,"k3":14,"k5":""},{"k2":15,"k3":16,"k5":""}]],"k4":["www"]} +{"k1":[[{"k2":0,"k3":20,"k5":"some"}]],"k4":[""]} +============= +{"obj":{"k1":"aa","k2":{"k3":"bb","k4":"c"}}} +{"obj":{"k1":"","k2":{"k3":"","k4":""}}} +{"obj":{"k1":"eee","k2":{"k3":"","k4":""}}} diff --git a/tests/queries/0_stateless/01825_type_json_in_other_types.sh b/tests/queries/0_stateless/01825_type_json_in_other_types.sh new file mode 100755 index 00000000000..e9cf0bcaca1 --- /dev/null +++ b/tests/queries/0_stateless/01825_type_json_in_other_types.sh @@ -0,0 +1,91 @@ +#!/usr/bin/env bash +# Tags: no-fasttest + +CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +# shellcheck source=../shell_config.sh +. "$CUR_DIR"/../shell_config.sh + +${CLICKHOUSE_CLIENT} -q "SET allow_experimental_object_type = 1" +${CLICKHOUSE_CLIENT} -q "DROP TABLE IF EXISTS t_json_nested" + +${CLICKHOUSE_CLIENT} -q " + CREATE TABLE t_json_nested + ( + id UInt32, + data Tuple(String, Map(String, Array(JSON)), JSON) + ) + ENGINE = MergeTree ORDER BY id" --allow_experimental_object_type 1 + +cat < Date: Wed, 11 May 2022 13:37:02 +0000 Subject: [PATCH 007/239] fix checks --- src/Columns/ColumnObject.cpp | 1 - src/DataTypes/ObjectUtils.cpp | 4 ++-- src/DataTypes/Serializations/SerializationObject.cpp | 1 - 3 files changed, 2 insertions(+), 4 deletions(-) diff --git a/src/Columns/ColumnObject.cpp b/src/Columns/ColumnObject.cpp index 34919ad2398..c6a63cc4172 100644 --- a/src/Columns/ColumnObject.cpp +++ b/src/Columns/ColumnObject.cpp @@ -22,7 +22,6 @@ namespace ErrorCodes extern const int ILLEGAL_COLUMN; extern const int DUPLICATE_COLUMN; extern const int NUMBER_OF_DIMENSIONS_MISMATHED; - extern const int NOT_IMPLEMENTED; extern const int SIZES_OF_COLUMNS_DOESNT_MATCH; extern const int ARGUMENT_OUT_OF_BOUND; } diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index 4aea9a5f7e3..e75e2b7bc62 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -188,7 +188,7 @@ static std::pair recursivlyConvertDynamicColumnToTuple( return { - ColumnArray::create(std::move(new_column), column_array.getOffsetsPtr()), + ColumnArray::create(new_column, column_array.getOffsetsPtr()), std::make_shared(std::move(new_type)), }; } @@ -201,7 +201,7 @@ static std::pair recursivlyConvertDynamicColumnToTuple( return { - ColumnMap::create(std::move(new_column)), + ColumnMap::create(new_column), std::make_shared(std::move(new_type)), }; } diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index 593a9a1bb02..af816568790 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -22,7 +22,6 @@ namespace ErrorCodes extern const int NOT_IMPLEMENTED; extern const int INCORRECT_DATA; extern const int CANNOT_READ_ALL_DATA; - extern const int LOGICAL_ERROR; extern const int ARGUMENT_OUT_OF_BOUND; } From 86b29b7f1a088fd0da3e2a72ac69b7aee9bf4357 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Thu, 8 Sep 2022 15:16:39 +0000 Subject: [PATCH 008/239] fix serilization of Object inside other types --- src/DataTypes/Serializations/ISerialization.h | 2 + .../Serializations/SerializationArray.cpp | 4 +- .../Serializations/SerializationArray.h | 1 + .../SerializationLowCardinality.cpp | 1 + .../SerializationLowCardinality.h | 1 + .../Serializations/SerializationMap.cpp | 3 +- .../Serializations/SerializationMap.h | 1 + .../Serializations/SerializationNamed.cpp | 3 +- .../Serializations/SerializationNamed.h | 1 + .../Serializations/SerializationNullable.cpp | 4 +- .../Serializations/SerializationNullable.h | 1 + .../Serializations/SerializationObject.cpp | 43 ++++++++----------- .../Serializations/SerializationObject.h | 1 + .../Serializations/SerializationSparse.cpp | 4 +- .../Serializations/SerializationSparse.h | 1 + .../Serializations/SerializationTuple.cpp | 3 +- .../Serializations/SerializationTuple.h | 1 + .../Serializations/SerializationWrapper.cpp | 3 +- .../Serializations/SerializationWrapper.h | 1 + .../tests/gtest_object_serialization.cpp | 2 +- src/Formats/NativeWriter.cpp | 2 +- src/Functions/blockSerializedSize.cpp | 2 +- .../MergeTreeDataPartWriterCompact.cpp | 2 +- .../MergeTree/MergeTreeDataPartWriterWide.cpp | 2 +- src/Storages/MergeTree/MergeTreeIndexSet.cpp | 5 ++- src/Storages/StorageLog.cpp | 2 +- 26 files changed, 57 insertions(+), 39 deletions(-) diff --git a/src/DataTypes/Serializations/ISerialization.h b/src/DataTypes/Serializations/ISerialization.h index b5d2082631e..f23b1095749 100644 --- a/src/DataTypes/Serializations/ISerialization.h +++ b/src/DataTypes/Serializations/ISerialization.h @@ -219,7 +219,9 @@ public: }; /// Call before serializeBinaryBulkWithMultipleStreams chain to write something before first mark. + /// Column may be used only to retrieve the structure. virtual void serializeBinaryBulkStatePrefix( + const IColumn & /*column*/, SerializeBinaryBulkSettings & /*settings*/, SerializeBinaryBulkStatePtr & /*state*/) const {} diff --git a/src/DataTypes/Serializations/SerializationArray.cpp b/src/DataTypes/Serializations/SerializationArray.cpp index 625f2dce0b0..13577633327 100644 --- a/src/DataTypes/Serializations/SerializationArray.cpp +++ b/src/DataTypes/Serializations/SerializationArray.cpp @@ -235,11 +235,13 @@ void SerializationArray::enumerateStreams( } void SerializationArray::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { settings.path.push_back(Substream::ArrayElements); - nested->serializeBinaryBulkStatePrefix(settings, state); + const auto & column_array = assert_cast(column); + nested->serializeBinaryBulkStatePrefix(column_array.getData(), settings, state); settings.path.pop_back(); } diff --git a/src/DataTypes/Serializations/SerializationArray.h b/src/DataTypes/Serializations/SerializationArray.h index 3769f8a4513..89bf50697fc 100644 --- a/src/DataTypes/Serializations/SerializationArray.h +++ b/src/DataTypes/Serializations/SerializationArray.h @@ -41,6 +41,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationLowCardinality.cpp b/src/DataTypes/Serializations/SerializationLowCardinality.cpp index 8e19c5a740b..942ebbbe595 100644 --- a/src/DataTypes/Serializations/SerializationLowCardinality.cpp +++ b/src/DataTypes/Serializations/SerializationLowCardinality.cpp @@ -225,6 +225,7 @@ struct DeserializeStateLowCardinality : public ISerialization::DeserializeBinary }; void SerializationLowCardinality::serializeBinaryBulkStatePrefix( + const IColumn & /*column*/, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { diff --git a/src/DataTypes/Serializations/SerializationLowCardinality.h b/src/DataTypes/Serializations/SerializationLowCardinality.h index 96e3a297d6a..c42f30f223a 100644 --- a/src/DataTypes/Serializations/SerializationLowCardinality.h +++ b/src/DataTypes/Serializations/SerializationLowCardinality.h @@ -23,6 +23,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationMap.cpp b/src/DataTypes/Serializations/SerializationMap.cpp index ea22070b5b1..824953da322 100644 --- a/src/DataTypes/Serializations/SerializationMap.cpp +++ b/src/DataTypes/Serializations/SerializationMap.cpp @@ -273,10 +273,11 @@ void SerializationMap::enumerateStreams( } void SerializationMap::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { - nested->serializeBinaryBulkStatePrefix(settings, state); + nested->serializeBinaryBulkStatePrefix(extractNestedColumn(column), settings, state); } void SerializationMap::serializeBinaryBulkStateSuffix( diff --git a/src/DataTypes/Serializations/SerializationMap.h b/src/DataTypes/Serializations/SerializationMap.h index 93b3e179499..746598c2cc4 100644 --- a/src/DataTypes/Serializations/SerializationMap.h +++ b/src/DataTypes/Serializations/SerializationMap.h @@ -37,6 +37,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationNamed.cpp b/src/DataTypes/Serializations/SerializationNamed.cpp index 097e9cedfbe..5217bb6c068 100644 --- a/src/DataTypes/Serializations/SerializationNamed.cpp +++ b/src/DataTypes/Serializations/SerializationNamed.cpp @@ -17,11 +17,12 @@ void SerializationNamed::enumerateStreams( } void SerializationNamed::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { addToPath(settings.path); - nested_serialization->serializeBinaryBulkStatePrefix(settings, state); + nested_serialization->serializeBinaryBulkStatePrefix(column, settings, state); settings.path.pop_back(); } diff --git a/src/DataTypes/Serializations/SerializationNamed.h b/src/DataTypes/Serializations/SerializationNamed.h index 343b96c16e3..12e5ce63e93 100644 --- a/src/DataTypes/Serializations/SerializationNamed.h +++ b/src/DataTypes/Serializations/SerializationNamed.h @@ -31,6 +31,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationNullable.cpp b/src/DataTypes/Serializations/SerializationNullable.cpp index a6273deaa30..0ee0443f251 100644 --- a/src/DataTypes/Serializations/SerializationNullable.cpp +++ b/src/DataTypes/Serializations/SerializationNullable.cpp @@ -73,11 +73,13 @@ void SerializationNullable::enumerateStreams( } void SerializationNullable::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { settings.path.push_back(Substream::NullableElements); - nested->serializeBinaryBulkStatePrefix(settings, state); + const auto & column_nullable = assert_cast(column); + nested->serializeBinaryBulkStatePrefix(column_nullable.getNestedColumn(), settings, state); settings.path.pop_back(); } diff --git a/src/DataTypes/Serializations/SerializationNullable.h b/src/DataTypes/Serializations/SerializationNullable.h index e6e0e4f33c2..bbf9a1e0c48 100644 --- a/src/DataTypes/Serializations/SerializationNullable.h +++ b/src/DataTypes/Serializations/SerializationNullable.h @@ -19,6 +19,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index e7a403106a8..5577cab5982 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -139,7 +139,6 @@ void SerializationObject::checkSerializationIsSupported(const TSettings template struct SerializationObject::SerializeStateObject : public ISerialization::SerializeBinaryBulkState { - bool is_first = true; DataTypePtr nested_type; SerializationPtr nested_serialization; SerializeBinaryBulkStatePtr nested_state; @@ -156,6 +155,7 @@ struct SerializationObject::DeserializeStateObject : public ISerializati template void SerializationObject::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { @@ -166,13 +166,25 @@ void SerializationObject::serializeBinaryBulkStatePrefix( settings.path.push_back(Substream::ObjectStructure); auto * stream = settings.getter(settings.path); - settings.path.pop_back(); if (!stream) throw Exception(ErrorCodes::LOGICAL_ERROR, "Missing stream for kind of binary serialization"); writeIntBinary(static_cast(BinarySerializationKind::TUPLE), *stream); - state = std::make_shared(); + + const auto & column_object = assert_cast(column); + auto [tuple_column, tuple_type] = unflattenObjectToTuple(column_object); + writeStringBinary(tuple_type->getName(), *stream); + + auto state_object = std::make_shared(); + state_object->nested_type = tuple_type; + state_object->nested_serialization = tuple_type->getDefaultSerialization(); + + settings.path.back() = Substream::ObjectData; + state_object->nested_serialization->serializeBinaryBulkStatePrefix(*tuple_column, settings, state_object->nested_state); + + state = std::move(state_object); + settings.path.pop_back(); } template @@ -266,25 +278,7 @@ void SerializationObject::serializeBinaryBulkWithMultipleStreams( auto [tuple_column, tuple_type] = unflattenObjectToTuple(column_object); - if (state_object->is_first) - { - /// Actually it's a part of serializeBinaryBulkStatePrefix, - /// but it cannot be done there, because we have to know the - /// structure of column. - - settings.path.push_back(Substream::ObjectStructure); - if (auto * stream = settings.getter(settings.path)) - writeStringBinary(tuple_type->getName(), *stream); - - state_object->nested_type = tuple_type; - state_object->nested_serialization = tuple_type->getDefaultSerialization(); - state_object->is_first = false; - - settings.path.back() = Substream::ObjectData; - state_object->nested_serialization->serializeBinaryBulkStatePrefix(settings, state_object->nested_state); - settings.path.pop_back(); - } - else if (!state_object->nested_type->equals(*tuple_type)) + if (!state_object->nested_type->equals(*tuple_type)) { throw Exception(ErrorCodes::LOGICAL_ERROR, "Types of internal column of Object mismatched. Expected: {}, Got: {}", @@ -452,8 +446,9 @@ void SerializationObject::serializeTextFromSubcolumn( { if (ind < part->size()) { - auto info = least_common_type->getSerializationInfo(*part); - auto serialization = least_common_type->getSerialization(*info); + auto part_type = getDataTypeByColumn(*part); + auto info = part_type->getSerializationInfo(*part); + auto serialization = part_type->getSerialization(*info); serialization->serializeTextJSON(*part, ind, ostr, settings); return; } diff --git a/src/DataTypes/Serializations/SerializationObject.h b/src/DataTypes/Serializations/SerializationObject.h index ef8019ead77..47a7127cd1c 100644 --- a/src/DataTypes/Serializations/SerializationObject.h +++ b/src/DataTypes/Serializations/SerializationObject.h @@ -31,6 +31,7 @@ public: */ void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationSparse.cpp b/src/DataTypes/Serializations/SerializationSparse.cpp index 6fa40e460c5..a3d7e071215 100644 --- a/src/DataTypes/Serializations/SerializationSparse.cpp +++ b/src/DataTypes/Serializations/SerializationSparse.cpp @@ -184,11 +184,13 @@ void SerializationSparse::enumerateStreams( } void SerializationSparse::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { settings.path.push_back(Substream::SparseElements); - nested->serializeBinaryBulkStatePrefix(settings, state); + const auto & column_sparse = assert_cast(column); + nested->serializeBinaryBulkStatePrefix(column_sparse.getValuesColumn(), settings, state); settings.path.pop_back(); } diff --git a/src/DataTypes/Serializations/SerializationSparse.h b/src/DataTypes/Serializations/SerializationSparse.h index 54ab4853360..f6082f9a949 100644 --- a/src/DataTypes/Serializations/SerializationSparse.h +++ b/src/DataTypes/Serializations/SerializationSparse.h @@ -33,6 +33,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationTuple.cpp b/src/DataTypes/Serializations/SerializationTuple.cpp index 8138b15c9af..21f6d47f1f9 100644 --- a/src/DataTypes/Serializations/SerializationTuple.cpp +++ b/src/DataTypes/Serializations/SerializationTuple.cpp @@ -317,6 +317,7 @@ struct DeserializeBinaryBulkStateTuple : public ISerialization::DeserializeBinar void SerializationTuple::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { @@ -324,7 +325,7 @@ void SerializationTuple::serializeBinaryBulkStatePrefix( tuple_state->states.resize(elems.size()); for (size_t i = 0; i < elems.size(); ++i) - elems[i]->serializeBinaryBulkStatePrefix(settings, tuple_state->states[i]); + elems[i]->serializeBinaryBulkStatePrefix(extractElementColumn(column, i), settings, tuple_state->states[i]); state = std::move(tuple_state); } diff --git a/src/DataTypes/Serializations/SerializationTuple.h b/src/DataTypes/Serializations/SerializationTuple.h index e82d8473645..0eab14431d3 100644 --- a/src/DataTypes/Serializations/SerializationTuple.h +++ b/src/DataTypes/Serializations/SerializationTuple.h @@ -39,6 +39,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/SerializationWrapper.cpp b/src/DataTypes/Serializations/SerializationWrapper.cpp index 271c53dfcf1..8d41746055d 100644 --- a/src/DataTypes/Serializations/SerializationWrapper.cpp +++ b/src/DataTypes/Serializations/SerializationWrapper.cpp @@ -13,10 +13,11 @@ void SerializationWrapper::enumerateStreams( } void SerializationWrapper::serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const { - nested_serialization->serializeBinaryBulkStatePrefix(settings, state); + nested_serialization->serializeBinaryBulkStatePrefix(column, settings, state); } void SerializationWrapper::serializeBinaryBulkStateSuffix( diff --git a/src/DataTypes/Serializations/SerializationWrapper.h b/src/DataTypes/Serializations/SerializationWrapper.h index 43fc7e9914a..a3a30b8f6b7 100644 --- a/src/DataTypes/Serializations/SerializationWrapper.h +++ b/src/DataTypes/Serializations/SerializationWrapper.h @@ -26,6 +26,7 @@ public: const SubstreamData & data) const override; void serializeBinaryBulkStatePrefix( + const IColumn & column, SerializeBinaryBulkSettings & settings, SerializeBinaryBulkStatePtr & state) const override; diff --git a/src/DataTypes/Serializations/tests/gtest_object_serialization.cpp b/src/DataTypes/Serializations/tests/gtest_object_serialization.cpp index f1fbbe115e2..fc7432d5bf6 100644 --- a/src/DataTypes/Serializations/tests/gtest_object_serialization.cpp +++ b/src/DataTypes/Serializations/tests/gtest_object_serialization.cpp @@ -31,7 +31,7 @@ TEST(SerializationObject, FromString) settings.getter = [&out](const auto &) { return &out; }; writeIntBinary(static_cast(1), out); - serialization->serializeBinaryBulkStatePrefix(settings, state); + serialization->serializeBinaryBulkStatePrefix(*column_string, settings, state); serialization->serializeBinaryBulkWithMultipleStreams(*column_string, 0, column_string->size(), settings, state); serialization->serializeBinaryBulkStateSuffix(settings, state); } diff --git a/src/Formats/NativeWriter.cpp b/src/Formats/NativeWriter.cpp index 9d4cfb68d56..499f3215d63 100644 --- a/src/Formats/NativeWriter.cpp +++ b/src/Formats/NativeWriter.cpp @@ -59,7 +59,7 @@ static void writeData(const ISerialization & serialization, const ColumnPtr & co settings.low_cardinality_max_dictionary_size = 0; //-V1048 ISerialization::SerializeBinaryBulkStatePtr state; - serialization.serializeBinaryBulkStatePrefix(settings, state); + serialization.serializeBinaryBulkStatePrefix(*full_column, settings, state); serialization.serializeBinaryBulkWithMultipleStreams(*full_column, offset, limit, settings, state); serialization.serializeBinaryBulkStateSuffix(settings, state); } diff --git a/src/Functions/blockSerializedSize.cpp b/src/Functions/blockSerializedSize.cpp index d406984c51c..35be65f3fed 100644 --- a/src/Functions/blockSerializedSize.cpp +++ b/src/Functions/blockSerializedSize.cpp @@ -54,7 +54,7 @@ public: auto serialization = elem.type->getDefaultSerialization(); - serialization->serializeBinaryBulkStatePrefix(settings, state); + serialization->serializeBinaryBulkStatePrefix(*full_column, settings, state); serialization->serializeBinaryBulkWithMultipleStreams(*full_column, 0 /** offset */, 0 /** limit */, settings, state); diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp index 771248b99c6..d0311f0250c 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp @@ -121,7 +121,7 @@ void writeColumnSingleGranule( serialize_settings.position_independent_encoding = true; //-V1048 serialize_settings.low_cardinality_max_dictionary_size = 0; //-V1048 - serialization->serializeBinaryBulkStatePrefix(serialize_settings, state); + serialization->serializeBinaryBulkStatePrefix(*column.column, serialize_settings, state); serialization->serializeBinaryBulkWithMultipleStreams(*column.column, from_row, number_of_rows, serialize_settings, state); serialization->serializeBinaryBulkStateSuffix(serialize_settings, state); } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp index 3d4aa0a7707..2f2a4fef9cb 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp @@ -349,7 +349,7 @@ void MergeTreeDataPartWriterWide::writeColumn( { ISerialization::SerializeBinaryBulkSettings serialize_settings; serialize_settings.getter = createStreamGetter(name_and_type, offset_columns); - serialization->serializeBinaryBulkStatePrefix(serialize_settings, it->second); + serialization->serializeBinaryBulkStatePrefix(column, serialize_settings, it->second); } const auto & global_settings = storage.getContext()->getSettingsRef(); diff --git a/src/Storages/MergeTree/MergeTreeIndexSet.cpp b/src/Storages/MergeTree/MergeTreeIndexSet.cpp index 3c31deda823..0e15f2c4cb6 100644 --- a/src/Storages/MergeTree/MergeTreeIndexSet.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexSet.cpp @@ -74,8 +74,9 @@ void MergeTreeIndexGranuleSet::serializeBinary(WriteBuffer & ostr) const auto serialization = type->getDefaultSerialization(); ISerialization::SerializeBinaryBulkStatePtr state; - serialization->serializeBinaryBulkStatePrefix(settings, state); - serialization->serializeBinaryBulkWithMultipleStreams(*block.getByPosition(i).column, 0, size(), settings, state); + const auto & column = *block.getByPosition(i).column; + serialization->serializeBinaryBulkStatePrefix(column, settings, state); + serialization->serializeBinaryBulkWithMultipleStreams(column, 0, size(), settings, state); serialization->serializeBinaryBulkStateSuffix(settings, state); } } diff --git a/src/Storages/StorageLog.cpp b/src/Storages/StorageLog.cpp index c6bc55fd620..1f623503f97 100644 --- a/src/Storages/StorageLog.cpp +++ b/src/Storages/StorageLog.cpp @@ -462,7 +462,7 @@ void LogSink::writeData(const NameAndTypePair & name_and_type, const IColumn & c settings.getter = createStreamGetter(name_and_type); if (!serialize_states.contains(name)) - serialization->serializeBinaryBulkStatePrefix(settings, serialize_states[name]); + serialization->serializeBinaryBulkStatePrefix(column, settings, serialize_states[name]); if (storage.use_marks_file) { From bbb7974dfbf1123dc648946c36749b72c8e3cebd Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Thu, 8 Sep 2022 15:47:43 +0000 Subject: [PATCH 009/239] fix style check --- src/DataTypes/ObjectUtils.cpp | 1 + src/DataTypes/ObjectUtils.h | 2 +- src/DataTypes/Serializations/SerializationObject.cpp | 1 + tests/queries/0_stateless/01825_type_json_in_array.reference | 4 ++-- 4 files changed, 5 insertions(+), 3 deletions(-) diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index 580583df5a4..b7f83ac6f63 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -939,6 +939,7 @@ size_t FieldVisitorToNumberOfDimensions::operator()(const Array & x) { const size_t size = x.size(); size_t dimensions = 0; + for (size_t i = 0; i < size; ++i) { size_t element_dimensions = applyVisitor(*this, x[i]); diff --git a/src/DataTypes/ObjectUtils.h b/src/DataTypes/ObjectUtils.h index 3c2812b6297..4160064c09b 100644 --- a/src/DataTypes/ObjectUtils.h +++ b/src/DataTypes/ObjectUtils.h @@ -39,7 +39,7 @@ Array createEmptyArrayField(size_t num_dimensions); DataTypePtr getDataTypeByColumn(const IColumn & column); /// Converts Object types and columns to Tuples in @columns_list and @block -/// and checks that types are consistent with types in @extended_storage_columns. +/// and checks that types are consistent with types in @storage_snapshot. void convertDynamicColumnsToTuples(Block & block, const StorageSnapshotPtr & storage_snapshot); /// Checks that each path is not the prefix of any other path. diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index 5577cab5982..25c458aad64 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -29,6 +29,7 @@ namespace ErrorCodes extern const int INCORRECT_DATA; extern const int CANNOT_READ_ALL_DATA; extern const int ARGUMENT_OUT_OF_BOUND; + extern const int LOGICAL_ERROR; } template diff --git a/tests/queries/0_stateless/01825_type_json_in_array.reference b/tests/queries/0_stateless/01825_type_json_in_array.reference index 3f6dc6ccd88..d06320ca291 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.reference +++ b/tests/queries/0_stateless/01825_type_json_in_array.reference @@ -8,8 +8,8 @@ Array(Tuple(k1 Int8, k2 Tuple(k3 Int8, k4 Int8, k5 String))) {"id":1,"arr":[{"k1":[{"k2":"aaa","k3":"bbb","k4":0},{"k2":"ccc","k3":"","k4":0}],"k5":{"k6":""}}]} {"id":2,"arr":[{"k1":[{"k2":"","k3":"ddd","k4":10},{"k2":"","k3":"","k4":20}],"k5":{"k6":"foo"}}]} -1 [['aaa','ccc']] [['bbb','']] [[]] [''] -2 [[]] [['ddd','']] [[10,20]] ['foo'] +1 [['aaa','ccc']] [['bbb','']] [[0,0]] [''] +2 [['','']] [['ddd','']] [[10,20]] ['foo'] {"k1":{"k2":"","k3":"","k4":20}} {"k1":{"k2":"","k3":"ddd","k4":10}} {"k1":{"k2":"aaa","k3":"bbb","k4":0}} From ad664cb6ddde7b7a7ebc8e2864df106b8cf37388 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 12 Sep 2022 18:14:02 +0000 Subject: [PATCH 010/239] fix serializePrefix for sparse columns --- src/DataTypes/Serializations/SerializationSparse.cpp | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/src/DataTypes/Serializations/SerializationSparse.cpp b/src/DataTypes/Serializations/SerializationSparse.cpp index 1580706b09f..cd09cd7be5a 100644 --- a/src/DataTypes/Serializations/SerializationSparse.cpp +++ b/src/DataTypes/Serializations/SerializationSparse.cpp @@ -183,8 +183,11 @@ void SerializationSparse::serializeBinaryBulkStatePrefix( SerializeBinaryBulkStatePtr & state) const { settings.path.push_back(Substream::SparseElements); - const auto & column_sparse = assert_cast(column); - nested->serializeBinaryBulkStatePrefix(column_sparse.getValuesColumn(), settings, state); + if (const auto * column_sparse = typeid_cast(&column)) + nested->serializeBinaryBulkStatePrefix(column_sparse->getValuesColumn(), settings, state); + else + nested->serializeBinaryBulkStatePrefix(column, settings, state); + settings.path.pop_back(); } From c74b5c8126a7b32e2df22f2c27a5627a448e31e1 Mon Sep 17 00:00:00 2001 From: avogar Date: Fri, 30 Sep 2022 16:59:17 +0000 Subject: [PATCH 011/239] Fix schema inference in s3Cluster and improve in hdfsCluster --- src/Interpreters/getTableExpressions.cpp | 17 ++++++ src/Interpreters/getTableExpressions.h | 3 ++ src/Storages/HDFS/StorageHDFSCluster.cpp | 37 ++++++++++++- src/Storages/HDFS/StorageHDFSCluster.h | 1 + src/Storages/StorageS3.cpp | 54 ++----------------- src/Storages/StorageS3.h | 16 ------ src/Storages/StorageS3Cluster.cpp | 47 ++++++++++++---- src/Storages/StorageS3Cluster.h | 1 + .../TableFunctionHDFSCluster.cpp | 2 +- src/TableFunctions/TableFunctionS3.cpp | 4 +- src/TableFunctions/TableFunctionS3Cluster.cpp | 1 + 11 files changed, 104 insertions(+), 79 deletions(-) diff --git a/src/Interpreters/getTableExpressions.cpp b/src/Interpreters/getTableExpressions.cpp index 830f0ea4411..43c9438ecec 100644 --- a/src/Interpreters/getTableExpressions.cpp +++ b/src/Interpreters/getTableExpressions.cpp @@ -3,6 +3,7 @@ #include #include #include +#include #include namespace DB @@ -155,4 +156,20 @@ TablesWithColumns getDatabaseAndTablesWithColumns( return tables_with_columns; } +ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & query) +{ + auto * select_query = query->as(); + if (!select_query || !select_query->tables()) + return nullptr; + + auto * tables = select_query->tables()->as(); + auto * table_expression = tables->children[0]->as()->table_expression->as(); + if (!table_expression->table_function) + return nullptr; + + auto * table_function = table_expression->table_function->as(); + return table_function->arguments->as(); +} + + } diff --git a/src/Interpreters/getTableExpressions.h b/src/Interpreters/getTableExpressions.h index c4ca01ee3c3..4497b376cc7 100644 --- a/src/Interpreters/getTableExpressions.h +++ b/src/Interpreters/getTableExpressions.h @@ -3,6 +3,7 @@ #include #include #include +#include namespace DB { @@ -23,4 +24,6 @@ ASTPtr extractTableExpression(const ASTSelectQuery & select, size_t table_number TablesWithColumns getDatabaseAndTablesWithColumns( const ASTTableExprConstPtrs & table_expressions, ContextPtr context, bool include_alias_cols, bool include_materialized_cols); +ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & query); + } diff --git a/src/Storages/HDFS/StorageHDFSCluster.cpp b/src/Storages/HDFS/StorageHDFSCluster.cpp index 47a6fbf5eaa..7851532a6a4 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.cpp +++ b/src/Storages/HDFS/StorageHDFSCluster.cpp @@ -25,6 +25,7 @@ #include #include #include +#include #include @@ -32,6 +33,34 @@ namespace DB { +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; +} + +static ASTPtr addColumnsStructureToQuery(const ASTPtr & query, const String & structure) +{ + /// Add argument with table structure to hdfsCluster table function in select query. + auto result_query = query->clone(); + ASTExpressionList * expression_list = extractTableFunctionArgumentsFromSelectQuery(result_query); + if (!expression_list) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected SELECT query from table function hdfsCluster, got '{}'", queryToString(query)); + + auto structure_literal = std::make_shared(structure); + + if (expression_list->children.size() != 2 && expression_list->children.size() != 3) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected 2 or 3 arguments in hdfsCluster table functions, got {}", expression_list->children.size()); + + if (expression_list->children.size() == 2) + { + auto format_literal = std::make_shared("auto"); + expression_list->children.push_back(format_literal); + } + + expression_list->children.push_back(structure_literal); + return result_query; +} + StorageHDFSCluster::StorageHDFSCluster( ContextPtr context_, String cluster_name_, @@ -56,6 +85,7 @@ StorageHDFSCluster::StorageHDFSCluster( { auto columns = StorageHDFS::getTableStructureFromData(format_name, uri_, compression_method, context_); storage_metadata.setColumns(columns); + need_to_add_structure_to_query = true; } else storage_metadata.setColumns(columns_); @@ -92,6 +122,11 @@ Pipe StorageHDFSCluster::read( const bool add_agg_info = processed_stage == QueryProcessingStage::WithMergeableState; + auto query_to_send = query_info.original_query; + if (need_to_add_structure_to_query) + query_to_send = addColumnsStructureToQuery( + query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll())); + for (const auto & replicas : cluster->getShardsAddresses()) { /// There will be only one replica, because we consider each replica as a shard @@ -110,7 +145,7 @@ Pipe StorageHDFSCluster::read( /// So, task_identifier is passed as constructor argument. It is more obvious. auto remote_query_executor = std::make_shared( connection, - queryToString(query_info.original_query), + queryToString(query_to_send), header, context, /*throttler=*/nullptr, diff --git a/src/Storages/HDFS/StorageHDFSCluster.h b/src/Storages/HDFS/StorageHDFSCluster.h index 21ae73c11ea..d5af257f803 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.h +++ b/src/Storages/HDFS/StorageHDFSCluster.h @@ -44,6 +44,7 @@ private: String uri; String format_name; String compression_method; + bool need_to_add_structure_to_query = false; }; diff --git a/src/Storages/StorageS3.cpp b/src/Storages/StorageS3.cpp index e15956f78be..8a6349007ec 100644 --- a/src/Storages/StorageS3.cpp +++ b/src/Storages/StorageS3.cpp @@ -361,39 +361,6 @@ String StorageS3Source::KeysIterator::next() return pimpl->next(); } -class StorageS3Source::ReadTasksIterator::Impl -{ -public: - explicit Impl(const std::vector & read_tasks_, const ReadTaskCallback & new_read_tasks_callback_) - : read_tasks(read_tasks_), new_read_tasks_callback(new_read_tasks_callback_) - { - } - - String next() - { - size_t current_index = index.fetch_add(1, std::memory_order_relaxed); - if (current_index >= read_tasks.size()) - return new_read_tasks_callback(); - return read_tasks[current_index]; - } - -private: - std::atomic_size_t index = 0; - std::vector read_tasks; - ReadTaskCallback new_read_tasks_callback; -}; - -StorageS3Source::ReadTasksIterator::ReadTasksIterator( - const std::vector & read_tasks_, const ReadTaskCallback & new_read_tasks_callback_) - : pimpl(std::make_shared(read_tasks_, new_read_tasks_callback_)) -{ -} - -String StorageS3Source::ReadTasksIterator::next() -{ - return pimpl->next(); -} - Block StorageS3Source::getHeader(Block sample_block, const std::vector & requested_virtual_columns) { for (const auto & virtual_column : requested_virtual_columns) @@ -802,8 +769,7 @@ StorageS3::StorageS3( distributed_processing_, is_key_with_globs, format_settings, - context_, - &read_tasks_used_in_schema_inference); + context_); storage_metadata.setColumns(columns); } else @@ -831,19 +797,14 @@ std::shared_ptr StorageS3::createFileIterator( ContextPtr local_context, ASTPtr query, const Block & virtual_block, - const std::vector & read_tasks, std::unordered_map * object_infos, Strings * read_keys) { if (distributed_processing) { return std::make_shared( - [read_tasks_iterator = std::make_shared(read_tasks, local_context->getReadTaskCallback()), read_keys]() -> String - { - auto key = read_tasks_iterator->next(); - if (read_keys) - read_keys->push_back(key); - return key; + [callback = local_context->getReadTaskCallback()]() -> String { + return callback(); }); } else if (is_key_with_globs) @@ -903,7 +864,6 @@ Pipe StorageS3::read( local_context, query_info.query, virtual_block, - read_tasks_used_in_schema_inference, &object_infos); ColumnsDescription columns_description; @@ -1201,7 +1161,7 @@ ColumnsDescription StorageS3::getTableStructureFromData( return getTableStructureFromDataImpl( configuration.format, s3_configuration, configuration.compression_method, distributed_processing, - s3_configuration.uri.key.find_first_of("*?{") != std::string::npos, format_settings, ctx, nullptr, object_infos); + s3_configuration.uri.key.find_first_of("*?{") != std::string::npos, format_settings, ctx, object_infos); } ColumnsDescription StorageS3::getTableStructureFromDataImpl( @@ -1212,13 +1172,12 @@ ColumnsDescription StorageS3::getTableStructureFromDataImpl( bool is_key_with_globs, const std::optional & format_settings, ContextPtr ctx, - std::vector * read_keys_in_distributed_processing, std::unordered_map * object_infos) { std::vector read_keys; auto file_iterator - = createFileIterator(s3_configuration, {s3_configuration.uri.key}, is_key_with_globs, distributed_processing, ctx, nullptr, {}, {}, object_infos, &read_keys); + = createFileIterator(s3_configuration, {s3_configuration.uri.key}, is_key_with_globs, distributed_processing, ctx, nullptr, {}, object_infos, &read_keys); std::optional columns_from_cache; size_t prev_read_keys_size = read_keys.size(); @@ -1271,9 +1230,6 @@ ColumnsDescription StorageS3::getTableStructureFromDataImpl( if (ctx->getSettingsRef().schema_inference_use_cache_for_s3) addColumnsToCache(read_keys, s3_configuration, columns, format, format_settings, ctx); - if (distributed_processing && read_keys_in_distributed_processing) - *read_keys_in_distributed_processing = std::move(read_keys); - return columns; } diff --git a/src/Storages/StorageS3.h b/src/Storages/StorageS3.h index c63508c8e6a..7c6970bab0e 100644 --- a/src/Storages/StorageS3.h +++ b/src/Storages/StorageS3.h @@ -66,18 +66,6 @@ public: std::shared_ptr pimpl; }; - class ReadTasksIterator - { - public: - ReadTasksIterator(const std::vector & read_tasks_, const ReadTaskCallback & new_read_tasks_callback_); - String next(); - - private: - class Impl; - /// shared_ptr to have copy constructor - std::shared_ptr pimpl; - }; - using IteratorWrapper = std::function; static Block getHeader(Block sample_block, const std::vector & requested_virtual_columns); @@ -238,8 +226,6 @@ private: ASTPtr partition_by; bool is_key_with_globs = false; - std::vector read_tasks_used_in_schema_inference; - std::unordered_map object_infos; static void updateS3Configuration(ContextPtr, S3Configuration &); @@ -252,7 +238,6 @@ private: ContextPtr local_context, ASTPtr query, const Block & virtual_block, - const std::vector & read_tasks = {}, std::unordered_map * object_infos = nullptr, Strings * read_keys = nullptr); @@ -264,7 +249,6 @@ private: bool is_key_with_globs, const std::optional & format_settings, ContextPtr ctx, - std::vector * read_keys_in_distributed_processing = nullptr, std::unordered_map * object_infos = nullptr); bool supportsSubsetOfColumns() const override; diff --git a/src/Storages/StorageS3Cluster.cpp b/src/Storages/StorageS3Cluster.cpp index 800bce0afde..245888f5ce3 100644 --- a/src/Storages/StorageS3Cluster.cpp +++ b/src/Storages/StorageS3Cluster.cpp @@ -5,14 +5,9 @@ #if USE_AWS_S3 #include "Common/Exception.h" -#include #include "Client/Connection.h" #include "Core/QueryProcessingStage.h" -#include -#include -#include #include -#include #include #include #include @@ -23,28 +18,54 @@ #include #include #include -#include "Processors/ISource.h" #include #include #include #include +#include #include #include #include +#include #include #include #include #include -#include #include #include -#include -#include namespace DB { + +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; +} + +static ASTPtr addColumnsStructureToQuery(const ASTPtr & query, const String & structure) +{ + /// Add argument with table structure to s3Cluster table function in select query. + auto result_query = query->clone(); + ASTExpressionList * expression_list = extractTableFunctionArgumentsFromSelectQuery(result_query); + if (!expression_list) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected SELECT query from table function from s3Cluster, got '{}'", queryToString(query)); + auto structure_literal = std::make_shared(structure); + + if (expression_list->children.size() < 2 || expression_list->children.size() > 5) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected 2 to 5 arguments in s3Cluster table functions, got {}", expression_list->children.size()); + + if (expression_list->children.size() == 2 || expression_list->children.size() == 4) + { + auto format_literal = std::make_shared("auto"); + expression_list->children.push_back(format_literal); + } + + expression_list->children.push_back(structure_literal); + return result_query; +} + StorageS3Cluster::StorageS3Cluster( const StorageS3ClusterConfiguration & configuration_, const StorageID & table_id_, @@ -72,6 +93,7 @@ StorageS3Cluster::StorageS3Cluster( auto columns = StorageS3::getTableStructureFromDataImpl(format_name, s3_configuration, compression_method, /*distributed_processing_*/false, is_key_with_globs, /*format_settings=*/std::nullopt, context_); storage_metadata.setColumns(columns); + need_to_add_structure_to_query = true; } else storage_metadata.setColumns(columns_); @@ -117,6 +139,11 @@ Pipe StorageS3Cluster::read( const bool add_agg_info = processed_stage == QueryProcessingStage::WithMergeableState; + ASTPtr query_to_send = query_info.original_query; + if (need_to_add_structure_to_query) + query_to_send = addColumnsStructureToQuery( + query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll())); + for (const auto & replicas : cluster->getShardsAddresses()) { /// There will be only one replica, because we consider each replica as a shard @@ -135,7 +162,7 @@ Pipe StorageS3Cluster::read( /// So, task_identifier is passed as constructor argument. It is more obvious. auto remote_query_executor = std::make_shared( connection, - queryToString(query_info.original_query), + queryToString(query_to_send), header, context, /*throttler=*/nullptr, diff --git a/src/Storages/StorageS3Cluster.h b/src/Storages/StorageS3Cluster.h index e5ca3b58123..7e4d4801090 100644 --- a/src/Storages/StorageS3Cluster.h +++ b/src/Storages/StorageS3Cluster.h @@ -46,6 +46,7 @@ private: String compression_method; NamesAndTypesList virtual_columns; Block virtual_block; + bool need_to_add_structure_to_query = false; }; diff --git a/src/TableFunctions/TableFunctionHDFSCluster.cpp b/src/TableFunctions/TableFunctionHDFSCluster.cpp index 385d280a100..fd2ef7bb052 100644 --- a/src/TableFunctions/TableFunctionHDFSCluster.cpp +++ b/src/TableFunctions/TableFunctionHDFSCluster.cpp @@ -48,7 +48,7 @@ void TableFunctionHDFSCluster::parseArguments(const ASTPtr & ast_function, Conte const auto message = fmt::format( "The signature of table function {} shall be the following:\n" \ " - cluster, uri\n",\ - " - cluster, format\n",\ + " - cluster, uri, format\n",\ " - cluster, uri, format, structure\n",\ " - cluster, uri, format, structure, compression_method", getName()); diff --git a/src/TableFunctions/TableFunctionS3.cpp b/src/TableFunctions/TableFunctionS3.cpp index 0bf33007760..0ceb7cd5e5b 100644 --- a/src/TableFunctions/TableFunctionS3.cpp +++ b/src/TableFunctions/TableFunctionS3.cpp @@ -64,7 +64,7 @@ void TableFunctionS3::parseArgumentsImpl(const String & error_message, ASTs & ar if (args.size() == 4) { auto second_arg = checkAndGetLiteralArgument(args[1], "format/access_key_id"); - if (FormatFactory::instance().getAllFormats().contains(second_arg)) + if (second_arg == "auto" || FormatFactory::instance().getAllFormats().contains(second_arg)) args_to_idx = {{"format", 1}, {"structure", 2}, {"compression_method", 3}}; else @@ -77,7 +77,7 @@ void TableFunctionS3::parseArgumentsImpl(const String & error_message, ASTs & ar { auto second_arg = checkAndGetLiteralArgument(args[1], "format/access_key_id"); - if (FormatFactory::instance().getAllFormats().contains(second_arg)) + if (second_arg == "auto" || FormatFactory::instance().getAllFormats().contains(second_arg)) args_to_idx = {{"format", 1}, {"structure", 2}}; else args_to_idx = {{"access_key_id", 1}, {"secret_access_key", 2}}; diff --git a/src/TableFunctions/TableFunctionS3Cluster.cpp b/src/TableFunctions/TableFunctionS3Cluster.cpp index 99c3ff85009..668cd67dd08 100644 --- a/src/TableFunctions/TableFunctionS3Cluster.cpp +++ b/src/TableFunctions/TableFunctionS3Cluster.cpp @@ -77,6 +77,7 @@ void TableFunctionS3Cluster::parseArguments(const ASTPtr & ast_function, Context /// StorageS3ClusterConfiguration inherints from StorageS3Configuration, so it is safe to upcast it. TableFunctionS3::parseArgumentsImpl(message, clipped_args, context, static_cast(configuration)); + LOG_DEBUG(&Poco::Logger::get("TableFunctionS3Cluster"), "Structure: {}", configuration.structure); } From addd36f4369ac0e45b08f048450eab73172f1301 Mon Sep 17 00:00:00 2001 From: avogar Date: Fri, 30 Sep 2022 17:00:25 +0000 Subject: [PATCH 012/239] Remove logging --- src/TableFunctions/TableFunctionS3Cluster.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/src/TableFunctions/TableFunctionS3Cluster.cpp b/src/TableFunctions/TableFunctionS3Cluster.cpp index 668cd67dd08..99c3ff85009 100644 --- a/src/TableFunctions/TableFunctionS3Cluster.cpp +++ b/src/TableFunctions/TableFunctionS3Cluster.cpp @@ -77,7 +77,6 @@ void TableFunctionS3Cluster::parseArguments(const ASTPtr & ast_function, Context /// StorageS3ClusterConfiguration inherints from StorageS3Configuration, so it is safe to upcast it. TableFunctionS3::parseArgumentsImpl(message, clipped_args, context, static_cast(configuration)); - LOG_DEBUG(&Poco::Logger::get("TableFunctionS3Cluster"), "Structure: {}", configuration.structure); } From e005b83d790e604bf5140819fc360963a5b76ca6 Mon Sep 17 00:00:00 2001 From: avogar Date: Fri, 30 Sep 2022 17:56:43 +0000 Subject: [PATCH 013/239] Add tests --- ...2457_s3_cluster_schema_inference.reference | 44 +++++++++++++++++++ .../02457_s3_cluster_schema_inference.sql | 13 ++++++ ...58_hdfs_cluster_schema_inference.reference | 10 +++++ .../02458_hdfs_cluster_schema_inference.sql | 12 +++++ 4 files changed, 79 insertions(+) create mode 100644 tests/queries/0_stateless/02457_s3_cluster_schema_inference.reference create mode 100644 tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql create mode 100644 tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.reference create mode 100644 tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.sql diff --git a/tests/queries/0_stateless/02457_s3_cluster_schema_inference.reference b/tests/queries/0_stateless/02457_s3_cluster_schema_inference.reference new file mode 100644 index 00000000000..b918bf2b155 --- /dev/null +++ b/tests/queries/0_stateless/02457_s3_cluster_schema_inference.reference @@ -0,0 +1,44 @@ +c1 Nullable(Int64) +c2 Nullable(Int64) +c3 Nullable(Int64) +c1 Nullable(Int64) +c2 Nullable(Int64) +c3 Nullable(Int64) +c1 Nullable(Int64) +c2 Nullable(Int64) +c3 Nullable(Int64) +c1 Nullable(Int64) +c2 Nullable(Int64) +c3 Nullable(Int64) +0 0 0 +0 0 0 +1 2 3 +4 5 6 +7 8 9 +10 11 12 +13 14 15 +16 17 18 +0 0 0 +0 0 0 +1 2 3 +4 5 6 +7 8 9 +10 11 12 +13 14 15 +16 17 18 +0 0 0 +0 0 0 +1 2 3 +4 5 6 +7 8 9 +10 11 12 +13 14 15 +16 17 18 +0 0 0 +0 0 0 +1 2 3 +4 5 6 +7 8 9 +10 11 12 +13 14 15 +16 17 18 diff --git a/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql b/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql new file mode 100644 index 00000000000..039c7a9fd7f --- /dev/null +++ b/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql @@ -0,0 +1,13 @@ +-- Tags: no-fasttest +-- Tag no-fasttest: Depends on AWS + +desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv'); +desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'TSV'); +desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest'); +desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest', 'TSV'); + +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv') order by c1, c2, c2; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'TSV') order by c1, c2, c2; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest') order by c1, c2, c2; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest', 'TSV') order by c1, c2, c2; + diff --git a/tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.reference b/tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.reference new file mode 100644 index 00000000000..a812e64a642 --- /dev/null +++ b/tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.reference @@ -0,0 +1,10 @@ +c1 Nullable(Int64) +c2 Nullable(Int64) +c3 Nullable(Int64) +c1 Nullable(Int64) +c2 Nullable(Int64) +c3 Nullable(Int64) +1 2 3 +4 5 6 +1 2 3 +4 5 6 diff --git a/tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.sql b/tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.sql new file mode 100644 index 00000000000..42e88fc44b2 --- /dev/null +++ b/tests/queries/0_stateless/02458_hdfs_cluster_schema_inference.sql @@ -0,0 +1,12 @@ +-- Tags: no-fasttest, no-parallel, no-cpu-aarch64 +-- Tag no-fasttest: Depends on Java + +insert into table function hdfs('hdfs://localhost:12222/test_02458_1.tsv', 'TSV', 'column1 UInt32, column2 UInt32, column3 UInt32') select 1, 2, 3 settings hdfs_truncate_on_insert=1; +insert into table function hdfs('hdfs://localhost:12222/test_02458_2.tsv', 'TSV', 'column1 UInt32, column2 UInt32, column3 UInt32') select 4, 5, 6 settings hdfs_truncate_on_insert=1; + +desc hdfsCluster('test_cluster_one_shard_three_replicas_localhost', 'hdfs://localhost:12222/test_02458_{1,2}.tsv'); +desc hdfsCluster('test_cluster_one_shard_three_replicas_localhost', 'hdfs://localhost:12222/test_02458_{1,2}.tsv', 'TSV'); + +select * from hdfsCluster('test_cluster_one_shard_three_replicas_localhost', 'hdfs://localhost:12222/test_02458_{1,2}.tsv') order by c1, c2, c3; +select * from hdfsCluster('test_cluster_one_shard_three_replicas_localhost', 'hdfs://localhost:12222/test_02458_{1,2}.tsv', 'TSV') order by c1, c2, c3; + From 8db899e550608a1c5b1854e88520d318ff5bb331 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 3 Oct 2022 14:44:33 +0000 Subject: [PATCH 014/239] fix assertion in fuzzer --- .../Serializations/SerializationObject.cpp | 13 ++++++++++--- .../0_stateless/01825_type_json_in_array.reference | 1 + .../0_stateless/01825_type_json_in_array.sql | 1 + 3 files changed, 12 insertions(+), 3 deletions(-) diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp index 25c458aad64..98a94886f67 100644 --- a/src/DataTypes/Serializations/SerializationObject.cpp +++ b/src/DataTypes/Serializations/SerializationObject.cpp @@ -165,16 +165,23 @@ void SerializationObject::serializeBinaryBulkStatePrefix( throw Exception(ErrorCodes::NOT_IMPLEMENTED, "DataTypeObject doesn't support serialization with non-trivial state"); + const auto & column_object = assert_cast(column); + if (!column_object.isFinalized()) + { + auto finalized = column_object.cloneFinalized(); + serializeBinaryBulkStatePrefix(*finalized, settings, state); + return; + } + settings.path.push_back(Substream::ObjectStructure); auto * stream = settings.getter(settings.path); if (!stream) throw Exception(ErrorCodes::LOGICAL_ERROR, "Missing stream for kind of binary serialization"); - writeIntBinary(static_cast(BinarySerializationKind::TUPLE), *stream); - - const auto & column_object = assert_cast(column); auto [tuple_column, tuple_type] = unflattenObjectToTuple(column_object); + + writeIntBinary(static_cast(BinarySerializationKind::TUPLE), *stream); writeStringBinary(tuple_type->getName(), *stream); auto state_object = std::make_shared(); diff --git a/tests/queries/0_stateless/01825_type_json_in_array.reference b/tests/queries/0_stateless/01825_type_json_in_array.reference index d06320ca291..c36a22e6951 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.reference +++ b/tests/queries/0_stateless/01825_type_json_in_array.reference @@ -20,3 +20,4 @@ Tuple(k2 String, k3 String, k4 Int8) {"arr":[1,{"y":1}]} {"arr":[2,{"y":2}]} {"arr":[{"x":"aaa","y":[1,2,3]}]} +{"arr":[{"x":1}]} diff --git a/tests/queries/0_stateless/01825_type_json_in_array.sql b/tests/queries/0_stateless/01825_type_json_in_array.sql index 935969272fa..e5c20d7ba6b 100644 --- a/tests/queries/0_stateless/01825_type_json_in_array.sql +++ b/tests/queries/0_stateless/01825_type_json_in_array.sql @@ -32,3 +32,4 @@ SELECT * FROM values('arr Array(JSON)', '[\'{"x" : 1}\']') FORMAT JSONEachRow; SELECT * FROM values('arr Map(String, JSON)', '{\'x\' : \'{"y" : 1}\', \'t\' : \'{"y" : 2}\'}') FORMAT JSONEachRow; SELECT * FROM values('arr Tuple(Int32, JSON)', '(1, \'{"y" : 1}\')', '(2, \'{"y" : 2}\')') FORMAT JSONEachRow; SELECT * FROM format(JSONEachRow, '{"arr" : [{"x" : "aaa", "y" : [1,2,3]}]}') FORMAT JSONEachRow; +SELECT * FROM values('arr Array(JSON)', '[\'{"x" : 1}\']') FORMAT JSONEachRow; From 5cb63575493092a26bec969f6415ef63efd76f3d Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 3 Oct 2022 19:37:27 +0000 Subject: [PATCH 015/239] fix style check --- src/DataTypes/ObjectUtils.cpp | 1 - src/Storages/MergeTree/MergeTreeSink.cpp | 1 - src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp | 1 - 3 files changed, 3 deletions(-) diff --git a/src/DataTypes/ObjectUtils.cpp b/src/DataTypes/ObjectUtils.cpp index 34ece825a16..e711b34ffa9 100644 --- a/src/DataTypes/ObjectUtils.cpp +++ b/src/DataTypes/ObjectUtils.cpp @@ -1,4 +1,3 @@ -#include #include #include #include diff --git a/src/Storages/MergeTree/MergeTreeSink.cpp b/src/Storages/MergeTree/MergeTreeSink.cpp index dcbd721267b..a3b0e4a0a78 100644 --- a/src/Storages/MergeTree/MergeTreeSink.cpp +++ b/src/Storages/MergeTree/MergeTreeSink.cpp @@ -1,7 +1,6 @@ #include #include #include -#include #include #include diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp index 2577144b3fb..8069e9b3f6a 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp +++ b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp @@ -1,7 +1,6 @@ #include #include #include -#include #include #include #include From 0a26afd1c8513074c9bef3c84d5e61eb7f41f3d6 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Mon, 10 Oct 2022 18:54:35 -0300 Subject: [PATCH 016/239] Add mutex around ares fd processing --- src/Common/CaresPTRResolver.cpp | 6 ++++++ src/Common/CaresPTRResolver.h | 3 ++- 2 files changed, 8 insertions(+), 1 deletion(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index a02909309b6..32d9fd65e1c 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -2,6 +2,7 @@ #include #include #include +#include #include "ares.h" #include "netdb.h" @@ -40,6 +41,8 @@ namespace DB } } + std::mutex CaresPTRResolver::mutex; + CaresPTRResolver::CaresPTRResolver(CaresPTRResolver::provider_token) : channel(nullptr) { /* @@ -119,6 +122,9 @@ namespace DB { FD_ZERO(&read_fds); FD_ZERO(&write_fds); + + std::lock_guard guard {mutex}; + nfds = ares_fds(channel, &read_fds,&write_fds); if (nfds == 0) { diff --git a/src/Common/CaresPTRResolver.h b/src/Common/CaresPTRResolver.h index e5182d34682..47c3aaf7bec 100644 --- a/src/Common/CaresPTRResolver.h +++ b/src/Common/CaresPTRResolver.h @@ -20,7 +20,6 @@ namespace DB * Allow only DNSPTRProvider to instantiate this class * */ struct provider_token {}; - public: explicit CaresPTRResolver(provider_token); ~CaresPTRResolver() override; @@ -37,6 +36,8 @@ namespace DB void resolve_v6(const std::string & ip, std::unordered_set & response); ares_channel channel; + + static std::mutex mutex; }; } From 6dae76ff8809e11a4e11350df173a735ab000bed Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Tue, 11 Oct 2022 12:47:27 -0300 Subject: [PATCH 017/239] add mutex include for release builds --- src/Common/CaresPTRResolver.h | 1 + 1 file changed, 1 insertion(+) diff --git a/src/Common/CaresPTRResolver.h b/src/Common/CaresPTRResolver.h index 47c3aaf7bec..a012e7bed35 100644 --- a/src/Common/CaresPTRResolver.h +++ b/src/Common/CaresPTRResolver.h @@ -1,5 +1,6 @@ #pragma once +#include #include "DNSPTRResolver.h" using ares_channel = struct ares_channeldata *; From 9abf13bf78cd3fe32c50c1e8b9096f1d98fa7c45 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Wed, 12 Oct 2022 13:56:00 -0300 Subject: [PATCH 018/239] Use poll instead of select for c-ares --- src/Common/CaresPTRResolver.cpp | 83 +++++++++++++++++++++++++++------ src/Common/CaresPTRResolver.h | 12 ++++- 2 files changed, 78 insertions(+), 17 deletions(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index 32d9fd65e1c..79b78a734aa 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -41,8 +41,6 @@ namespace DB } } - std::mutex CaresPTRResolver::mutex; - CaresPTRResolver::CaresPTRResolver(CaresPTRResolver::provider_token) : channel(nullptr) { /* @@ -113,26 +111,81 @@ namespace DB void CaresPTRResolver::wait() { - timeval * tvp, tv; - fd_set read_fds; - fd_set write_fds; - int nfds; + int sockets[ARES_GETSOCK_MAXNUM]; + pollfd pollfd[ARES_GETSOCK_MAXNUM]; - for (;;) + while(true) { - FD_ZERO(&read_fds); - FD_ZERO(&write_fds); + auto readable_sockets = get_readable_sockets(sockets, pollfd); + auto timeout = calculate_timeout(); - std::lock_guard guard {mutex}; + int number_of_fds_ready = 0; + if(!readable_sockets.empty()) + number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), timeout); - nfds = ares_fds(channel, &read_fds,&write_fds); - if (nfds == 0) + if(!number_of_fds_ready) + { + process_possible_timeout(); + break; + } + else + { + process_readable_sockets(readable_sockets); + } + } + } + + std::span CaresPTRResolver::get_readable_sockets(int * sockets, pollfd * pollfd) + { + int sockets_bitmask = ares_getsock(channel, sockets, ARES_GETSOCK_MAXNUM); + + int number_of_sockets_to_poll = 0; + + for(int i = 0; i < ARES_GETSOCK_MAXNUM; i++, number_of_sockets_to_poll++) + { + pollfd[i].events = 0; + pollfd[i].revents = 0; + + if(ARES_GETSOCK_READABLE(sockets_bitmask, i)) + { + pollfd[i].fd = sockets[i]; + pollfd[i].events = POLLIN; + } + else { break; } - tvp = ares_timeout(channel, nullptr, &tv); - select(nfds, &read_fds, &write_fds, nullptr, tvp); - ares_process(channel, &read_fds, &write_fds); + } + + return std::span(pollfd, number_of_sockets_to_poll); + } + + int64_t CaresPTRResolver::calculate_timeout() + { + timeval tv; + if (auto * tvp = ares_timeout(channel, nullptr, &tv)) + { + auto timeout = tvp->tv_sec * 1000 + tvp->tv_usec / 1000; + + return timeout; + } + + return 0; + } + + void CaresPTRResolver::process_possible_timeout() + { + /* Call ares_process() unconditonally here, even if we simply timed out + above, as otherwise the ares name resolve won't timeout! */ + ares_process_fd(channel, ARES_SOCKET_BAD, ARES_SOCKET_BAD); + } + + void CaresPTRResolver::process_readable_sockets(std::span readable_sockets) + { + for (auto readable_socket : readable_sockets) + { + auto fd = readable_socket.revents & POLLIN ? readable_socket.fd : ARES_SOCKET_BAD; + ares_process_fd(channel, fd, ARES_SOCKET_BAD); } } } diff --git a/src/Common/CaresPTRResolver.h b/src/Common/CaresPTRResolver.h index a012e7bed35..654f2af9008 100644 --- a/src/Common/CaresPTRResolver.h +++ b/src/Common/CaresPTRResolver.h @@ -1,6 +1,7 @@ #pragma once -#include +#include +#include #include "DNSPTRResolver.h" using ares_channel = struct ares_channeldata *; @@ -36,8 +37,15 @@ namespace DB void resolve_v6(const std::string & ip, std::unordered_set & response); - ares_channel channel; + std::span get_readable_sockets(int * sockets, pollfd * pollfd); + int64_t calculate_timeout(); + + void process_possible_timeout(); + + void process_readable_sockets(std::span readable_sockets); + + ares_channel channel; static std::mutex mutex; }; } From d98eac11eb5684d91fd94df737bd0e61f9c6fd0d Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Wed, 12 Oct 2022 17:47:56 -0300 Subject: [PATCH 019/239] remove mutex declaration --- src/Common/CaresPTRResolver.h | 1 - 1 file changed, 1 deletion(-) diff --git a/src/Common/CaresPTRResolver.h b/src/Common/CaresPTRResolver.h index 654f2af9008..38344d75ade 100644 --- a/src/Common/CaresPTRResolver.h +++ b/src/Common/CaresPTRResolver.h @@ -46,7 +46,6 @@ namespace DB void process_readable_sockets(std::span readable_sockets); ares_channel channel; - static std::mutex mutex; }; } From 72ba210b24e35adcd5a6ebd52f2c60e6ef0d7ed3 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Wed, 12 Oct 2022 20:23:29 -0300 Subject: [PATCH 020/239] working on style --- src/Common/CaresPTRResolver.cpp | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index 79b78a734aa..cf0e1661322 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -121,16 +121,18 @@ namespace DB int number_of_fds_ready = 0; if(!readable_sockets.empty()) - number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), timeout); - - if(!number_of_fds_ready) { - process_possible_timeout(); - break; + number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), timeout); + } + + if(number_of_fds_ready > 0) + { + process_readable_sockets(readable_sockets); } else { - process_readable_sockets(readable_sockets); + process_possible_timeout(); + break; } } } From f588830e245a0b1a564a01ad9965c4a5780706d0 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Wed, 12 Oct 2022 20:35:17 -0300 Subject: [PATCH 021/239] working on style --- src/Common/CaresPTRResolver.cpp | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index cf0e1661322..a3eeb3d24ad 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -114,18 +114,18 @@ namespace DB int sockets[ARES_GETSOCK_MAXNUM]; pollfd pollfd[ARES_GETSOCK_MAXNUM]; - while(true) + while (true) { auto readable_sockets = get_readable_sockets(sockets, pollfd); auto timeout = calculate_timeout(); int number_of_fds_ready = 0; - if(!readable_sockets.empty()) + if (!readable_sockets.empty()) { number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), timeout); } - if(number_of_fds_ready > 0) + if (number_of_fds_ready > 0) { process_readable_sockets(readable_sockets); } @@ -143,7 +143,7 @@ namespace DB int number_of_sockets_to_poll = 0; - for(int i = 0; i < ARES_GETSOCK_MAXNUM; i++, number_of_sockets_to_poll++) + for (int i = 0; i < ARES_GETSOCK_MAXNUM; i++, number_of_sockets_to_poll++) { pollfd[i].events = 0; pollfd[i].revents = 0; From 91560dd60708af4fded8bf55c00fdafc0057fbe8 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Wed, 12 Oct 2022 21:28:32 -0300 Subject: [PATCH 022/239] working on style --- src/Common/CaresPTRResolver.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index a3eeb3d24ad..f04c6fb6de7 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -148,7 +148,7 @@ namespace DB pollfd[i].events = 0; pollfd[i].revents = 0; - if(ARES_GETSOCK_READABLE(sockets_bitmask, i)) + if (ARES_GETSOCK_READABLE(sockets_bitmask, i)) { pollfd[i].fd = sockets[i]; pollfd[i].events = POLLIN; From 6d22bb78e0db5deebe88444cee5475516c7d322d Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Fri, 14 Oct 2022 12:39:08 -0300 Subject: [PATCH 023/239] custom py script in integ tests --- docker/test/integration/base/Dockerfile | 5 ++ .../__init__.py | 0 .../configs/host_regexp.xml | 11 +++ .../configs/listen_host.xml | 5 ++ .../coredns_config/Corefile | 8 ++ .../coredns_config/example.com | 1 + .../scripts/stress_test.py | 56 +++++++++++++ .../test.py | 83 +++++++++++++++++++ 8 files changed, 169 insertions(+) create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/__init__.py create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/host_regexp.xml create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/listen_host.xml create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/Corefile create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/example.com create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py diff --git a/docker/test/integration/base/Dockerfile b/docker/test/integration/base/Dockerfile index 9b6318a5426..a2d86187a23 100644 --- a/docker/test/integration/base/Dockerfile +++ b/docker/test/integration/base/Dockerfile @@ -27,9 +27,14 @@ RUN apt-get update \ tar \ tzdata \ unixodbc \ + python3-pip \ + libcurl4-openssl-dev \ + libssl-dev \ && apt-get clean \ && rm -rf /var/lib/apt/lists/* /var/cache/debconf /tmp/* +RUN pip3 install pycurl + # Architecture of the image when BuildKit/buildx is used ARG TARGETARCH diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/__init__.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/__init__.py new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/host_regexp.xml b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/host_regexp.xml new file mode 100644 index 00000000000..7a2141e6c7e --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/host_regexp.xml @@ -0,0 +1,11 @@ + + + + + + test1\.example\.com$ + + default + + + \ No newline at end of file diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/listen_host.xml b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/listen_host.xml new file mode 100644 index 00000000000..58ef55cd3f3 --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/listen_host.xml @@ -0,0 +1,5 @@ + + :: + 0.0.0.0 + 1 + diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/Corefile b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/Corefile new file mode 100644 index 00000000000..0dd198441dc --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/Corefile @@ -0,0 +1,8 @@ +. { + hosts /example.com { + reload "200ms" + fallthrough + } + forward . 127.0.0.11 + log +} diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/example.com b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/example.com new file mode 100644 index 00000000000..9beb415c290 --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/coredns_config/example.com @@ -0,0 +1 @@ +filled in runtime, but needs to exist in order to be volume mapped in docker \ No newline at end of file diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py new file mode 100644 index 00000000000..1a840246ccf --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py @@ -0,0 +1,56 @@ +import pycurl +import threading +from io import BytesIO +import sys + +server_ip = sys.argv[1] + +mutex = threading.Lock() +success_counter = 0 +number_of_threads = 1 +number_of_iterations = 400 + +def perform_request(): + + buffer = BytesIO() + crl = pycurl.Curl() + crl.setopt(pycurl.INTERFACE, '192.168.0.157') + crl.setopt(crl.WRITEDATA, buffer) + crl.setopt(crl.URL, f'http://{server_ip}:8123/?query=select+1&user=test_dns') + + crl.perform() + + # End curl session + crl.close() + + str_response = buffer.getvalue().decode('iso-8859-1') + expected_response = "1\n" + + mutex.acquire() + + global success_counter + + if (str_response == expected_response): + success_counter += 1 + + mutex.release() + + # print(buffer.getvalue().decode('iso-8859-1')) + +def perform_multiple_requests(n): + for i in range(n): + perform_request() + +threads = [] + + +for i in range(number_of_threads): + thread = threading.Thread(target=perform_multiple_requests, args=(number_of_iterations,)) + thread.start() + threads.append(thread) + + +for thread in threads: + thread.join() + +exit(success_counter == number_of_threads * number_of_iterations) diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py new file mode 100644 index 00000000000..566eb5c0b2b --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py @@ -0,0 +1,83 @@ +import pytest +from helpers.cluster import ClickHouseCluster, get_docker_compose_path, run_and_check +from time import sleep +import os + +DOCKER_COMPOSE_PATH = get_docker_compose_path() +SCRIPT_DIR = os.path.dirname(os.path.realpath(__file__)) + +cluster = ClickHouseCluster(__file__) + +ch_server = cluster.add_instance( + "clickhouse-server", + with_coredns=True, + main_configs=["configs/listen_host.xml"], + user_configs=["configs/host_regexp.xml"], + ipv6_address="2001:3984:3989::1:1111", +) + +client = cluster.add_instance( + "clickhouse-client", + ipv6_address="2001:3984:3989::1:1112", +) + + +@pytest.fixture(scope="module") +def started_cluster(): + global cluster + try: + cluster.start() + yield cluster + + finally: + cluster.shutdown() + + +def setup_dns_server(ip): + domains_string = "test3.example.com test2.example.com test1.example.com" + example_file_path = f'{ch_server.env_variables["COREDNS_CONFIG_DIR"]}/example.com' + run_and_check(f"echo '{ip} {domains_string}' > {example_file_path}", shell=True) + + +def setup_ch_server(dns_server_ip): + ch_server.exec_in_container( + (["bash", "-c", f"echo 'nameserver {dns_server_ip}' > /etc/resolv.conf"]) + ) + ch_server.exec_in_container( + (["bash", "-c", "echo 'options ndots:0' >> /etc/resolv.conf"]) + ) + ch_server.query("SYSTEM DROP DNS CACHE") + + +def build_endpoint_v4(ip): + return f"'http://{ip}:8123/?query=SELECT+1&user=test_dns'" + + +def build_endpoint_v6(ip): + return build_endpoint_v4(f"[{ip}]") + + +def test_host_regexp_multiple_ptr_v4(started_cluster): + server_ip = cluster.get_instance_ip("clickhouse-server") + client_ip = cluster.get_instance_ip("clickhouse-client") + dns_server_ip = cluster.get_instance_ip(cluster.coredns_host) + + setup_dns_server(client_ip) + setup_ch_server(dns_server_ip) + + current_dir = os.path.dirname(__file__) + client.copy_file_to_container(os.path.join(current_dir, "scripts", "stress_test.py"), "stress_test.py") + + assert "1\n" == client.exec_in_container(["python3", f"stress_test.py", server_ip]) + + # benchmark_command = f"echo 'select 1' | clickhouse benchmark -h {server_ip} --user test_dns -c 10 --reconnect" + + # assert "1\n" == client.exec_in_container((["bash", "-c", benchmark_command])) +# container_id = cluster.get_container_id("resolver") +# current_dir = os.path.dirname(__file__) +# cluster.copy_file_to_container( +# container_id, +# os.path.join(current_dir, "s3_endpoint", "endpoint.py"), +# "endpoint.py", +# ) +# cluster.exec_in_container(container_id, ["python", "endpoint.py"], detach=True) From 299d8484737f436166bc05ca3b864368b90cde90 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Fri, 14 Oct 2022 15:23:10 -0300 Subject: [PATCH 024/239] style fix? --- .../scripts/stress_test.py | 18 ++++++++++-------- .../test.py | 15 +++------------ 2 files changed, 13 insertions(+), 20 deletions(-) diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py index 1a840246ccf..2d8a70edc57 100644 --- a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py @@ -10,45 +10,47 @@ success_counter = 0 number_of_threads = 1 number_of_iterations = 400 + def perform_request(): buffer = BytesIO() crl = pycurl.Curl() - crl.setopt(pycurl.INTERFACE, '192.168.0.157') + crl.setopt(pycurl.INTERFACE, "192.168.0.157") crl.setopt(crl.WRITEDATA, buffer) - crl.setopt(crl.URL, f'http://{server_ip}:8123/?query=select+1&user=test_dns') + crl.setopt(crl.URL, f"http://{server_ip}:8123/?query=select+1&user=test_dns") crl.perform() # End curl session crl.close() - str_response = buffer.getvalue().decode('iso-8859-1') + str_response = buffer.getvalue().decode("iso-8859-1") expected_response = "1\n" mutex.acquire() global success_counter - if (str_response == expected_response): + if str_response == expected_response: success_counter += 1 mutex.release() - - # print(buffer.getvalue().decode('iso-8859-1')) + def perform_multiple_requests(n): for i in range(n): perform_request() + threads = [] for i in range(number_of_threads): - thread = threading.Thread(target=perform_multiple_requests, args=(number_of_iterations,)) + thread = threading.Thread( + target=perform_multiple_requests, args=(number_of_iterations,) + ) thread.start() threads.append(thread) - for thread in threads: thread.join() diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py index 566eb5c0b2b..ef1c615023d 100644 --- a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py @@ -66,18 +66,9 @@ def test_host_regexp_multiple_ptr_v4(started_cluster): setup_ch_server(dns_server_ip) current_dir = os.path.dirname(__file__) - client.copy_file_to_container(os.path.join(current_dir, "scripts", "stress_test.py"), "stress_test.py") + client.copy_file_to_container( + os.path.join(current_dir, "scripts", "stress_test.py"), "stress_test.py" + ) assert "1\n" == client.exec_in_container(["python3", f"stress_test.py", server_ip]) - # benchmark_command = f"echo 'select 1' | clickhouse benchmark -h {server_ip} --user test_dns -c 10 --reconnect" - - # assert "1\n" == client.exec_in_container((["bash", "-c", benchmark_command])) -# container_id = cluster.get_container_id("resolver") -# current_dir = os.path.dirname(__file__) -# cluster.copy_file_to_container( -# container_id, -# os.path.join(current_dir, "s3_endpoint", "endpoint.py"), -# "endpoint.py", -# ) -# cluster.exec_in_container(container_id, ["python", "endpoint.py"], detach=True) From f414edebae69c5d247f2657c5a312c2e06b13e98 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Fri, 14 Oct 2022 15:54:48 -0300 Subject: [PATCH 025/239] fix black style --- .../test_host_regexp_multiple_ptr_records_concurrent/test.py | 1 - 1 file changed, 1 deletion(-) diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py index ef1c615023d..7ead30275a5 100644 --- a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py @@ -71,4 +71,3 @@ def test_host_regexp_multiple_ptr_v4(started_cluster): ) assert "1\n" == client.exec_in_container(["python3", f"stress_test.py", server_ip]) - From 91433e5b9cf3d82dbb766043819c84aa7155cb41 Mon Sep 17 00:00:00 2001 From: Smita Kulkarni Date: Mon, 17 Oct 2022 09:04:25 +0200 Subject: [PATCH 026/239] =?UTF-8?q?Added=20**=20glob=20support=20for=20rec?= =?UTF-8?q?ursive=20directory=20traversal=20to=20filesystem=20and=20=20S3.?= =?UTF-8?q?=20Implementation:=20*=20Updated=20parseGlob=20to=20not=20add?= =?UTF-8?q?=20=E2=80=98/=E2=80=98=20restriction=20when=20**=20is=20used.?= =?UTF-8?q?=20*=20Updated=20S3=20&=20filesystem=20to=20fetch=20files=20and?= =?UTF-8?q?=20not=20use=20regex=20match=20if=20glob=20is=20**.=20Testing:?= =?UTF-8?q?=20*=20Added=20a=20test=20for=20filesystem=20tests/queries/0=5F?= =?UTF-8?q?stateless/02459=5Fglob=5Ffor=5Frecursive=5Fdirectory=5Ftraversa?= =?UTF-8?q?l.sh?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- src/Common/parseGlobs.cpp | 10 ++++- src/Storages/StorageFile.cpp | 24 ++++++++--- src/Storages/StorageS3.cpp | 7 ++- ...or_recursive_directory_traversal.reference | 14 ++++++ ..._glob_for_recursive_directory_traversal.sh | 43 +++++++++++++++++++ 5 files changed, 89 insertions(+), 9 deletions(-) create mode 100644 tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.reference create mode 100755 tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.sh diff --git a/src/Common/parseGlobs.cpp b/src/Common/parseGlobs.cpp index 8e9195f9842..b7c9dbe3a70 100644 --- a/src/Common/parseGlobs.cpp +++ b/src/Common/parseGlobs.cpp @@ -90,17 +90,23 @@ std::string makeRegexpPatternFromGlobs(const std::string & initial_str_with_glob oss_for_replacing << escaped_with_globs.substr(current_index); std::string almost_res = oss_for_replacing.str(); WriteBufferFromOwnString buf_final_processing; + char previous = ' '; for (const auto & letter : almost_res) { - if ((letter == '?') || (letter == '*')) + if (previous == '*' && letter == '*') + { + buf_final_processing << "[^{}]"; + } + else if ((letter == '?') || (letter == '*')) { buf_final_processing << "[^/]"; /// '?' is any symbol except '/' if (letter == '?') continue; } - if ((letter == '.') || (letter == '{') || (letter == '}')) + else if ((letter == '.') || (letter == '{') || (letter == '}')) buf_final_processing << '\\'; buf_final_processing << letter; + previous = letter; } return buf_final_processing.str(); } diff --git a/src/Storages/StorageFile.cpp b/src/Storages/StorageFile.cpp index 29f2d0667d9..1f0704b37fa 100644 --- a/src/Storages/StorageFile.cpp +++ b/src/Storages/StorageFile.cpp @@ -81,7 +81,8 @@ void listFilesWithRegexpMatchingImpl( const std::string & path_for_ls, const std::string & for_match, size_t & total_bytes_to_read, - std::vector & result) + std::vector & result, + bool recursive = false) { const size_t first_glob = for_match.find_first_of("*?{"); @@ -89,10 +90,17 @@ void listFilesWithRegexpMatchingImpl( const std::string suffix_with_globs = for_match.substr(end_of_path_without_globs); /// begin with '/' const size_t next_slash = suffix_with_globs.find('/', 1); - auto regexp = makeRegexpPatternFromGlobs(suffix_with_globs.substr(0, next_slash)); + const std::string current_glob = suffix_with_globs.substr(0, next_slash); + auto regexp = makeRegexpPatternFromGlobs(current_glob); + re2::RE2 matcher(regexp); + bool skip_regex = current_glob == "/*" ? true : false; + if (!recursive) + recursive = current_glob == "/**" ; + const std::string prefix_without_globs = path_for_ls + for_match.substr(1, end_of_path_without_globs); + if (!fs::exists(prefix_without_globs)) return; @@ -107,15 +115,21 @@ void listFilesWithRegexpMatchingImpl( /// Condition is_directory means what kind of path is it in current iteration of ls if (!it->is_directory() && !looking_for_directory) { - if (re2::RE2::FullMatch(file_name, matcher)) + if (skip_regex || re2::RE2::FullMatch(file_name, matcher)) { total_bytes_to_read += it->file_size(); result.push_back(it->path().string()); } } - else if (it->is_directory() && looking_for_directory) + else if (it->is_directory()) { - if (re2::RE2::FullMatch(file_name, matcher)) + if (recursive) + { + listFilesWithRegexpMatchingImpl(fs::path(full_path).append(it->path().string()) / "" , + looking_for_directory ? suffix_with_globs.substr(next_slash) : current_glob , + total_bytes_to_read, result, recursive); + } + else if (looking_for_directory && re2::RE2::FullMatch(file_name, matcher)) { /// Recursion depth is limited by pattern. '*' works only for depth = 1, for depth = 2 pattern path is '*/*'. So we do not need additional check. listFilesWithRegexpMatchingImpl(fs::path(full_path) / "", suffix_with_globs.substr(next_slash), total_bytes_to_read, result); diff --git a/src/Storages/StorageS3.cpp b/src/Storages/StorageS3.cpp index 675dd548088..a7e0bfb98ad 100644 --- a/src/Storages/StorageS3.cpp +++ b/src/Storages/StorageS3.cpp @@ -139,7 +139,9 @@ public: request.SetBucket(globbed_uri.bucket); request.SetPrefix(key_prefix); + matcher = std::make_unique(makeRegexpPatternFromGlobs(globbed_uri.key)); + recursive = globbed_uri.key == "/**" ? true : false; fillInternalBufferAssumeLocked(); } @@ -197,7 +199,7 @@ private: for (const auto & row : result_batch) { const String & key = row.GetKey(); - if (re2::RE2::FullMatch(key, *matcher)) + if (recursive || re2::RE2::FullMatch(key, *matcher)) { String path = fs::path(globbed_uri.bucket) / key; if (object_infos) @@ -224,7 +226,7 @@ private: for (const auto & row : result_batch) { String key = row.GetKey(); - if (re2::RE2::FullMatch(key, *matcher)) + if (recursive || re2::RE2::FullMatch(key, *matcher)) buffer.emplace_back(std::move(key)); } } @@ -252,6 +254,7 @@ private: Aws::S3::Model::ListObjectsV2Request request; Aws::S3::Model::ListObjectsV2Outcome outcome; std::unique_ptr matcher; + bool recursive{false}; bool is_finished{false}; std::unordered_map * object_infos; Strings * read_keys; diff --git a/tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.reference b/tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.reference new file mode 100644 index 00000000000..8e146946955 --- /dev/null +++ b/tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.reference @@ -0,0 +1,14 @@ +1 1 +2 2 +1 1 +2 2 +3 3 +4 4 +5 5 +6 6 +3 3 +4 4 +3 3 +4 4 +5 5 +6 6 diff --git a/tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.sh b/tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.sh new file mode 100755 index 00000000000..2c3deda2328 --- /dev/null +++ b/tests/queries/0_stateless/02459_glob_for_recursive_directory_traversal.sh @@ -0,0 +1,43 @@ +#!/usr/bin/env bash +# Tags: no-fasttest, no-parallel + +CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd) +# shellcheck source=../shell_config.sh +. "$CUR_DIR"/../shell_config.sh + +user_files_path=$(clickhouse-client --query "select _path,_file from file('nonexist.txt', 'CSV', 'val1 char')" 2>&1 | grep Exception | awk '{gsub("/nonexist.txt","",$9); print $9}') + +mkdir $user_files_path/d1 +touch $user_files_path/d1/text1.txt + +for i in {1..2} +do + echo $i$'\t'$i >> $user_files_path/d1/text1.txt +done + +mkdir $user_files_path/d1/d2 +touch $user_files_path/d1/d2/text2.txt +for i in {3..4} +do + echo $i$'\t'$i >> $user_files_path/d1/d2/text2.txt +done + +mkdir $user_files_path/d1/d2/d3 +touch $user_files_path/d1/d2/d3/text3.txt +for i in {5..6} +do + echo $i$'\t'$i >> $user_files_path/d1/d2/d3/text3.txt +done + +${CLICKHOUSE_CLIENT} -q "SELECT * from file ('d1/*','TSV', 'Index UInt8, Number UInt8')" | sort --numeric-sort +${CLICKHOUSE_CLIENT} -q "SELECT * from file ('d1/**','TSV', 'Index UInt8, Number UInt8')" | sort --numeric-sort +${CLICKHOUSE_CLIENT} -q "SELECT * from file ('d1/*/tex*','TSV', 'Index UInt8, Number UInt8')" | sort --numeric-sort +${CLICKHOUSE_CLIENT} -q "SELECT * from file ('d1/**/tex*','TSV', 'Index UInt8, Number UInt8')" | sort --numeric-sort + + +rm $user_files_path/d1/d2/d3/text3.txt +rmdir $user_files_path/d1/d2/d3 +rm $user_files_path/d1/d2/text2.txt +rmdir $user_files_path/d1/d2 +rm $user_files_path/d1/text1.txt +rmdir $user_files_path/d1 \ No newline at end of file From 3d1ed2969ae2d25038e4f6b1e974225f90e94625 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Mon, 17 Oct 2022 18:21:03 +0000 Subject: [PATCH 027/239] Fix possible LOGICAL_ERROR in binary ariphmetics monotonicity. Fix invalid evaluation of binary monotonic function in KeyCondition. --- src/Functions/FunctionBinaryArithmetic.h | 18 ++++++++++-------- src/Storages/MergeTree/KeyCondition.cpp | 2 ++ ...2461_mullable_pk_monotonicity_bug.reference | 4 ++++ .../02461_mullable_pk_monotonicity_bug.sql | 6 ++++++ 4 files changed, 22 insertions(+), 8 deletions(-) create mode 100644 tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference create mode 100644 tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index baa3c65537d..cf752bbcbf8 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -1782,21 +1782,24 @@ public: { ColumnsWithTypeAndName columns_with_constant = {{left.column->cloneResized(1), left.type, left.name}, - {right.type->createColumnConst(1, point), right.type, right.name}}; + {removeNullable(right.type)->createColumnConst(1, point), removeNullable(right.type), right.name}}; + /// This is a bit dangerous to call Base::executeImpl cause it ignores `use Default Implementation For XXX` flags. + /// It was possible to check monotonicity for nullable right type which result to exception. + /// Adding removeNullable above fixes the issue, but some other inconsistency may left. auto col = Base::executeImpl(columns_with_constant, return_type, 1); Field point_transformed; col->get(0, point_transformed); return point_transformed; }; - transform(left_point); - transform(right_point); + + bool is_positive_monotonicity = applyVisitor(FieldVisitorAccurateLess(), left_point, right_point) + == applyVisitor(FieldVisitorAccurateLess(), transform(left_point), transform(right_point)); if (name_view == "plus") { // Check if there is an overflow - if (applyVisitor(FieldVisitorAccurateLess(), left_point, right_point) - == applyVisitor(FieldVisitorAccurateLess(), transform(left_point), transform(right_point))) + if (is_positive_monotonicity) return {true, true, false, true}; else return {false, true, false, false}; @@ -1804,8 +1807,7 @@ public: else { // Check if there is an overflow - if (applyVisitor(FieldVisitorAccurateLess(), left_point, right_point) - != applyVisitor(FieldVisitorAccurateLess(), transform(left_point), transform(right_point))) + if (!is_positive_monotonicity) return {true, false, false, true}; else return {false, false, false, false}; @@ -1817,7 +1819,7 @@ public: auto transform = [&](const Field & point) { ColumnsWithTypeAndName columns_with_constant - = {{left.type->createColumnConst(1, point), left.type, left.name}, + = {{removeNullable(left.type)->createColumnConst(1, point), removeNullable(left.type), left.name}, {right.column->cloneResized(1), right.type, right.name}}; auto col = Base::executeImpl(columns_with_constant, return_type, 1); diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp index d7c33c8663b..ec5454973c7 100644 --- a/src/Storages/MergeTree/KeyCondition.cpp +++ b/src/Storages/MergeTree/KeyCondition.cpp @@ -1407,6 +1407,7 @@ public: ColumnsWithTypeAndName new_arguments; new_arguments.reserve(arguments.size() + 1); new_arguments.push_back(const_arg); + new_arguments.front().column = new_arguments.front().column->cloneResized(input_rows_count); for (const auto & arg : arguments) new_arguments.push_back(arg); return func->prepare(new_arguments)->execute(new_arguments, result_type, input_rows_count, dry_run); @@ -1415,6 +1416,7 @@ public: { auto new_arguments = arguments; new_arguments.push_back(const_arg); + new_arguments.back().column = new_arguments.back().column->cloneResized(input_rows_count); return func->prepare(new_arguments)->execute(new_arguments, result_type, input_rows_count, dry_run); } else diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference new file mode 100644 index 00000000000..099b7d91c92 --- /dev/null +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference @@ -0,0 +1,4 @@ +1 +2 +1 +2 diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql new file mode 100644 index 00000000000..56ea054cc47 --- /dev/null +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql @@ -0,0 +1,6 @@ +create table tab (x Nullable(UInt8)) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; +insert into tab select number from numbers(4); + +set max_rows_to_read = 2; +SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; From c8642514d0a9c6a5f583902f3aacee87b4b50452 Mon Sep 17 00:00:00 2001 From: Suzy Wang Date: Mon, 17 Oct 2022 12:02:28 -0700 Subject: [PATCH 028/239] format ipv6 in s390x, endianess support --- src/Common/formatIPv6.h | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/src/Common/formatIPv6.h b/src/Common/formatIPv6.h index 83b9d6e9fb1..31d5e83760a 100644 --- a/src/Common/formatIPv6.h +++ b/src/Common/formatIPv6.h @@ -5,6 +5,7 @@ #include #include #include +#include #include #include @@ -55,8 +56,11 @@ inline bool parseIPv4(const char * src, unsigned char * dst) } if (*(src - 1) != '\0') return false; - +#if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ + reverseMemcpy(dst, &result, sizeof(result)); +#else memcpy(dst, &result, sizeof(result)); +#endif return true; } From 32fad205f0a43ed43c5d97499b92b751d3a1e19e Mon Sep 17 00:00:00 2001 From: Smita Kulkarni Date: Tue, 18 Oct 2022 14:08:26 +0200 Subject: [PATCH 029/239] Added documentation for file and s3 - 36316 Support ** glob for recursive directory traversal --- docs/en/sql-reference/table-functions/file.md | 17 +++++++++++++++++ docs/en/sql-reference/table-functions/s3.md | 12 ++++++++++++ 2 files changed, 29 insertions(+) diff --git a/docs/en/sql-reference/table-functions/file.md b/docs/en/sql-reference/table-functions/file.md index f40107aaaca..fc81e7cf649 100644 --- a/docs/en/sql-reference/table-functions/file.md +++ b/docs/en/sql-reference/table-functions/file.md @@ -81,6 +81,7 @@ Multiple path components can have globs. For being processed file must exist and - `?` — Substitutes any single character. - `{some_string,another_string,yet_another_one}` — Substitutes any of strings `'some_string', 'another_string', 'yet_another_one'`. - `{N..M}` — Substitutes any number in range from N to M including both borders. +- `**` - Fetches all files inside the folder recursively. Constructions with `{}` are similar to the [remote](remote.md) table function. @@ -119,6 +120,22 @@ Query the data from files named `file000`, `file001`, … , `file999`: SELECT count(*) FROM file('big_dir/file{0..9}{0..9}{0..9}', 'CSV', 'name String, value UInt32'); ``` +**Example** + +Query the data from all files inside `big_dir` directory recursively: + +``` sql +SELECT count(*) FROM file('big_dir/**', 'CSV', 'name String, value UInt32'); +``` + +**Example** + +Query the data from all `file002` files from any folder inside `big_dir` directory recursively: + +``` sql +SELECT count(*) FROM file('big_dir/**/file002', 'CSV', 'name String, value UInt32'); +``` + ## Virtual Columns - `_path` — Path to the file. diff --git a/docs/en/sql-reference/table-functions/s3.md b/docs/en/sql-reference/table-functions/s3.md index 545037665bb..545a89223bf 100644 --- a/docs/en/sql-reference/table-functions/s3.md +++ b/docs/en/sql-reference/table-functions/s3.md @@ -127,6 +127,18 @@ INSERT INTO FUNCTION s3('https://clickhouse-public-datasets.s3.amazonaws.com/my- SELECT name, value FROM existing_table; ``` +Glob ** can be used for recursive directory traversal. Consider the below example, it will fetch all files from `my-test-bucket-768` directory recursively: + +``` sql +SELECT * FROM s3('https://clickhouse-public-datasets.s3.amazonaws.com/my-test-bucket-768/**', 'CSV', 'name String, value UInt32', 'gzip'); +``` + +The below get data from all `test-data.csv.gz` files from any folder inside `my-test-bucket` directory recursively: + +``` sql +SELECT * FROM s3('https://clickhouse-public-datasets.s3.amazonaws.com/my-test-bucket-768/**/test-data.csv.gz', 'CSV', 'name String, value UInt32', 'gzip'); +``` + ## Partitioned Write If you specify `PARTITION BY` expression when inserting data into `S3` table, a separate file is created for each partition value. Splitting the data into separate files helps to improve reading operations efficiency. From c457f7dff3d100c29c24e1592f665a9d03ecffed Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Tue, 18 Oct 2022 12:21:16 +0000 Subject: [PATCH 030/239] Fix and more tests. --- src/Functions/FunctionBinaryArithmetic.h | 15 +++++++++---- ...461_mullable_pk_monotonicity_bug.reference | 20 ++++++++++++++++++ .../02461_mullable_pk_monotonicity_bug.sql | 21 +++++++++++++++++++ 3 files changed, 52 insertions(+), 4 deletions(-) diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index cf752bbcbf8..05f092b5060 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -39,6 +39,7 @@ #include #include #include +#include #include #if USE_EMBEDDED_COMPILER @@ -1778,11 +1779,14 @@ public: // const +|- variable if (left.column && isColumnConst(*left.column)) { + auto left_type = removeLowCardinality(removeNullable(left.type)); + auto right_type = removeLowCardinality(removeNullable(right.type)); + auto transform = [&](const Field & point) { ColumnsWithTypeAndName columns_with_constant - = {{left.column->cloneResized(1), left.type, left.name}, - {removeNullable(right.type)->createColumnConst(1, point), removeNullable(right.type), right.name}}; + = {{left_type->createColumnConst(1, (*left.column)[0]), left_type, left.name}, + {right_type->createColumnConst(1, point), right_type, right.name}}; /// This is a bit dangerous to call Base::executeImpl cause it ignores `use Default Implementation For XXX` flags. /// It was possible to check monotonicity for nullable right type which result to exception. @@ -1816,11 +1820,14 @@ public: // variable +|- constant else if (right.column && isColumnConst(*right.column)) { + auto left_type = removeLowCardinality(removeNullable(left.type)); + auto right_type = removeLowCardinality(removeNullable(right.type)); + auto transform = [&](const Field & point) { ColumnsWithTypeAndName columns_with_constant - = {{removeNullable(left.type)->createColumnConst(1, point), removeNullable(left.type), left.name}, - {right.column->cloneResized(1), right.type, right.name}}; + = {{left_type->createColumnConst(1, point), left_type, left.name}, + {right_type->createColumnConst(1, (*right.column)[0]), right_type, right.name}}; auto col = Base::executeImpl(columns_with_constant, return_type, 1); Field point_transformed; diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference index 099b7d91c92..c2983d46447 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference @@ -2,3 +2,23 @@ 2 1 2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql index 56ea054cc47..798868fe566 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql @@ -1,6 +1,27 @@ create table tab (x Nullable(UInt8)) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; insert into tab select number from numbers(4); +set allow_suspicious_low_cardinality_types=1; +set max_rows_to_read = 2; + +SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; + +drop table tab; + +set max_rows_to_read = 100; + +create table tab (x LowCardinality(UInt8)) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; +insert into tab select number from numbers(4); + set max_rows_to_read = 2; SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; From 2c2f977096570207529166bbec2cf942bb0100c7 Mon Sep 17 00:00:00 2001 From: avogar Date: Tue, 18 Oct 2022 13:03:55 +0000 Subject: [PATCH 031/239] Make better --- src/Interpreters/getTableExpressions.cpp | 16 ------- src/Interpreters/getTableExpressions.h | 2 - src/Storages/HDFS/StorageHDFSCluster.cpp | 35 +++----------- src/Storages/HDFS/StorageHDFSCluster.h | 2 +- src/Storages/StorageS3Cluster.cpp | 30 ++---------- ...lumnsStructureToQueryWithClusterEngine.cpp | 47 +++++++++++++++++++ ...ColumnsStructureToQueryWithClusterEngine.h | 11 +++++ 7 files changed, 69 insertions(+), 74 deletions(-) create mode 100644 src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp create mode 100644 src/Storages/addColumnsStructureToQueryWithClusterEngine.h diff --git a/src/Interpreters/getTableExpressions.cpp b/src/Interpreters/getTableExpressions.cpp index 43c9438ecec..26fdab5b836 100644 --- a/src/Interpreters/getTableExpressions.cpp +++ b/src/Interpreters/getTableExpressions.cpp @@ -156,20 +156,4 @@ TablesWithColumns getDatabaseAndTablesWithColumns( return tables_with_columns; } -ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & query) -{ - auto * select_query = query->as(); - if (!select_query || !select_query->tables()) - return nullptr; - - auto * tables = select_query->tables()->as(); - auto * table_expression = tables->children[0]->as()->table_expression->as(); - if (!table_expression->table_function) - return nullptr; - - auto * table_function = table_expression->table_function->as(); - return table_function->arguments->as(); -} - - } diff --git a/src/Interpreters/getTableExpressions.h b/src/Interpreters/getTableExpressions.h index 4497b376cc7..22fb30fbad3 100644 --- a/src/Interpreters/getTableExpressions.h +++ b/src/Interpreters/getTableExpressions.h @@ -24,6 +24,4 @@ ASTPtr extractTableExpression(const ASTSelectQuery & select, size_t table_number TablesWithColumns getDatabaseAndTablesWithColumns( const ASTTableExprConstPtrs & table_expressions, ContextPtr context, bool include_alias_cols, bool include_materialized_cols); -ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & query); - } diff --git a/src/Storages/HDFS/StorageHDFSCluster.cpp b/src/Storages/HDFS/StorageHDFSCluster.cpp index 53170ee2439..7b110bb09fc 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.cpp +++ b/src/Storages/HDFS/StorageHDFSCluster.cpp @@ -11,7 +11,6 @@ #include #include #include -#include #include #include #include @@ -26,6 +25,7 @@ #include #include #include +#include #include @@ -38,29 +38,6 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } -static ASTPtr addColumnsStructureToQuery(const ASTPtr & query, const String & structure) -{ - /// Add argument with table structure to hdfsCluster table function in select query. - auto result_query = query->clone(); - ASTExpressionList * expression_list = extractTableFunctionArgumentsFromSelectQuery(result_query); - if (!expression_list) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected SELECT query from table function hdfsCluster, got '{}'", queryToString(query)); - - auto structure_literal = std::make_shared(structure); - - if (expression_list->children.size() != 2 && expression_list->children.size() != 3) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected 2 or 3 arguments in hdfsCluster table functions, got {}", expression_list->children.size()); - - if (expression_list->children.size() == 2) - { - auto format_literal = std::make_shared("auto"); - expression_list->children.push_back(format_literal); - } - - expression_list->children.push_back(structure_literal); - return result_query; -} - StorageHDFSCluster::StorageHDFSCluster( ContextPtr context_, String cluster_name_, @@ -85,7 +62,7 @@ StorageHDFSCluster::StorageHDFSCluster( { auto columns = StorageHDFS::getTableStructureFromData(format_name, uri_, compression_method, context_); storage_metadata.setColumns(columns); - need_to_add_structure_to_query = true; + add_columns_structure_to_query = true; } else storage_metadata.setColumns(columns_); @@ -122,10 +99,10 @@ Pipe StorageHDFSCluster::read( const bool add_agg_info = processed_stage == QueryProcessingStage::WithMergeableState; - auto query_to_send = query_info.original_query; - if (need_to_add_structure_to_query) - query_to_send = addColumnsStructureToQuery( - query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll())); + auto query_to_send = query_info.original_query->clone(); + if (add_columns_structure_to_query) + addColumnsStructureToQueryWithClusterEngine( + query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll()), 3, getName()); for (const auto & replicas : cluster->getShardsAddresses()) { diff --git a/src/Storages/HDFS/StorageHDFSCluster.h b/src/Storages/HDFS/StorageHDFSCluster.h index 88ad2c8871c..1eb823582d5 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.h +++ b/src/Storages/HDFS/StorageHDFSCluster.h @@ -44,7 +44,7 @@ private: String uri; String format_name; String compression_method; - bool need_to_add_structure_to_query = false; + bool add_columns_structure_to_query = false; }; diff --git a/src/Storages/StorageS3Cluster.cpp b/src/Storages/StorageS3Cluster.cpp index d4bd551aa56..eaa2c9a46d2 100644 --- a/src/Storages/StorageS3Cluster.cpp +++ b/src/Storages/StorageS3Cluster.cpp @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include @@ -27,6 +26,7 @@ #include #include #include +#include #include #include @@ -44,28 +44,6 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } -static ASTPtr addColumnsStructureToQuery(const ASTPtr & query, const String & structure) -{ - /// Add argument with table structure to s3Cluster table function in select query. - auto result_query = query->clone(); - ASTExpressionList * expression_list = extractTableFunctionArgumentsFromSelectQuery(result_query); - if (!expression_list) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected SELECT query from table function from s3Cluster, got '{}'", queryToString(query)); - auto structure_literal = std::make_shared(structure); - - if (expression_list->children.size() < 2 || expression_list->children.size() > 5) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected 2 to 5 arguments in s3Cluster table functions, got {}", expression_list->children.size()); - - if (expression_list->children.size() == 2 || expression_list->children.size() == 4) - { - auto format_literal = std::make_shared("auto"); - expression_list->children.push_back(format_literal); - } - - expression_list->children.push_back(structure_literal); - return result_query; -} - StorageS3Cluster::StorageS3Cluster( const StorageS3ClusterConfiguration & configuration_, const StorageID & table_id_, @@ -139,10 +117,10 @@ Pipe StorageS3Cluster::read( const bool add_agg_info = processed_stage == QueryProcessingStage::WithMergeableState; - ASTPtr query_to_send = query_info.original_query; + ASTPtr query_to_send = query_info.original_query->clone(); if (need_to_add_structure_to_query) - query_to_send = addColumnsStructureToQuery( - query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll())); + addColumnsStructureToQueryWithClusterEngine( + query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll()), 5, getName()); for (const auto & replicas : cluster->getShardsAddresses()) { diff --git a/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp b/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp new file mode 100644 index 00000000000..34b05445b0e --- /dev/null +++ b/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp @@ -0,0 +1,47 @@ +#include +#include +#include +#include +#include +#include +#include + +namespace DB +{ + +static ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & query) +{ + auto * select_query = query->as(); + if (!select_query || !select_query->tables()) + return nullptr; + + auto * tables = select_query->tables()->as(); + auto * table_expression = tables->children[0]->as()->table_expression->as(); + if (!table_expression->table_function) + return nullptr; + + auto * table_function = table_expression->table_function->as(); + return table_function->arguments->as(); +} + +void addColumnsStructureToQueryWithClusterEngine(ASTPtr & query, const String & structure, size_t max_arguments, const String & function_name) +{ + /// Add argument with table structure to s3Cluster table function in select query. + ASTExpressionList * expression_list = extractTableFunctionArgumentsFromSelectQuery(query); + if (!expression_list) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected SELECT query from table function {}, got '{}'", function_name, queryToString(query)); + auto structure_literal = std::make_shared(structure); + + if (expression_list->children.size() < 2 || expression_list->children.size() > max_arguments) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected 2 to {} arguments in {} table functions, got {}", function_name, max_arguments, expression_list->children.size()); + + if (expression_list->children.size() == 2 || expression_list->children.size() == max_arguments - 1) + { + auto format_literal = std::make_shared("auto"); + expression_list->children.push_back(format_literal); + } + + expression_list->children.push_back(structure_literal); +} + +} diff --git a/src/Storages/addColumnsStructureToQueryWithClusterEngine.h b/src/Storages/addColumnsStructureToQueryWithClusterEngine.h new file mode 100644 index 00000000000..f39f3a31630 --- /dev/null +++ b/src/Storages/addColumnsStructureToQueryWithClusterEngine.h @@ -0,0 +1,11 @@ +#pragma once + +#include + +namespace DB +{ + +/// Add structure argument for queries with s3Cluster/hdfsCluster table function. +void addColumnsStructureToQueryWithClusterEngine(ASTPtr & query, const String & structure, size_t max_arguments, const String & function_name); + +} From 9a7fbcecb00c7419f14f656ad2481e1717eac3ae Mon Sep 17 00:00:00 2001 From: avogar Date: Tue, 18 Oct 2022 13:04:42 +0000 Subject: [PATCH 032/239] Remove unneded include --- src/Interpreters/getTableExpressions.cpp | 1 - src/Interpreters/getTableExpressions.h | 1 - 2 files changed, 2 deletions(-) diff --git a/src/Interpreters/getTableExpressions.cpp b/src/Interpreters/getTableExpressions.cpp index 26fdab5b836..830f0ea4411 100644 --- a/src/Interpreters/getTableExpressions.cpp +++ b/src/Interpreters/getTableExpressions.cpp @@ -3,7 +3,6 @@ #include #include #include -#include #include namespace DB diff --git a/src/Interpreters/getTableExpressions.h b/src/Interpreters/getTableExpressions.h index 22fb30fbad3..c4ca01ee3c3 100644 --- a/src/Interpreters/getTableExpressions.h +++ b/src/Interpreters/getTableExpressions.h @@ -3,7 +3,6 @@ #include #include #include -#include namespace DB { From 3bb11618597facbfbdedd4522b6292c7829e4293 Mon Sep 17 00:00:00 2001 From: avogar Date: Tue, 18 Oct 2022 13:05:40 +0000 Subject: [PATCH 033/239] Better naming --- src/Storages/StorageS3Cluster.cpp | 4 ++-- src/Storages/StorageS3Cluster.h | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/src/Storages/StorageS3Cluster.cpp b/src/Storages/StorageS3Cluster.cpp index eaa2c9a46d2..6376fac4cfc 100644 --- a/src/Storages/StorageS3Cluster.cpp +++ b/src/Storages/StorageS3Cluster.cpp @@ -71,7 +71,7 @@ StorageS3Cluster::StorageS3Cluster( auto columns = StorageS3::getTableStructureFromDataImpl(format_name, s3_configuration, compression_method, /*distributed_processing_*/false, is_key_with_globs, /*format_settings=*/std::nullopt, context_); storage_metadata.setColumns(columns); - need_to_add_structure_to_query = true; + add_columns_structure_to_query = true; } else storage_metadata.setColumns(columns_); @@ -118,7 +118,7 @@ Pipe StorageS3Cluster::read( const bool add_agg_info = processed_stage == QueryProcessingStage::WithMergeableState; ASTPtr query_to_send = query_info.original_query->clone(); - if (need_to_add_structure_to_query) + if (add_columns_structure_to_query) addColumnsStructureToQueryWithClusterEngine( query_to_send, StorageDictionary::generateNamesAndTypesDescription(storage_snapshot->metadata->getColumns().getAll()), 5, getName()); diff --git a/src/Storages/StorageS3Cluster.h b/src/Storages/StorageS3Cluster.h index 3f77152833d..c3ff4bdb94b 100644 --- a/src/Storages/StorageS3Cluster.h +++ b/src/Storages/StorageS3Cluster.h @@ -46,7 +46,7 @@ private: String compression_method; NamesAndTypesList virtual_columns; Block virtual_block; - bool need_to_add_structure_to_query = false; + bool add_columns_structure_to_query = false; }; From be9c6b2c47e35f4bd382be652a9c79cef0ff22e7 Mon Sep 17 00:00:00 2001 From: avogar Date: Tue, 18 Oct 2022 13:08:45 +0000 Subject: [PATCH 034/239] Fix typo in test --- .../0_stateless/02457_s3_cluster_schema_inference.sql | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql b/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql index 039c7a9fd7f..03e8785b24b 100644 --- a/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql +++ b/tests/queries/0_stateless/02457_s3_cluster_schema_inference.sql @@ -6,8 +6,8 @@ desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localh desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest'); desc s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest', 'TSV'); -select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv') order by c1, c2, c2; -select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'TSV') order by c1, c2, c2; -select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest') order by c1, c2, c2; -select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest', 'TSV') order by c1, c2, c2; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv') order by c1, c2, c3; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'TSV') order by c1, c2, c3; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest') order by c1, c2, c3; +select * from s3Cluster('test_cluster_one_shard_three_replicas_localhost', 'http://localhost:11111/test/{a,b}.tsv', 'test', 'testtest', 'TSV') order by c1, c2, c3; From ba3e213a8392f6205374522270d83182dc87d864 Mon Sep 17 00:00:00 2001 From: Kruglov Pavel <48961922+Avogar@users.noreply.github.com> Date: Tue, 18 Oct 2022 17:03:04 +0200 Subject: [PATCH 035/239] Remove outdated comment --- src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp b/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp index 34b05445b0e..9f09c61ec2f 100644 --- a/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp +++ b/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp @@ -26,7 +26,6 @@ static ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & void addColumnsStructureToQueryWithClusterEngine(ASTPtr & query, const String & structure, size_t max_arguments, const String & function_name) { - /// Add argument with table structure to s3Cluster table function in select query. ASTExpressionList * expression_list = extractTableFunctionArgumentsFromSelectQuery(query); if (!expression_list) throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected SELECT query from table function {}, got '{}'", function_name, queryToString(query)); From 4645d465b12013d99d88956020034cbfa47d9455 Mon Sep 17 00:00:00 2001 From: Boris Kuschel Date: Tue, 18 Oct 2022 17:38:06 -0500 Subject: [PATCH 036/239] update zlib-ng to latest --- contrib/zlib-ng | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/zlib-ng b/contrib/zlib-ng index bffad6f6fe7..50f0eae1a41 160000 --- a/contrib/zlib-ng +++ b/contrib/zlib-ng @@ -1 +1 @@ -Subproject commit bffad6f6fe74d6a2f92e2668390664a926c68733 +Subproject commit 50f0eae1a411764cd6d1e85b3ce471438acd3c1c From 7f0a2d32c852874ebb521f7883542e6adbc63dad Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Tue, 18 Oct 2022 20:15:58 -0300 Subject: [PATCH 037/239] fix test errors --- .../configs/config.xml | 4 ++++ .../scripts/stress_test.py | 17 +++++++++++------ .../test.py | 6 ++---- 3 files changed, 17 insertions(+), 10 deletions(-) create mode 100644 tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/config.xml diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/config.xml b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/config.xml new file mode 100644 index 00000000000..42a1f962705 --- /dev/null +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/configs/config.xml @@ -0,0 +1,4 @@ + + 1 + 250 + diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py index 2d8a70edc57..81dfb4cc511 100644 --- a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py @@ -3,19 +3,20 @@ import threading from io import BytesIO import sys -server_ip = sys.argv[1] +client_ip = sys.argv[1] +server_ip = sys.argv[2] mutex = threading.Lock() success_counter = 0 -number_of_threads = 1 -number_of_iterations = 400 +number_of_threads = 200 +number_of_iterations = 200 def perform_request(): buffer = BytesIO() crl = pycurl.Curl() - crl.setopt(pycurl.INTERFACE, "192.168.0.157") + crl.setopt(pycurl.INTERFACE, client_ip) crl.setopt(crl.WRITEDATA, buffer) crl.setopt(crl.URL, f"http://{server_ip}:8123/?query=select+1&user=test_dns") @@ -38,7 +39,7 @@ def perform_request(): def perform_multiple_requests(n): - for i in range(n): + for request_number in range(n): perform_request() @@ -55,4 +56,8 @@ for i in range(number_of_threads): for thread in threads: thread.join() -exit(success_counter == number_of_threads * number_of_iterations) + +if success_counter == number_of_threads * number_of_iterations: + exit(0) + +exit(1) diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py index 7ead30275a5..62f47579612 100644 --- a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/test.py @@ -11,14 +11,12 @@ cluster = ClickHouseCluster(__file__) ch_server = cluster.add_instance( "clickhouse-server", with_coredns=True, - main_configs=["configs/listen_host.xml"], + main_configs=["configs/config.xml", "configs/listen_host.xml"], user_configs=["configs/host_regexp.xml"], - ipv6_address="2001:3984:3989::1:1111", ) client = cluster.add_instance( "clickhouse-client", - ipv6_address="2001:3984:3989::1:1112", ) @@ -70,4 +68,4 @@ def test_host_regexp_multiple_ptr_v4(started_cluster): os.path.join(current_dir, "scripts", "stress_test.py"), "stress_test.py" ) - assert "1\n" == client.exec_in_container(["python3", f"stress_test.py", server_ip]) + client.exec_in_container(["python3", f"stress_test.py", client_ip, server_ip]) From 549597fe804353800b6505c2b03c029a073f9782 Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Wed, 19 Oct 2022 12:15:27 +0800 Subject: [PATCH 038/239] Fix misbehavior of key analysis. When key types cannot be inside Nullable, it can break key analysis. See https://github.com/ClickHouse/ClickHouse/issues/42456. --- src/Interpreters/castColumn.cpp | 10 ++++++++++ src/Interpreters/castColumn.h | 1 + src/Storages/MergeTree/KeyCondition.cpp | 6 ++++-- ...dition_with_types_that_cannot_be_nullable.reference | 1 + ...ey_condition_with_types_that_cannot_be_nullable.sql | 9 +++++++++ 5 files changed, 25 insertions(+), 2 deletions(-) create mode 100644 tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.reference create mode 100644 tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.sql diff --git a/src/Interpreters/castColumn.cpp b/src/Interpreters/castColumn.cpp index dc9882b84b0..744cfff2527 100644 --- a/src/Interpreters/castColumn.cpp +++ b/src/Interpreters/castColumn.cpp @@ -51,4 +51,14 @@ ColumnPtr castColumnAccurateOrNull(const ColumnWithTypeAndName & arg, const Data return castColumn(arg, type); } +ColumnPtr tryCastColumnAccurate(const ColumnWithTypeAndName & arg, const DataTypePtr & type) +try +{ + return castColumn(arg, type); +} +catch (...) +{ + return nullptr; +} + } diff --git a/src/Interpreters/castColumn.h b/src/Interpreters/castColumn.h index fcbea0f4646..c0a2cfbefbc 100644 --- a/src/Interpreters/castColumn.h +++ b/src/Interpreters/castColumn.h @@ -8,5 +8,6 @@ namespace DB ColumnPtr castColumn(const ColumnWithTypeAndName & arg, const DataTypePtr & type); ColumnPtr castColumnAccurate(const ColumnWithTypeAndName & arg, const DataTypePtr & type); ColumnPtr castColumnAccurateOrNull(const ColumnWithTypeAndName & arg, const DataTypePtr & type); +ColumnPtr tryCastColumnAccurate(const ColumnWithTypeAndName & arg, const DataTypePtr & type); } diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp index 9f5f2873b98..3ecaff2c19a 100644 --- a/src/Storages/MergeTree/KeyCondition.cpp +++ b/src/Storages/MergeTree/KeyCondition.cpp @@ -1182,11 +1182,13 @@ bool KeyCondition::transformConstantWithValidFunctions( { auto const_type = cur_node->result_type; auto const_column = out_type->createColumnConst(1, out_value); - auto const_value = (*castColumnAccurateOrNull({const_column, out_type, ""}, const_type))[0]; + auto maybe_const_value_column = tryCastColumnAccurate({const_column, out_type, ""}, const_type); - if (const_value.isNull()) + if (maybe_const_value_column == nullptr) return false; + auto const_value = (*maybe_const_value_column)[0]; + while (!chain.empty()) { const auto * func = chain.top(); diff --git a/tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.reference b/tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.reference new file mode 100644 index 00000000000..13b65c29f05 --- /dev/null +++ b/tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.reference @@ -0,0 +1 @@ +printer1 diff --git a/tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.sql b/tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.sql new file mode 100644 index 00000000000..690ec6c70e0 --- /dev/null +++ b/tests/queries/0_stateless/02457_key_condition_with_types_that_cannot_be_nullable.sql @@ -0,0 +1,9 @@ +drop table if exists test; + +create table test (Printer LowCardinality(String), IntervalStart DateTime) engine MergeTree partition by (hiveHash(Printer), toYear(IntervalStart)) order by (Printer, IntervalStart); + +insert into test values ('printer1', '2006-02-07 06:28:15'); + +select Printer from test where Printer='printer1'; + +drop table test; From c8f9e7bd1f028be15d233571c5cff4ed2442ad64 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Wed, 19 Oct 2022 14:33:37 +0000 Subject: [PATCH 039/239] Update test. --- src/Functions/FunctionBinaryArithmetic.h | 8 ++++---- .../02461_mullable_pk_monotonicity_bug.reference | 8 ++++++++ .../0_stateless/02461_mullable_pk_monotonicity_bug.sql | 4 ++++ 3 files changed, 16 insertions(+), 4 deletions(-) diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index 05f092b5060..3ffe054a439 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -1779,8 +1779,8 @@ public: // const +|- variable if (left.column && isColumnConst(*left.column)) { - auto left_type = removeLowCardinality(removeNullable(left.type)); - auto right_type = removeLowCardinality(removeNullable(right.type)); + auto left_type = removeNullable(removeLowCardinality(left.type)); + auto right_type = removeNullable(removeLowCardinality(right.type)); auto transform = [&](const Field & point) { @@ -1820,8 +1820,8 @@ public: // variable +|- constant else if (right.column && isColumnConst(*right.column)) { - auto left_type = removeLowCardinality(removeNullable(left.type)); - auto right_type = removeLowCardinality(removeNullable(right.type)); + auto left_type = removeNullable(removeLowCardinality(left.type)); + auto right_type = removeNullable(removeLowCardinality(right.type)); auto transform = [&](const Field & point) { diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference index c2983d46447..5ee6e6c67c4 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference @@ -22,3 +22,11 @@ 2 1 2 +1 +2 +1 +2 +1 +2 +1 +2 diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql index 798868fe566..4f20ec0cd35 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql @@ -7,9 +7,11 @@ set max_rows_to_read = 2; SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(Nullable(UInt8))) <= 2 order by x; SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; drop table tab; @@ -22,6 +24,8 @@ set max_rows_to_read = 2; SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(Nullable(UInt8))) <= 2 order by x; SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; From 58b6194a2d3ff7361617217538c573db7c529b10 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Wed, 19 Oct 2022 15:08:00 +0000 Subject: [PATCH 040/239] Fix another one case. --- src/Functions/FunctionsConversion.h | 4 ++++ .../0_stateless/02461_mullable_pk_monotonicity_bug.sql | 9 +++++++++ 2 files changed, 13 insertions(+) diff --git a/src/Functions/FunctionsConversion.h b/src/Functions/FunctionsConversion.h index 8cbe3b0e532..ebf8417aea6 100644 --- a/src/Functions/FunctionsConversion.h +++ b/src/Functions/FunctionsConversion.h @@ -2209,6 +2209,10 @@ struct ToNumberMonotonicity /// Integer cases. + /// Do not support 128 bit integers and decimals for now. + if (!isNativeInteger(type)) + return {}; + const bool from_is_unsigned = type.isValueRepresentedByUnsignedInteger(); const bool to_is_unsigned = is_unsigned_v; diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql index 4f20ec0cd35..ce190694514 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql @@ -29,3 +29,12 @@ SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; + +drop table tab; + +set max_rows_to_read = 100; + +create table tab (x UInt128) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; +insert into tab select number from numbers(4); + +SELECT x + 1 FROM tab WHERE (x + 1::LowCardinality(UInt8)) <= -9223372036854775808 order by x; From 830ab3a5a959134ff3fad278b36b160d5e185f1e Mon Sep 17 00:00:00 2001 From: Kruglov Pavel <48961922+Avogar@users.noreply.github.com> Date: Wed, 19 Oct 2022 21:48:01 +0200 Subject: [PATCH 041/239] Fix style --- src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp b/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp index 9f09c61ec2f..31f49fa5490 100644 --- a/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp +++ b/src/Storages/addColumnsStructureToQueryWithClusterEngine.cpp @@ -9,6 +9,11 @@ namespace DB { +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; +} + static ASTExpressionList * extractTableFunctionArgumentsFromSelectQuery(ASTPtr & query) { auto * select_query = query->as(); From ae3fa705946745a4340bd0fe0a29acea7cb0a955 Mon Sep 17 00:00:00 2001 From: Kruglov Pavel <48961922+Avogar@users.noreply.github.com> Date: Wed, 19 Oct 2022 21:48:20 +0200 Subject: [PATCH 042/239] Fix style --- src/Storages/StorageS3Cluster.cpp | 5 ----- 1 file changed, 5 deletions(-) diff --git a/src/Storages/StorageS3Cluster.cpp b/src/Storages/StorageS3Cluster.cpp index 6376fac4cfc..bdf08327a97 100644 --- a/src/Storages/StorageS3Cluster.cpp +++ b/src/Storages/StorageS3Cluster.cpp @@ -39,11 +39,6 @@ namespace DB { -namespace ErrorCodes -{ - extern const int LOGICAL_ERROR; -} - StorageS3Cluster::StorageS3Cluster( const StorageS3ClusterConfiguration & configuration_, const StorageID & table_id_, From df96acc80f087a876387798e6107c5aeb13161ff Mon Sep 17 00:00:00 2001 From: Kruglov Pavel <48961922+Avogar@users.noreply.github.com> Date: Wed, 19 Oct 2022 21:48:45 +0200 Subject: [PATCH 043/239] Fix style --- src/Storages/HDFS/StorageHDFSCluster.cpp | 5 ----- 1 file changed, 5 deletions(-) diff --git a/src/Storages/HDFS/StorageHDFSCluster.cpp b/src/Storages/HDFS/StorageHDFSCluster.cpp index 7b110bb09fc..2b0222b7349 100644 --- a/src/Storages/HDFS/StorageHDFSCluster.cpp +++ b/src/Storages/HDFS/StorageHDFSCluster.cpp @@ -33,11 +33,6 @@ namespace DB { -namespace ErrorCodes -{ - extern const int LOGICAL_ERROR; -} - StorageHDFSCluster::StorageHDFSCluster( ContextPtr context_, String cluster_name_, From 2499ab3b3a9bb9fcf9ff09172fb0b869f7db4e19 Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Thu, 20 Oct 2022 17:53:52 +0800 Subject: [PATCH 044/239] Use convertFieldToType --- src/Storages/MergeTree/KeyCondition.cpp | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp index 3ecaff2c19a..194cfbdabfc 100644 --- a/src/Storages/MergeTree/KeyCondition.cpp +++ b/src/Storages/MergeTree/KeyCondition.cpp @@ -6,7 +6,6 @@ #include #include #include -#include #include #include #include @@ -1181,14 +1180,10 @@ bool KeyCondition::transformConstantWithValidFunctions( if (is_valid_chain) { auto const_type = cur_node->result_type; - auto const_column = out_type->createColumnConst(1, out_value); - auto maybe_const_value_column = tryCastColumnAccurate({const_column, out_type, ""}, const_type); - - if (maybe_const_value_column == nullptr) + auto const_value = convertFieldToType(out_value, *const_type); + if (const_value.isNull()) return false; - auto const_value = (*maybe_const_value_column)[0]; - while (!chain.empty()) { const auto * func = chain.top(); From 820e6b427697211d15988a38aaefbc6fd6bb7e39 Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 10:52:43 +0000 Subject: [PATCH 045/239] Build with libcxx(abi) 15 --- contrib/libcxx | 2 +- contrib/libcxx-cmake/CMakeLists.txt | 2 + contrib/libcxxabi | 2 +- contrib/libcxxabi-cmake/CMakeLists.txt | 1 + contrib/rocksdb | 2 +- .../tests/gtest_compressionCodec.cpp | 2 +- .../tests/gtest_json_parser.cpp | 2 +- .../Algorithms/AggregatingSortedAlgorithm.cpp | 66 ------------------ .../Algorithms/AggregatingSortedAlgorithm.h | 67 ++++++++++++++++++- .../Algorithms/SummingSortedAlgorithm.cpp | 8 +-- 10 files changed, 77 insertions(+), 77 deletions(-) diff --git a/contrib/libcxx b/contrib/libcxx index 172b2ae074f..19330eed499 160000 --- a/contrib/libcxx +++ b/contrib/libcxx @@ -1 +1 @@ -Subproject commit 172b2ae074f6755145b91c53a95c8540c1468239 +Subproject commit 19330eed499f0f2011437a92b7e8567ea36fe082 diff --git a/contrib/libcxx-cmake/CMakeLists.txt b/contrib/libcxx-cmake/CMakeLists.txt index 6f42a479588..53c6ff58f83 100644 --- a/contrib/libcxx-cmake/CMakeLists.txt +++ b/contrib/libcxx-cmake/CMakeLists.txt @@ -25,6 +25,7 @@ set(SRCS "${LIBCXX_SOURCE_DIR}/src/ios.cpp" "${LIBCXX_SOURCE_DIR}/src/ios.instantiations.cpp" "${LIBCXX_SOURCE_DIR}/src/iostream.cpp" +"${LIBCXX_SOURCE_DIR}/src/legacy_debug_handler.cpp" "${LIBCXX_SOURCE_DIR}/src/legacy_pointer_safety.cpp" "${LIBCXX_SOURCE_DIR}/src/locale.cpp" "${LIBCXX_SOURCE_DIR}/src/memory.cpp" @@ -49,6 +50,7 @@ set(SRCS "${LIBCXX_SOURCE_DIR}/src/valarray.cpp" "${LIBCXX_SOURCE_DIR}/src/variant.cpp" "${LIBCXX_SOURCE_DIR}/src/vector.cpp" +"${LIBCXX_SOURCE_DIR}/src/verbose_abort.cpp" ) add_library(cxx ${SRCS}) diff --git a/contrib/libcxxabi b/contrib/libcxxabi index 6eb7cc7a7bd..a736a6b3c6a 160000 --- a/contrib/libcxxabi +++ b/contrib/libcxxabi @@ -1 +1 @@ -Subproject commit 6eb7cc7a7bdd779e6734d1b9fb451df2274462d7 +Subproject commit a736a6b3c6a7b8aae2ebad629ca21b2c55b4820e diff --git a/contrib/libcxxabi-cmake/CMakeLists.txt b/contrib/libcxxabi-cmake/CMakeLists.txt index bf1ede8a60e..221a18de6e5 100644 --- a/contrib/libcxxabi-cmake/CMakeLists.txt +++ b/contrib/libcxxabi-cmake/CMakeLists.txt @@ -9,6 +9,7 @@ set(SRCS "${LIBCXXABI_SOURCE_DIR}/src/cxa_exception_storage.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_guard.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_handlers.cpp" +"${LIBCXXABI_SOURCE_DIR}/src/cxa_noexception.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_personality.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_thread_atexit.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_vector.cpp" diff --git a/contrib/rocksdb b/contrib/rocksdb index e7c2b2f7bcf..2c8998e26c6 160000 --- a/contrib/rocksdb +++ b/contrib/rocksdb @@ -1 +1 @@ -Subproject commit e7c2b2f7bcf3b4b33892a1a6d25c32a93edfbdb9 +Subproject commit 2c8998e26c6d46b27c710d7829c3a15e34959f70 diff --git a/src/Compression/tests/gtest_compressionCodec.cpp b/src/Compression/tests/gtest_compressionCodec.cpp index 2df3edb23ad..633c2426d2c 100644 --- a/src/Compression/tests/gtest_compressionCodec.cpp +++ b/src/Compression/tests/gtest_compressionCodec.cpp @@ -1043,7 +1043,7 @@ INSTANTIATE_TEST_SUITE_P(RandomInt, ::testing::Combine( DefaultCodecsToTest, ::testing::Values( - generateSeq(G(RandomGenerator(0))), + generateSeq(G(RandomGenerator(0))), generateSeq(G(RandomGenerator(0))), generateSeq(G(RandomGenerator(0, 0, 1000'000'000))), generateSeq(G(RandomGenerator(0, 0, 1000'000'000))) diff --git a/src/DataTypes/Serializations/tests/gtest_json_parser.cpp b/src/DataTypes/Serializations/tests/gtest_json_parser.cpp index 4dddb3cd03d..9b0c8e44d02 100644 --- a/src/DataTypes/Serializations/tests/gtest_json_parser.cpp +++ b/src/DataTypes/Serializations/tests/gtest_json_parser.cpp @@ -69,7 +69,7 @@ static std::ostream & operator<<(std::ostream & ostr, const JSONPathAndValue & p bool first = true; for (const auto & part : path_and_value.path.getParts()) { - ostr << (first ? "{" : ", {") << part.key << ", " << part.is_nested << ", " << part.anonymous_array_level << "}"; + ostr << (first ? "{" : ", {") << part.key << ", " << part.is_nested << ", " << static_cast(part.anonymous_array_level) << "}"; first = false; } diff --git a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp index ebc1b37074b..7a0847bcbc5 100644 --- a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp +++ b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp @@ -1,7 +1,5 @@ #include -#include -#include #include #include #include @@ -18,70 +16,6 @@ AggregatingSortedAlgorithm::ColumnsDefinition::ColumnsDefinition() = default; AggregatingSortedAlgorithm::ColumnsDefinition::ColumnsDefinition(ColumnsDefinition &&) noexcept = default; AggregatingSortedAlgorithm::ColumnsDefinition::~ColumnsDefinition() = default; -/// Stores information for aggregation of AggregateFunction columns -struct AggregatingSortedAlgorithm::AggregateDescription -{ - ColumnAggregateFunction * column = nullptr; - const size_t column_number = 0; /// Position in header. - - AggregateDescription() = default; - explicit AggregateDescription(size_t col_number) : column_number(col_number) {} -}; - -/// Stores information for aggregation of SimpleAggregateFunction columns -struct AggregatingSortedAlgorithm::SimpleAggregateDescription -{ - /// An aggregate function 'anyLast', 'sum'... - AggregateFunctionPtr function; - IAggregateFunction::AddFunc add_function = nullptr; - - size_t column_number = 0; - IColumn * column = nullptr; - - /// For LowCardinality, convert is converted to nested type. nested_type is nullptr if no conversion needed. - const DataTypePtr nested_type; /// Nested type for LowCardinality, if it is. - const DataTypePtr real_type; /// Type in header. - - AlignedBuffer state; - bool created = false; - - SimpleAggregateDescription( - AggregateFunctionPtr function_, const size_t column_number_, - DataTypePtr nested_type_, DataTypePtr real_type_) - : function(std::move(function_)), column_number(column_number_) - , nested_type(std::move(nested_type_)), real_type(std::move(real_type_)) - { - add_function = function->getAddressOfAddFunction(); - state.reset(function->sizeOfData(), function->alignOfData()); - } - - void createState() - { - if (created) - return; - function->create(state.data()); - created = true; - } - - void destroyState() - { - if (!created) - return; - function->destroy(state.data()); - created = false; - } - - /// Explicitly destroy aggregation state if the stream is terminated - ~SimpleAggregateDescription() - { - destroyState(); - } - - SimpleAggregateDescription() = default; - SimpleAggregateDescription(SimpleAggregateDescription &&) = default; - SimpleAggregateDescription(const SimpleAggregateDescription &) = delete; -}; - static AggregatingSortedAlgorithm::ColumnsDefinition defineColumns( const Block & header, const SortDescription & description) { diff --git a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h index e572ed7d526..0ffbd1262d3 100644 --- a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h +++ b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h @@ -1,5 +1,7 @@ #pragma once +#include +#include #include #include @@ -23,8 +25,69 @@ public: void consume(Input & input, size_t source_num) override; Status merge() override; - struct SimpleAggregateDescription; - struct AggregateDescription; + /// Stores information for aggregation of SimpleAggregateFunction columns + struct SimpleAggregateDescription + { + /// An aggregate function 'anyLast', 'sum'... + AggregateFunctionPtr function; + IAggregateFunction::AddFunc add_function = nullptr; + + size_t column_number = 0; + IColumn * column = nullptr; + + /// For LowCardinality, convert is converted to nested type. nested_type is nullptr if no conversion needed. + const DataTypePtr nested_type; /// Nested type for LowCardinality, if it is. + const DataTypePtr real_type; /// Type in header. + + AlignedBuffer state; + bool created = false; + + SimpleAggregateDescription( + AggregateFunctionPtr function_, const size_t column_number_, + DataTypePtr nested_type_, DataTypePtr real_type_) + : function(std::move(function_)), column_number(column_number_) + , nested_type(std::move(nested_type_)), real_type(std::move(real_type_)) + { + add_function = function->getAddressOfAddFunction(); + state.reset(function->sizeOfData(), function->alignOfData()); + } + + void createState() + { + if (created) + return; + function->create(state.data()); + created = true; + } + + void destroyState() + { + if (!created) + return; + function->destroy(state.data()); + created = false; + } + + /// Explicitly destroy aggregation state if the stream is terminated + ~SimpleAggregateDescription() + { + destroyState(); + } + + SimpleAggregateDescription() = default; + SimpleAggregateDescription(SimpleAggregateDescription &&) = default; + SimpleAggregateDescription(const SimpleAggregateDescription &) = delete; + }; + + /// Stores information for aggregation of AggregateFunction columns + struct AggregateDescription + { + ColumnAggregateFunction * column = nullptr; + const size_t column_number = 0; /// Position in header. + + AggregateDescription() = default; + explicit AggregateDescription(size_t col_number) : column_number(col_number) {} + }; /// This structure define columns into one of three types: /// * columns which are not aggregate functions and not needed to be aggregated diff --git a/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp b/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp index 8636813132d..c79c667a988 100644 --- a/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp +++ b/src/Processors/Merges/Algorithms/SummingSortedAlgorithm.cpp @@ -23,10 +23,6 @@ namespace ErrorCodes extern const int CORRUPTED_DATA; } -SummingSortedAlgorithm::ColumnsDefinition::ColumnsDefinition() = default; -SummingSortedAlgorithm::ColumnsDefinition::ColumnsDefinition(ColumnsDefinition &&) noexcept = default; -SummingSortedAlgorithm::ColumnsDefinition::~ColumnsDefinition() = default; - /// Stores numbers of key-columns and value-columns. struct SummingSortedAlgorithm::MapDescription { @@ -777,4 +773,8 @@ IMergingAlgorithm::Status SummingSortedAlgorithm::merge() return Status(merged_data.pull(), true); } +SummingSortedAlgorithm::ColumnsDefinition::ColumnsDefinition() = default; +SummingSortedAlgorithm::ColumnsDefinition::ColumnsDefinition(ColumnsDefinition &&) noexcept = default; +SummingSortedAlgorithm::ColumnsDefinition::~ColumnsDefinition() = default; + } From cdae7082a30da62e8b2df4430360887494a44148 Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 11:09:32 +0000 Subject: [PATCH 046/239] Up-inline ShardPartition, ShardPartitionPiece and TaskCluster --- programs/copier/CMakeLists.txt | 5 +- programs/copier/ShardPartition.cpp | 63 +++++++++++++++++++++ programs/copier/ShardPartition.h | 57 +------------------ programs/copier/ShardPartitionPiece.cpp | 56 +++++++++++++++++++ programs/copier/ShardPartitionPiece.h | 53 +----------------- programs/copier/TaskCluster.cpp | 74 +++++++++++++++++++++++++ programs/copier/TaskCluster.h | 66 +--------------------- 7 files changed, 200 insertions(+), 174 deletions(-) create mode 100644 programs/copier/ShardPartition.cpp create mode 100644 programs/copier/ShardPartitionPiece.cpp create mode 100644 programs/copier/TaskCluster.cpp diff --git a/programs/copier/CMakeLists.txt b/programs/copier/CMakeLists.txt index 57e0996ed78..394983d51b2 100644 --- a/programs/copier/CMakeLists.txt +++ b/programs/copier/CMakeLists.txt @@ -1,7 +1,10 @@ set(CLICKHOUSE_COPIER_SOURCES "${CMAKE_CURRENT_SOURCE_DIR}/ClusterCopierApp.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/ClusterCopier.cpp" - "${CMAKE_CURRENT_SOURCE_DIR}/Internals.cpp") + "${CMAKE_CURRENT_SOURCE_DIR}/Internals.cpp" + "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartition.cpp" + "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartitionPiece.cpp" + "${CMAKE_CURRENT_SOURCE_DIR}/TaskCluster.cpp") set (CLICKHOUSE_COPIER_LINK PRIVATE diff --git a/programs/copier/ShardPartition.cpp b/programs/copier/ShardPartition.cpp new file mode 100644 index 00000000000..fb482163377 --- /dev/null +++ b/programs/copier/ShardPartition.cpp @@ -0,0 +1,63 @@ +#include "ShardPartition.h" + +namespace DB +{ + +ShardPartition::ShardPartition(TaskShard &parent, String name_quoted_, size_t number_of_splits) + : task_shard(parent), name(std::move(name_quoted_)) { pieces.reserve(number_of_splits); } + +String ShardPartition::getPartitionCleanStartPath() const +{ + return getPartitionPath() + "/clean_start"; +} + +String ShardPartition::getPartitionPieceCleanStartPath(size_t current_piece_number) const +{ + assert(current_piece_number < task_shard.task_table.number_of_splits); + return getPartitionPiecePath(current_piece_number) + "/clean_start"; +} + +String ShardPartition::getPartitionPath() const +{ + return task_shard.task_table.getPartitionPath(name); +} + +String ShardPartition::getPartitionPiecePath(size_t current_piece_number) const +{ + assert(current_piece_number < task_shard.task_table.number_of_splits); + return task_shard.task_table.getPartitionPiecePath(name, current_piece_number); +} + +String ShardPartition::getShardStatusPath() const +{ + // schema: //tables///shards/ + // e.g. /root/table_test.hits/201701/shards/1 + return getPartitionShardsPath() + "/" + toString(task_shard.numberInCluster()); +} + +String ShardPartition::getPartitionShardsPath() const +{ + return getPartitionPath() + "/shards"; +} + +String ShardPartition::getPartitionActiveWorkersPath() const +{ + return getPartitionPath() + "/partition_active_workers"; +} + +String ShardPartition::getActiveWorkerPath() const +{ + return getPartitionActiveWorkersPath() + "/" + toString(task_shard.numberInCluster()); +} + +String ShardPartition::getCommonPartitionIsDirtyPath() const +{ + return getPartitionPath() + "/is_dirty"; +} + +String ShardPartition::getCommonPartitionIsCleanedPath() const +{ + return getCommonPartitionIsDirtyPath() + "/cleaned"; +} + +} diff --git a/programs/copier/ShardPartition.h b/programs/copier/ShardPartition.h index 7de381977f9..a78790c0db1 100644 --- a/programs/copier/ShardPartition.h +++ b/programs/copier/ShardPartition.h @@ -12,8 +12,7 @@ namespace DB /// This class describes a partition (name) that is stored on the shard (parent). struct ShardPartition { - ShardPartition(TaskShard &parent, String name_quoted_, size_t number_of_splits = 10) - : task_shard(parent), name(std::move(name_quoted_)) { pieces.reserve(number_of_splits); } + ShardPartition(TaskShard &parent, String name_quoted_, size_t number_of_splits = 10); String getPartitionPath() const; @@ -45,58 +44,4 @@ struct ShardPartition String name; }; -inline String ShardPartition::getPartitionCleanStartPath() const -{ - return getPartitionPath() + "/clean_start"; -} - -inline String ShardPartition::getPartitionPieceCleanStartPath(size_t current_piece_number) const -{ - assert(current_piece_number < task_shard.task_table.number_of_splits); - return getPartitionPiecePath(current_piece_number) + "/clean_start"; -} - -inline String ShardPartition::getPartitionPath() const -{ - return task_shard.task_table.getPartitionPath(name); -} - -inline String ShardPartition::getPartitionPiecePath(size_t current_piece_number) const -{ - assert(current_piece_number < task_shard.task_table.number_of_splits); - return task_shard.task_table.getPartitionPiecePath(name, current_piece_number); -} - -inline String ShardPartition::getShardStatusPath() const -{ - // schema: //tables/
//shards/ - // e.g. /root/table_test.hits/201701/shards/1 - return getPartitionShardsPath() + "/" + toString(task_shard.numberInCluster()); -} - -inline String ShardPartition::getPartitionShardsPath() const -{ - return getPartitionPath() + "/shards"; -} - -inline String ShardPartition::getPartitionActiveWorkersPath() const -{ - return getPartitionPath() + "/partition_active_workers"; -} - -inline String ShardPartition::getActiveWorkerPath() const -{ - return getPartitionActiveWorkersPath() + "/" + toString(task_shard.numberInCluster()); -} - -inline String ShardPartition::getCommonPartitionIsDirtyPath() const -{ - return getPartitionPath() + "/is_dirty"; -} - -inline String ShardPartition::getCommonPartitionIsCleanedPath() const -{ - return getCommonPartitionIsDirtyPath() + "/cleaned"; -} - } diff --git a/programs/copier/ShardPartitionPiece.cpp b/programs/copier/ShardPartitionPiece.cpp new file mode 100644 index 00000000000..c060a955c9a --- /dev/null +++ b/programs/copier/ShardPartitionPiece.cpp @@ -0,0 +1,56 @@ +#include "ShardPartitionPiece.h" + +namespace DB +{ + +ShardPartitionPiece::ShardPartitionPiece(ShardPartition &parent, size_t current_piece_number_, bool is_present_piece_) + : is_absent_piece(!is_present_piece_), current_piece_number(current_piece_number_), + shard_partition(parent) {} + +String ShardPartitionPiece::getPartitionPiecePath() const +{ + return shard_partition.getPartitionPath() + "/piece_" + toString(current_piece_number); +} + +String ShardPartitionPiece::getPartitionPieceCleanStartPath() const +{ + return getPartitionPiecePath() + "/clean_start"; +} + +String ShardPartitionPiece::getPartitionPieceIsDirtyPath() const +{ + return getPartitionPiecePath() + "/is_dirty"; +} + +String ShardPartitionPiece::getPartitionPieceIsCleanedPath() const +{ + return getPartitionPieceIsDirtyPath() + "/cleaned"; +} + +String ShardPartitionPiece::getPartitionPieceActiveWorkersPath() const +{ + return getPartitionPiecePath() + "/partition_piece_active_workers"; +} + +String ShardPartitionPiece::getActiveWorkerPath() const +{ + return getPartitionPieceActiveWorkersPath() + "/" + toString(shard_partition.task_shard.numberInCluster()); +} + +/// On what shards do we have current partition. +String ShardPartitionPiece::getPartitionPieceShardsPath() const +{ + return getPartitionPiecePath() + "/shards"; +} + +String ShardPartitionPiece::getShardStatusPath() const +{ + return getPartitionPieceShardsPath() + "/" + toString(shard_partition.task_shard.numberInCluster()); +} + +String ShardPartitionPiece::getPartitionPieceCleanerPath() const +{ + return getPartitionPieceIsDirtyPath() + "/cleaner"; +} + +} diff --git a/programs/copier/ShardPartitionPiece.h b/programs/copier/ShardPartitionPiece.h index a21fd531da4..243cb7439b8 100644 --- a/programs/copier/ShardPartitionPiece.h +++ b/programs/copier/ShardPartitionPiece.h @@ -7,10 +7,7 @@ namespace DB struct ShardPartitionPiece { - - ShardPartitionPiece(ShardPartition &parent, size_t current_piece_number_, bool is_present_piece_) - : is_absent_piece(!is_present_piece_), current_piece_number(current_piece_number_), - shard_partition(parent) {} + ShardPartitionPiece(ShardPartition &parent, size_t current_piece_number_, bool is_present_piece_); String getPartitionPiecePath() const; @@ -37,52 +34,4 @@ struct ShardPartitionPiece ShardPartition & shard_partition; }; - -inline String ShardPartitionPiece::getPartitionPiecePath() const -{ - return shard_partition.getPartitionPath() + "/piece_" + toString(current_piece_number); -} - -inline String ShardPartitionPiece::getPartitionPieceCleanStartPath() const -{ - return getPartitionPiecePath() + "/clean_start"; -} - -inline String ShardPartitionPiece::getPartitionPieceIsDirtyPath() const -{ - return getPartitionPiecePath() + "/is_dirty"; -} - -inline String ShardPartitionPiece::getPartitionPieceIsCleanedPath() const -{ - return getPartitionPieceIsDirtyPath() + "/cleaned"; -} - -inline String ShardPartitionPiece::getPartitionPieceActiveWorkersPath() const -{ - return getPartitionPiecePath() + "/partition_piece_active_workers"; -} - -inline String ShardPartitionPiece::getActiveWorkerPath() const -{ - return getPartitionPieceActiveWorkersPath() + "/" + toString(shard_partition.task_shard.numberInCluster()); -} - -/// On what shards do we have current partition. -inline String ShardPartitionPiece::getPartitionPieceShardsPath() const -{ - return getPartitionPiecePath() + "/shards"; -} - -inline String ShardPartitionPiece::getShardStatusPath() const -{ - return getPartitionPieceShardsPath() + "/" + toString(shard_partition.task_shard.numberInCluster()); -} - -inline String ShardPartitionPiece::getPartitionPieceCleanerPath() const -{ - return getPartitionPieceIsDirtyPath() + "/cleaner"; -} - - } diff --git a/programs/copier/TaskCluster.cpp b/programs/copier/TaskCluster.cpp new file mode 100644 index 00000000000..6b7911f56f2 --- /dev/null +++ b/programs/copier/TaskCluster.cpp @@ -0,0 +1,74 @@ +#include "TaskCluster.h" + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int BAD_ARGUMENTS; +} + +TaskCluster::TaskCluster(const String & task_zookeeper_path_, const String & default_local_database_) + : task_zookeeper_path(task_zookeeper_path_) + , default_local_database(default_local_database_) +{} + +void DB::TaskCluster::loadTasks(const Poco::Util::AbstractConfiguration & config, const String & base_key) +{ + String prefix = base_key.empty() ? "" : base_key + "."; + + clusters_prefix = prefix + "remote_servers"; + if (!config.has(clusters_prefix)) + throw Exception("You should specify list of clusters in " + clusters_prefix, ErrorCodes::BAD_ARGUMENTS); + + Poco::Util::AbstractConfiguration::Keys tables_keys; + config.keys(prefix + "tables", tables_keys); + + for (const auto & table_key : tables_keys) + { + table_tasks.emplace_back(*this, config, prefix + "tables", table_key); + } +} + +void DB::TaskCluster::reloadSettings(const Poco::Util::AbstractConfiguration & config, const String & base_key) +{ + String prefix = base_key.empty() ? "" : base_key + "."; + + max_workers = config.getUInt64(prefix + "max_workers"); + + settings_common = Settings(); + if (config.has(prefix + "settings")) + settings_common.loadSettingsFromConfig(prefix + "settings", config); + + settings_common.prefer_localhost_replica = 0; + + settings_pull = settings_common; + if (config.has(prefix + "settings_pull")) + settings_pull.loadSettingsFromConfig(prefix + "settings_pull", config); + + settings_push = settings_common; + if (config.has(prefix + "settings_push")) + settings_push.loadSettingsFromConfig(prefix + "settings_push", config); + + auto set_default_value = [] (auto && setting, auto && default_value) + { + setting = setting.changed ? setting.value : default_value; + }; + + /// Override important settings + settings_pull.readonly = 1; + settings_pull.prefer_localhost_replica = false; + settings_push.insert_distributed_sync = true; + settings_push.prefer_localhost_replica = false; + + set_default_value(settings_pull.load_balancing, LoadBalancing::NEAREST_HOSTNAME); + set_default_value(settings_pull.max_threads, 1); + set_default_value(settings_pull.max_block_size, 8192UL); + set_default_value(settings_pull.preferred_block_size_bytes, 0); + + set_default_value(settings_push.insert_distributed_timeout, 0); + set_default_value(settings_push.replication_alter_partitions_sync, 2); +} + +} + diff --git a/programs/copier/TaskCluster.h b/programs/copier/TaskCluster.h index 7d8f01ba15f..c59c7bff4d7 100644 --- a/programs/copier/TaskCluster.h +++ b/programs/copier/TaskCluster.h @@ -5,17 +5,10 @@ namespace DB { -namespace ErrorCodes -{ - extern const int BAD_ARGUMENTS; -} struct TaskCluster { - TaskCluster(const String & task_zookeeper_path_, const String & default_local_database_) - : task_zookeeper_path(task_zookeeper_path_) - , default_local_database(default_local_database_) - {} + TaskCluster(const String & task_zookeeper_path_, const String & default_local_database_); void loadTasks(const Poco::Util::AbstractConfiguration & config, const String & base_key = ""); @@ -50,61 +43,4 @@ struct TaskCluster pcg64 random_engine; }; -inline void DB::TaskCluster::loadTasks(const Poco::Util::AbstractConfiguration & config, const String & base_key) -{ - String prefix = base_key.empty() ? "" : base_key + "."; - - clusters_prefix = prefix + "remote_servers"; - if (!config.has(clusters_prefix)) - throw Exception("You should specify list of clusters in " + clusters_prefix, ErrorCodes::BAD_ARGUMENTS); - - Poco::Util::AbstractConfiguration::Keys tables_keys; - config.keys(prefix + "tables", tables_keys); - - for (const auto & table_key : tables_keys) - { - table_tasks.emplace_back(*this, config, prefix + "tables", table_key); - } -} - -inline void DB::TaskCluster::reloadSettings(const Poco::Util::AbstractConfiguration & config, const String & base_key) -{ - String prefix = base_key.empty() ? "" : base_key + "."; - - max_workers = config.getUInt64(prefix + "max_workers"); - - settings_common = Settings(); - if (config.has(prefix + "settings")) - settings_common.loadSettingsFromConfig(prefix + "settings", config); - - settings_common.prefer_localhost_replica = 0; - - settings_pull = settings_common; - if (config.has(prefix + "settings_pull")) - settings_pull.loadSettingsFromConfig(prefix + "settings_pull", config); - - settings_push = settings_common; - if (config.has(prefix + "settings_push")) - settings_push.loadSettingsFromConfig(prefix + "settings_push", config); - - auto set_default_value = [] (auto && setting, auto && default_value) - { - setting = setting.changed ? setting.value : default_value; - }; - - /// Override important settings - settings_pull.readonly = 1; - settings_pull.prefer_localhost_replica = false; - settings_push.insert_distributed_sync = true; - settings_push.prefer_localhost_replica = false; - - set_default_value(settings_pull.load_balancing, LoadBalancing::NEAREST_HOSTNAME); - set_default_value(settings_pull.max_threads, 1); - set_default_value(settings_pull.max_block_size, 8192UL); - set_default_value(settings_pull.preferred_block_size_bytes, 0); - - set_default_value(settings_push.insert_distributed_timeout, 0); - set_default_value(settings_push.replication_alter_partitions_sync, 2); -} - } From f6de964eaa8dd4f14bcfbb28a999a9ae7bed384f Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 11:20:18 +0000 Subject: [PATCH 047/239] Fix static_assert in random generator --- programs/copier/TaskTableAndShard.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/programs/copier/TaskTableAndShard.h b/programs/copier/TaskTableAndShard.h index cef9b669971..e643d66d1c6 100644 --- a/programs/copier/TaskTableAndShard.h +++ b/programs/copier/TaskTableAndShard.h @@ -361,7 +361,7 @@ template inline void TaskTable::initShards(RandomEngine && random_engine) { const String & fqdn_name = getFQDNOrHostName(); - std::uniform_int_distribution get_urand(0, std::numeric_limits::max()); + std::uniform_int_distribution get_urand(0, std::numeric_limits::max()); // Compute the priority for (const auto & shard_info : cluster_pull->getShardsInfo()) From 0df9d2f5ccd8f88e967e830d6a7bf5c9e07e830e Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Thu, 20 Oct 2022 13:41:27 +0000 Subject: [PATCH 048/239] Fix tests. --- src/Functions/FunctionsConversion.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/Functions/FunctionsConversion.h b/src/Functions/FunctionsConversion.h index ebf8417aea6..b801cc070b6 100644 --- a/src/Functions/FunctionsConversion.h +++ b/src/Functions/FunctionsConversion.h @@ -2209,16 +2209,16 @@ struct ToNumberMonotonicity /// Integer cases. - /// Do not support 128 bit integers and decimals for now. - if (!isNativeInteger(type)) - return {}; - const bool from_is_unsigned = type.isValueRepresentedByUnsignedInteger(); const bool to_is_unsigned = is_unsigned_v; const size_t size_of_from = type.getSizeOfValueInMemory(); const size_t size_of_to = sizeof(T); + /// Do not support 128 bit integers and decimals for now. + if (size_of_from > sizeof(Int64)) + return {}; + const bool left_in_first_half = left.isNull() ? from_is_unsigned : (left.get() >= 0); From df318d871964c30f0c303ba50077c491c54d89b3 Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 11:45:15 +0000 Subject: [PATCH 049/239] Un-inline TaskTableAndShard --- programs/copier/CMakeLists.txt | 3 +- programs/copier/TaskTableAndShard.cpp | 241 ++++++++++++++++++++++++++ programs/copier/TaskTableAndShard.h | 229 +----------------------- 3 files changed, 249 insertions(+), 224 deletions(-) create mode 100644 programs/copier/TaskTableAndShard.cpp diff --git a/programs/copier/CMakeLists.txt b/programs/copier/CMakeLists.txt index 394983d51b2..69019843d05 100644 --- a/programs/copier/CMakeLists.txt +++ b/programs/copier/CMakeLists.txt @@ -4,7 +4,8 @@ set(CLICKHOUSE_COPIER_SOURCES "${CMAKE_CURRENT_SOURCE_DIR}/Internals.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartition.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartitionPiece.cpp" - "${CMAKE_CURRENT_SOURCE_DIR}/TaskCluster.cpp") + "${CMAKE_CURRENT_SOURCE_DIR}/TaskCluster.cpp" + "${CMAKE_CURRENT_SOURCE_DIR}/TaskTableAndShard.cpp") set (CLICKHOUSE_COPIER_LINK PRIVATE diff --git a/programs/copier/TaskTableAndShard.cpp b/programs/copier/TaskTableAndShard.cpp new file mode 100644 index 00000000000..36aff4f4cd8 --- /dev/null +++ b/programs/copier/TaskTableAndShard.cpp @@ -0,0 +1,241 @@ +#include "TaskTableAndShard.h" + +namespace DB +{ +namespace ErrorCodes +{ + extern const int UNKNOWN_ELEMENT_IN_CONFIG; + extern const int LOGICAL_ERROR; +} + +TaskTable::TaskTable(TaskCluster & parent, const Poco::Util::AbstractConfiguration & config, + const String & prefix_, const String & table_key) + : task_cluster(parent) +{ + String table_prefix = prefix_ + "." + table_key + "."; + + name_in_config = table_key; + + number_of_splits = config.getUInt64(table_prefix + "number_of_splits", 3); + + allow_to_copy_alias_and_materialized_columns = config.getBool(table_prefix + "allow_to_copy_alias_and_materialized_columns", false); + allow_to_drop_target_partitions = config.getBool(table_prefix + "allow_to_drop_target_partitions", false); + + cluster_pull_name = config.getString(table_prefix + "cluster_pull"); + cluster_push_name = config.getString(table_prefix + "cluster_push"); + + table_pull.first = config.getString(table_prefix + "database_pull"); + table_pull.second = config.getString(table_prefix + "table_pull"); + + table_push.first = config.getString(table_prefix + "database_push"); + table_push.second = config.getString(table_prefix + "table_push"); + + /// Used as node name in ZooKeeper + table_id = escapeForFileName(cluster_push_name) + + "." + escapeForFileName(table_push.first) + + "." + escapeForFileName(table_push.second); + + engine_push_str = config.getString(table_prefix + "engine", "rand()"); + + { + ParserStorage parser_storage; + engine_push_ast = parseQuery(parser_storage, engine_push_str, 0, DBMS_DEFAULT_MAX_PARSER_DEPTH); + engine_push_partition_key_ast = extractPartitionKey(engine_push_ast); + primary_key_comma_separated = boost::algorithm::join(extractPrimaryKeyColumnNames(engine_push_ast), ", "); + is_replicated_table = isReplicatedTableEngine(engine_push_ast); + } + + sharding_key_str = config.getString(table_prefix + "sharding_key"); + + auxiliary_engine_split_asts.reserve(number_of_splits); + { + ParserExpressionWithOptionalAlias parser_expression(false); + sharding_key_ast = parseQuery(parser_expression, sharding_key_str, 0, DBMS_DEFAULT_MAX_PARSER_DEPTH); + main_engine_split_ast = createASTStorageDistributed(cluster_push_name, table_push.first, table_push.second, + sharding_key_ast); + + for (const auto piece_number : collections::range(0, number_of_splits)) + { + auxiliary_engine_split_asts.emplace_back + ( + createASTStorageDistributed(cluster_push_name, table_push.first, + table_push.second + "_piece_" + toString(piece_number), sharding_key_ast) + ); + } + } + + where_condition_str = config.getString(table_prefix + "where_condition", ""); + if (!where_condition_str.empty()) + { + ParserExpressionWithOptionalAlias parser_expression(false); + where_condition_ast = parseQuery(parser_expression, where_condition_str, 0, DBMS_DEFAULT_MAX_PARSER_DEPTH); + + // Will use canonical expression form + where_condition_str = queryToString(where_condition_ast); + } + + String enabled_partitions_prefix = table_prefix + "enabled_partitions"; + has_enabled_partitions = config.has(enabled_partitions_prefix); + + if (has_enabled_partitions) + { + Strings keys; + config.keys(enabled_partitions_prefix, keys); + + if (keys.empty()) + { + /// Parse list of partition from space-separated string + String partitions_str = config.getString(table_prefix + "enabled_partitions"); + boost::trim_if(partitions_str, isWhitespaceASCII); + boost::split(enabled_partitions, partitions_str, isWhitespaceASCII, boost::token_compress_on); + } + else + { + /// Parse sequence of ... + for (const String &key : keys) + { + if (!startsWith(key, "partition")) + throw Exception("Unknown key " + key + " in " + enabled_partitions_prefix, ErrorCodes::UNKNOWN_ELEMENT_IN_CONFIG); + + enabled_partitions.emplace_back(config.getString(enabled_partitions_prefix + "." + key)); + } + } + + std::copy(enabled_partitions.begin(), enabled_partitions.end(), std::inserter(enabled_partitions_set, enabled_partitions_set.begin())); + } +} + + +String TaskTable::getPartitionPath(const String & partition_name) const +{ + return task_cluster.task_zookeeper_path // root + + "/tables/" + table_id // tables/dst_cluster.merge.hits + + "/" + escapeForFileName(partition_name); // 201701 +} + +String TaskTable::getPartitionAttachIsActivePath(const String & partition_name) const +{ + return getPartitionPath(partition_name) + "/attach_active"; +} + +String TaskTable::getPartitionAttachIsDonePath(const String & partition_name) const +{ + return getPartitionPath(partition_name) + "/attach_is_done"; +} + +String TaskTable::getPartitionPiecePath(const String & partition_name, size_t piece_number) const +{ + assert(piece_number < number_of_splits); + return getPartitionPath(partition_name) + "/piece_" + toString(piece_number); // 1...number_of_splits +} + +String TaskTable::getCertainPartitionIsDirtyPath(const String &partition_name) const +{ + return getPartitionPath(partition_name) + "/is_dirty"; +} + +String TaskTable::getCertainPartitionPieceIsDirtyPath(const String & partition_name, const size_t piece_number) const +{ + return getPartitionPiecePath(partition_name, piece_number) + "/is_dirty"; +} + +String TaskTable::getCertainPartitionIsCleanedPath(const String & partition_name) const +{ + return getCertainPartitionIsDirtyPath(partition_name) + "/cleaned"; +} + +String TaskTable::getCertainPartitionPieceIsCleanedPath(const String & partition_name, const size_t piece_number) const +{ + return getCertainPartitionPieceIsDirtyPath(partition_name, piece_number) + "/cleaned"; +} + +String TaskTable::getCertainPartitionTaskStatusPath(const String & partition_name) const +{ + return getPartitionPath(partition_name) + "/shards"; +} + +String TaskTable::getCertainPartitionPieceTaskStatusPath(const String & partition_name, const size_t piece_number) const +{ + return getPartitionPiecePath(partition_name, piece_number) + "/shards"; +} + +bool TaskTable::isReplicatedTable() const +{ + return is_replicated_table; +} + +String TaskTable::getStatusAllPartitionCount() const +{ + return task_cluster.task_zookeeper_path + "/status/all_partitions_count"; +} + +String TaskTable::getStatusProcessedPartitionsCount() const +{ + return task_cluster.task_zookeeper_path + "/status/processed_partitions_count"; +} + +ASTPtr TaskTable::rewriteReplicatedCreateQueryToPlain() const +{ + ASTPtr prev_engine_push_ast = engine_push_ast->clone(); + + auto & new_storage_ast = prev_engine_push_ast->as(); + auto & new_engine_ast = new_storage_ast.engine->as(); + + /// Remove "Replicated" from name + new_engine_ast.name = new_engine_ast.name.substr(10); + + if (new_engine_ast.arguments) + { + auto & replicated_table_arguments = new_engine_ast.arguments->children; + + + /// In some cases of Atomic database engine usage ReplicatedMergeTree tables + /// could be created without arguments. + if (!replicated_table_arguments.empty()) + { + /// Delete first two arguments of Replicated...MergeTree() table. + replicated_table_arguments.erase(replicated_table_arguments.begin()); + replicated_table_arguments.erase(replicated_table_arguments.begin()); + } + } + + return new_storage_ast.clone(); +} + +ClusterPartition & TaskTable::getClusterPartition(const String & partition_name) +{ + auto it = cluster_partitions.find(partition_name); + if (it == cluster_partitions.end()) + throw Exception("There are no cluster partition " + partition_name + " in " + table_id, + ErrorCodes::LOGICAL_ERROR); + return it->second; +} + +TaskShard::TaskShard(TaskTable & parent, const ShardInfo & info_) : task_table(parent), info(info_) +{ + list_of_split_tables_on_shard.assign(task_table.number_of_splits, DatabaseAndTableName()); +} + +UInt32 TaskShard::numberInCluster() const +{ + return info.shard_num; +} + +UInt32 TaskShard::indexInCluster() const +{ + return info.shard_num - 1; +} + +String DB::TaskShard::getDescription() const +{ + return fmt::format("N{} (having a replica {}, pull table {} of cluster {}", + numberInCluster(), getHostNameExample(), getQuotedTable(task_table.table_pull), task_table.cluster_pull_name); +} + +String DB::TaskShard::getHostNameExample() const +{ + const auto & replicas = task_table.cluster_pull->getShardsAddresses().at(indexInCluster()); + return replicas.at(0).readableString(); +} + +} diff --git a/programs/copier/TaskTableAndShard.h b/programs/copier/TaskTableAndShard.h index e643d66d1c6..1fe5b0a255e 100644 --- a/programs/copier/TaskTableAndShard.h +++ b/programs/copier/TaskTableAndShard.h @@ -13,18 +13,12 @@ namespace DB { -namespace ErrorCodes -{ - extern const int UNKNOWN_ELEMENT_IN_CONFIG; - extern const int LOGICAL_ERROR; -} struct TaskShard; struct TaskTable { - TaskTable(TaskCluster & parent, const Poco::Util::AbstractConfiguration & config, const String & prefix, - const String & table_key); + TaskTable(TaskCluster & parent, const Poco::Util::AbstractConfiguration & config, const String & prefix, const String & table_key); TaskCluster & task_cluster; @@ -51,7 +45,7 @@ struct TaskTable String getCertainPartitionPieceTaskStatusPath(const String & partition_name, size_t piece_number) const; - bool isReplicatedTable() const { return is_replicated_table; } + bool isReplicatedTable() const; /// These nodes are used for check-status option String getStatusAllPartitionCount() const; @@ -136,14 +130,7 @@ struct TaskTable /// Partition names to process in user-specified order Strings ordered_partition_names; - ClusterPartition & getClusterPartition(const String & partition_name) - { - auto it = cluster_partitions.find(partition_name); - if (it == cluster_partitions.end()) - throw Exception("There are no cluster partition " + partition_name + " in " + table_id, - ErrorCodes::LOGICAL_ERROR); - return it->second; - } + ClusterPartition & getClusterPartition(const String & partition_name); Stopwatch watch; UInt64 bytes_copied = 0; @@ -156,18 +143,15 @@ struct TaskTable struct TaskShard { - TaskShard(TaskTable & parent, const ShardInfo & info_) : task_table(parent), info(info_) - { - list_of_split_tables_on_shard.assign(task_table.number_of_splits, DatabaseAndTableName()); - } + TaskShard(TaskTable & parent, const ShardInfo & info_); TaskTable & task_table; ShardInfo info; - UInt32 numberInCluster() const { return info.shard_num; } + UInt32 numberInCluster() const; - UInt32 indexInCluster() const { return info.shard_num - 1; } + UInt32 indexInCluster() const; String getDescription() const; @@ -197,166 +181,6 @@ struct TaskShard }; -inline String TaskTable::getPartitionPath(const String & partition_name) const -{ - return task_cluster.task_zookeeper_path // root - + "/tables/" + table_id // tables/dst_cluster.merge.hits - + "/" + escapeForFileName(partition_name); // 201701 -} - -inline String TaskTable::getPartitionAttachIsActivePath(const String & partition_name) const -{ - return getPartitionPath(partition_name) + "/attach_active"; -} - -inline String TaskTable::getPartitionAttachIsDonePath(const String & partition_name) const -{ - return getPartitionPath(partition_name) + "/attach_is_done"; -} - -inline String TaskTable::getPartitionPiecePath(const String & partition_name, size_t piece_number) const -{ - assert(piece_number < number_of_splits); - return getPartitionPath(partition_name) + "/piece_" + toString(piece_number); // 1...number_of_splits -} - -inline String TaskTable::getCertainPartitionIsDirtyPath(const String &partition_name) const -{ - return getPartitionPath(partition_name) + "/is_dirty"; -} - -inline String TaskTable::getCertainPartitionPieceIsDirtyPath(const String & partition_name, const size_t piece_number) const -{ - return getPartitionPiecePath(partition_name, piece_number) + "/is_dirty"; -} - -inline String TaskTable::getCertainPartitionIsCleanedPath(const String & partition_name) const -{ - return getCertainPartitionIsDirtyPath(partition_name) + "/cleaned"; -} - -inline String TaskTable::getCertainPartitionPieceIsCleanedPath(const String & partition_name, const size_t piece_number) const -{ - return getCertainPartitionPieceIsDirtyPath(partition_name, piece_number) + "/cleaned"; -} - -inline String TaskTable::getCertainPartitionTaskStatusPath(const String & partition_name) const -{ - return getPartitionPath(partition_name) + "/shards"; -} - -inline String TaskTable::getCertainPartitionPieceTaskStatusPath(const String & partition_name, const size_t piece_number) const -{ - return getPartitionPiecePath(partition_name, piece_number) + "/shards"; -} - -inline String TaskTable::getStatusAllPartitionCount() const -{ - return task_cluster.task_zookeeper_path + "/status/all_partitions_count"; -} - -inline String TaskTable::getStatusProcessedPartitionsCount() const -{ - return task_cluster.task_zookeeper_path + "/status/processed_partitions_count"; -} - -inline TaskTable::TaskTable(TaskCluster & parent, const Poco::Util::AbstractConfiguration & config, - const String & prefix_, const String & table_key) - : task_cluster(parent) -{ - String table_prefix = prefix_ + "." + table_key + "."; - - name_in_config = table_key; - - number_of_splits = config.getUInt64(table_prefix + "number_of_splits", 3); - - allow_to_copy_alias_and_materialized_columns = config.getBool(table_prefix + "allow_to_copy_alias_and_materialized_columns", false); - allow_to_drop_target_partitions = config.getBool(table_prefix + "allow_to_drop_target_partitions", false); - - cluster_pull_name = config.getString(table_prefix + "cluster_pull"); - cluster_push_name = config.getString(table_prefix + "cluster_push"); - - table_pull.first = config.getString(table_prefix + "database_pull"); - table_pull.second = config.getString(table_prefix + "table_pull"); - - table_push.first = config.getString(table_prefix + "database_push"); - table_push.second = config.getString(table_prefix + "table_push"); - - /// Used as node name in ZooKeeper - table_id = escapeForFileName(cluster_push_name) - + "." + escapeForFileName(table_push.first) - + "." + escapeForFileName(table_push.second); - - engine_push_str = config.getString(table_prefix + "engine", "rand()"); - - { - ParserStorage parser_storage; - engine_push_ast = parseQuery(parser_storage, engine_push_str, 0, DBMS_DEFAULT_MAX_PARSER_DEPTH); - engine_push_partition_key_ast = extractPartitionKey(engine_push_ast); - primary_key_comma_separated = boost::algorithm::join(extractPrimaryKeyColumnNames(engine_push_ast), ", "); - is_replicated_table = isReplicatedTableEngine(engine_push_ast); - } - - sharding_key_str = config.getString(table_prefix + "sharding_key"); - - auxiliary_engine_split_asts.reserve(number_of_splits); - { - ParserExpressionWithOptionalAlias parser_expression(false); - sharding_key_ast = parseQuery(parser_expression, sharding_key_str, 0, DBMS_DEFAULT_MAX_PARSER_DEPTH); - main_engine_split_ast = createASTStorageDistributed(cluster_push_name, table_push.first, table_push.second, - sharding_key_ast); - - for (const auto piece_number : collections::range(0, number_of_splits)) - { - auxiliary_engine_split_asts.emplace_back - ( - createASTStorageDistributed(cluster_push_name, table_push.first, - table_push.second + "_piece_" + toString(piece_number), sharding_key_ast) - ); - } - } - - where_condition_str = config.getString(table_prefix + "where_condition", ""); - if (!where_condition_str.empty()) - { - ParserExpressionWithOptionalAlias parser_expression(false); - where_condition_ast = parseQuery(parser_expression, where_condition_str, 0, DBMS_DEFAULT_MAX_PARSER_DEPTH); - - // Will use canonical expression form - where_condition_str = queryToString(where_condition_ast); - } - - String enabled_partitions_prefix = table_prefix + "enabled_partitions"; - has_enabled_partitions = config.has(enabled_partitions_prefix); - - if (has_enabled_partitions) - { - Strings keys; - config.keys(enabled_partitions_prefix, keys); - - if (keys.empty()) - { - /// Parse list of partition from space-separated string - String partitions_str = config.getString(table_prefix + "enabled_partitions"); - boost::trim_if(partitions_str, isWhitespaceASCII); - boost::split(enabled_partitions, partitions_str, isWhitespaceASCII, boost::token_compress_on); - } - else - { - /// Parse sequence of ... - for (const String &key : keys) - { - if (!startsWith(key, "partition")) - throw Exception("Unknown key " + key + " in " + enabled_partitions_prefix, ErrorCodes::UNKNOWN_ELEMENT_IN_CONFIG); - - enabled_partitions.emplace_back(config.getString(enabled_partitions_prefix + "." + key)); - } - } - - std::copy(enabled_partitions.begin(), enabled_partitions.end(), std::inserter(enabled_partitions_set, enabled_partitions_set.begin())); - } -} - template inline void TaskTable::initShards(RandomEngine && random_engine) { @@ -390,45 +214,4 @@ inline void TaskTable::initShards(RandomEngine && random_engine) local_shards.assign(all_shards.begin(), it_first_remote); } -inline ASTPtr TaskTable::rewriteReplicatedCreateQueryToPlain() const -{ - ASTPtr prev_engine_push_ast = engine_push_ast->clone(); - - auto & new_storage_ast = prev_engine_push_ast->as(); - auto & new_engine_ast = new_storage_ast.engine->as(); - - /// Remove "Replicated" from name - new_engine_ast.name = new_engine_ast.name.substr(10); - - if (new_engine_ast.arguments) - { - auto & replicated_table_arguments = new_engine_ast.arguments->children; - - - /// In some cases of Atomic database engine usage ReplicatedMergeTree tables - /// could be created without arguments. - if (!replicated_table_arguments.empty()) - { - /// Delete first two arguments of Replicated...MergeTree() table. - replicated_table_arguments.erase(replicated_table_arguments.begin()); - replicated_table_arguments.erase(replicated_table_arguments.begin()); - } - } - - return new_storage_ast.clone(); -} - - -inline String DB::TaskShard::getDescription() const -{ - return fmt::format("N{} (having a replica {}, pull table {} of cluster {}", - numberInCluster(), getHostNameExample(), getQuotedTable(task_table.table_pull), task_table.cluster_pull_name); -} - -inline String DB::TaskShard::getHostNameExample() const -{ - const auto & replicas = task_table.cluster_pull->getShardsAddresses().at(indexInCluster()); - return replicas.at(0).readableString(); -} - } From 50c932e566cc9b268afe5709a50cb6650ead8d7a Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 13:56:40 +0000 Subject: [PATCH 050/239] Separate TaskShard and TaskTable --- programs/copier/CMakeLists.txt | 3 +- programs/copier/ClusterCopier.h | 3 +- programs/copier/ShardPartition.h | 2 +- programs/copier/TaskShard.cpp | 37 ++++++++++++ programs/copier/TaskShard.h | 58 +++++++++++++++++++ .../{TaskTableAndShard.cpp => TaskTable.cpp} | 29 +--------- .../{TaskTableAndShard.h => TaskTable.h} | 43 -------------- 7 files changed, 101 insertions(+), 74 deletions(-) create mode 100644 programs/copier/TaskShard.cpp create mode 100644 programs/copier/TaskShard.h rename programs/copier/{TaskTableAndShard.cpp => TaskTable.cpp} (90%) rename programs/copier/{TaskTableAndShard.h => TaskTable.h} (83%) diff --git a/programs/copier/CMakeLists.txt b/programs/copier/CMakeLists.txt index 69019843d05..fb4c848270f 100644 --- a/programs/copier/CMakeLists.txt +++ b/programs/copier/CMakeLists.txt @@ -5,7 +5,8 @@ set(CLICKHOUSE_COPIER_SOURCES "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartition.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartitionPiece.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/TaskCluster.cpp" - "${CMAKE_CURRENT_SOURCE_DIR}/TaskTableAndShard.cpp") + "${CMAKE_CURRENT_SOURCE_DIR}/TaskShard.cpp" + "${CMAKE_CURRENT_SOURCE_DIR}/TaskTable.cpp") set (CLICKHOUSE_COPIER_LINK PRIVATE diff --git a/programs/copier/ClusterCopier.h b/programs/copier/ClusterCopier.h index b354fc59eee..063b13e9078 100644 --- a/programs/copier/ClusterCopier.h +++ b/programs/copier/ClusterCopier.h @@ -3,7 +3,8 @@ #include "Aliases.h" #include "Internals.h" #include "TaskCluster.h" -#include "TaskTableAndShard.h" +#include "TaskShard.h" +#include "TaskTable.h" #include "ShardPartition.h" #include "ShardPartitionPiece.h" #include "ZooKeeperStaff.h" diff --git a/programs/copier/ShardPartition.h b/programs/copier/ShardPartition.h index a78790c0db1..3487c0e51ac 100644 --- a/programs/copier/ShardPartition.h +++ b/programs/copier/ShardPartition.h @@ -1,7 +1,7 @@ #pragma once #include "Aliases.h" -#include "TaskTableAndShard.h" +#include "TaskShard.h" namespace DB { diff --git a/programs/copier/TaskShard.cpp b/programs/copier/TaskShard.cpp new file mode 100644 index 00000000000..af21848b384 --- /dev/null +++ b/programs/copier/TaskShard.cpp @@ -0,0 +1,37 @@ +#include "TaskShard.h" + +#include "TaskTable.h" + +namespace DB +{ + +TaskShard::TaskShard(TaskTable & parent, const ShardInfo & info_) + : task_table(parent) + , info(info_) +{ + list_of_split_tables_on_shard.assign(task_table.number_of_splits, DatabaseAndTableName()); +} + +UInt32 TaskShard::numberInCluster() const +{ + return info.shard_num; +} + +UInt32 TaskShard::indexInCluster() const +{ + return info.shard_num - 1; +} + +String DB::TaskShard::getDescription() const +{ + return fmt::format("N{} (having a replica {}, pull table {} of cluster {}", + numberInCluster(), getHostNameExample(), getQuotedTable(task_table.table_pull), task_table.cluster_pull_name); +} + +String DB::TaskShard::getHostNameExample() const +{ + const auto & replicas = task_table.cluster_pull->getShardsAddresses().at(indexInCluster()); + return replicas.at(0).readableString(); +} + +} diff --git a/programs/copier/TaskShard.h b/programs/copier/TaskShard.h new file mode 100644 index 00000000000..970ca21612f --- /dev/null +++ b/programs/copier/TaskShard.h @@ -0,0 +1,58 @@ +#pragma once + +#include "Aliases.h" +#include "Internals.h" +#include "ClusterPartition.h" + +#include +#include + +#include +#include + + +namespace DB +{ + +struct TaskTable; + +struct TaskShard +{ + TaskShard(TaskTable & parent, const ShardInfo & info_); + + TaskTable & task_table; + + ShardInfo info; + + UInt32 numberInCluster() const; + + UInt32 indexInCluster() const; + + String getDescription() const; + + String getHostNameExample() const; + + /// Used to sort clusters by their proximity + ShardPriority priority; + + /// Column with unique destination partitions (computed from engine_push_partition_key expr.) in the shard + ColumnWithTypeAndName partition_key_column; + + /// There is a task for each destination partition + TasksPartition partition_tasks; + + /// Which partitions have been checked for existence + /// If some partition from this lists is exists, it is in partition_tasks + std::set checked_partitions; + + /// Last CREATE TABLE query of the table of the shard + ASTPtr current_pull_table_create_query; + ASTPtr current_push_table_create_query; + + /// Internal distributed tables + DatabaseAndTableName table_read_shard; + DatabaseAndTableName main_table_split_shard; + ListOfDatabasesAndTableNames list_of_split_tables_on_shard; +}; + +} diff --git a/programs/copier/TaskTableAndShard.cpp b/programs/copier/TaskTable.cpp similarity index 90% rename from programs/copier/TaskTableAndShard.cpp rename to programs/copier/TaskTable.cpp index 36aff4f4cd8..2f282842db6 100644 --- a/programs/copier/TaskTableAndShard.cpp +++ b/programs/copier/TaskTable.cpp @@ -1,4 +1,4 @@ -#include "TaskTableAndShard.h" +#include "TaskTable.h" namespace DB { @@ -211,31 +211,4 @@ ClusterPartition & TaskTable::getClusterPartition(const String & partition_name) return it->second; } -TaskShard::TaskShard(TaskTable & parent, const ShardInfo & info_) : task_table(parent), info(info_) -{ - list_of_split_tables_on_shard.assign(task_table.number_of_splits, DatabaseAndTableName()); -} - -UInt32 TaskShard::numberInCluster() const -{ - return info.shard_num; -} - -UInt32 TaskShard::indexInCluster() const -{ - return info.shard_num - 1; -} - -String DB::TaskShard::getDescription() const -{ - return fmt::format("N{} (having a replica {}, pull table {} of cluster {}", - numberInCluster(), getHostNameExample(), getQuotedTable(task_table.table_pull), task_table.cluster_pull_name); -} - -String DB::TaskShard::getHostNameExample() const -{ - const auto & replicas = task_table.cluster_pull->getShardsAddresses().at(indexInCluster()); - return replicas.at(0).readableString(); -} - } diff --git a/programs/copier/TaskTableAndShard.h b/programs/copier/TaskTable.h similarity index 83% rename from programs/copier/TaskTableAndShard.h rename to programs/copier/TaskTable.h index 1fe5b0a255e..32b8759cec1 100644 --- a/programs/copier/TaskTableAndShard.h +++ b/programs/copier/TaskTable.h @@ -14,8 +14,6 @@ namespace DB { -struct TaskShard; - struct TaskTable { TaskTable(TaskCluster & parent, const Poco::Util::AbstractConfiguration & config, const String & prefix, const String & table_key); @@ -140,47 +138,6 @@ struct TaskTable void initShards(RandomEngine &&random_engine); }; - -struct TaskShard -{ - TaskShard(TaskTable & parent, const ShardInfo & info_); - - TaskTable & task_table; - - ShardInfo info; - - UInt32 numberInCluster() const; - - UInt32 indexInCluster() const; - - String getDescription() const; - - String getHostNameExample() const; - - /// Used to sort clusters by their proximity - ShardPriority priority; - - /// Column with unique destination partitions (computed from engine_push_partition_key expr.) in the shard - ColumnWithTypeAndName partition_key_column; - - /// There is a task for each destination partition - TasksPartition partition_tasks; - - /// Which partitions have been checked for existence - /// If some partition from this lists is exists, it is in partition_tasks - std::set checked_partitions; - - /// Last CREATE TABLE query of the table of the shard - ASTPtr current_pull_table_create_query; - ASTPtr current_push_table_create_query; - - /// Internal distributed tables - DatabaseAndTableName table_read_shard; - DatabaseAndTableName main_table_split_shard; - ListOfDatabasesAndTableNames list_of_split_tables_on_shard; -}; - - template inline void TaskTable::initShards(RandomEngine && random_engine) { From 867bcdbb1c5f433db91ea5c7141b168993c75425 Mon Sep 17 00:00:00 2001 From: Kruglov Pavel <48961922+Avogar@users.noreply.github.com> Date: Thu, 20 Oct 2022 16:46:25 +0200 Subject: [PATCH 051/239] Fix typo in setting name that led to bad usage of schema inference cache --- src/Formats/EscapingRuleUtils.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Formats/EscapingRuleUtils.cpp b/src/Formats/EscapingRuleUtils.cpp index e47525d089a..e80ab50968d 100644 --- a/src/Formats/EscapingRuleUtils.cpp +++ b/src/Formats/EscapingRuleUtils.cpp @@ -859,7 +859,7 @@ String getAdditionalFormatInfoByEscapingRule(const FormatSettings & settings, Fo result += fmt::format( ", use_best_effort_in_schema_inference={}, bool_true_representation={}, bool_false_representation={}," " null_representation={}, delimiter={}, tuple_delimiter={}", - settings.tsv.use_best_effort_in_schema_inference, + settings.csv.use_best_effort_in_schema_inference, settings.bool_true_representation, settings.bool_false_representation, settings.csv.null_representation, From f435e4fbe2c270ae9cbf5b54ede38045d0398f37 Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 14:37:27 +0000 Subject: [PATCH 052/239] Un-clusterf$ck incomplete type errors --- programs/copier/Aliases.h | 23 ++------- programs/copier/CMakeLists.txt | 1 + programs/copier/ClusterPartition.h | 25 ++++++---- programs/copier/ShardPartition.cpp | 11 ++++- programs/copier/ShardPartition.h | 11 ++++- programs/copier/ShardPartitionPiece.cpp | 14 ++++-- programs/copier/ShardPartitionPiece.h | 8 +++- programs/copier/StatusAccumulator.cpp | 48 +++++++++++++++++++ programs/copier/StatusAccumulator.h | 62 +++++-------------------- programs/copier/TaskCluster.h | 8 +++- programs/copier/TaskShard.cpp | 2 +- programs/copier/TaskShard.h | 14 +++--- programs/copier/TaskTable.cpp | 7 +++ programs/copier/TaskTable.h | 15 +++--- 14 files changed, 144 insertions(+), 105 deletions(-) create mode 100644 programs/copier/StatusAccumulator.cpp diff --git a/programs/copier/Aliases.h b/programs/copier/Aliases.h index c4d9c40d9f1..02be3441acd 100644 --- a/programs/copier/Aliases.h +++ b/programs/copier/Aliases.h @@ -1,6 +1,10 @@ #pragma once -#include +#include + +#include + +#include namespace DB { @@ -8,21 +12,4 @@ namespace DB using DatabaseAndTableName = std::pair; using ListOfDatabasesAndTableNames = std::vector; - - /// Hierarchical description of the tasks - struct ShardPartitionPiece; - struct ShardPartition; - struct TaskShard; - struct TaskTable; - struct TaskCluster; - struct ClusterPartition; - - using PartitionPieces = std::vector; - using TasksPartition = std::map>; - using ShardInfo = Cluster::ShardInfo; - using TaskShardPtr = std::shared_ptr; - using TasksShard = std::vector; - using TasksTable = std::list; - using ClusterPartitions = std::map>; } - diff --git a/programs/copier/CMakeLists.txt b/programs/copier/CMakeLists.txt index fb4c848270f..2c17e70bc5e 100644 --- a/programs/copier/CMakeLists.txt +++ b/programs/copier/CMakeLists.txt @@ -4,6 +4,7 @@ set(CLICKHOUSE_COPIER_SOURCES "${CMAKE_CURRENT_SOURCE_DIR}/Internals.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartition.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/ShardPartitionPiece.cpp" + "${CMAKE_CURRENT_SOURCE_DIR}/StatusAccumulator.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/TaskCluster.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/TaskShard.cpp" "${CMAKE_CURRENT_SOURCE_DIR}/TaskTable.cpp") diff --git a/programs/copier/ClusterPartition.h b/programs/copier/ClusterPartition.h index ed69bfa8c26..22063989e22 100644 --- a/programs/copier/ClusterPartition.h +++ b/programs/copier/ClusterPartition.h @@ -1,17 +1,22 @@ #pragma once -#include "Aliases.h" +#include +#include namespace DB { - /// Contains info about all shards that contain a partition - struct ClusterPartition - { - double elapsed_time_seconds = 0; - UInt64 bytes_copied = 0; - UInt64 rows_copied = 0; - UInt64 blocks_copied = 0; - UInt64 total_tries = 0; - }; +/// Contains info about all shards that contain a partition +struct ClusterPartition +{ + double elapsed_time_seconds = 0; + UInt64 bytes_copied = 0; + UInt64 rows_copied = 0; + UInt64 blocks_copied = 0; + + UInt64 total_tries = 0; +}; + +using ClusterPartitions = std::map>; + } diff --git a/programs/copier/ShardPartition.cpp b/programs/copier/ShardPartition.cpp index fb482163377..4c962fc807d 100644 --- a/programs/copier/ShardPartition.cpp +++ b/programs/copier/ShardPartition.cpp @@ -1,10 +1,17 @@ #include "ShardPartition.h" +#include "TaskShard.h" +#include "TaskTable.h" + namespace DB { -ShardPartition::ShardPartition(TaskShard &parent, String name_quoted_, size_t number_of_splits) - : task_shard(parent), name(std::move(name_quoted_)) { pieces.reserve(number_of_splits); } +ShardPartition::ShardPartition(TaskShard & parent, String name_quoted_, size_t number_of_splits) + : task_shard(parent) + , name(std::move(name_quoted_)) +{ + pieces.reserve(number_of_splits); +} String ShardPartition::getPartitionCleanStartPath() const { diff --git a/programs/copier/ShardPartition.h b/programs/copier/ShardPartition.h index 3487c0e51ac..2457213733c 100644 --- a/programs/copier/ShardPartition.h +++ b/programs/copier/ShardPartition.h @@ -1,11 +1,16 @@ #pragma once -#include "Aliases.h" -#include "TaskShard.h" +#include "ShardPartitionPiece.h" + +#include + +#include namespace DB { +struct TaskShard; + /// Just destination partition of a shard /// I don't know what this comment means. /// In short, when we discovered what shards contain currently processing partition, @@ -44,4 +49,6 @@ struct ShardPartition String name; }; +using TasksPartition = std::map>; + } diff --git a/programs/copier/ShardPartitionPiece.cpp b/programs/copier/ShardPartitionPiece.cpp index c060a955c9a..36d1621e012 100644 --- a/programs/copier/ShardPartitionPiece.cpp +++ b/programs/copier/ShardPartitionPiece.cpp @@ -1,11 +1,19 @@ #include "ShardPartitionPiece.h" +#include "ShardPartition.h" +#include "TaskShard.h" + +#include + namespace DB { -ShardPartitionPiece::ShardPartitionPiece(ShardPartition &parent, size_t current_piece_number_, bool is_present_piece_) - : is_absent_piece(!is_present_piece_), current_piece_number(current_piece_number_), - shard_partition(parent) {} +ShardPartitionPiece::ShardPartitionPiece(ShardPartition & parent, size_t current_piece_number_, bool is_present_piece_) + : is_absent_piece(!is_present_piece_) + , current_piece_number(current_piece_number_) + , shard_partition(parent) +{ +} String ShardPartitionPiece::getPartitionPiecePath() const { diff --git a/programs/copier/ShardPartitionPiece.h b/programs/copier/ShardPartitionPiece.h index 243cb7439b8..aba378d466d 100644 --- a/programs/copier/ShardPartitionPiece.h +++ b/programs/copier/ShardPartitionPiece.h @@ -1,13 +1,15 @@ #pragma once -#include "Internals.h" +#include namespace DB { +struct ShardPartition; + struct ShardPartitionPiece { - ShardPartitionPiece(ShardPartition &parent, size_t current_piece_number_, bool is_present_piece_); + ShardPartitionPiece(ShardPartition & parent, size_t current_piece_number_, bool is_present_piece_); String getPartitionPiecePath() const; @@ -34,4 +36,6 @@ struct ShardPartitionPiece ShardPartition & shard_partition; }; +using PartitionPieces = std::vector; + } diff --git a/programs/copier/StatusAccumulator.cpp b/programs/copier/StatusAccumulator.cpp new file mode 100644 index 00000000000..77adeac708c --- /dev/null +++ b/programs/copier/StatusAccumulator.cpp @@ -0,0 +1,48 @@ +#include "StatusAccumulator.h" + +#include +#include +#include +#include + +#include + +namespace DB +{ + +StatusAccumulator::MapPtr StatusAccumulator::fromJSON(String state_json) +{ + Poco::JSON::Parser parser; + auto state = parser.parse(state_json).extract(); + MapPtr result_ptr = std::make_shared(); + for (const auto & table_name : state->getNames()) + { + auto table_status_json = state->getValue(table_name); + auto table_status = parser.parse(table_status_json).extract(); + /// Map entry will be created if it is absent + auto & map_table_status = (*result_ptr)[table_name]; + map_table_status.all_partitions_count += table_status->getValue("all_partitions_count"); + map_table_status.processed_partitions_count += table_status->getValue("processed_partitions_count"); + } + return result_ptr; +} + +String StatusAccumulator::serializeToJSON(MapPtr statuses) +{ + Poco::JSON::Object result_json; + for (const auto & [table_name, table_status] : *statuses) + { + Poco::JSON::Object status_json; + status_json.set("all_partitions_count", table_status.all_partitions_count); + status_json.set("processed_partitions_count", table_status.processed_partitions_count); + + result_json.set(table_name, status_json); + } + std::ostringstream oss; // STYLE_CHECK_ALLOW_STD_STRING_STREAM + oss.exceptions(std::ios::failbit); + Poco::JSON::Stringifier::stringify(result_json, oss); + auto result = oss.str(); + return result; +} + +} diff --git a/programs/copier/StatusAccumulator.h b/programs/copier/StatusAccumulator.h index 6e20e3dc95d..d420b611602 100644 --- a/programs/copier/StatusAccumulator.h +++ b/programs/copier/StatusAccumulator.h @@ -1,65 +1,27 @@ #pragma once +#include -#include -#include -#include -#include - -#include #include -#include -#include +#include namespace DB { class StatusAccumulator { - public: - struct TableStatus - { - size_t all_partitions_count; - size_t processed_partitions_count; - }; +public: + struct TableStatus + { + size_t all_partitions_count; + size_t processed_partitions_count; + }; - using Map = std::unordered_map; - using MapPtr = std::shared_ptr; + using Map = std::unordered_map; + using MapPtr = std::shared_ptr; - static MapPtr fromJSON(std::string state_json) - { - Poco::JSON::Parser parser; - auto state = parser.parse(state_json).extract(); - MapPtr result_ptr = std::make_shared(); - for (const auto & table_name : state->getNames()) - { - auto table_status_json = state->getValue(table_name); - auto table_status = parser.parse(table_status_json).extract(); - /// Map entry will be created if it is absent - auto & map_table_status = (*result_ptr)[table_name]; - map_table_status.all_partitions_count += table_status->getValue("all_partitions_count"); - map_table_status.processed_partitions_count += table_status->getValue("processed_partitions_count"); - } - return result_ptr; - } - - static std::string serializeToJSON(MapPtr statuses) - { - Poco::JSON::Object result_json; - for (const auto & [table_name, table_status] : *statuses) - { - Poco::JSON::Object status_json; - status_json.set("all_partitions_count", table_status.all_partitions_count); - status_json.set("processed_partitions_count", table_status.processed_partitions_count); - - result_json.set(table_name, status_json); - } - std::ostringstream oss; // STYLE_CHECK_ALLOW_STD_STRING_STREAM - oss.exceptions(std::ios::failbit); - Poco::JSON::Stringifier::stringify(result_json, oss); - auto result = oss.str(); - return result; - } + static MapPtr fromJSON(String state_json); + static String serializeToJSON(MapPtr statuses); }; } diff --git a/programs/copier/TaskCluster.h b/programs/copier/TaskCluster.h index c59c7bff4d7..fc1c8a663ec 100644 --- a/programs/copier/TaskCluster.h +++ b/programs/copier/TaskCluster.h @@ -1,8 +1,14 @@ #pragma once -#include "Aliases.h" +#include "TaskTable.h" + +#include +#include + #include +#include + namespace DB { diff --git a/programs/copier/TaskShard.cpp b/programs/copier/TaskShard.cpp index af21848b384..d156f451a84 100644 --- a/programs/copier/TaskShard.cpp +++ b/programs/copier/TaskShard.cpp @@ -5,7 +5,7 @@ namespace DB { -TaskShard::TaskShard(TaskTable & parent, const ShardInfo & info_) +TaskShard::TaskShard(TaskTable & parent, const Cluster::ShardInfo & info_) : task_table(parent) , info(info_) { diff --git a/programs/copier/TaskShard.h b/programs/copier/TaskShard.h index 970ca21612f..05d652077ea 100644 --- a/programs/copier/TaskShard.h +++ b/programs/copier/TaskShard.h @@ -3,12 +3,7 @@ #include "Aliases.h" #include "Internals.h" #include "ClusterPartition.h" - -#include -#include - -#include -#include +#include "ShardPartition.h" namespace DB @@ -18,11 +13,11 @@ struct TaskTable; struct TaskShard { - TaskShard(TaskTable & parent, const ShardInfo & info_); + TaskShard(TaskTable & parent, const Cluster::ShardInfo & info_); TaskTable & task_table; - ShardInfo info; + Cluster::ShardInfo info; UInt32 numberInCluster() const; @@ -55,4 +50,7 @@ struct TaskShard ListOfDatabasesAndTableNames list_of_split_tables_on_shard; }; +using TaskShardPtr = std::shared_ptr; +using TasksShard = std::vector; + } diff --git a/programs/copier/TaskTable.cpp b/programs/copier/TaskTable.cpp index 2f282842db6..5b09a9c99a7 100644 --- a/programs/copier/TaskTable.cpp +++ b/programs/copier/TaskTable.cpp @@ -1,5 +1,12 @@ #include "TaskTable.h" +#include "ClusterPartition.h" +#include "TaskCluster.h" + +#include + +#include + namespace DB { namespace ErrorCodes diff --git a/programs/copier/TaskTable.h b/programs/copier/TaskTable.h index 32b8759cec1..2bb7f078bc6 100644 --- a/programs/copier/TaskTable.h +++ b/programs/copier/TaskTable.h @@ -1,19 +1,15 @@ #pragma once #include "Aliases.h" -#include "Internals.h" -#include "ClusterPartition.h" - -#include -#include - -#include -#include +#include "TaskShard.h" namespace DB { +struct ClusterPartition; +struct TaskCluster; + struct TaskTable { TaskTable(TaskCluster & parent, const Poco::Util::AbstractConfiguration & config, const String & prefix, const String & table_key); @@ -138,6 +134,9 @@ struct TaskTable void initShards(RandomEngine &&random_engine); }; +using TasksTable = std::list; + + template inline void TaskTable::initShards(RandomEngine && random_engine) { From eda3955f010ddb49136633043cf0e9adbfe297be Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Thu, 20 Oct 2022 14:50:51 +0000 Subject: [PATCH 053/239] Update libcxx submodule --- contrib/libcxx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/libcxx b/contrib/libcxx index 19330eed499..4db7f838afd 160000 --- a/contrib/libcxx +++ b/contrib/libcxx @@ -1 +1 @@ -Subproject commit 19330eed499f0f2011437a92b7e8567ea36fe082 +Subproject commit 4db7f838afd3139eb3761694b04d31275df45d2d From 2fe59e7e8f03db75c42c40f7e266581d307bc97a Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Thu, 20 Oct 2022 13:30:20 -0300 Subject: [PATCH 054/239] lower thread and iterations number --- .../scripts/stress_test.py | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py index 81dfb4cc511..b8bafb3d0c1 100644 --- a/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py +++ b/tests/integration/test_host_regexp_multiple_ptr_records_concurrent/scripts/stress_test.py @@ -8,8 +8,8 @@ server_ip = sys.argv[2] mutex = threading.Lock() success_counter = 0 -number_of_threads = 200 -number_of_iterations = 200 +number_of_threads = 100 +number_of_iterations = 100 def perform_request(): From 1a462fddea3c518e51c68f74a2f7c6193644042a Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Fri, 21 Oct 2022 13:28:24 +0800 Subject: [PATCH 055/239] Fix test --- src/Interpreters/convertFieldToType.cpp | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/Interpreters/convertFieldToType.cpp b/src/Interpreters/convertFieldToType.cpp index 4e7562ef451..fdbae838ab4 100644 --- a/src/Interpreters/convertFieldToType.cpp +++ b/src/Interpreters/convertFieldToType.cpp @@ -218,10 +218,11 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID } if (which_type.isDateTime64() - && (which_from_type.isNativeInt() || which_from_type.isNativeUInt() || which_from_type.isDate() || which_from_type.isDate32() || which_from_type.isDateTime() || which_from_type.isDateTime64())) + && (src.getType() == Field::Types::UInt64 || src.getType() == Field::Types::Int64 || src.getType() == Field::Types::Decimal64)) { const auto scale = static_cast(type).getScale(); - const auto decimal_value = DecimalUtils::decimalFromComponents(applyVisitor(FieldVisitorConvertToNumber(), src), 0, scale); + const auto decimal_value + = DecimalUtils::decimalFromComponents(applyVisitor(FieldVisitorConvertToNumber(), src), 0, scale); return Field(DecimalField(decimal_value, scale)); } } From 2c41c0357d4b25e81479af487dd66e71b34e99d4 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Fri, 21 Oct 2022 14:20:49 -0300 Subject: [PATCH 056/239] Make CaresPTRResolver completely thread-safe --- src/Common/CaresPTRResolver.h | 3 +++ src/Common/DNSPTRResolverProvider.cpp | 5 ++++- src/Common/LockedDNSPTRResolver.cpp | 25 +++++++++++++++++++++++++ src/Common/LockedDNSPTRResolver.h | 25 +++++++++++++++++++++++++ 4 files changed, 57 insertions(+), 1 deletion(-) create mode 100644 src/Common/LockedDNSPTRResolver.cpp create mode 100644 src/Common/LockedDNSPTRResolver.h diff --git a/src/Common/CaresPTRResolver.h b/src/Common/CaresPTRResolver.h index 38344d75ade..9df6d7aeb72 100644 --- a/src/Common/CaresPTRResolver.h +++ b/src/Common/CaresPTRResolver.h @@ -2,6 +2,7 @@ #include #include +#include #include "DNSPTRResolver.h" using ares_channel = struct ares_channeldata *; @@ -46,6 +47,8 @@ namespace DB void process_readable_sockets(std::span readable_sockets); ares_channel channel; + + static std::mutex mutex; }; } diff --git a/src/Common/DNSPTRResolverProvider.cpp b/src/Common/DNSPTRResolverProvider.cpp index 97d601a3a78..63d23612fba 100644 --- a/src/Common/DNSPTRResolverProvider.cpp +++ b/src/Common/DNSPTRResolverProvider.cpp @@ -1,12 +1,15 @@ #include "DNSPTRResolverProvider.h" +#include "LockedDNSPTRResolver.h" #include "CaresPTRResolver.h" namespace DB { std::shared_ptr DNSPTRResolverProvider::get() { - return std::make_shared( + static auto resolver = std::make_shared( CaresPTRResolver::provider_token {} ); + + return resolver; } } diff --git a/src/Common/LockedDNSPTRResolver.cpp b/src/Common/LockedDNSPTRResolver.cpp new file mode 100644 index 00000000000..947a109860d --- /dev/null +++ b/src/Common/LockedDNSPTRResolver.cpp @@ -0,0 +1,25 @@ +#include "LockedDNSPTRResolver.h" + +namespace DB { + + std::mutex LockedPTRResolver::mutex; + + LockedPTRResolver::LockedPTRResolver(std::unique_ptr resolver_) + : resolver(std::move(resolver_)) + {} + + std::unordered_set LockedPTRResolver::resolve(const std::string & ip) + { + std::lock_guard guard(mutex); + + return resolver->resolve(ip); + } + + std::unordered_set LockedPTRResolver::resolve_v6(const std::string & ip) + { + std::lock_guard guard(mutex); + + return resolver->resolve_v6(ip); + } + +} diff --git a/src/Common/LockedDNSPTRResolver.h b/src/Common/LockedDNSPTRResolver.h new file mode 100644 index 00000000000..82fdd189744 --- /dev/null +++ b/src/Common/LockedDNSPTRResolver.h @@ -0,0 +1,25 @@ +#pragma once + +#include + +#include + +namespace DB { + class LockedPTRResolver : public DNSPTRResolver + { + public: + + LockedPTRResolver(std::unique_ptr resolver); + + std::unordered_set resolve(const std::string & ip) override; + + std::unordered_set resolve_v6(const std::string & ip) override; + + private: + // this needs to be owned + std::unique_ptr resolver; + + static std::mutex mutex; + }; +} + From e2fcd51c935ce5eb46b9af7e6a4c4ef49d93a106 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Fri, 21 Oct 2022 14:43:08 -0300 Subject: [PATCH 057/239] what --- src/Common/CaresPTRResolver.cpp | 6 ++++++ src/Common/DNSPTRResolverProvider.cpp | 1 - 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index f04c6fb6de7..c6fe70a09fa 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -41,6 +41,8 @@ namespace DB } } + std::mutex CaresPTRResolver::mutex; + CaresPTRResolver::CaresPTRResolver(CaresPTRResolver::provider_token) : channel(nullptr) { /* @@ -74,6 +76,8 @@ namespace DB std::unordered_set CaresPTRResolver::resolve(const std::string & ip) { + std::lock_guard guard(mutex); + std::unordered_set ptr_records; resolve(ip, ptr_records); @@ -84,6 +88,8 @@ namespace DB std::unordered_set CaresPTRResolver::resolve_v6(const std::string & ip) { + std::lock_guard guard(mutex); + std::unordered_set ptr_records; resolve_v6(ip, ptr_records); diff --git a/src/Common/DNSPTRResolverProvider.cpp b/src/Common/DNSPTRResolverProvider.cpp index 63d23612fba..91ce4dbb938 100644 --- a/src/Common/DNSPTRResolverProvider.cpp +++ b/src/Common/DNSPTRResolverProvider.cpp @@ -1,5 +1,4 @@ #include "DNSPTRResolverProvider.h" -#include "LockedDNSPTRResolver.h" #include "CaresPTRResolver.h" namespace DB From 615b73773ae321a0c024dd5bcc05fcf96d55a4f3 Mon Sep 17 00:00:00 2001 From: Arthur Passos Date: Fri, 21 Oct 2022 14:46:10 -0300 Subject: [PATCH 058/239] removing unwanted files --- src/Common/LockedDNSPTRResolver.cpp | 25 ------------------------- src/Common/LockedDNSPTRResolver.h | 25 ------------------------- 2 files changed, 50 deletions(-) delete mode 100644 src/Common/LockedDNSPTRResolver.cpp delete mode 100644 src/Common/LockedDNSPTRResolver.h diff --git a/src/Common/LockedDNSPTRResolver.cpp b/src/Common/LockedDNSPTRResolver.cpp deleted file mode 100644 index 947a109860d..00000000000 --- a/src/Common/LockedDNSPTRResolver.cpp +++ /dev/null @@ -1,25 +0,0 @@ -#include "LockedDNSPTRResolver.h" - -namespace DB { - - std::mutex LockedPTRResolver::mutex; - - LockedPTRResolver::LockedPTRResolver(std::unique_ptr resolver_) - : resolver(std::move(resolver_)) - {} - - std::unordered_set LockedPTRResolver::resolve(const std::string & ip) - { - std::lock_guard guard(mutex); - - return resolver->resolve(ip); - } - - std::unordered_set LockedPTRResolver::resolve_v6(const std::string & ip) - { - std::lock_guard guard(mutex); - - return resolver->resolve_v6(ip); - } - -} diff --git a/src/Common/LockedDNSPTRResolver.h b/src/Common/LockedDNSPTRResolver.h deleted file mode 100644 index 82fdd189744..00000000000 --- a/src/Common/LockedDNSPTRResolver.h +++ /dev/null @@ -1,25 +0,0 @@ -#pragma once - -#include - -#include - -namespace DB { - class LockedPTRResolver : public DNSPTRResolver - { - public: - - LockedPTRResolver(std::unique_ptr resolver); - - std::unordered_set resolve(const std::string & ip) override; - - std::unordered_set resolve_v6(const std::string & ip) override; - - private: - // this needs to be owned - std::unique_ptr resolver; - - static std::mutex mutex; - }; -} - From f104dd08cdf820382a375a32fb5b53ab2507fd63 Mon Sep 17 00:00:00 2001 From: Amos Bird Date: Sat, 22 Oct 2022 15:47:13 +0800 Subject: [PATCH 059/239] Only fix LowCardinality for now --- src/Interpreters/castColumn.cpp | 10 ---------- src/Interpreters/castColumn.h | 1 - src/Storages/MergeTree/KeyCondition.cpp | 8 ++++++-- 3 files changed, 6 insertions(+), 13 deletions(-) diff --git a/src/Interpreters/castColumn.cpp b/src/Interpreters/castColumn.cpp index 744cfff2527..dc9882b84b0 100644 --- a/src/Interpreters/castColumn.cpp +++ b/src/Interpreters/castColumn.cpp @@ -51,14 +51,4 @@ ColumnPtr castColumnAccurateOrNull(const ColumnWithTypeAndName & arg, const Data return castColumn(arg, type); } -ColumnPtr tryCastColumnAccurate(const ColumnWithTypeAndName & arg, const DataTypePtr & type) -try -{ - return castColumn(arg, type); -} -catch (...) -{ - return nullptr; -} - } diff --git a/src/Interpreters/castColumn.h b/src/Interpreters/castColumn.h index c0a2cfbefbc..fcbea0f4646 100644 --- a/src/Interpreters/castColumn.h +++ b/src/Interpreters/castColumn.h @@ -8,6 +8,5 @@ namespace DB ColumnPtr castColumn(const ColumnWithTypeAndName & arg, const DataTypePtr & type); ColumnPtr castColumnAccurate(const ColumnWithTypeAndName & arg, const DataTypePtr & type); ColumnPtr castColumnAccurateOrNull(const ColumnWithTypeAndName & arg, const DataTypePtr & type); -ColumnPtr tryCastColumnAccurate(const ColumnWithTypeAndName & arg, const DataTypePtr & type); } diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp index 194cfbdabfc..cc5f6998955 100644 --- a/src/Storages/MergeTree/KeyCondition.cpp +++ b/src/Storages/MergeTree/KeyCondition.cpp @@ -6,6 +6,7 @@ #include #include #include +#include #include #include #include @@ -1179,8 +1180,11 @@ bool KeyCondition::transformConstantWithValidFunctions( if (is_valid_chain) { - auto const_type = cur_node->result_type; - auto const_value = convertFieldToType(out_value, *const_type); + out_type = removeLowCardinality(out_type); + auto const_type = removeLowCardinality(cur_node->result_type); + auto const_column = out_type->createColumnConst(1, out_value); + auto const_value = (*castColumnAccurateOrNull({const_column, out_type, ""}, const_type))[0]; + if (const_value.isNull()) return false; From 715b923bf9bbace82be7414b59c52afc9e5b7dc0 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Sat, 22 Oct 2022 13:09:47 +0000 Subject: [PATCH 060/239] Add fuzzed test. --- .../0_stateless/01902_table_function_merge_db_repr.sql | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql b/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql index 3801a83e247..460ce16ccad 100644 --- a/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql +++ b/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql @@ -62,6 +62,11 @@ SHOW CREATE TABLE 01902_db.t_merge_1; SELECT 'SELECT _database, _table, n FROM merge(currentDatabase(), ^t) ORDER BY _database, _table, n'; SELECT _database, _table, n FROM merge(currentDatabase(), '^t') ORDER BY _database, _table, n; +--fuzzed LOGICAL_ERROR +CREATE TABLE 01902_db.t4 (n Date) ENGINE=MergeTree ORDER BY n; +INSERT INTO 01902_db.t4 SELECT * FROM numbers(10); +SELECT NULL FROM 01902_db.t_merge WHERE n ORDER BY _table DESC; + DROP DATABASE 01902_db; DROP DATABASE 01902_db1; DROP DATABASE 01902_db2; From cd7a48215ccbff4bbce45a1e43f9c69e59e88c93 Mon Sep 17 00:00:00 2001 From: flynn Date: Sat, 22 Oct 2022 16:17:05 +0000 Subject: [PATCH 061/239] Fix create set with wrong header when data type is lowcardinality --- src/Interpreters/Set.cpp | 1 + ...467_set_with_lowcardinality_type.reference | 2 ++ .../02467_set_with_lowcardinality_type.sql | 30 +++++++++++++++++++ 3 files changed, 33 insertions(+) create mode 100644 tests/queries/0_stateless/02467_set_with_lowcardinality_type.reference create mode 100644 tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql diff --git a/src/Interpreters/Set.cpp b/src/Interpreters/Set.cpp index ded8b04a589..e75232aa0f5 100644 --- a/src/Interpreters/Set.cpp +++ b/src/Interpreters/Set.cpp @@ -131,6 +131,7 @@ void Set::setHeader(const ColumnsWithTypeAndName & header) if (const auto * low_cardinality_type = typeid_cast(data_types.back().get())) { data_types.back() = low_cardinality_type->getDictionaryType(); + set_elements_types.back() = low_cardinality_type->getDictionaryType(); materialized_columns.emplace_back(key_columns.back()->convertToFullColumnIfLowCardinality()); key_columns.back() = materialized_columns.back().get(); } diff --git a/tests/queries/0_stateless/02467_set_with_lowcardinality_type.reference b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.reference new file mode 100644 index 00000000000..b3f28057554 --- /dev/null +++ b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.reference @@ -0,0 +1,2 @@ +1 test +1 test diff --git a/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql new file mode 100644 index 00000000000..7b572df73f5 --- /dev/null +++ b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql @@ -0,0 +1,30 @@ +DROP TABLE IF EXISTS bloom_filter_nullable_index__fuzz_0; +CREATE TABLE bloom_filter_nullable_index__fuzz_0 +( + `order_key` UInt64, + `str` Nullable(String), + INDEX idx str TYPE bloom_filter GRANULARITY 1 +) +ENGINE = MergeTree ORDER BY order_key SETTINGS index_granularity = 6; + +INSERT INTO bloom_filter_nullable_index__fuzz_0 VALUES (1, 'test'); +INSERT INTO bloom_filter_nullable_index__fuzz_0 VALUES (2, 'test2'); + +DROP TABLE IF EXISTS bloom_filter_nullable_index__fuzz_1; +CREATE TABLE bloom_filter_nullable_index__fuzz_1 +( + `order_key` UInt64, + `str` Nullable(String), + INDEX idx str TYPE bloom_filter GRANULARITY 1 +) +ENGINE = MergeTree ORDER BY order_key SETTINGS index_granularity = 6; + +INSERT INTO bloom_filter_nullable_index__fuzz_0 VALUES (1, 'test'); +INSERT INTO bloom_filter_nullable_index__fuzz_0 VALUES (2, 'test2'); + +DROP TABLE IF EXISTS nullable_string_value__fuzz_2; +CREATE TABLE nullable_string_value__fuzz_2 (`value` LowCardinality(String)) ENGINE = TinyLog; +INSERT INTO nullable_string_value__fuzz_2 VALUES ('test'); + +SELECT * FROM bloom_filter_nullable_index__fuzz_0 WHERE str IN (SELECT value FROM nullable_string_value__fuzz_2); +SELECT * FROM bloom_filter_nullable_index__fuzz_1 WHERE str IN (SELECT value FROM nullable_string_value__fuzz_2); From 1aa8a986dd8f7a7b2a17f5466cc94fac1637e9db Mon Sep 17 00:00:00 2001 From: flynn Date: Sat, 22 Oct 2022 16:19:49 +0000 Subject: [PATCH 062/239] update test --- tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql index 7b572df73f5..92519a48bea 100644 --- a/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql +++ b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql @@ -1,3 +1,4 @@ +-- https://github.com/ClickHouse/ClickHouse/issues/42460 DROP TABLE IF EXISTS bloom_filter_nullable_index__fuzz_0; CREATE TABLE bloom_filter_nullable_index__fuzz_0 ( From 5ac7538e92e32b63a635587aa7ee34f9cc5493fe Mon Sep 17 00:00:00 2001 From: flynn Date: Sat, 22 Oct 2022 16:21:28 +0000 Subject: [PATCH 063/239] update test --- .../queries/0_stateless/02467_set_with_lowcardinality_type.sql | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql index 92519a48bea..dee6f7de74a 100644 --- a/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql +++ b/tests/queries/0_stateless/02467_set_with_lowcardinality_type.sql @@ -15,7 +15,7 @@ DROP TABLE IF EXISTS bloom_filter_nullable_index__fuzz_1; CREATE TABLE bloom_filter_nullable_index__fuzz_1 ( `order_key` UInt64, - `str` Nullable(String), + `str` String, INDEX idx str TYPE bloom_filter GRANULARITY 1 ) ENGINE = MergeTree ORDER BY order_key SETTINGS index_granularity = 6; From 2fc91fd33802ba69193e3d8cb1caa154dcefcd24 Mon Sep 17 00:00:00 2001 From: Han Fei Date: Sat, 22 Oct 2022 18:27:24 +0200 Subject: [PATCH 064/239] fix behaviour of max_rows_to_read for trival limit queries --- src/Interpreters/InterpreterSelectQuery.cpp | 17 +++++++++----- .../QueryPlan/ReadFromMergeTree.cpp | 19 +++++++++++++++- .../MergeTree/MergeTreeDataSelectExecutor.cpp | 4 ++++ .../MergeTree/MergeTreeSelectProcessor.cpp | 8 ------- src/Storages/SelectQueryInfo.h | 3 +++ src/Storages/System/StorageSystemNumbers.cpp | 17 +++++++++++--- ...5_limit_trivial_max_rows_to_read.reference | 7 ++++++ .../02465_limit_trivial_max_rows_to_read.sql | 22 +++++++++++++++++++ 8 files changed, 80 insertions(+), 17 deletions(-) create mode 100644 tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.reference create mode 100644 tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.sql diff --git a/src/Interpreters/InterpreterSelectQuery.cpp b/src/Interpreters/InterpreterSelectQuery.cpp index 79deb38317c..adf341f5ffd 100644 --- a/src/Interpreters/InterpreterSelectQuery.cpp +++ b/src/Interpreters/InterpreterSelectQuery.cpp @@ -2143,6 +2143,8 @@ void InterpreterSelectQuery::executeFetchColumns(QueryProcessingStage::Enum proc auto [limit_length, limit_offset] = getLimitLengthAndOffset(query, context); + auto local_limits = getStorageLimits(*context, options); + /** Optimization - if not specified DISTINCT, WHERE, GROUP, HAVING, ORDER, JOIN, LIMIT BY, WITH TIES * but LIMIT is specified, and limit + offset < max_block_size, * then as the block size we will use limit + offset (not to read more from the table than requested), @@ -2161,17 +2163,22 @@ void InterpreterSelectQuery::executeFetchColumns(QueryProcessingStage::Enum proc && !query_analyzer->hasAggregation() && !query_analyzer->hasWindow() && query.limitLength() - && limit_length <= std::numeric_limits::max() - limit_offset - && limit_length + limit_offset < max_block_size) + && limit_length <= std::numeric_limits::max() - limit_offset) { - max_block_size = std::max(1, limit_length + limit_offset); - max_threads_execute_query = max_streams = 1; + if (limit_length + limit_offset < max_block_size) + { + max_block_size = std::max(1, limit_length + limit_offset); + max_threads_execute_query = max_streams = 1; + } + if (limit_length + limit_offset < local_limits.local_limits.size_limits.max_rows) + { + query_info.limit = limit_length + limit_offset; + } } if (!max_block_size) throw Exception("Setting 'max_block_size' cannot be zero", ErrorCodes::PARAMETER_OUT_OF_BOUND); - auto local_limits = getStorageLimits(*context, options); storage_limits.emplace_back(local_limits); /// Initialize the initial data streams to which the query transforms are superimposed. Table or subquery or prepared input? diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.cpp b/src/Processors/QueryPlan/ReadFromMergeTree.cpp index b340073e73d..164ec8777de 100644 --- a/src/Processors/QueryPlan/ReadFromMergeTree.cpp +++ b/src/Processors/QueryPlan/ReadFromMergeTree.cpp @@ -173,6 +173,9 @@ Pipe ReadFromMergeTree::readFromPool( total_rows += part.getRowsCount(); } + if (query_info.limit > 0 && query_info.limit < total_rows) + total_rows = query_info.limit; + const auto & settings = context->getSettingsRef(); const auto & client_info = context->getClientInfo(); MergeTreeReadPool::BackoffSettings backoff_settings(settings); @@ -246,10 +249,24 @@ ProcessorPtr ReadFromMergeTree::createSource( }; } - return std::make_shared( + auto total_rows = part.getRowsCount(); + if (query_info.limit > 0 && query_info.limit < total_rows) + total_rows = query_info.limit; + + auto source = std::make_shared( data, storage_snapshot, part.data_part, max_block_size, preferred_block_size_bytes, preferred_max_column_in_block_size_bytes, required_columns, part.ranges, use_uncompressed_cache, prewhere_info, actions_settings, reader_settings, virt_column_names, part.part_index_in_query, has_limit_below_one_block, std::move(extension)); + + /// Actually it means that parallel reading from replicas enabled + /// and we have to collaborate with initiator. + /// In this case we won't set approximate rows, because it will be accounted multiple times. + /// Also do not count amount of read rows if we read in order of sorting key, + /// because we don't know actual amount of read rows in case when limit is set. + if (!extension.has_value() && !reader_settings.read_in_order) + source -> addTotalRowsApprox(total_rows); + + return source; } Pipe ReadFromMergeTree::readInOrder( diff --git a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp index 91ecb3a37a0..0eddaac2fac 100644 --- a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp +++ b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp @@ -1061,6 +1061,10 @@ RangesInDataParts MergeTreeDataSelectExecutor::filterPartsByPrimaryKeyAndSkipInd auto current_rows_estimate = ranges.getRowsCount(); size_t prev_total_rows_estimate = total_rows.fetch_add(current_rows_estimate); size_t total_rows_estimate = current_rows_estimate + prev_total_rows_estimate; + if (query_info.limit > 0 && total_rows_estimate > query_info.limit) + { + total_rows_estimate = query_info.limit; + } limits.check(total_rows_estimate, 0, "rows (controlled by 'max_rows_to_read' setting)", ErrorCodes::TOO_MANY_ROWS); leaf_limits.check( total_rows_estimate, 0, "rows (controlled by 'max_rows_to_read_leaf' setting)", ErrorCodes::TOO_MANY_ROWS); diff --git a/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp b/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp index 59cbae3f914..2490eb77772 100644 --- a/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp +++ b/src/Storages/MergeTree/MergeTreeSelectProcessor.cpp @@ -38,14 +38,6 @@ MergeTreeSelectProcessor::MergeTreeSelectProcessor( has_limit_below_one_block(has_limit_below_one_block_), total_rows(data_part->index_granularity.getRowsCountInRanges(all_mark_ranges)) { - /// Actually it means that parallel reading from replicas enabled - /// and we have to collaborate with initiator. - /// In this case we won't set approximate rows, because it will be accounted multiple times. - /// Also do not count amount of read rows if we read in order of sorting key, - /// because we don't know actual amount of read rows in case when limit is set. - if (!extension_.has_value() && !reader_settings.read_in_order) - addTotalRowsApprox(total_rows); - ordered_names = header_without_virtual_columns.getNames(); } diff --git a/src/Storages/SelectQueryInfo.h b/src/Storages/SelectQueryInfo.h index f2835ab4dbf..94a8c1143f3 100644 --- a/src/Storages/SelectQueryInfo.h +++ b/src/Storages/SelectQueryInfo.h @@ -220,6 +220,9 @@ struct SelectQueryInfo Block minmax_count_projection_block; MergeTreeDataSelectAnalysisResultPtr merge_tree_select_result_ptr; + // If limit is not 0, that means it's a trival limit query. + UInt64 limit = 0; + InputOrderInfoPtr getInputOrderInfo() const { return input_order_info ? input_order_info : (projection ? projection->input_order_info : nullptr); diff --git a/src/Storages/System/StorageSystemNumbers.cpp b/src/Storages/System/StorageSystemNumbers.cpp index 523ec25b89c..fbcd449051f 100644 --- a/src/Storages/System/StorageSystemNumbers.cpp +++ b/src/Storages/System/StorageSystemNumbers.cpp @@ -2,6 +2,7 @@ #include #include #include +#include #include #include @@ -125,7 +126,7 @@ StorageSystemNumbers::StorageSystemNumbers(const StorageID & table_id, bool mult Pipe StorageSystemNumbers::read( const Names & column_names, const StorageSnapshotPtr & storage_snapshot, - SelectQueryInfo &, + SelectQueryInfo & query_info, ContextPtr /*context*/, QueryProcessingStage::Enum /*processed_stage*/, size_t max_block_size, @@ -154,7 +155,12 @@ Pipe StorageSystemNumbers::read( auto source = std::make_shared(state, max_block_size, max_counter); if (i == 0) - source->addTotalRowsApprox(*limit); + { + auto rows_appr = *limit; + if (query_info.limit > 0 && query_info.limit < rows_appr) + rows_appr = query_info.limit; + source->addTotalRowsApprox(rows_appr); + } pipe.addSource(std::move(source)); } @@ -167,7 +173,12 @@ Pipe StorageSystemNumbers::read( auto source = std::make_shared(max_block_size, offset + i * max_block_size, num_streams * max_block_size); if (limit && i == 0) - source->addTotalRowsApprox(*limit); + { + auto rows_appr = *limit; + if (query_info.limit > 0 && query_info.limit < rows_appr) + rows_appr = query_info.limit; + source->addTotalRowsApprox(rows_appr); + } pipe.addSource(std::move(source)); } diff --git a/tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.reference b/tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.reference new file mode 100644 index 00000000000..87370760038 --- /dev/null +++ b/tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.reference @@ -0,0 +1,7 @@ +0 +0 +1 +2 +3 +4 +0 diff --git a/tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.sql b/tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.sql new file mode 100644 index 00000000000..ee7a4e6b6b5 --- /dev/null +++ b/tests/queries/0_stateless/02465_limit_trivial_max_rows_to_read.sql @@ -0,0 +1,22 @@ +DROP TABLE IF EXISTS t_max_rows_to_read; + +CREATE TABLE t_max_rows_to_read (a UInt64) +ENGINE = MergeTree ORDER BY a +SETTINGS index_granularity = 4; + +INSERT INTO t_max_rows_to_read SELECT number FROM numbers(100); + +SET max_block_size = 10; +SET max_rows_to_read = 20; +SET read_overflow_mode = 'throw'; + +SELECT number FROM numbers(30); -- { serverError 158 } +SELECT number FROM numbers(30) LIMIT 21; -- { serverError 158 } +SELECT number FROM numbers(30) LIMIT 1; +SELECT number FROM numbers(5); + +SELECT a FROM t_max_rows_to_read LIMIT 1; +SELECT a FROM t_max_rows_to_read LIMIT 11 offset 11; -- { serverError 158 } +SELECT a FROM t_max_rows_to_read WHERE a > 50 LIMIT 1; -- { serverError 158 } + +DROP TABLE t_max_rows_to_read; From 56e5daba0c97f6e55ad556c895670ea42efc0296 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Sat, 22 Oct 2022 22:51:59 +0000 Subject: [PATCH 065/239] remove DataPartStorageBuilder --- .../DiskObjectStorageTransaction.cpp | 2 +- .../MergeTree/DataPartStorageOnDisk.cpp | 182 ++++++++---------- .../MergeTree/DataPartStorageOnDisk.h | 67 +++---- src/Storages/MergeTree/DataPartsExchange.cpp | 73 +++---- src/Storages/MergeTree/DataPartsExchange.h | 5 +- src/Storages/MergeTree/IDataPartStorage.h | 55 ++---- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 29 ++- src/Storages/MergeTree/IMergeTreeDataPart.h | 20 +- .../IMergeTreeDataPartInfoForReader.h | 5 +- .../MergeTree/IMergeTreeDataPartWriter.cpp | 4 +- .../MergeTree/IMergeTreeDataPartWriter.h | 6 +- .../MergeTree/IMergedBlockOutputStream.cpp | 5 +- .../MergeTree/IMergedBlockOutputStream.h | 6 +- .../LoadedMergeTreeDataPartInfoForReader.h | 5 +- .../MergeTree/MergeFromLogEntryTask.cpp | 4 +- .../MergeTree/MergePlainMergeTreeTask.cpp | 3 +- src/Storages/MergeTree/MergeTask.cpp | 19 +- src/Storages/MergeTree/MergeTask.h | 9 - src/Storages/MergeTree/MergeTreeData.cpp | 59 ++---- src/Storages/MergeTree/MergeTreeData.h | 24 +-- .../MergeTree/MergeTreeDataMergerMutator.cpp | 7 +- .../MergeTree/MergeTreeDataMergerMutator.h | 4 +- .../MergeTree/MergeTreeDataPartCompact.cpp | 9 +- .../MergeTree/MergeTreeDataPartCompact.h | 7 +- .../MergeTree/MergeTreeDataPartInMemory.cpp | 50 ++--- .../MergeTree/MergeTreeDataPartInMemory.h | 14 +- .../MergeTree/MergeTreeDataPartWide.cpp | 11 +- .../MergeTree/MergeTreeDataPartWide.h | 7 +- .../MergeTreeDataPartWriterCompact.cpp | 9 +- .../MergeTreeDataPartWriterCompact.h | 3 +- .../MergeTreeDataPartWriterInMemory.cpp | 4 +- .../MergeTreeDataPartWriterInMemory.h | 4 +- .../MergeTreeDataPartWriterOnDisk.cpp | 20 +- .../MergeTree/MergeTreeDataPartWriterOnDisk.h | 5 +- .../MergeTree/MergeTreeDataPartWriterWide.cpp | 17 +- .../MergeTree/MergeTreeDataPartWriterWide.h | 3 +- .../MergeTree/MergeTreeDataWriter.cpp | 34 +--- src/Storages/MergeTree/MergeTreeDataWriter.h | 5 - src/Storages/MergeTree/MergeTreePartition.cpp | 8 +- src/Storages/MergeTree/MergeTreePartition.h | 8 +- .../MergeTree/MergeTreePartsMover.cpp | 21 +- src/Storages/MergeTree/MergeTreeSink.cpp | 2 +- .../MergeTree/MergeTreeWriteAheadLog.cpp | 6 +- .../MergeTree/MergedBlockOutputStream.cpp | 41 ++-- .../MergeTree/MergedBlockOutputStream.h | 3 +- .../MergedColumnOnlyOutputStream.cpp | 8 +- .../MergeTree/MergedColumnOnlyOutputStream.h | 3 +- .../MergeTree/MutateFromLogEntryTask.cpp | 7 +- .../MergeTree/MutatePlainMergeTreeTask.cpp | 7 +- src/Storages/MergeTree/MutateTask.cpp | 58 ++---- src/Storages/MergeTree/MutateTask.h | 2 +- .../MergeTree/ReplicatedMergeTreeSink.cpp | 10 +- .../MergeTree/ReplicatedMergeTreeSink.h | 1 - src/Storages/StorageMergeTree.cpp | 14 +- src/Storages/StorageReplicatedMergeTree.cpp | 72 +++---- src/Storages/StorageReplicatedMergeTree.h | 6 +- 56 files changed, 432 insertions(+), 640 deletions(-) diff --git a/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp b/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp index 2a75668dd76..b55fb2c4fa5 100644 --- a/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp +++ b/src/Disks/ObjectStorages/DiskObjectStorageTransaction.cpp @@ -599,7 +599,7 @@ std::unique_ptr DiskObjectStorageTransaction::writeFile auto write_operation = std::make_unique(object_storage, metadata_storage, object); std::function create_metadata_callback; - if (autocommit) + if (autocommit) { create_metadata_callback = [tx = shared_from_this(), mode, path, blob_name] (size_t count) { diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index efc7710f640..765c62ba903 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -12,6 +12,7 @@ #include #include #include +#include namespace DB { @@ -29,6 +30,11 @@ DataPartStorageOnDisk::DataPartStorageOnDisk(VolumePtr volume_, std::string root { } +std::shared_ptr DataPartStorageOnDisk::clone() const +{ + return std::make_shared(volume, root_path, part_dir); +} + std::string DataPartStorageOnDisk::getFullPath() const { return fs::path(volume->getDisk()->getPath()) / root_path / part_dir / ""; @@ -54,6 +60,11 @@ DataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); } +MutableDataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) +{ + return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); +} + bool DataPartStorageOnDisk::exists() const { return volume->getDisk()->exists(fs::path(root_path) / part_dir); @@ -199,11 +210,6 @@ void DataPartStorageOnDisk::checkConsistency(const MergeTreeDataPartChecksums & checksums.checkSizes(volume->getDisk(), getRelativePath()); } -DataPartStorageBuilderPtr DataPartStorageOnDisk::getBuilder() const -{ - return std::make_shared(volume, root_path, part_dir); -} - void DataPartStorageOnDisk::remove( CanRemoveCallback && can_remove_callback, const MergeTreeDataPartChecksums & checksums, @@ -488,11 +494,6 @@ bool DataPartStorageOnDisk::looksLikeBrokenDetachedPartHasTheSameContent(const S return original_files_list == detached_files_list; } -void DataPartStorageBuilderOnDisk::setRelativePath(const std::string & path) -{ - part_dir = path; -} - std::string DataPartStorageOnDisk::getDiskName() const { return volume->getDisk()->getName(); @@ -548,7 +549,7 @@ DataPartStorageOnDisk::DisksSet::const_iterator DataPartStorageOnDisk::isStoredO return disks.find(volume->getDisk()); } -ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) const +ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) { auto res = volume->reserve(bytes); if (!res) @@ -557,7 +558,7 @@ ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) const return res; } -ReservationPtr DataPartStorageOnDisk::tryReserve(UInt64 bytes) const +ReservationPtr DataPartStorageOnDisk::tryReserve(UInt64 bytes) { return volume->reserve(bytes); } @@ -798,7 +799,7 @@ void DataPartStorageOnDisk::backup( } } -DataPartStoragePtr DataPartStorageOnDisk::freeze( +MutableDataPartStoragePtr DataPartStorageOnDisk::freeze( const std::string & to, const std::string & dir_path, bool make_source_readonly, @@ -822,7 +823,7 @@ DataPartStoragePtr DataPartStorageOnDisk::freeze( return std::make_shared(single_disk_volume, to, dir_path); } -DataPartStoragePtr DataPartStorageOnDisk::clone( +MutableDataPartStoragePtr DataPartStorageOnDisk::clonePart( const std::string & to, const std::string & dir_path, const DiskPtr & disk, @@ -835,6 +836,7 @@ DataPartStoragePtr DataPartStorageOnDisk::clone( LOG_WARNING(log, "Path {} already exists. Will remove it and clone again.", fullPath(disk, path_to_clone)); disk->removeRecursive(path_to_clone); } + disk->createDirectories(to); volume->getDisk()->copy(getRelativePath(), disk, to); volume->getDisk()->removeFileIfExists(fs::path(path_to_clone) / "delete-on-destroy.txt"); @@ -849,7 +851,7 @@ void DataPartStorageOnDisk::onRename(const std::string & new_root_path, const st root_path = new_root_path; } -void DataPartStorageBuilderOnDisk::rename( +void DataPartStorageOnDisk::rename( const std::string & new_root_path, const std::string & new_part_dir, Poco::Logger * log, @@ -870,7 +872,7 @@ void DataPartStorageBuilderOnDisk::rename( "Part directory {} already exists and contains {} files. Removing it.", fullPath(volume->getDisk(), to), files.size()); - transaction->removeRecursive(to); + executeOperation([&](auto & disk) { disk.removeRecursive(to); }); } else { @@ -884,8 +886,12 @@ void DataPartStorageBuilderOnDisk::rename( String from = getRelativePath(); /// Why? - transaction->setLastModified(from, Poco::Timestamp::fromEpochTime(time(nullptr))); - transaction->moveDirectory(from, to); + executeOperation([&](auto & disk) + { + disk.setLastModified(from, Poco::Timestamp::fromEpochTime(time(nullptr))); + disk.moveDirectory(from, to); + }); + part_dir = new_part_dir; root_path = new_root_path; @@ -920,51 +926,52 @@ void DataPartStorageOnDisk::changeRootPath(const std::string & from_root, const root_path = to_root.substr(0, dst_size) + root_path.substr(prefix_size); } -DataPartStorageBuilderOnDisk::DataPartStorageBuilderOnDisk( - VolumePtr volume_, - std::string root_path_, - std::string part_dir_) - : volume(std::move(volume_)) - , root_path(std::move(root_path_)) - , part_dir(std::move(part_dir_)) - , transaction(volume->getDisk()->createTransaction()) -{ -} - -std::unique_ptr DataPartStorageBuilderOnDisk::writeFile( - const String & name, - size_t buf_size, - const WriteSettings & settings) -{ - return transaction->writeFile(fs::path(root_path) / part_dir / name, buf_size, WriteMode::Rewrite, settings, /* autocommit = */ false); -} - -void DataPartStorageBuilderOnDisk::removeFile(const String & name) -{ - transaction->removeFile(fs::path(root_path) / part_dir / name); -} - -void DataPartStorageBuilderOnDisk::removeFileIfExists(const String & name) -{ - transaction->removeFileIfExists(fs::path(root_path) / part_dir / name); -} - -void DataPartStorageBuilderOnDisk::removeRecursive() -{ - transaction->removeRecursive(fs::path(root_path) / part_dir); -} - -void DataPartStorageBuilderOnDisk::removeSharedRecursive(bool keep_in_remote_fs) -{ - transaction->removeSharedRecursive(fs::path(root_path) / part_dir, keep_in_remote_fs, {}); -} - -SyncGuardPtr DataPartStorageBuilderOnDisk::getDirectorySyncGuard() const +SyncGuardPtr DataPartStorageOnDisk::getDirectorySyncGuard() const { return volume->getDisk()->getDirectorySyncGuard(fs::path(root_path) / part_dir); } -void DataPartStorageBuilderOnDisk::createHardLinkFrom(const IDataPartStorage & source, const std::string & from, const std::string & to) const +template +void DataPartStorageOnDisk::executeOperation(Op && op) +{ + if (transaction) + op(*transaction); + else + op(*volume->getDisk()); +} + +std::unique_ptr DataPartStorageOnDisk::writeFile( + const String & name, + size_t buf_size, + const WriteSettings & settings) +{ + if (transaction) + return transaction->writeFile(fs::path(root_path) / part_dir / name, buf_size, WriteMode::Rewrite, settings, /* autocommit = */ false); + + return volume->getDisk()->writeFile(fs::path(root_path) / part_dir / name, buf_size, WriteMode::Rewrite, settings); +} + +void DataPartStorageOnDisk::removeFile(const String & name) +{ + executeOperation([&](auto & disk) { disk.removeFile(fs::path(root_path) / part_dir / name); }); +} + +void DataPartStorageOnDisk::removeFileIfExists(const String & name) +{ + executeOperation([&](auto & disk) { disk.removeFileIfExists(fs::path(root_path) / part_dir / name); }); +} + +void DataPartStorageOnDisk::removeRecursive() +{ + executeOperation([&](auto & disk) { disk.removeRecursive(fs::path(root_path) / part_dir); }); +} + +void DataPartStorageOnDisk::removeSharedRecursive(bool keep_in_remote_fs) +{ + executeOperation([&](auto & disk) { disk.removeSharedRecursive(fs::path(root_path) / part_dir, keep_in_remote_fs, {}); }); +} + +void DataPartStorageOnDisk::createHardLinkFrom(const IDataPartStorage & source, const std::string & from, const std::string & to) { const auto * source_on_disk = typeid_cast(&source); if (!source_on_disk) @@ -973,58 +980,39 @@ void DataPartStorageBuilderOnDisk::createHardLinkFrom(const IDataPartStorage & s "Cannot create hardlink from different storage. Expected DataPartStorageOnDisk, got {}", typeid(source).name()); - transaction->createHardLink( - fs::path(source_on_disk->getRelativePath()) / from, - fs::path(root_path) / part_dir / to); + executeOperation([&](auto & disk) + { + disk.createHardLink( + fs::path(source_on_disk->getRelativePath()) / from, + fs::path(root_path) / part_dir / to); + }); } -bool DataPartStorageBuilderOnDisk::exists() const +void DataPartStorageOnDisk::createDirectories() { - return volume->getDisk()->exists(fs::path(root_path) / part_dir); + executeOperation([&](auto & disk) { disk.createDirectories(fs::path(root_path) / part_dir); }); } -std::string DataPartStorageBuilderOnDisk::getFullPath() const +void DataPartStorageOnDisk::createProjection(const std::string & name) { - return fs::path(volume->getDisk()->getPath()) / root_path / part_dir; + executeOperation([&](auto & disk) { disk.createDirectory(fs::path(root_path) / part_dir / name); }); } -std::string DataPartStorageBuilderOnDisk::getRelativePath() const +void DataPartStorageOnDisk::beginTransaction() { - return fs::path(root_path) / part_dir; + if (transaction) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Uncommitted transaction already exists"); + + transaction = volume->getDisk()->createTransaction(); } -void DataPartStorageBuilderOnDisk::createDirectories() +void DataPartStorageOnDisk::commitTransaction() { - transaction->createDirectories(fs::path(root_path) / part_dir); -} + if (!transaction) + throw Exception(ErrorCodes::LOGICAL_ERROR, "There is no uncommitted transaction"); -void DataPartStorageBuilderOnDisk::createProjection(const std::string & name) -{ - transaction->createDirectory(fs::path(root_path) / part_dir / name); -} - -ReservationPtr DataPartStorageBuilderOnDisk::reserve(UInt64 bytes) -{ - auto res = volume->reserve(bytes); - if (!res) - throw Exception(ErrorCodes::NOT_ENOUGH_SPACE, "Cannot reserve {}, not enough space", ReadableSize(bytes)); - - return res; -} - -DataPartStorageBuilderPtr DataPartStorageBuilderOnDisk::getProjection(const std::string & name) const -{ - return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); -} - -DataPartStoragePtr DataPartStorageBuilderOnDisk::getStorage() const -{ - return std::make_shared(volume, root_path, part_dir); -} - -void DataPartStorageBuilderOnDisk::commit() -{ transaction->commit(); + transaction.reset(); } } diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.h b/src/Storages/MergeTree/DataPartStorageOnDisk.h index d325049f056..3ce063ca990 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.h +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.h @@ -15,6 +15,7 @@ class DataPartStorageOnDisk final : public IDataPartStorage { public: DataPartStorageOnDisk(VolumePtr volume_, std::string root_path_, std::string part_dir_); + std::shared_ptr clone() const override; std::string getFullPath() const override; std::string getRelativePath() const override; @@ -22,6 +23,7 @@ public: std::string getFullRootPath() const override; DataPartStoragePtr getProjection(const std::string & name) const override; + MutableDataPartStoragePtr getProjection(const std::string & name) override; bool exists() const override; bool exists(const std::string & name) const override; @@ -75,8 +77,8 @@ public: DisksSet::const_iterator isStoredOnDisk(const DisksSet & disks) const override; - ReservationPtr reserve(UInt64 bytes) const override; - ReservationPtr tryReserve(UInt64 bytes) const override; + ReservationPtr reserve(UInt64 bytes) override; + ReservationPtr tryReserve(UInt64 bytes) override; size_t getVolumeIndex(const IStoragePolicy &) const override; void writeChecksums(const MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const override; @@ -100,7 +102,7 @@ public: bool make_temporary_hard_links, TemporaryFilesOnDisks * temp_dirs) const override; - DataPartStoragePtr freeze( + MutableDataPartStoragePtr freeze( const std::string & to, const std::string & dir_path, bool make_source_readonly, @@ -108,7 +110,7 @@ public: bool copy_instead_of_hardlink, const NameSet & files_to_copy_instead_of_hardlinks) const override; - DataPartStoragePtr clone( + MutableDataPartStoragePtr clonePart( const std::string & to, const std::string & dir_path, const DiskPtr & disk, @@ -116,40 +118,9 @@ public: void changeRootPath(const std::string & from_root, const std::string & to_root) override; - DataPartStorageBuilderPtr getBuilder() const override; -private: - VolumePtr volume; - std::string root_path; - std::string part_dir; - - void clearDirectory( - const std::string & dir, - bool can_remove_shared_data, - const NameSet & names_not_to_remove, - const MergeTreeDataPartChecksums & checksums, - const std::unordered_set & skip_directories, - bool is_temp, - MergeTreeDataPartState state, - Poco::Logger * log, - bool is_projection) const; -}; - -class DataPartStorageBuilderOnDisk final : public IDataPartStorageBuilder -{ -public: - DataPartStorageBuilderOnDisk(VolumePtr volume_, std::string root_path_, std::string part_dir_); - - void setRelativePath(const std::string & path) override; - - bool exists() const override; - void createDirectories() override; void createProjection(const std::string & name) override; - std::string getPartDirectory() const override { return part_dir; } - std::string getFullPath() const override; - std::string getRelativePath() const override; - std::unique_ptr writeFile( const String & name, size_t buf_size, @@ -162,13 +133,7 @@ public: SyncGuardPtr getDirectorySyncGuard() const override; - void createHardLinkFrom(const IDataPartStorage & source, const std::string & from, const std::string & to) const override; - - ReservationPtr reserve(UInt64 bytes) override; - - DataPartStorageBuilderPtr getProjection(const std::string & name) const override; - - DataPartStoragePtr getStorage() const override; + void createHardLinkFrom(const IDataPartStorage & source, const std::string & from, const std::string & to) override; void rename( const std::string & new_root_path, @@ -177,13 +142,29 @@ public: bool remove_new_dir_if_exists, bool fsync_part_dir) override; - void commit() override; + void beginTransaction() override; + void commitTransaction() override; + bool hasActiveTransaction() const override { return transaction != nullptr; } private: VolumePtr volume; std::string root_path; std::string part_dir; DiskTransactionPtr transaction; + + template + void executeOperation(Op && op); + + void clearDirectory( + const std::string & dir, + bool can_remove_shared_data, + const NameSet & names_not_to_remove, + const MergeTreeDataPartChecksums & checksums, + const std::unordered_set & skip_directories, + bool is_temp, + MergeTreeDataPartState state, + Poco::Logger * log, + bool is_projection) const; }; } diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index 475461aa0d6..3398839131c 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -13,9 +13,9 @@ #include #include #include +#include #include #include -#include #include #include #include @@ -728,13 +728,9 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( data.getRelativeDataPath(), part_name); - auto data_part_storage_builder = std::make_shared( - volume, - data.getRelativeDataPath(), - part_name); - MergeTreeData::MutableDataPartPtr new_data_part = std::make_shared(data, part_name, data_part_storage); + new_data_part->version.setCreationTID(Tx::PrehistoricTID, nullptr); for (auto i = 0ul; i < projections; ++i) @@ -750,7 +746,6 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( throttler->add(block.bytes()); auto projection_part_storage = data_part_storage->getProjection(projection_name + ".proj"); - auto projection_part_storage_builder = data_part_storage_builder->getProjection(projection_name + ".proj"); MergeTreePartInfo new_part_info("all", 0, 0, 0); MergeTreeData::MutableDataPartPtr new_projection_part = @@ -764,7 +759,6 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( MergedBlockOutputStream part_out( new_projection_part, - projection_part_storage_builder, metadata_snapshot->projections.get(projection_name).metadata, block.getNamesAndTypesList(), {}, @@ -792,7 +786,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( new_data_part->partition.create(metadata_snapshot, block, 0, context); MergedBlockOutputStream part_out( - new_data_part, data_part_storage_builder, metadata_snapshot, block.getNamesAndTypesList(), {}, + new_data_part, metadata_snapshot, block.getNamesAndTypesList(), {}, CompressionCodecFactory::instance().get("NONE", {}), NO_TRANSACTION_PTR); part_out.write(block); @@ -804,7 +798,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( void Fetcher::downloadBasePartOrProjectionPartToDiskRemoteMeta( const String & replica_path, - DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & data_part_storage, PooledReadWriteBufferFromHTTP & in, MergeTreeData::DataPart::Checksums & checksums, ThrottlerPtr throttler) const @@ -820,7 +814,7 @@ void Fetcher::downloadBasePartOrProjectionPartToDiskRemoteMeta( readStringBinary(file_name, in); readBinary(file_size, in); - String metadata_file = fs::path(data_part_storage_builder->getFullPath()) / file_name; + String metadata_file = fs::path(data_part_storage->getFullPath()) / file_name; { auto file_out = std::make_unique(metadata_file, DBMS_DEFAULT_BUFFER_SIZE, -1, 0666, nullptr, 0); @@ -834,8 +828,8 @@ void Fetcher::downloadBasePartOrProjectionPartToDiskRemoteMeta( /// NOTE The is_cancelled flag also makes sense to check every time you read over the network, /// performing a poll with a not very large timeout. /// And now we check it only between read chunks (in the `copyData` function). - data_part_storage_builder->removeSharedRecursive(true); - data_part_storage_builder->commit(); + data_part_storage->removeSharedRecursive(true); + data_part_storage->commitTransaction(); throw Exception("Fetching of part was cancelled", ErrorCodes::ABORTED); } @@ -861,7 +855,7 @@ void Fetcher::downloadBasePartOrProjectionPartToDiskRemoteMeta( void Fetcher::downloadBaseOrProjectionPartToDisk( const String & replica_path, - DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & data_part_storage, bool sync, PooledReadWriteBufferFromHTTP & in, MergeTreeData::DataPart::Checksums & checksums, @@ -880,14 +874,14 @@ void Fetcher::downloadBaseOrProjectionPartToDisk( /// File must be inside "absolute_part_path" directory. /// Otherwise malicious ClickHouse replica may force us to write to arbitrary path. - String absolute_file_path = fs::weakly_canonical(fs::path(data_part_storage_builder->getRelativePath()) / file_name); - if (!startsWith(absolute_file_path, fs::weakly_canonical(data_part_storage_builder->getRelativePath()).string())) + String absolute_file_path = fs::weakly_canonical(fs::path(data_part_storage->getRelativePath()) / file_name); + if (!startsWith(absolute_file_path, fs::weakly_canonical(data_part_storage->getRelativePath()).string())) throw Exception(ErrorCodes::INSECURE_PATH, "File path ({}) doesn't appear to be inside part path ({}). " "This may happen if we are trying to download part from malicious replica or logical error.", - absolute_file_path, data_part_storage_builder->getRelativePath()); + absolute_file_path, data_part_storage->getRelativePath()); - auto file_out = data_part_storage_builder->writeFile(file_name, std::min(file_size, DBMS_DEFAULT_BUFFER_SIZE), {}); + auto file_out = data_part_storage->writeFile(file_name, std::min(file_size, DBMS_DEFAULT_BUFFER_SIZE), {}); HashingWriteBuffer hashing_out(*file_out); copyDataWithThrottler(in, hashing_out, file_size, blocker.getCounter(), throttler); @@ -896,7 +890,7 @@ void Fetcher::downloadBaseOrProjectionPartToDisk( /// NOTE The is_cancelled flag also makes sense to check every time you read over the network, /// performing a poll with a not very large timeout. /// And now we check it only between read chunks (in the `copyData` function). - data_part_storage_builder->removeRecursive(); + data_part_storage->removeRecursive(); throw Exception("Fetching of part was cancelled", ErrorCodes::ABORTED); } @@ -906,7 +900,7 @@ void Fetcher::downloadBaseOrProjectionPartToDisk( if (expected_hash != hashing_out.getHash()) throw Exception(ErrorCodes::CHECKSUM_DOESNT_MATCH, "Checksum mismatch for file {} transferred from {}", - (fs::path(data_part_storage_builder->getFullPath()) / file_name).string(), + (fs::path(data_part_storage->getFullPath()) / file_name).string(), replica_path); if (file_name != "checksums.txt" && @@ -951,15 +945,12 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDisk( part_relative_path, part_dir); - DataPartStorageBuilderPtr data_part_storage_builder = std::make_shared( - volume, - part_relative_path, - part_dir); + data_part_storage->beginTransaction(); - if (data_part_storage_builder->exists()) + if (data_part_storage->exists()) { LOG_WARNING(log, "Directory {} already exists, probably result of a failed fetch. Will remove it before fetching part.", - data_part_storage_builder->getFullPath()); + data_part_storage->getFullPath()); /// Even if it's a temporary part it could be downloaded with zero copy replication and this function /// is executed as a callback. @@ -967,10 +958,10 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDisk( /// We don't control the amount of refs for temporary parts so we cannot decide can we remove blobs /// or not. So we are not doing it bool keep_shared = disk->supportZeroCopyReplication() && data_settings->allow_remote_fs_zero_copy_replication; - data_part_storage_builder->removeSharedRecursive(keep_shared); + data_part_storage->removeSharedRecursive(keep_shared); } - data_part_storage_builder->createDirectories(); + data_part_storage->createDirectories(); SyncGuardPtr sync_guard; if (data.getSettings()->fsync_part_directory) @@ -985,19 +976,18 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDisk( MergeTreeData::DataPart::Checksums projection_checksum; auto projection_part_storage = data_part_storage->getProjection(projection_name + ".proj"); - auto projection_part_storage_builder = data_part_storage_builder->getProjection(projection_name + ".proj"); - - projection_part_storage_builder->createDirectories(); + projection_part_storage->createDirectories(); downloadBaseOrProjectionPartToDisk( - replica_path, projection_part_storage_builder, sync, in, projection_checksum, throttler); + replica_path, projection_part_storage, sync, in, projection_checksum, throttler); checksums.addFile( projection_name + ".proj", projection_checksum.getTotalSizeOnDisk(), projection_checksum.getTotalChecksumUInt128()); } // Download the base part - downloadBaseOrProjectionPartToDisk(replica_path, data_part_storage_builder, sync, in, checksums, throttler); + downloadBaseOrProjectionPartToDisk(replica_path, data_part_storage, sync, in, checksums, throttler); assertEOF(in); + data_part_storage->commitTransaction(); MergeTreeData::MutableDataPartPtr new_data_part = data.createPart(part_name, data_part_storage); new_data_part->version.setCreationTID(Tx::PrehistoricTID, nullptr); new_data_part->is_temp = true; @@ -1043,17 +1033,14 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDiskRemoteMeta( part_relative_path, part_dir); - DataPartStorageBuilderPtr data_part_storage_builder = std::make_shared( - volume, - part_relative_path, - part_dir); + data_part_storage->beginTransaction(); if (data_part_storage->exists()) throw Exception(ErrorCodes::DIRECTORY_ALREADY_EXISTS, "Directory {} already exists.", data_part_storage->getFullPath()); CurrentMetrics::Increment metric_increment{CurrentMetrics::ReplicatedFetch}; - volume->getDisk()->createDirectories(data_part_storage->getFullPath()); + data_part_storage->createDirectories(); for (auto i = 0ul; i < projections; ++i) { @@ -1062,24 +1049,22 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDiskRemoteMeta( MergeTreeData::DataPart::Checksums projection_checksum; auto projection_part_storage = data_part_storage->getProjection(projection_name + ".proj"); - auto projection_part_storage_builder = data_part_storage_builder->getProjection(projection_name + ".proj"); - - projection_part_storage_builder->createDirectories(); + projection_part_storage->createDirectories(); downloadBasePartOrProjectionPartToDiskRemoteMeta( - replica_path, projection_part_storage_builder, in, projection_checksum, throttler); + replica_path, projection_part_storage, in, projection_checksum, throttler); checksums.addFile( projection_name + ".proj", projection_checksum.getTotalSizeOnDisk(), projection_checksum.getTotalChecksumUInt128()); } downloadBasePartOrProjectionPartToDiskRemoteMeta( - replica_path, data_part_storage_builder, in, checksums, throttler); + replica_path, data_part_storage, in, checksums, throttler); assertEOF(in); MergeTreeData::MutableDataPartPtr new_data_part; try { - data_part_storage_builder->commit(); + data_part_storage->commitTransaction(); new_data_part = data.createPart(part_name, data_part_storage); new_data_part->version.setCreationTID(Tx::PrehistoricTID, nullptr); diff --git a/src/Storages/MergeTree/DataPartsExchange.h b/src/Storages/MergeTree/DataPartsExchange.h index 9e453ffb422..59800756c34 100644 --- a/src/Storages/MergeTree/DataPartsExchange.h +++ b/src/Storages/MergeTree/DataPartsExchange.h @@ -94,7 +94,7 @@ public: private: void downloadBaseOrProjectionPartToDisk( const String & replica_path, - DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & data_part_storage, bool sync, PooledReadWriteBufferFromHTTP & in, MergeTreeData::DataPart::Checksums & checksums, @@ -102,12 +102,11 @@ private: void downloadBasePartOrProjectionPartToDiskRemoteMeta( const String & replica_path, - DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & data_part_storage, PooledReadWriteBufferFromHTTP & in, MergeTreeData::DataPart::Checksums & checksums, ThrottlerPtr throttler) const; - MergeTreeData::MutableDataPartPtr downloadPartToDisk( const String & part_name, const String & replica_path, diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index 03627938348..3b005942c54 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -4,6 +4,7 @@ #include #include #include +#include #include namespace DB @@ -18,6 +19,7 @@ struct CanRemoveDescription NameSet files_not_to_remove; }; + using CanRemoveCallback = std::function; class IDataPartStorageIterator @@ -61,16 +63,14 @@ struct WriteSettings; class TemporaryFileOnDisk; -class IDataPartStorageBuilder; -using DataPartStorageBuilderPtr = std::shared_ptr; - /// This is an abstraction of storage for data part files. /// Ideally, it is assumed to contains read-only methods from IDisk. /// It is not fulfilled now, but let's try our best. -class IDataPartStorage +class IDataPartStorage : public boost::noncopyable { public: virtual ~IDataPartStorage() = default; + virtual std::shared_ptr clone() const = 0; /// Methods to get path components of a data part. virtual std::string getFullPath() const = 0; /// '/var/lib/clickhouse/data/database/table/moving/all_1_5_1' @@ -81,7 +81,8 @@ public: /// virtual std::string getRelativeRootPath() const = 0; /// Get a storage for projection. - virtual std::shared_ptr getProjection(const std::string & name) const = 0; + virtual std::shared_ptr getProjection(const std::string & name) const = 0; + virtual std::shared_ptr getProjection(const std::string & name) = 0; /// Part directory exists. virtual bool exists() const = 0; @@ -155,8 +156,8 @@ public: /// Reserve space on the same disk. /// Probably we should try to remove it later. - virtual ReservationPtr reserve(UInt64 /*bytes*/) const { return nullptr; } - virtual ReservationPtr tryReserve(UInt64 /*bytes*/) const { return nullptr; } + virtual ReservationPtr reserve(UInt64 /*bytes*/) { return nullptr; } + virtual ReservationPtr tryReserve(UInt64 /*bytes*/) { return nullptr; } virtual size_t getVolumeIndex(const IStoragePolicy &) const { return 0; } /// Some methods which change data part internals possibly after creation. @@ -205,7 +206,7 @@ public: const NameSet & files_to_copy_instead_of_hardlinks) const = 0; /// Make a full copy of a data part into 'to/dir_path' (possibly to a different disk). - virtual std::shared_ptr clone( + virtual std::shared_ptr clonePart( const std::string & to, const std::string & dir_path, const DiskPtr & disk, @@ -215,29 +216,6 @@ public: /// Right now, this is needed for rename table query. virtual void changeRootPath(const std::string & from_root, const std::string & to_root) = 0; - /// Leak of abstraction as well. We should use builder as one-time object which allow - /// us to build parts, while storage should be read-only method to access part properties - /// related to disk. However our code is really tricky and sometimes we need ad-hoc builders. - virtual DataPartStorageBuilderPtr getBuilder() const = 0; -}; - -using DataPartStoragePtr = std::shared_ptr; - -/// This interface is needed to write data part. -class IDataPartStorageBuilder -{ -public: - virtual ~IDataPartStorageBuilder() = default; - - /// Reset part directory, used for im-memory parts - virtual void setRelativePath(const std::string & path) = 0; - - virtual std::string getPartDirectory() const = 0; - virtual std::string getFullPath() const = 0; - virtual std::string getRelativePath() const = 0; - - virtual bool exists() const = 0; - virtual void createDirectories() = 0; virtual void createProjection(const std::string & name) = 0; @@ -250,13 +228,7 @@ public: virtual SyncGuardPtr getDirectorySyncGuard() const { return nullptr; } - virtual void createHardLinkFrom(const IDataPartStorage & source, const std::string & from, const std::string & to) const = 0; - - virtual ReservationPtr reserve(UInt64 /*bytes*/) { return nullptr; } - - virtual std::shared_ptr getProjection(const std::string & name) const = 0; - - virtual DataPartStoragePtr getStorage() const = 0; + virtual void createHardLinkFrom(const IDataPartStorage & source, const std::string & from, const std::string & to) = 0; /// Rename part. /// Ideally, new_root_path should be the same as current root (but it is not true). @@ -271,7 +243,12 @@ public: bool remove_new_dir_if_exists, bool fsync_part_dir) = 0; - virtual void commit() = 0; + virtual void beginTransaction() = 0; + virtual void commitTransaction() = 0; + virtual bool hasActiveTransaction() const = 0; }; +using DataPartStoragePtr = std::shared_ptr; +using MutableDataPartStoragePtr = std::shared_ptr; + } diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index cc9a14162f8..6a641f0c94e 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -101,7 +101,7 @@ void IMergeTreeDataPart::MinMaxIndex::load(const MergeTreeData & data, const Par } IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::store( - const MergeTreeData & data, const DataPartStorageBuilderPtr & data_part_storage_builder, Checksums & out_checksums) const + const MergeTreeData & data, const MutableDataPartStoragePtr & part_storage, Checksums & out_checksums) const { auto metadata_snapshot = data.getInMemoryMetadataPtr(); const auto & partition_key = metadata_snapshot->getPartitionKey(); @@ -109,20 +109,20 @@ IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::s auto minmax_column_names = data.getMinMaxColumnsNames(partition_key); auto minmax_column_types = data.getMinMaxColumnsTypes(partition_key); - return store(minmax_column_names, minmax_column_types, data_part_storage_builder, out_checksums); + return store(minmax_column_names, minmax_column_types, part_storage, out_checksums); } IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::store( const Names & column_names, const DataTypes & data_types, - const DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & part_storage, Checksums & out_checksums) const { if (!initialized) throw Exception( ErrorCodes::LOGICAL_ERROR, "Attempt to store uninitialized MinMax index for part {}. This is a bug", - data_part_storage_builder->getFullPath()); + part_storage->getFullPath()); WrittenFiles written_files; @@ -131,7 +131,7 @@ IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::s String file_name = "minmax_" + escapeForFileName(column_names[i]) + ".idx"; auto serialization = data_types.at(i)->getDefaultSerialization(); - auto out = data_part_storage_builder->writeFile(file_name, DBMS_DEFAULT_BUFFER_SIZE, {}); + auto out = part_storage->writeFile(file_name, DBMS_DEFAULT_BUFFER_SIZE, {}); HashingWriteBuffer out_hashing(*out); serialization->serializeBinary(hyperrectangle[i].left, out_hashing); serialization->serializeBinary(hyperrectangle[i].right, out_hashing); @@ -301,7 +301,7 @@ static void decrementTypeMetric(MergeTreeDataPartType type) IMergeTreeDataPart::IMergeTreeDataPart( const MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, Type part_type_, const IMergeTreeDataPart * parent_part_) : storage(storage_) @@ -315,6 +315,7 @@ IMergeTreeDataPart::IMergeTreeDataPart( { if (parent_part) state = MergeTreeDataPartState::Active; + incrementStateMetric(state); incrementTypeMetric(part_type); @@ -328,7 +329,7 @@ IMergeTreeDataPart::IMergeTreeDataPart( const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, Type part_type_, const IMergeTreeDataPart * parent_part_) : storage(storage_) @@ -1365,7 +1366,7 @@ bool IMergeTreeDataPart::shallParticipateInMerges(const StoragePolicyPtr & stora return data_part_storage->shallParticipateInMerges(*storage_policy); } -void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_new_dir_if_exists, DataPartStorageBuilderPtr builder) const +void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) const try { assertOnDisk(); @@ -1384,14 +1385,12 @@ try metadata_manager->deleteAll(true); metadata_manager->assertAllDeleted(true); - builder->rename(to.parent_path(), to.filename(), storage.log, remove_new_dir_if_exists, fsync_dir); + data_part_storage->rename(to.parent_path(), to.filename(), storage.log, remove_new_dir_if_exists, fsync_dir); data_part_storage->onRename(to.parent_path(), to.filename()); metadata_manager->updateAll(true); for (const auto & [p_name, part] : projection_parts) - { part->data_part_storage = data_part_storage->getProjection(p_name + ".proj"); - } } catch (...) { @@ -1507,11 +1506,11 @@ std::optional IMergeTreeDataPart::getRelativePathForDetachedPart(const S return {}; } -void IMergeTreeDataPart::renameToDetached(const String & prefix, DataPartStorageBuilderPtr builder) const +void IMergeTreeDataPart::renameToDetached(const String & prefix) const { auto path_to_detach = getRelativePathForDetachedPart(prefix, /* broken */ false); assert(path_to_detach); - renameTo(path_to_detach.value(), true, builder); + renameTo(path_to_detach.value(), true); part_is_probably_removed_from_disk = true; } @@ -1539,7 +1538,7 @@ void IMergeTreeDataPart::makeCloneInDetached(const String & prefix, const Storag {}); } -DataPartStoragePtr IMergeTreeDataPart::makeCloneOnDisk(const DiskPtr & disk, const String & directory_name) const +MutableDataPartStoragePtr IMergeTreeDataPart::makeCloneOnDisk(const DiskPtr & disk, const String & directory_name) const { assertOnDisk(); @@ -1549,7 +1548,7 @@ DataPartStoragePtr IMergeTreeDataPart::makeCloneOnDisk(const DiskPtr & disk, con throw Exception("Can not clone data part " + name + " to empty directory.", ErrorCodes::LOGICAL_ERROR); String path_to_clone = fs::path(storage.relative_data_path) / directory_name / ""; - return data_part_storage->clone(path_to_clone, data_part_storage->getPartDirectory(), disk, storage.log); + return data_part_storage->clonePart(path_to_clone, data_part_storage->getPartDirectory(), disk, storage.log); } void IMergeTreeDataPart::checkConsistencyBase() const diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.h b/src/Storages/MergeTree/IMergeTreeDataPart.h index 6f034574fb4..0fe94b666b6 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.h +++ b/src/Storages/MergeTree/IMergeTreeDataPart.h @@ -67,19 +67,18 @@ public: using uint128 = IPartMetadataManager::uint128; - IMergeTreeDataPart( const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, Type part_type_, const IMergeTreeDataPart * parent_part_); IMergeTreeDataPart( const MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, Type part_type_, const IMergeTreeDataPart * parent_part_); @@ -94,13 +93,12 @@ public: const ReadBufferFromFileBase::ProfileCallback & profile_callback_) const = 0; virtual MergeTreeWriterPtr getWriter( - DataPartStorageBuilderPtr data_part_storage_builder, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & computed_index_granularity) const = 0; + const MergeTreeIndexGranularity & computed_index_granularity) = 0; virtual bool isStoredOnDisk() const = 0; @@ -202,7 +200,7 @@ public: /// This is an object which encapsulates all the operations with disk. /// Contains a path to stored data. - DataPartStoragePtr data_part_storage; + MutableDataPartStoragePtr data_part_storage; MergeTreeIndexGranularityInfo index_granularity_info; @@ -289,8 +287,8 @@ public: using WrittenFiles = std::vector>; - [[nodiscard]] WrittenFiles store(const MergeTreeData & data, const DataPartStorageBuilderPtr & data_part_storage_builder, Checksums & checksums) const; - [[nodiscard]] WrittenFiles store(const Names & column_names, const DataTypes & data_types, const DataPartStorageBuilderPtr & data_part_storage_builder, Checksums & checksums) const; + [[nodiscard]] WrittenFiles store(const MergeTreeData & data, const MutableDataPartStoragePtr & part_storage, Checksums & checksums) const; + [[nodiscard]] WrittenFiles store(const Names & column_names, const DataTypes & data_types, const MutableDataPartStoragePtr & part_storage, Checksums & checksums) const; void update(const Block & block, const Names & column_names); void merge(const MinMaxIndex & other); @@ -321,17 +319,17 @@ public: size_t getFileSizeOrZero(const String & file_name) const; /// Moves a part to detached/ directory and adds prefix to its name - void renameToDetached(const String & prefix, DataPartStorageBuilderPtr builder) const; + void renameToDetached(const String & prefix) const; /// Makes checks and move part to new directory /// Changes only relative_dir_name, you need to update other metadata (name, is_temp) explicitly - virtual void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists, DataPartStorageBuilderPtr builder) const; + virtual void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) const; /// Makes clone of a part in detached/ directory via hard links virtual void makeCloneInDetached(const String & prefix, const StorageMetadataPtr & metadata_snapshot) const; /// Makes full clone of part in specified subdirectory (relative to storage data directory, e.g. "detached") on another disk - DataPartStoragePtr makeCloneOnDisk(const DiskPtr & disk, const String & directory_name) const; + MutableDataPartStoragePtr makeCloneOnDisk(const DiskPtr & disk, const String & directory_name) const; /// Checks that .bin and .mrk files exist. /// diff --git a/src/Storages/MergeTree/IMergeTreeDataPartInfoForReader.h b/src/Storages/MergeTree/IMergeTreeDataPartInfoForReader.h index 28f834d661d..2e4972c2788 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPartInfoForReader.h +++ b/src/Storages/MergeTree/IMergeTreeDataPartInfoForReader.h @@ -7,7 +7,8 @@ namespace DB { class IDataPartStorage; -using DataPartStoragePtr = std::shared_ptr; +using DataPartStoragePtr = std::shared_ptr; + class MergeTreeIndexGranularity; struct MergeTreeDataPartChecksums; struct MergeTreeIndexGranularityInfo; @@ -36,7 +37,7 @@ public: virtual bool isProjectionPart() const = 0; - virtual const DataPartStoragePtr & getDataPartStorage() const = 0; + virtual DataPartStoragePtr getDataPartStorage() const = 0; virtual const NamesAndTypesList & getColumns() const = 0; diff --git a/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp b/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp index 84d0b50ae2f..2488c63e309 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPartWriter.cpp @@ -38,14 +38,12 @@ Block permuteBlockIfNeeded(const Block & block, const IColumn::Permutation * per } IMergeTreeDataPartWriter::IMergeTreeDataPartWriter( - const MergeTreeData::DataPartPtr & data_part_, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot_, const MergeTreeWriterSettings & settings_, const MergeTreeIndexGranularity & index_granularity_) : data_part(data_part_) - , data_part_storage_builder(std::move(data_part_storage_builder_)) , storage(data_part_->storage) , metadata_snapshot(metadata_snapshot_) , columns_list(columns_list_) diff --git a/src/Storages/MergeTree/IMergeTreeDataPartWriter.h b/src/Storages/MergeTree/IMergeTreeDataPartWriter.h index 417e2713180..fa3c675f7da 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPartWriter.h +++ b/src/Storages/MergeTree/IMergeTreeDataPartWriter.h @@ -22,8 +22,7 @@ class IMergeTreeDataPartWriter : private boost::noncopyable { public: IMergeTreeDataPartWriter( - const MergeTreeData::DataPartPtr & data_part_, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot_, const MergeTreeWriterSettings & settings_, @@ -42,8 +41,7 @@ public: protected: - const MergeTreeData::DataPartPtr data_part; - DataPartStorageBuilderPtr data_part_storage_builder; + const MergeTreeMutableDataPartPtr data_part; const MergeTreeData & storage; const StorageMetadataPtr metadata_snapshot; const NamesAndTypesList columns_list; diff --git a/src/Storages/MergeTree/IMergedBlockOutputStream.cpp b/src/Storages/MergeTree/IMergedBlockOutputStream.cpp index 5af9bbd3ed8..54f393a65a2 100644 --- a/src/Storages/MergeTree/IMergedBlockOutputStream.cpp +++ b/src/Storages/MergeTree/IMergedBlockOutputStream.cpp @@ -6,14 +6,13 @@ namespace DB { IMergedBlockOutputStream::IMergedBlockOutputStream( - DataPartStorageBuilderPtr data_part_storage_builder_, - const MergeTreeDataPartPtr & data_part, + const MergeTreeMutableDataPartPtr & data_part, const StorageMetadataPtr & metadata_snapshot_, const NamesAndTypesList & columns_list, bool reset_columns_) : storage(data_part->storage) , metadata_snapshot(metadata_snapshot_) - , data_part_storage_builder(std::move(data_part_storage_builder_)) + , data_part_storage(data_part->data_part_storage) , reset_columns(reset_columns_) { if (reset_columns) diff --git a/src/Storages/MergeTree/IMergedBlockOutputStream.h b/src/Storages/MergeTree/IMergedBlockOutputStream.h index dbcca1443b5..ca4e3899b29 100644 --- a/src/Storages/MergeTree/IMergedBlockOutputStream.h +++ b/src/Storages/MergeTree/IMergedBlockOutputStream.h @@ -1,5 +1,6 @@ #pragma once +#include "Storages/MergeTree/IDataPartStorage.h" #include #include #include @@ -12,8 +13,7 @@ class IMergedBlockOutputStream { public: IMergedBlockOutputStream( - DataPartStorageBuilderPtr data_part_storage_builder_, - const MergeTreeDataPartPtr & data_part, + const MergeTreeMutableDataPartPtr & data_part, const StorageMetadataPtr & metadata_snapshot_, const NamesAndTypesList & columns_list, bool reset_columns_); @@ -42,7 +42,7 @@ protected: const MergeTreeData & storage; StorageMetadataPtr metadata_snapshot; - DataPartStorageBuilderPtr data_part_storage_builder; + MutableDataPartStoragePtr data_part_storage; IMergeTreeDataPart::MergeTreeWriterPtr writer; bool reset_columns = false; diff --git a/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h b/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h index a16aaa728ae..ee265ee6fb1 100644 --- a/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h +++ b/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h @@ -12,7 +12,8 @@ public: explicit LoadedMergeTreeDataPartInfoForReader(MergeTreeData::DataPartPtr data_part_) : IMergeTreeDataPartInfoForReader(data_part_->storage.getContext()) , data_part(data_part_) - {} + { + } bool isCompactPart() const override { return DB::isCompactPart(data_part); } @@ -22,7 +23,7 @@ public: bool isProjectionPart() const override { return data_part->isProjectionPart(); } - const DataPartStoragePtr & getDataPartStorage() const override { return data_part->data_part_storage; } + DataPartStoragePtr getDataPartStorage() const override { return data_part->data_part_storage; } const NamesAndTypesList & getColumns() const override { return data_part->getColumns(); } diff --git a/src/Storages/MergeTree/MergeFromLogEntryTask.cpp b/src/Storages/MergeTree/MergeFromLogEntryTask.cpp index 18982c3bbf4..182d5df4960 100644 --- a/src/Storages/MergeTree/MergeFromLogEntryTask.cpp +++ b/src/Storages/MergeTree/MergeFromLogEntryTask.cpp @@ -294,12 +294,10 @@ ReplicatedMergeMutateTaskBase::PrepareResult MergeFromLogEntryTask::prepare() bool MergeFromLogEntryTask::finalize(ReplicatedMergeMutateTaskBase::PartLogWriter write_part_log) { part = merge_task->getFuture().get(); - auto builder = merge_task->getBuilder(); /// Task is not needed merge_task.reset(); - - storage.merger_mutator.renameMergedTemporaryPart(part, parts, NO_TRANSACTION_PTR, *transaction_ptr, builder); + storage.merger_mutator.renameMergedTemporaryPart(part, parts, NO_TRANSACTION_PTR, *transaction_ptr); try { diff --git a/src/Storages/MergeTree/MergePlainMergeTreeTask.cpp b/src/Storages/MergeTree/MergePlainMergeTreeTask.cpp index 0dcdd927e7b..cc5e87956a1 100644 --- a/src/Storages/MergeTree/MergePlainMergeTreeTask.cpp +++ b/src/Storages/MergeTree/MergePlainMergeTreeTask.cpp @@ -115,10 +115,9 @@ void MergePlainMergeTreeTask::prepare() void MergePlainMergeTreeTask::finish() { new_part = merge_task->getFuture().get(); - auto builder = merge_task->getBuilder(); MergeTreeData::Transaction transaction(storage, txn.get()); - storage.merger_mutator.renameMergedTemporaryPart(new_part, future_part->parts, txn, transaction, builder); + storage.merger_mutator.renameMergedTemporaryPart(new_part, future_part->parts, txn, transaction); transaction.commit(); write_part_log({}); diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp index c247d2d2476..130d156e53c 100644 --- a/src/Storages/MergeTree/MergeTask.cpp +++ b/src/Storages/MergeTree/MergeTask.cpp @@ -1,3 +1,4 @@ +#include "Storages/MergeTree/IDataPartStorage.h" #include #include @@ -125,23 +126,26 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() ctx->disk = global_ctx->space_reservation->getDisk(); String local_tmp_part_basename = local_tmp_prefix + global_ctx->future_part->name + local_tmp_suffix; + MutableDataPartStoragePtr data_part_storage; - if (global_ctx->parent_path_storage_builder) + if (global_ctx->parent_part) { - global_ctx->data_part_storage_builder = global_ctx->parent_path_storage_builder->getProjection(local_tmp_part_basename); + data_part_storage = global_ctx->parent_part->data_part_storage->getProjection(local_tmp_part_basename); } else { auto local_single_disk_volume = std::make_shared("volume_" + global_ctx->future_part->name, ctx->disk, 0); - global_ctx->data_part_storage_builder = std::make_shared( + data_part_storage = std::make_shared( local_single_disk_volume, global_ctx->data->relative_data_path, local_tmp_part_basename); } - if (global_ctx->data_part_storage_builder->exists()) - throw Exception("Directory " + global_ctx->data_part_storage_builder->getFullPath() + " already exists", ErrorCodes::DIRECTORY_ALREADY_EXISTS); + data_part_storage->beginTransaction(); + + if (data_part_storage->exists()) + throw Exception("Directory " + data_part_storage->getFullPath() + " already exists", ErrorCodes::DIRECTORY_ALREADY_EXISTS); if (!global_ctx->parent_part) global_ctx->temporary_directory_lock = global_ctx->data->getTemporaryPartDirectoryHolder(local_tmp_part_basename); @@ -163,8 +167,6 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() global_ctx->merging_columns, global_ctx->merging_column_names); - auto data_part_storage = global_ctx->data_part_storage_builder->getStorage(); - global_ctx->new_data_part = global_ctx->data->createPart( global_ctx->future_part->name, global_ctx->future_part->type, @@ -302,7 +304,6 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() global_ctx->to = std::make_shared( global_ctx->new_data_part, - global_ctx->data_part_storage_builder, global_ctx->metadata_snapshot, global_ctx->merging_columns, MergeTreeIndexFactory::instance().getMany(global_ctx->metadata_snapshot->getSecondaryIndices()), @@ -501,7 +502,6 @@ void MergeTask::VerticalMergeStage::prepareVerticalMergeForOneColumn() const ctx->executor = std::make_unique(ctx->column_parts_pipeline); ctx->column_to = std::make_unique( - global_ctx->data_part_storage_builder, global_ctx->new_data_part, global_ctx->metadata_snapshot, ctx->executor->getHeader(), @@ -654,7 +654,6 @@ bool MergeTask::MergeProjectionsStage::mergeMinMaxIndexAndPrepareProjections() c global_ctx->deduplicate_by_columns, projection_merging_params, global_ctx->new_data_part.get(), - global_ctx->data_part_storage_builder.get(), ".proj", NO_TRANSACTION_PTR, global_ctx->data, diff --git a/src/Storages/MergeTree/MergeTask.h b/src/Storages/MergeTree/MergeTask.h index 43aba602052..07d46460423 100644 --- a/src/Storages/MergeTree/MergeTask.h +++ b/src/Storages/MergeTree/MergeTask.h @@ -60,7 +60,6 @@ public: Names deduplicate_by_columns_, MergeTreeData::MergingParams merging_params_, const IMergeTreeDataPart * parent_part_, - const IDataPartStorageBuilder * parent_path_storage_builder_, String suffix_, MergeTreeTransactionPtr txn, MergeTreeData * data_, @@ -82,7 +81,6 @@ public: global_ctx->deduplicate = std::move(deduplicate_); global_ctx->deduplicate_by_columns = std::move(deduplicate_by_columns_); global_ctx->parent_part = std::move(parent_part_); - global_ctx->parent_path_storage_builder = std::move(parent_path_storage_builder_); global_ctx->data = std::move(data_); global_ctx->mutator = std::move(mutator_); global_ctx->merges_blocker = std::move(merges_blocker_); @@ -102,11 +100,6 @@ public: return global_ctx->promise.get_future(); } - DataPartStorageBuilderPtr getBuilder() - { - return global_ctx->data_part_storage_builder; - } - bool execute(); private: @@ -142,7 +135,6 @@ private: FutureMergedMutatedPartPtr future_part{nullptr}; /// This will be either nullptr or new_data_part, so raw pointer is ok. const IMergeTreeDataPart * parent_part{nullptr}; - const IDataPartStorageBuilder * parent_path_storage_builder{nullptr}; ContextPtr context{nullptr}; time_t time_of_merge{0}; ReservationSharedPtr space_reservation{nullptr}; @@ -168,7 +160,6 @@ private: std::unique_ptr merging_executor; MergeTreeData::MutableDataPartPtr new_data_part{nullptr}; - DataPartStorageBuilderPtr data_part_storage_builder; /// If lightweight delete mask is present then some input rows are filtered out right after reading. std::shared_ptr> input_rows_filtered{std::make_shared>(0)}; diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 66950734d5f..9e9d90b6b1e 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -1082,7 +1082,6 @@ void MergeTreeData::loadDataPartsFromDisk( if (size_of_part.has_value()) part_size_str = formatReadableSizeWithBinarySuffix(*size_of_part); - LOG_ERROR(log, "Detaching broken part {}{} (size: {}). " "If it happened after update, it is likely because of backward incompatibility. " @@ -1397,11 +1396,7 @@ void MergeTreeData::loadDataParts(bool skip_sanity_checks) } for (auto & part : broken_parts_to_detach) - { - auto builder = part->data_part_storage->getBuilder(); - part->renameToDetached("broken-on-start", builder); /// detached parts must not have '_' in prefixes - builder->commit(); - } + part->renameToDetached("broken-on-start"); /// detached parts must not have '_' in prefixes for (auto & part : duplicate_parts_to_remove) part->remove(); @@ -2726,7 +2721,7 @@ MergeTreeDataPartType MergeTreeData::choosePartTypeOnDisk(size_t bytes_uncompres MergeTreeData::MutableDataPartPtr MergeTreeData::createPart(const String & name, MergeTreeDataPartType type, const MergeTreePartInfo & part_info, - const DataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const + const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const { if (type == MergeTreeDataPartType::Compact) return std::make_shared(*this, name, part_info, data_part_storage, parent_part); @@ -2739,14 +2734,14 @@ MergeTreeData::MutableDataPartPtr MergeTreeData::createPart(const String & name, } MergeTreeData::MutableDataPartPtr MergeTreeData::createPart( - const String & name, const DataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const + const String & name, const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const { return createPart(name, MergeTreePartInfo::fromPartName(name, format_version), data_part_storage, parent_part); } MergeTreeData::MutableDataPartPtr MergeTreeData::createPart( const String & name, const MergeTreePartInfo & part_info, - const DataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const + const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const { MergeTreeDataPartType type; auto mrk_ext = MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(data_part_storage); @@ -2943,12 +2938,11 @@ MergeTreeData::DataPartsVector MergeTreeData::getActivePartsToReplace( bool MergeTreeData::renameTempPartAndAdd( MutableDataPartPtr & part, Transaction & out_transaction, - DataPartStorageBuilderPtr builder, DataPartsLock & lock) { DataPartsVector covered_parts; - if (!renameTempPartAndReplaceImpl(part, out_transaction, lock, builder, &covered_parts)) + if (!renameTempPartAndReplaceImpl(part, out_transaction, lock, &covered_parts)) return false; if (!covered_parts.empty()) @@ -2982,7 +2976,7 @@ void MergeTreeData::checkPartCanBeAddedToTable(MutableDataPartPtr & part, DataPa } } -void MergeTreeData::preparePartForCommit(MutableDataPartPtr & part, Transaction & out_transaction, DataPartStorageBuilderPtr builder) +void MergeTreeData::preparePartForCommit(MutableDataPartPtr & part, Transaction & out_transaction) { part->is_temp = false; part->setState(DataPartState::PreActive); @@ -2994,17 +2988,16 @@ void MergeTreeData::preparePartForCommit(MutableDataPartPtr & part, Transaction return !may_be_cleaned_up || temporary_parts.contains(dir_name); }()); - part->renameTo(part->name, true, builder); + part->renameTo(part->name, true); data_parts_indexes.insert(part); - out_transaction.addPart(part, builder); + out_transaction.addPart(part); } bool MergeTreeData::renameTempPartAndReplaceImpl( MutableDataPartPtr & part, Transaction & out_transaction, DataPartsLock & lock, - DataPartStorageBuilderPtr builder, DataPartsVector * out_covered_parts) { LOG_TRACE(log, "Renaming temporary part {} to {}.", part->data_part_storage->getPartDirectory(), part->name); @@ -3029,7 +3022,7 @@ bool MergeTreeData::renameTempPartAndReplaceImpl( /// All checks are passed. Now we can rename the part on disk. /// So, we maintain invariant: if a non-temporary part in filesystem then it is in data_parts - preparePartForCommit(part, out_transaction, builder); + preparePartForCommit(part, out_transaction); if (out_covered_parts) { @@ -3045,21 +3038,19 @@ bool MergeTreeData::renameTempPartAndReplaceImpl( MergeTreeData::DataPartsVector MergeTreeData::renameTempPartAndReplaceUnlocked( MutableDataPartPtr & part, Transaction & out_transaction, - DataPartStorageBuilderPtr builder, DataPartsLock & lock) { DataPartsVector covered_parts; - renameTempPartAndReplaceImpl(part, out_transaction, lock, builder, &covered_parts); + renameTempPartAndReplaceImpl(part, out_transaction, lock, &covered_parts); return covered_parts; } MergeTreeData::DataPartsVector MergeTreeData::renameTempPartAndReplace( MutableDataPartPtr & part, - Transaction & out_transaction, - DataPartStorageBuilderPtr builder) + Transaction & out_transaction) { auto part_lock = lockParts(); - return renameTempPartAndReplaceUnlocked(part, out_transaction, builder, part_lock); + return renameTempPartAndReplaceUnlocked(part, out_transaction, part_lock); } void MergeTreeData::removePartsFromWorkingSet(MergeTreeTransaction * txn, const MergeTreeData::DataPartsVector & remove, bool clear_without_timeout, DataPartsLock & acquired_lock) @@ -3280,9 +3271,7 @@ void MergeTreeData::forcefullyMovePartToDetachedAndRemoveFromMemory(const MergeT modifyPartState(it_part, DataPartState::Deleting); - auto builder = part->data_part_storage->getBuilder(); - part->renameToDetached(prefix, builder); - builder->commit(); + part->renameToDetached(prefix); data_parts_indexes.erase(it_part); @@ -4911,19 +4900,13 @@ ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, SpacePtr space) return checkAndReturnReservation(expected_size, std::move(reservation)); } -ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, const DataPartStoragePtr & data_part_storage) +ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage) { expected_size = std::max(RESERVATION_MIN_ESTIMATION_SIZE, expected_size); return data_part_storage->reserve(expected_size); } -ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, const DataPartStorageBuilderPtr & data_part_storage_builder) -{ - expected_size = std::max(RESERVATION_MIN_ESTIMATION_SIZE, expected_size); - return data_part_storage_builder->reserve(expected_size); -} - -ReservationPtr MergeTreeData::tryReserveSpace(UInt64 expected_size, const DataPartStoragePtr & data_part_storage) +ReservationPtr MergeTreeData::tryReserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage) { expected_size = std::max(RESERVATION_MIN_ESTIMATION_SIZE, expected_size); return data_part_storage->tryReserve(expected_size); @@ -5162,12 +5145,11 @@ void MergeTreeData::Transaction::rollbackPartsToTemporaryState() clear(); } -void MergeTreeData::Transaction::addPart(MutableDataPartPtr & part, DataPartStorageBuilderPtr builder) +void MergeTreeData::Transaction::addPart(MutableDataPartPtr & part) { precommitted_parts.insert(part); if (asInMemoryPart(part)) has_in_memory_parts = true; - part_builders.push_back(builder); } void MergeTreeData::Transaction::rollback() @@ -5205,8 +5187,9 @@ MergeTreeData::DataPartsVector MergeTreeData::Transaction::commit(MergeTreeData: auto parts_lock = acquired_parts_lock ? MergeTreeData::DataPartsLock() : data.lockParts(); auto * owing_parts_lock = acquired_parts_lock ? acquired_parts_lock : &parts_lock; - for (auto & builder : part_builders) - builder->commit(); + for (const auto & part : precommitted_parts) + if (part->data_part_storage->hasActiveTransaction()) + part->data_part_storage->commitTransaction(); bool commit_to_wal = has_in_memory_parts && settings->in_memory_parts_enable_wal; if (txn || commit_to_wal) @@ -5215,7 +5198,7 @@ MergeTreeData::DataPartsVector MergeTreeData::Transaction::commit(MergeTreeData: if (commit_to_wal) wal = data.getWriteAheadLog(); - for (const DataPartPtr & part : precommitted_parts) + for (const auto & part : precommitted_parts) { if (txn) { @@ -5240,7 +5223,7 @@ MergeTreeData::DataPartsVector MergeTreeData::Transaction::commit(MergeTreeData: size_t reduce_rows = 0; size_t reduce_parts = 0; - for (const DataPartPtr & part : precommitted_parts) + for (const auto & part : precommitted_parts) { DataPartPtr covering_part; DataPartsVector covered_parts = data.getActivePartsToReplace(part->info, part->name, covering_part, *owing_parts_lock); diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index c4a5d66ccbe..327718d15ed 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -225,15 +225,15 @@ public: /// After this method setColumns must be called MutableDataPartPtr createPart(const String & name, MergeTreeDataPartType type, const MergeTreePartInfo & part_info, - const DataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part = nullptr) const; + const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part = nullptr) const; /// Create part, that already exists on filesystem. /// After this methods 'loadColumnsChecksumsIndexes' must be called. MutableDataPartPtr createPart(const String & name, - const DataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part = nullptr) const; + const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part = nullptr) const; MutableDataPartPtr createPart(const String & name, const MergeTreePartInfo & part_info, - const DataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part = nullptr) const; + const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part = nullptr) const; /// Auxiliary object to add a set of parts into the working set in two steps: /// * First, as PreActive parts (the parts are ready, but not yet in the active set). @@ -247,7 +247,7 @@ public: DataPartsVector commit(MergeTreeData::DataPartsLock * acquired_parts_lock = nullptr); - void addPart(MutableDataPartPtr & part, DataPartStorageBuilderPtr builder); + void addPart(MutableDataPartPtr & part); void rollback(); @@ -276,7 +276,6 @@ public: MergeTreeData & data; MergeTreeTransaction * txn; DataParts precommitted_parts; - std::vector part_builders; DataParts locked_parts; bool has_in_memory_parts = false; @@ -414,9 +413,8 @@ public: SelectQueryInfo & info) const override; ReservationPtr reserveSpace(UInt64 expected_size, VolumePtr & volume) const; - static ReservationPtr tryReserveSpace(UInt64 expected_size, const DataPartStoragePtr & data_part_storage); - static ReservationPtr reserveSpace(UInt64 expected_size, const DataPartStoragePtr & data_part_storage); - static ReservationPtr reserveSpace(UInt64 expected_size, const DataPartStorageBuilderPtr & data_part_storage_builder); + static ReservationPtr tryReserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage); + static ReservationPtr reserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage); static bool partsContainSameProjections(const DataPartPtr & left, const DataPartPtr & right); @@ -555,21 +553,18 @@ public: bool renameTempPartAndAdd( MutableDataPartPtr & part, Transaction & transaction, - DataPartStorageBuilderPtr builder, DataPartsLock & lock); /// The same as renameTempPartAndAdd but the block range of the part can contain existing parts. /// Returns all parts covered by the added part (in ascending order). DataPartsVector renameTempPartAndReplace( MutableDataPartPtr & part, - Transaction & out_transaction, - DataPartStorageBuilderPtr builder); + Transaction & out_transaction); /// Unlocked version of previous one. Useful when added multiple parts with a single lock. DataPartsVector renameTempPartAndReplaceUnlocked( MutableDataPartPtr & part, Transaction & out_transaction, - DataPartStorageBuilderPtr builder, DataPartsLock & lock); /// Remove parts from working set immediately (without wait for background @@ -979,7 +974,7 @@ public: /// Fetch part only if some replica has it on shared storage like S3 /// Overridden in StorageReplicatedMergeTree - virtual DataPartStoragePtr tryToFetchIfShared(const IMergeTreeDataPart &, const DiskPtr &, const String &) { return nullptr; } + virtual bool tryToFetchIfShared(const IMergeTreeDataPart &, const DiskPtr &, const String &) { return false; } /// Check shared data usage on other replicas for detached/freezed part /// Remove local files and remote files if needed @@ -1270,7 +1265,7 @@ private: /// Preparing itself to be committed in memory: fill some fields inside part, add it to data_parts_indexes /// in precommitted state and to transaction - void preparePartForCommit(MutableDataPartPtr & part, Transaction & out_transaction, DataPartStorageBuilderPtr builder); + void preparePartForCommit(MutableDataPartPtr & part, Transaction & out_transaction); /// Low-level method for preparing parts for commit (in-memory). /// FIXME Merge MergeTreeTransaction and Transaction @@ -1278,7 +1273,6 @@ private: MutableDataPartPtr & part, Transaction & out_transaction, DataPartsLock & lock, - DataPartStorageBuilderPtr builder, DataPartsVector * out_covered_parts); /// RAII Wrapper for atomic work with currently moving parts diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index 1a5c94a2e26..c96003c8938 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -483,7 +483,6 @@ MergeTaskPtr MergeTreeDataMergerMutator::mergePartsToTemporaryPart( const MergeTreeData::MergingParams & merging_params, const MergeTreeTransactionPtr & txn, const IMergeTreeDataPart * parent_part, - const IDataPartStorageBuilder * parent_path_storage_builder, const String & suffix) { return std::make_shared( @@ -498,7 +497,6 @@ MergeTaskPtr MergeTreeDataMergerMutator::mergePartsToTemporaryPart( deduplicate_by_columns, merging_params, parent_part, - parent_path_storage_builder, suffix, txn, &data, @@ -540,8 +538,7 @@ MergeTreeData::DataPartPtr MergeTreeDataMergerMutator::renameMergedTemporaryPart MergeTreeData::MutableDataPartPtr & new_data_part, const MergeTreeData::DataPartsVector & parts, const MergeTreeTransactionPtr & txn, - MergeTreeData::Transaction & out_transaction, - DataPartStorageBuilderPtr builder) + MergeTreeData::Transaction & out_transaction) { /// Some of source parts was possibly created in transaction, so non-transactional merge may break isolation. if (data.transactions_enabled.load(std::memory_order_relaxed) && !txn) @@ -549,7 +546,7 @@ MergeTreeData::DataPartPtr MergeTreeDataMergerMutator::renameMergedTemporaryPart "but transactions were enabled for this table"); /// Rename new part, add to the set and remove original parts. - auto replaced_parts = data.renameTempPartAndReplace(new_data_part, out_transaction, builder); + auto replaced_parts = data.renameTempPartAndReplace(new_data_part, out_transaction); /// Let's check that all original parts have been deleted and only them. if (replaced_parts.size() != parts.size()) diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h index 14eb82c641c..d1b9d3c99e7 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h @@ -114,7 +114,6 @@ public: const MergeTreeData::MergingParams & merging_params, const MergeTreeTransactionPtr & txn, const IMergeTreeDataPart * parent_part = nullptr, - const IDataPartStorageBuilder * parent_path_storage_builder = nullptr, const String & suffix = ""); /// Mutate a single data part with the specified commands. Will create and return a temporary part. @@ -133,8 +132,7 @@ public: MergeTreeData::MutableDataPartPtr & new_data_part, const MergeTreeData::DataPartsVector & parts, const MergeTreeTransactionPtr & txn, - MergeTreeData::Transaction & out_transaction, - DataPartStorageBuilderPtr builder); + MergeTreeData::Transaction & out_transaction); /// The approximate amount of disk space needed for merge or mutation. With a surplus. diff --git a/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp index 9298e841072..4fedacee13c 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp @@ -22,7 +22,7 @@ namespace ErrorCodes MergeTreeDataPartCompact::MergeTreeDataPartCompact( MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_) : IMergeTreeDataPart(storage_, name_, data_part_storage_, Type::Compact, parent_part_) { @@ -32,7 +32,7 @@ MergeTreeDataPartCompact::MergeTreeDataPartCompact( const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_) : IMergeTreeDataPart(storage_, name_, info_, data_part_storage_, Type::Compact, parent_part_) { @@ -58,13 +58,12 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartCompact::getReader( } IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartCompact::getWriter( - DataPartStorageBuilderPtr data_part_storage_builder, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & computed_index_granularity) const + const MergeTreeIndexGranularity & computed_index_granularity) { NamesAndTypesList ordered_columns_list; std::copy_if(columns_list.begin(), columns_list.end(), std::back_inserter(ordered_columns_list), @@ -75,7 +74,7 @@ IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartCompact::getWriter( { return *getColumnPosition(lhs.name) < *getColumnPosition(rhs.name); }); return std::make_unique( - shared_from_this(), std::move(data_part_storage_builder), ordered_columns_list, metadata_snapshot, + shared_from_this(), ordered_columns_list, metadata_snapshot, indices_to_recalc, getMarksFileExtension(), default_codec_, writer_settings, computed_index_granularity); } diff --git a/src/Storages/MergeTree/MergeTreeDataPartCompact.h b/src/Storages/MergeTree/MergeTreeDataPartCompact.h index d3ac71cb02a..7c3fe012616 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartCompact.h +++ b/src/Storages/MergeTree/MergeTreeDataPartCompact.h @@ -25,13 +25,13 @@ public: const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_ = nullptr); MergeTreeDataPartCompact( MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_ = nullptr); MergeTreeReaderPtr getReader( @@ -45,13 +45,12 @@ public: const ReadBufferFromFileBase::ProfileCallback & profile_callback) const override; MergeTreeWriterPtr getWriter( - DataPartStorageBuilderPtr data_part_storage_builder, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & computed_index_granularity) const override; + const MergeTreeIndexGranularity & computed_index_granularity) override; bool isStoredOnDisk() const override { return true; } diff --git a/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp b/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp index 7a3c5f11c81..c30efbc7969 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp @@ -1,9 +1,10 @@ -#include "MergeTreeDataPartInMemory.h" +#include #include #include #include #include #include +#include #include #include #include @@ -21,7 +22,7 @@ namespace ErrorCodes MergeTreeDataPartInMemory::MergeTreeDataPartInMemory( MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_) : IMergeTreeDataPart(storage_, name_, data_part_storage_, Type::InMemory, parent_part_) { @@ -32,7 +33,7 @@ MergeTreeDataPartInMemory::MergeTreeDataPartInMemory( const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_) : IMergeTreeDataPart(storage_, name_, info_, data_part_storage_, Type::InMemory, parent_part_) { @@ -56,27 +57,27 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartInMemory::getReader( } IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartInMemory::getWriter( - DataPartStorageBuilderPtr data_part_storage_builder_, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & /* indices_to_recalc */, const CompressionCodecPtr & /* default_codec */, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & /* computed_index_granularity */) const + const MergeTreeIndexGranularity & /* computed_index_granularity */) { - data_part_storage_builder = data_part_storage_builder_; - auto ptr = std::static_pointer_cast(shared_from_this()); + auto ptr = std::static_pointer_cast(shared_from_this()); return std::make_unique( ptr, columns_list, metadata_snapshot, writer_settings); } -DataPartStoragePtr MergeTreeDataPartInMemory::flushToDisk(const String & new_relative_path, const StorageMetadataPtr & metadata_snapshot) const +MutableDataPartStoragePtr MergeTreeDataPartInMemory::flushToDisk(const String & new_relative_path, const StorageMetadataPtr & metadata_snapshot) const { - auto current_full_path = data_part_storage_builder->getFullPath(); - data_part_storage_builder->setRelativePath(new_relative_path); + auto current_full_path = data_part_storage->getFullPath(); + auto new_data_part_storage = data_part_storage->clone(); + + new_data_part_storage->setRelativePath(new_relative_path); + new_data_part_storage->beginTransaction(); auto new_type = storage.choosePartTypeOnDisk(block.bytes(), rows_count); - auto new_data_part_storage = data_part_storage_builder->getStorage(); auto new_data_part = storage.createPart(name, new_type, info, new_data_part_storage); new_data_part->uuid = uuid; @@ -84,50 +85,50 @@ DataPartStoragePtr MergeTreeDataPartInMemory::flushToDisk(const String & new_rel new_data_part->partition.value = partition.value; new_data_part->minmax_idx = minmax_idx; - if (data_part_storage_builder->exists()) + if (new_data_part_storage->exists()) { throw Exception( ErrorCodes::DIRECTORY_ALREADY_EXISTS, "Could not flush part {}. Part in {} already exists", quoteString(current_full_path), - data_part_storage_builder->getFullPath()); + new_data_part_storage->getFullPath()); } - data_part_storage_builder->createDirectories(); + new_data_part_storage->createDirectories(); auto compression_codec = storage.getContext()->chooseCompressionCodec(0, 0); auto indices = MergeTreeIndexFactory::instance().getMany(metadata_snapshot->getSecondaryIndices()); - MergedBlockOutputStream out(new_data_part, data_part_storage_builder, metadata_snapshot, columns, indices, compression_codec, NO_TRANSACTION_PTR); + MergedBlockOutputStream out(new_data_part, metadata_snapshot, columns, indices, compression_codec, NO_TRANSACTION_PTR); out.write(block); const auto & projections = metadata_snapshot->getProjections(); for (const auto & [projection_name, projection] : projection_parts) { if (projections.has(projection_name)) { - auto projection_part_storage_builder = data_part_storage_builder->getProjection(projection_name + ".proj"); - if (projection_part_storage_builder->exists()) + auto projection_part_storage = new_data_part_storage->getProjection(projection_name + ".proj"); + if (projection_part_storage->exists()) { throw Exception( ErrorCodes::DIRECTORY_ALREADY_EXISTS, "Could not flush projection part {}. Projection part in {} already exists", projection_name, - projection_part_storage_builder->getFullPath()); + projection_part_storage->getFullPath()); } auto projection_part = asInMemoryPart(projection); auto projection_type = storage.choosePartTypeOnDisk(projection_part->block.bytes(), rows_count); MergeTreePartInfo projection_info("all", 0, 0, 0); auto projection_data_part - = storage.createPart(projection_name, projection_type, projection_info, projection_part_storage_builder->getStorage(), parent_part); + = storage.createPart(projection_name, projection_type, projection_info, projection_part_storage, parent_part); projection_data_part->is_temp = false; // clean up will be done on parent part projection_data_part->setColumns(projection->getColumns(), {}); - projection_part_storage_builder->createDirectories(); + projection_part_storage->createDirectories(); const auto & desc = projections.get(name); auto projection_compression_codec = storage.getContext()->chooseCompressionCodec(0, 0); auto projection_indices = MergeTreeIndexFactory::instance().getMany(desc.metadata->getSecondaryIndices()); MergedBlockOutputStream projection_out( - projection_data_part, projection_part_storage_builder, desc.metadata, projection_part->columns, projection_indices, + projection_data_part, desc.metadata, projection_part->columns, projection_indices, projection_compression_codec, NO_TRANSACTION_PTR); projection_out.write(projection_part->block); @@ -137,6 +138,7 @@ DataPartStoragePtr MergeTreeDataPartInMemory::flushToDisk(const String & new_rel } out.finalizePart(new_data_part, false); + new_data_part_storage->commitTransaction(); return new_data_part_storage; } @@ -146,12 +148,12 @@ void MergeTreeDataPartInMemory::makeCloneInDetached(const String & prefix, const flushToDisk(detached_path, metadata_snapshot); } -void MergeTreeDataPartInMemory::renameTo(const String & new_relative_path, bool /* remove_new_dir_if_exists */, DataPartStorageBuilderPtr) const +void MergeTreeDataPartInMemory::renameTo(const String & new_relative_path, bool /* remove_new_dir_if_exists */) const { data_part_storage->setRelativePath(new_relative_path); - if (data_part_storage_builder) - data_part_storage_builder->setRelativePath(new_relative_path); + if (data_part_storage) + data_part_storage->setRelativePath(new_relative_path); } void MergeTreeDataPartInMemory::calculateEachColumnSizes(ColumnSizeByName & each_columns_size, ColumnSize & total_size) const diff --git a/src/Storages/MergeTree/MergeTreeDataPartInMemory.h b/src/Storages/MergeTree/MergeTreeDataPartInMemory.h index d985c7f055e..49bc5eff1ea 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartInMemory.h +++ b/src/Storages/MergeTree/MergeTreeDataPartInMemory.h @@ -14,13 +14,13 @@ public: const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_ = nullptr); MergeTreeDataPartInMemory( MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_ = nullptr); MergeTreeReaderPtr getReader( @@ -34,29 +34,27 @@ public: const ReadBufferFromFileBase::ProfileCallback & profile_callback) const override; MergeTreeWriterPtr getWriter( - DataPartStorageBuilderPtr data_part_storage_builder_, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & computed_index_granularity) const override; + const MergeTreeIndexGranularity & computed_index_granularity) override; bool isStoredOnDisk() const override { return false; } bool isStoredOnRemoteDisk() const override { return false; } bool isStoredOnRemoteDiskWithZeroCopySupport() const override { return false; } bool hasColumnFiles(const NameAndTypePair & column) const override { return !!getColumnPosition(column.getNameInStorage()); } String getFileNameForColumn(const NameAndTypePair & /* column */) const override { return ""; } - void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists, DataPartStorageBuilderPtr) const override; + void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) const override; void makeCloneInDetached(const String & prefix, const StorageMetadataPtr & metadata_snapshot) const override; - DataPartStoragePtr flushToDisk(const String & new_relative_path, const StorageMetadataPtr & metadata_snapshot) const; + MutableDataPartStoragePtr flushToDisk(const String & new_relative_path, const StorageMetadataPtr & metadata_snapshot) const; /// Returns hash of parts's block Checksum calculateBlockChecksum() const; mutable Block block; - mutable DataPartStorageBuilderPtr data_part_storage_builder; private: mutable std::condition_variable is_merged; @@ -66,6 +64,8 @@ private: }; using DataPartInMemoryPtr = std::shared_ptr; +using MutableDataPartInMemoryPtr = std::shared_ptr; + DataPartInMemoryPtr asInMemoryPart(const MergeTreeDataPartPtr & part); } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWide.cpp index 170d1b1d703..18467f2cef7 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWide.cpp @@ -21,7 +21,7 @@ namespace ErrorCodes MergeTreeDataPartWide::MergeTreeDataPartWide( MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_) : IMergeTreeDataPart(storage_, name_, data_part_storage_, Type::Wide, parent_part_) { @@ -31,7 +31,7 @@ MergeTreeDataPartWide::MergeTreeDataPartWide( const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_) : IMergeTreeDataPart(storage_, name_, info_, data_part_storage_, Type::Wide, parent_part_) { @@ -56,17 +56,16 @@ IMergeTreeDataPart::MergeTreeReaderPtr MergeTreeDataPartWide::getReader( } IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartWide::getWriter( - DataPartStorageBuilderPtr data_part_storage_builder, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & computed_index_granularity) const + const MergeTreeIndexGranularity & computed_index_granularity) { return std::make_unique( - shared_from_this(), data_part_storage_builder, - columns_list, metadata_snapshot, indices_to_recalc, + shared_from_this(), columns_list, + metadata_snapshot, indices_to_recalc, getMarksFileExtension(), default_codec_, writer_settings, computed_index_granularity); } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWide.h b/src/Storages/MergeTree/MergeTreeDataPartWide.h index 52afa9e82d4..4343148b175 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWide.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWide.h @@ -19,13 +19,13 @@ public: const MergeTreeData & storage_, const String & name_, const MergeTreePartInfo & info_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_ = nullptr); MergeTreeDataPartWide( MergeTreeData & storage_, const String & name_, - const DataPartStoragePtr & data_part_storage_, + const MutableDataPartStoragePtr & data_part_storage_, const IMergeTreeDataPart * parent_part_ = nullptr); MergeTreeReaderPtr getReader( @@ -39,13 +39,12 @@ public: const ReadBufferFromFileBase::ProfileCallback & profile_callback) const override; MergeTreeWriterPtr getWriter( - DataPartStorageBuilderPtr data_part_storage_builder, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & writer_settings, - const MergeTreeIndexGranularity & computed_index_granularity) const override; + const MergeTreeIndexGranularity & computed_index_granularity) override; bool isStoredOnDisk() const override { return true; } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp index 457aad55023..6821c52f0d2 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp @@ -10,8 +10,7 @@ namespace ErrorCodes } MergeTreeDataPartWriterCompact::MergeTreeDataPartWriterCompact( - const MergeTreeData::DataPartPtr & data_part_, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot_, const std::vector & indices_to_recalc_, @@ -19,16 +18,16 @@ MergeTreeDataPartWriterCompact::MergeTreeDataPartWriterCompact( const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & settings_, const MergeTreeIndexGranularity & index_granularity_) - : MergeTreeDataPartWriterOnDisk(data_part_, std::move(data_part_storage_builder_), columns_list_, metadata_snapshot_, + : MergeTreeDataPartWriterOnDisk(data_part_, columns_list_, metadata_snapshot_, indices_to_recalc_, marks_file_extension_, default_codec_, settings_, index_granularity_) - , plain_file(data_part_storage_builder->writeFile( + , plain_file(data_part_->data_part_storage->writeFile( MergeTreeDataPartCompact::DATA_FILE_NAME_WITH_EXTENSION, settings.max_compress_block_size, settings_.query_write_settings)) , plain_hashing(*plain_file) { - marks_file = data_part_storage_builder->writeFile( + marks_file = data_part_->data_part_storage->writeFile( MergeTreeDataPartCompact::DATA_FILE_NAME + marks_file_extension_, 4096, settings_.query_write_settings); diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h index 7b68f61925f..06f8122393f 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.h @@ -11,8 +11,7 @@ class MergeTreeDataPartWriterCompact : public MergeTreeDataPartWriterOnDisk { public: MergeTreeDataPartWriterCompact( - const MergeTreeData::DataPartPtr & data_part, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot_, const std::vector & indices_to_recalc, diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp index e1145868ce2..8066a097499 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.cpp @@ -11,11 +11,11 @@ namespace ErrorCodes } MergeTreeDataPartWriterInMemory::MergeTreeDataPartWriterInMemory( - const DataPartInMemoryPtr & part_, + const MutableDataPartInMemoryPtr & part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot_, const MergeTreeWriterSettings & settings_) - : IMergeTreeDataPartWriter(part_, nullptr, columns_list_, metadata_snapshot_, settings_) + : IMergeTreeDataPartWriter(part_, columns_list_, metadata_snapshot_, settings_) , part_in_memory(part_) {} void MergeTreeDataPartWriterInMemory::write( diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h index 233ca81a697..9e1e868beac 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterInMemory.h @@ -10,7 +10,7 @@ class MergeTreeDataPartWriterInMemory : public IMergeTreeDataPartWriter { public: MergeTreeDataPartWriterInMemory( - const DataPartInMemoryPtr & part_, + const MutableDataPartInMemoryPtr & part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot, const MergeTreeWriterSettings & settings_); @@ -24,7 +24,7 @@ public: private: void calculateAndSerializePrimaryIndex(const Block & primary_index_block); - DataPartInMemoryPtr part_in_memory; + MutableDataPartInMemoryPtr part_in_memory; }; } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp index 1d2b095330e..383cd79734f 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp @@ -48,7 +48,7 @@ void MergeTreeDataPartWriterOnDisk::Stream::sync() const MergeTreeDataPartWriterOnDisk::Stream::Stream( const String & escaped_column_name_, - const DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & data_part_storage, const String & data_path_, const std::string & data_file_extension_, const std::string & marks_path_, @@ -61,11 +61,11 @@ MergeTreeDataPartWriterOnDisk::Stream::Stream( escaped_column_name(escaped_column_name_), data_file_extension{data_file_extension_}, marks_file_extension{marks_file_extension_}, - plain_file(data_part_storage_builder->writeFile(data_path_ + data_file_extension, max_compress_block_size_, query_write_settings)), + plain_file(data_part_storage->writeFile(data_path_ + data_file_extension, max_compress_block_size_, query_write_settings)), plain_hashing(*plain_file), compressor(plain_hashing, compression_codec_, max_compress_block_size_), compressed_hashing(compressor), - marks_file(data_part_storage_builder->writeFile(marks_path_ + marks_file_extension, 4096, query_write_settings)), + marks_file(data_part_storage->writeFile(marks_path_ + marks_file_extension, 4096, query_write_settings)), marks_hashing(*marks_file), marks_compressor(marks_hashing, marks_compression_codec_, marks_compress_block_size_), marks_compressed_hashing(marks_compressor), @@ -96,8 +96,7 @@ void MergeTreeDataPartWriterOnDisk::Stream::addToChecksums(MergeTreeData::DataPa MergeTreeDataPartWriterOnDisk::MergeTreeDataPartWriterOnDisk( - const MergeTreeData::DataPartPtr & data_part_, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot_, const MergeTreeIndices & indices_to_recalc_, @@ -105,8 +104,7 @@ MergeTreeDataPartWriterOnDisk::MergeTreeDataPartWriterOnDisk( const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & settings_, const MergeTreeIndexGranularity & index_granularity_) - : IMergeTreeDataPartWriter(data_part_, std::move(data_part_storage_builder_), - columns_list_, metadata_snapshot_, settings_, index_granularity_) + : IMergeTreeDataPartWriter(data_part_, columns_list_, metadata_snapshot_, settings_, index_granularity_) , skip_indices(indices_to_recalc_) , marks_file_extension(marks_file_extension_) , default_codec(default_codec_) @@ -116,8 +114,8 @@ MergeTreeDataPartWriterOnDisk::MergeTreeDataPartWriterOnDisk( if (settings.blocks_are_granules_size && !index_granularity.empty()) throw Exception("Can't take information about index granularity from blocks, when non empty index_granularity array specified", ErrorCodes::LOGICAL_ERROR); - if (!data_part_storage_builder->exists()) - data_part_storage_builder->createDirectories(); + if (!data_part->data_part_storage->exists()) + data_part->data_part_storage->createDirectories(); if (settings.rewrite_primary_key) initPrimaryIndex(); @@ -178,7 +176,7 @@ void MergeTreeDataPartWriterOnDisk::initPrimaryIndex() if (metadata_snapshot->hasPrimaryKey()) { String index_name = "primary" + getIndexExtension(compress_primary_key); - index_file_stream = data_part_storage_builder->writeFile(index_name, DBMS_DEFAULT_BUFFER_SIZE, settings.query_write_settings); + index_file_stream = data_part->data_part_storage->writeFile(index_name, DBMS_DEFAULT_BUFFER_SIZE, settings.query_write_settings); index_file_hashing_stream = std::make_unique(*index_file_stream); if (compress_primary_key) @@ -204,7 +202,7 @@ void MergeTreeDataPartWriterOnDisk::initSkipIndices() skip_indices_streams.emplace_back( std::make_unique( stream_name, - data_part_storage_builder, + data_part->data_part_storage, stream_name, index_helper->getSerializedFileExtension(), stream_name, marks_file_extension, default_codec, settings.max_compress_block_size, diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.h b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.h index 4b58224de78..ab1adfe7f59 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.h @@ -50,7 +50,7 @@ public: { Stream( const String & escaped_column_name_, - const DataPartStorageBuilderPtr & data_part_storage_builder, + const MutableDataPartStoragePtr & data_part_storage, const String & data_path_, const std::string & data_file_extension_, const std::string & marks_path_, @@ -92,8 +92,7 @@ public: using StreamPtr = std::unique_ptr; MergeTreeDataPartWriterOnDisk( - const MergeTreeData::DataPartPtr & data_part_, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part_, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot_, const std::vector & indices_to_recalc, diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp index 70654f521a1..5c6e69abe4d 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp @@ -71,8 +71,7 @@ Granules getGranulesToWrite(const MergeTreeIndexGranularity & index_granularity, } MergeTreeDataPartWriterWide::MergeTreeDataPartWriterWide( - const MergeTreeData::DataPartPtr & data_part_, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part_, const NamesAndTypesList & columns_list_, const StorageMetadataPtr & metadata_snapshot_, const std::vector & indices_to_recalc_, @@ -80,7 +79,7 @@ MergeTreeDataPartWriterWide::MergeTreeDataPartWriterWide( const CompressionCodecPtr & default_codec_, const MergeTreeWriterSettings & settings_, const MergeTreeIndexGranularity & index_granularity_) - : MergeTreeDataPartWriterOnDisk(data_part_, std::move(data_part_storage_builder_), columns_list_, metadata_snapshot_, + : MergeTreeDataPartWriterOnDisk(data_part_, columns_list_, metadata_snapshot_, indices_to_recalc_, marks_file_extension_, default_codec_, settings_, index_granularity_) { @@ -117,7 +116,7 @@ void MergeTreeDataPartWriterWide::addStreams( column_streams[stream_name] = std::make_unique( stream_name, - data_part_storage_builder, + data_part->data_part_storage, stream_name, DATA_FILE_EXTENSION, stream_name, marks_file_extension, compression_codec, @@ -421,20 +420,18 @@ void MergeTreeDataPartWriterWide::validateColumnOfFixedSize(const NameAndTypePai String mrk_path = escaped_name + marks_file_extension; String bin_path = escaped_name + DATA_FILE_EXTENSION; - auto data_part_storage = data_part_storage_builder->getStorage(); - /// Some columns may be removed because of ttl. Skip them. - if (!data_part_storage->exists(mrk_path)) + if (!data_part->data_part_storage->exists(mrk_path)) return; - auto mrk_file_in = data_part_storage->readFile(mrk_path, {}, std::nullopt, std::nullopt); + auto mrk_file_in = data_part->data_part_storage->readFile(mrk_path, {}, std::nullopt, std::nullopt); std::unique_ptr mrk_in; if (data_part->index_granularity_info.mark_type.compressed) mrk_in = std::make_unique(std::move(mrk_file_in)); else mrk_in = std::move(mrk_file_in); - DB::CompressedReadBufferFromFile bin_in(data_part_storage->readFile(bin_path, {}, std::nullopt, std::nullopt)); + DB::CompressedReadBufferFromFile bin_in(data_part->data_part_storage->readFile(bin_path, {}, std::nullopt, std::nullopt)); bool must_be_last = false; UInt64 offset_in_compressed_file = 0; UInt64 offset_in_decompressed_block = 0; @@ -485,7 +482,7 @@ void MergeTreeDataPartWriterWide::validateColumnOfFixedSize(const NameAndTypePai if (index_granularity_rows != index_granularity.getMarkRows(mark_num)) throw Exception( ErrorCodes::LOGICAL_ERROR, "Incorrect mark rows for part {} for mark #{} (compressed offset {}, decompressed offset {}), in-memory {}, on disk {}, total marks {}", - data_part_storage_builder->getFullPath(), mark_num, offset_in_compressed_file, offset_in_decompressed_block, index_granularity.getMarkRows(mark_num), index_granularity_rows, index_granularity.getMarksCount()); + data_part->data_part_storage->getFullPath(), mark_num, offset_in_compressed_file, offset_in_decompressed_block, index_granularity.getMarkRows(mark_num), index_granularity_rows, index_granularity.getMarksCount()); auto column = type->createColumn(); diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h index 08815d9930a..633b5119474 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.h @@ -18,8 +18,7 @@ class MergeTreeDataPartWriterWide : public MergeTreeDataPartWriterOnDisk { public: MergeTreeDataPartWriterWide( - const MergeTreeData::DataPartPtr & data_part, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part, const NamesAndTypesList & columns_list, const StorageMetadataPtr & metadata_snapshot, const std::vector & indices_to_recalc, diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 7b99819340e..62d01a2d555 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -378,10 +378,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( data.relative_data_path, TMP_PREFIX + part_name); - auto data_part_storage_builder = std::make_shared( - data_part_volume, - data.relative_data_path, - TMP_PREFIX + part_name); + data_part_storage->beginTransaction(); auto new_data_part = data.createPart( part_name, @@ -413,10 +410,10 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( if (new_data_part->data_part_storage->exists()) { LOG_WARNING(log, "Removing old temporary directory {}", full_path); - data_part_storage_builder->removeRecursive(); + data_part_storage->removeRecursive(); } - data_part_storage_builder->createDirectories(); + data_part_storage->createDirectories(); if (data.getSettings()->fsync_part_directory) { @@ -448,7 +445,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( auto compression_codec = data.getContext()->chooseCompressionCodec(0, 0); const auto & index_factory = MergeTreeIndexFactory::instance(); - auto out = std::make_unique(new_data_part, data_part_storage_builder, metadata_snapshot, columns, + auto out = std::make_unique(new_data_part, metadata_snapshot, columns, index_factory.getMany(metadata_snapshot->getSecondaryIndices()), compression_codec, context->getCurrentTransaction(), false, false, context->getWriteSettings()); @@ -459,9 +456,8 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( auto projection_block = projection.calculate(block, context); if (projection_block.rows()) { - auto proj_temp_part = writeProjectionPart(data, log, projection_block, projection, data_part_storage_builder, new_data_part.get()); + auto proj_temp_part = writeProjectionPart(data, log, projection_block, projection, new_data_part.get()); new_data_part->addProjectionPart(projection.name, std::move(proj_temp_part.part)); - proj_temp_part.builder->commit(); for (auto & stream : proj_temp_part.streams) temp_part.streams.emplace_back(std::move(stream)); } @@ -473,7 +469,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( nullptr, nullptr); temp_part.part = new_data_part; - temp_part.builder = data_part_storage_builder; temp_part.streams.emplace_back(TemporaryPart::Stream{.stream = std::move(out), .finalizer = std::move(finalizer)}); ProfileEvents::increment(ProfileEvents::MergeTreeDataWriterRows, block.rows()); @@ -487,7 +482,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( const String & part_name, MergeTreeDataPartType part_type, const String & relative_path, - const DataPartStorageBuilderPtr & data_part_storage_builder, bool is_temp, const IMergeTreeDataPart * parent_part, const MergeTreeData & data, @@ -498,6 +492,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( TemporaryPart temp_part; const StorageMetadataPtr & metadata_snapshot = projection.metadata; MergeTreePartInfo new_part_info("all", 0, 0, 0); + auto projection_part_storage = parent_part->data_part_storage->getProjection(relative_path); auto new_data_part = data.createPart( part_name, @@ -506,7 +501,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( projection_part_storage, parent_part); - auto projection_part_storage_builder = data_part_storage_builder->getProjection(relative_path); new_data_part->is_temp = is_temp; NamesAndTypesList columns = metadata_snapshot->getColumns().getAllPhysical().filter(block.getNames()); @@ -522,10 +516,10 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( if (projection_part_storage->exists()) { LOG_WARNING(log, "Removing old temporary directory {}", projection_part_storage->getFullPath()); - projection_part_storage_builder->removeRecursive(); + projection_part_storage->removeRecursive(); } - projection_part_storage_builder->createDirectories(); + projection_part_storage->createDirectories(); } /// If we need to calculate some columns to sort. @@ -569,7 +563,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( auto out = std::make_unique( new_data_part, - projection_part_storage_builder, metadata_snapshot, columns, MergeTreeIndices{}, @@ -580,7 +573,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( out->writeWithPermutation(block, perm_ptr); auto finalizer = out->finalizePartAsync(new_data_part, false); temp_part.part = new_data_part; - temp_part.builder = projection_part_storage_builder; temp_part.streams.emplace_back(TemporaryPart::Stream{.stream = std::move(out), .finalizer = std::move(finalizer)}); ProfileEvents::increment(ProfileEvents::MergeTreeDataProjectionWriterRows, block.rows()); @@ -595,7 +587,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPart( Poco::Logger * log, Block block, const ProjectionDescription & projection, - const DataPartStorageBuilderPtr & data_part_storage_builder, const IMergeTreeDataPart * parent_part) { String part_name = projection.name; @@ -609,7 +600,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPart( /// Size of part would not be greater than block.bytes() + epsilon size_t expected_size = block.bytes(); // just check if there is enough space on parent volume - data.reserveSpace(expected_size, data_part_storage_builder); + data.reserveSpace(expected_size, parent_part->data_part_storage); part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); } @@ -617,7 +608,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPart( part_name, part_type, part_name + ".proj" /* relative_path */, - data_part_storage_builder, false /* is_temp */, parent_part, data, @@ -633,7 +623,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( Poco::Logger * log, Block block, const ProjectionDescription & projection, - const DataPartStorageBuilderPtr & data_part_storage_builder, const IMergeTreeDataPart * parent_part, size_t block_num) { @@ -648,7 +637,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( /// Size of part would not be greater than block.bytes() + epsilon size_t expected_size = block.bytes(); // just check if there is enough space on parent volume - data.reserveSpace(expected_size, data_part_storage_builder); + data.reserveSpace(expected_size, parent_part->data_part_storage); part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); } @@ -656,7 +645,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( part_name, part_type, part_name + ".tmp_proj" /* relative_path */, - data_part_storage_builder, true /* is_temp */, parent_part, data, @@ -670,14 +658,12 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeInMemoryProjectionP Poco::Logger * log, Block block, const ProjectionDescription & projection, - const DataPartStorageBuilderPtr & data_part_storage_builder, const IMergeTreeDataPart * parent_part) { return writeProjectionPartImpl( projection.name, MergeTreeDataPartType::InMemory, projection.name + ".proj" /* relative_path */, - data_part_storage_builder, false /* is_temp */, parent_part, data, diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.h b/src/Storages/MergeTree/MergeTreeDataWriter.h index 00438a29fa1..72ceb8b38e3 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.h +++ b/src/Storages/MergeTree/MergeTreeDataWriter.h @@ -52,7 +52,6 @@ public: struct TemporaryPart { MergeTreeData::MutableDataPartPtr part; - DataPartStorageBuilderPtr builder; struct Stream { @@ -78,7 +77,6 @@ public: Poco::Logger * log, Block block, const ProjectionDescription & projection, - const DataPartStorageBuilderPtr & data_part_storage_builder, const IMergeTreeDataPart * parent_part); /// For mutation: MATERIALIZE PROJECTION. @@ -87,7 +85,6 @@ public: Poco::Logger * log, Block block, const ProjectionDescription & projection, - const DataPartStorageBuilderPtr & data_part_storage_builder, const IMergeTreeDataPart * parent_part, size_t block_num); @@ -97,7 +94,6 @@ public: Poco::Logger * log, Block block, const ProjectionDescription & projection, - const DataPartStorageBuilderPtr & data_part_storage_builder, const IMergeTreeDataPart * parent_part); static Block mergeBlock( @@ -112,7 +108,6 @@ private: const String & part_name, MergeTreeDataPartType part_type, const String & relative_path, - const DataPartStorageBuilderPtr & data_part_storage_builder, bool is_temp, const IMergeTreeDataPart * parent_part, const MergeTreeData & data, diff --git a/src/Storages/MergeTree/MergeTreePartition.cpp b/src/Storages/MergeTree/MergeTreePartition.cpp index 4ea6ec11ecc..47f07e7bb08 100644 --- a/src/Storages/MergeTree/MergeTreePartition.cpp +++ b/src/Storages/MergeTree/MergeTreePartition.cpp @@ -382,20 +382,20 @@ void MergeTreePartition::load(const MergeTreeData & storage, const PartMetadataM partition_key_sample.getByPosition(i).type->getDefaultSerialization()->deserializeBinary(value[i], *file); } -std::unique_ptr MergeTreePartition::store(const MergeTreeData & storage, const DataPartStorageBuilderPtr & data_part_storage_builder, MergeTreeDataPartChecksums & checksums) const +std::unique_ptr MergeTreePartition::store(const MergeTreeData & storage, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums) const { auto metadata_snapshot = storage.getInMemoryMetadataPtr(); const auto & context = storage.getContext(); const auto & partition_key_sample = adjustPartitionKey(metadata_snapshot, storage.getContext()).sample_block; - return store(partition_key_sample, data_part_storage_builder, checksums, context->getWriteSettings()); + return store(partition_key_sample, data_part_storage, checksums, context->getWriteSettings()); } -std::unique_ptr MergeTreePartition::store(const Block & partition_key_sample, const DataPartStorageBuilderPtr & data_part_storage_builder, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const +std::unique_ptr MergeTreePartition::store(const Block & partition_key_sample, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const { if (!partition_key_sample) return nullptr; - auto out = data_part_storage_builder->writeFile("partition.dat", DBMS_DEFAULT_BUFFER_SIZE, settings); + auto out = data_part_storage->writeFile("partition.dat", DBMS_DEFAULT_BUFFER_SIZE, settings); HashingWriteBuffer out_hashing(*out); for (size_t i = 0; i < value.size(); ++i) { diff --git a/src/Storages/MergeTree/MergeTreePartition.h b/src/Storages/MergeTree/MergeTreePartition.h index 6394641dfa3..c9c6723df5a 100644 --- a/src/Storages/MergeTree/MergeTreePartition.h +++ b/src/Storages/MergeTree/MergeTreePartition.h @@ -15,10 +15,10 @@ class MergeTreeData; struct FormatSettings; struct MergeTreeDataPartChecksums; struct StorageInMemoryMetadata; -class IDataPartStorageBuilder; +class IDataPartStorage; using StorageMetadataPtr = std::shared_ptr; -using DataPartStorageBuilderPtr = std::shared_ptr; +using MutableDataPartStoragePtr = std::shared_ptr; /// This class represents a partition value of a single part and encapsulates its loading/storing logic. struct MergeTreePartition @@ -44,8 +44,8 @@ public: /// Store functions return write buffer with written but not finalized data. /// User must call finish() for returned object. - [[nodiscard]] std::unique_ptr store(const MergeTreeData & storage, const DataPartStorageBuilderPtr & data_part_storage_builder, MergeTreeDataPartChecksums & checksums) const; - [[nodiscard]] std::unique_ptr store(const Block & partition_key_sample, const DataPartStorageBuilderPtr & data_part_storage_builder, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const; + [[nodiscard]] std::unique_ptr store(const MergeTreeData & storage, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums) const; + [[nodiscard]] std::unique_ptr store(const Block & partition_key_sample, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const; void assign(const MergeTreePartition & other) { value = other.value; } diff --git a/src/Storages/MergeTree/MergeTreePartsMover.cpp b/src/Storages/MergeTree/MergeTreePartsMover.cpp index afeeacbe5d6..5a3f138d727 100644 --- a/src/Storages/MergeTree/MergeTreePartsMover.cpp +++ b/src/Storages/MergeTree/MergeTreePartsMover.cpp @@ -209,8 +209,7 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt auto disk = moving_part.reserved_space->getDisk(); LOG_DEBUG(log, "Cloning part {} from '{}' to '{}'", part->name, part->data_part_storage->getDiskName(), disk->getName()); - DataPartStoragePtr cloned_part_storage; - + MutableDataPartStoragePtr cloned_part_storage; if (disk->supportZeroCopyReplication() && settings->allow_remote_fs_zero_copy_replication) { /// Try zero-copy replication and fallback to default copy if it's not possible @@ -225,12 +224,16 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt disk->createDirectories(path_to_clone); - cloned_part_storage = data->tryToFetchIfShared(*part, disk, fs::path(path_to_clone) / part->name); + bool is_fetched = data->tryToFetchIfShared(*part, disk, fs::path(path_to_clone) / part->name); - if (!cloned_part_storage) + if (!is_fetched) { LOG_INFO(log, "Part {} was not fetched, we are the first who move it to another disk, so we will copy it", part->name); - cloned_part_storage = part->data_part_storage->clone(path_to_clone, part->data_part_storage->getPartDirectory(), disk, log); + cloned_part_storage = part->data_part_storage->clonePart(path_to_clone, part->data_part_storage->getPartDirectory(), disk, log); + } + else + { + cloned_part_storage = part->data_part_storage->clone(); } } else @@ -238,14 +241,13 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt cloned_part_storage = part->makeCloneOnDisk(disk, MergeTreeData::MOVING_DIR_NAME); } - MergeTreeData::MutableDataPartPtr cloned_part = data->createPart(part->name, cloned_part_storage); + auto cloned_part = data->createPart(part->name, cloned_part_storage); LOG_TRACE(log, "Part {} was cloned to {}", part->name, cloned_part->data_part_storage->getFullPath()); cloned_part->loadColumnsChecksumsIndexes(true, true); cloned_part->loadVersionMetadata(); cloned_part->modification_time = cloned_part->data_part_storage->getLastModified().epochTime(); return cloned_part; - } @@ -263,11 +265,8 @@ void MergeTreePartsMover::swapClonedPart(const MergeTreeData::DataPartPtr & clon return; } - auto builder = cloned_part->data_part_storage->getBuilder(); /// Don't remove new directory but throw an error because it may contain part which is currently in use. - cloned_part->renameTo(active_part->name, false, builder); - - builder->commit(); + cloned_part->renameTo(active_part->name, false); /// TODO what happen if server goes down here? data->swapActivePart(cloned_part); diff --git a/src/Storages/MergeTree/MergeTreeSink.cpp b/src/Storages/MergeTree/MergeTreeSink.cpp index 5d00db861a8..6178af83913 100644 --- a/src/Storages/MergeTree/MergeTreeSink.cpp +++ b/src/Storages/MergeTree/MergeTreeSink.cpp @@ -160,7 +160,7 @@ void MergeTreeSink::finishDelayedChunk() } } - added = storage.renameTempPartAndAdd(part, transaction, partition.temp_part.builder, lock); + added = storage.renameTempPartAndAdd(part, transaction, lock); transaction.commit(&lock); } diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp index 4735eae8fdd..85b41f6dd83 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp @@ -150,7 +150,6 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore( while (!in->eof()) { MergeTreeData::MutableDataPartPtr part; - DataPartStorageBuilderPtr data_part_storage_builder; UInt8 version; String part_name; Block block; @@ -177,7 +176,6 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore( { auto single_disk_volume = std::make_shared("volume_" + part_name, disk, 0); auto data_part_storage = std::make_shared(single_disk_volume, storage.getRelativeDataPath(), part_name); - data_part_storage_builder = std::make_shared(single_disk_volume, storage.getRelativeDataPath(), part_name); part = storage.createPart( part_name, @@ -222,7 +220,6 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore( { MergedBlockOutputStream part_out( part, - data_part_storage_builder, metadata_snapshot, block.getNamesAndTypesList(), {}, @@ -240,11 +237,12 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore( for (const auto & projection : metadata_snapshot->getProjections()) { auto projection_block = projection.calculate(block, context); - auto temp_part = MergeTreeDataWriter::writeInMemoryProjectionPart(storage, log, projection_block, projection, data_part_storage_builder, part.get()); + auto temp_part = MergeTreeDataWriter::writeInMemoryProjectionPart(storage, log, projection_block, projection, part.get()); temp_part.finalize(); if (projection_block.rows()) part->addProjectionPart(projection.name, std::move(temp_part.part)); } + part_out.finalizePart(part, false); min_block_number = std::min(min_block_number, part->info.min_block); diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.cpp b/src/Storages/MergeTree/MergedBlockOutputStream.cpp index 269a78977ad..1d0743a0429 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.cpp +++ b/src/Storages/MergeTree/MergedBlockOutputStream.cpp @@ -14,8 +14,7 @@ namespace ErrorCodes MergedBlockOutputStream::MergedBlockOutputStream( - const MergeTreeDataPartPtr & data_part, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part, const StorageMetadataPtr & metadata_snapshot_, const NamesAndTypesList & columns_list_, const MergeTreeIndices & skip_indices, @@ -24,7 +23,7 @@ MergedBlockOutputStream::MergedBlockOutputStream( bool reset_columns_, bool blocks_are_granules_size, const WriteSettings & write_settings_) - : IMergedBlockOutputStream(std::move(data_part_storage_builder_), data_part, metadata_snapshot_, columns_list_, reset_columns_) + : IMergedBlockOutputStream(data_part, metadata_snapshot_, columns_list_, reset_columns_) , columns_list(columns_list_) , default_codec(default_codec_) , write_settings(write_settings_) @@ -38,7 +37,7 @@ MergedBlockOutputStream::MergedBlockOutputStream( blocks_are_granules_size); if (data_part->isStoredOnDisk()) - data_part_storage_builder->createDirectories(); + data_part_storage->createDirectories(); /// We should write version metadata on part creation to distinguish it from parts that were created without transaction. TransactionID tid = txn ? txn->tid : Tx::PrehistoricTID; @@ -47,7 +46,7 @@ MergedBlockOutputStream::MergedBlockOutputStream( data_part->version.setCreationTID(tid, nullptr); data_part->storeVersionMetadata(); - writer = data_part->getWriter(data_part_storage_builder, columns_list, metadata_snapshot, skip_indices, default_codec, writer_settings, {}); + writer = data_part->getWriter(columns_list, metadata_snapshot, skip_indices, default_codec, writer_settings, {}); } /// If data is pre-sorted. @@ -68,17 +67,17 @@ struct MergedBlockOutputStream::Finalizer::Impl { IMergeTreeDataPartWriter & writer; MergeTreeData::MutableDataPartPtr part; - DataPartStorageBuilderPtr data_part_storage_builder; NameSet files_to_remove_after_finish; std::vector> written_files; bool sync; - Impl(IMergeTreeDataPartWriter & writer_, MergeTreeData::MutableDataPartPtr part_, DataPartStorageBuilderPtr data_part_storage_builder_, const NameSet & files_to_remove_after_finish_, bool sync_) + Impl(IMergeTreeDataPartWriter & writer_, MergeTreeData::MutableDataPartPtr part_, const NameSet & files_to_remove_after_finish_, bool sync_) : writer(writer_) , part(std::move(part_)) - , data_part_storage_builder(std::move(data_part_storage_builder_)) , files_to_remove_after_finish(files_to_remove_after_finish_) - , sync(sync_) {} + , sync(sync_) + { + } void finish(); }; @@ -95,7 +94,7 @@ void MergedBlockOutputStream::Finalizer::Impl::finish() writer.finish(sync); for (const auto & file_name : files_to_remove_after_finish) - data_part_storage_builder->removeFile(file_name); + part->data_part_storage->removeFile(file_name); for (auto & file : written_files) { @@ -165,7 +164,7 @@ MergedBlockOutputStream::Finalizer MergedBlockOutputStream::finalizePartAsync( new_part->setColumns(part_columns, serialization_infos); } - auto finalizer = std::make_unique(*writer, new_part, data_part_storage_builder, files_to_remove_after_sync, sync); + auto finalizer = std::make_unique(*writer, new_part, files_to_remove_after_sync, sync); if (new_part->isStoredOnDisk()) finalizer->written_files = finalizePartOnDisk(new_part, checksums); @@ -192,7 +191,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { if (storage.format_version >= MERGE_TREE_DATA_MIN_FORMAT_VERSION_WITH_CUSTOM_PARTITIONING || isCompactPart(new_part)) { - auto count_out = data_part_storage_builder->writeFile("count.txt", 4096, write_settings); + auto count_out = new_part->data_part_storage->writeFile("count.txt", 4096, write_settings); HashingWriteBuffer count_out_hashing(*count_out); writeIntText(rows_count, count_out_hashing); count_out_hashing.next(); @@ -206,7 +205,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { if (new_part->uuid != UUIDHelpers::Nil) { - auto out = data_part_storage_builder->writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, write_settings); + auto out = new_part->data_part_storage->writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, write_settings); HashingWriteBuffer out_hashing(*out); writeUUIDText(new_part->uuid, out_hashing); checksums.files[IMergeTreeDataPart::UUID_FILE_NAME].file_size = out_hashing.count(); @@ -217,12 +216,12 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (storage.format_version >= MERGE_TREE_DATA_MIN_FORMAT_VERSION_WITH_CUSTOM_PARTITIONING) { - if (auto file = new_part->partition.store(storage, data_part_storage_builder, checksums)) + if (auto file = new_part->partition.store(storage, new_part->data_part_storage, checksums)) written_files.emplace_back(std::move(file)); if (new_part->minmax_idx->initialized) { - auto files = new_part->minmax_idx->store(storage, data_part_storage_builder, checksums); + auto files = new_part->minmax_idx->store(storage, new_part->data_part_storage, checksums); for (auto & file : files) written_files.emplace_back(std::move(file)); } @@ -232,7 +231,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis } { - auto count_out = data_part_storage_builder->writeFile("count.txt", 4096, write_settings); + auto count_out = new_part->data_part_storage->writeFile("count.txt", 4096, write_settings); HashingWriteBuffer count_out_hashing(*count_out); writeIntText(rows_count, count_out_hashing); count_out_hashing.next(); @@ -246,7 +245,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (!new_part->ttl_infos.empty()) { /// Write a file with ttl infos in json format. - auto out = data_part_storage_builder->writeFile("ttl.txt", 4096, write_settings); + auto out = new_part->data_part_storage->writeFile("ttl.txt", 4096, write_settings); HashingWriteBuffer out_hashing(*out); new_part->ttl_infos.write(out_hashing); checksums.files["ttl.txt"].file_size = out_hashing.count(); @@ -257,7 +256,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (!new_part->getSerializationInfos().empty()) { - auto out = data_part_storage_builder->writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, write_settings); + auto out = new_part->data_part_storage->writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, write_settings); HashingWriteBuffer out_hashing(*out); new_part->getSerializationInfos().writeJSON(out_hashing); checksums.files[IMergeTreeDataPart::SERIALIZATION_FILE_NAME].file_size = out_hashing.count(); @@ -268,7 +267,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { /// Write a file with a description of columns. - auto out = data_part_storage_builder->writeFile("columns.txt", 4096, write_settings); + auto out = new_part->data_part_storage->writeFile("columns.txt", 4096, write_settings); new_part->getColumns().writeText(*out); out->preFinalize(); written_files.emplace_back(std::move(out)); @@ -276,7 +275,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (default_codec != nullptr) { - auto out = data_part_storage_builder->writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, write_settings); + auto out = new_part->data_part_storage->writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, write_settings); DB::writeText(queryToString(default_codec->getFullCodecDesc()), *out); out->preFinalize(); written_files.emplace_back(std::move(out)); @@ -289,7 +288,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { /// Write file with checksums. - auto out = data_part_storage_builder->writeFile("checksums.txt", 4096, write_settings); + auto out = new_part->data_part_storage->writeFile("checksums.txt", 4096, write_settings); checksums.write(*out); out->preFinalize(); written_files.emplace_back(std::move(out)); diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.h b/src/Storages/MergeTree/MergedBlockOutputStream.h index 92dcd8dd272..799bae8e94b 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.h +++ b/src/Storages/MergeTree/MergedBlockOutputStream.h @@ -15,8 +15,7 @@ class MergedBlockOutputStream final : public IMergedBlockOutputStream { public: MergedBlockOutputStream( - const MergeTreeDataPartPtr & data_part, - DataPartStorageBuilderPtr data_part_storage_builder_, + const MergeTreeMutableDataPartPtr & data_part, const StorageMetadataPtr & metadata_snapshot_, const NamesAndTypesList & columns_list_, const MergeTreeIndices & skip_indices, diff --git a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp index dd75cddd380..310bc849ffe 100644 --- a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp +++ b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp @@ -11,8 +11,7 @@ namespace ErrorCodes } MergedColumnOnlyOutputStream::MergedColumnOnlyOutputStream( - DataPartStorageBuilderPtr data_part_storage_builder_, - const MergeTreeDataPartPtr & data_part, + const MergeTreeMutableDataPartPtr & data_part, const StorageMetadataPtr & metadata_snapshot_, const Block & header_, CompressionCodecPtr default_codec, @@ -20,7 +19,7 @@ MergedColumnOnlyOutputStream::MergedColumnOnlyOutputStream( WrittenOffsetColumns * offset_columns_, const MergeTreeIndexGranularity & index_granularity, const MergeTreeIndexGranularityInfo * index_granularity_info) - : IMergedBlockOutputStream(std::move(data_part_storage_builder_), data_part, metadata_snapshot_, header_.getNamesAndTypesList(), /*reset_columns=*/ true) + : IMergedBlockOutputStream(data_part, metadata_snapshot_, header_.getNamesAndTypesList(), /*reset_columns=*/ true) , header(header_) { const auto & global_settings = data_part->storage.getContext()->getSettings(); @@ -34,7 +33,6 @@ MergedColumnOnlyOutputStream::MergedColumnOnlyOutputStream( /* rewrite_primary_key = */ false); writer = data_part->getWriter( - data_part_storage_builder, header.getNamesAndTypesList(), metadata_snapshot_, indices_to_recalc, @@ -81,7 +79,7 @@ MergedColumnOnlyOutputStream::fillChecksums( for (const String & removed_file : removed_files) { - data_part_storage_builder->removeFileIfExists(removed_file); + new_part->data_part_storage->removeFileIfExists(removed_file); if (all_checksums.files.contains(removed_file)) all_checksums.files.erase(removed_file); diff --git a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h index 1fd1c752226..f382b0fef60 100644 --- a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h +++ b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.h @@ -14,8 +14,7 @@ public: /// Pass empty 'already_written_offset_columns' first time then and pass the same object to subsequent instances of MergedColumnOnlyOutputStream /// if you want to serialize elements of Nested data structure in different instances of MergedColumnOnlyOutputStream. MergedColumnOnlyOutputStream( - DataPartStorageBuilderPtr data_part_storage_builder_, - const MergeTreeDataPartPtr & data_part, + const MergeTreeMutableDataPartPtr & data_part, const StorageMetadataPtr & metadata_snapshot_, const Block & header_, CompressionCodecPtr default_codec_, diff --git a/src/Storages/MergeTree/MutateFromLogEntryTask.cpp b/src/Storages/MergeTree/MutateFromLogEntryTask.cpp index 549c4e7373f..a4a75a637a4 100644 --- a/src/Storages/MergeTree/MutateFromLogEntryTask.cpp +++ b/src/Storages/MergeTree/MutateFromLogEntryTask.cpp @@ -193,12 +193,7 @@ ReplicatedMergeMutateTaskBase::PrepareResult MutateFromLogEntryTask::prepare() bool MutateFromLogEntryTask::finalize(ReplicatedMergeMutateTaskBase::PartLogWriter write_part_log) { new_part = mutate_task->getFuture().get(); - auto builder = mutate_task->getBuilder(); - - if (!builder) - builder = new_part->data_part_storage->getBuilder(); - - storage.renameTempPartAndReplace(new_part, *transaction_ptr, builder); + storage.renameTempPartAndReplace(new_part, *transaction_ptr); try { diff --git a/src/Storages/MergeTree/MutatePlainMergeTreeTask.cpp b/src/Storages/MergeTree/MutatePlainMergeTreeTask.cpp index 0cf10ee1935..b1714076a46 100644 --- a/src/Storages/MergeTree/MutatePlainMergeTreeTask.cpp +++ b/src/Storages/MergeTree/MutatePlainMergeTreeTask.cpp @@ -83,14 +83,9 @@ bool MutatePlainMergeTreeTask::executeStep() new_part = mutate_task->getFuture().get(); - auto builder = mutate_task->getBuilder(); - if (!builder) - builder = new_part->data_part_storage->getBuilder(); - - MergeTreeData::Transaction transaction(storage, merge_mutate_entry->txn.get()); /// FIXME Transactions: it's too optimistic, better to lock parts before starting transaction - storage.renameTempPartAndReplace(new_part, transaction, builder); + storage.renameTempPartAndReplace(new_part, transaction); transaction.commit(); storage.updateMutationEntriesErrors(future_part, true, ""); diff --git a/src/Storages/MergeTree/MutateTask.cpp b/src/Storages/MergeTree/MutateTask.cpp index 3d964e60798..80bf7e3c7d9 100644 --- a/src/Storages/MergeTree/MutateTask.cpp +++ b/src/Storages/MergeTree/MutateTask.cpp @@ -619,7 +619,6 @@ static NameToNameVector collectFilesForRenames( /// Initialize and write to disk new part fields like checksums, columns, etc. void finalizeMutatedPart( const MergeTreeDataPartPtr & source_part, - const DataPartStorageBuilderPtr & data_part_storage_builder, MergeTreeData::MutableDataPartPtr new_data_part, ExecuteTTLType execute_ttl_type, const CompressionCodecPtr & codec, @@ -627,7 +626,7 @@ void finalizeMutatedPart( { if (new_data_part->uuid != UUIDHelpers::Nil) { - auto out = data_part_storage_builder->writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, context->getWriteSettings()); + auto out = new_data_part->data_part_storage->writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, context->getWriteSettings()); HashingWriteBuffer out_hashing(*out); writeUUIDText(new_data_part->uuid, out_hashing); new_data_part->checksums.files[IMergeTreeDataPart::UUID_FILE_NAME].file_size = out_hashing.count(); @@ -637,7 +636,7 @@ void finalizeMutatedPart( if (execute_ttl_type != ExecuteTTLType::NONE) { /// Write a file with ttl infos in json format. - auto out_ttl = data_part_storage_builder->writeFile("ttl.txt", 4096, context->getWriteSettings()); + auto out_ttl = new_data_part->data_part_storage->writeFile("ttl.txt", 4096, context->getWriteSettings()); HashingWriteBuffer out_hashing(*out_ttl); new_data_part->ttl_infos.write(out_hashing); new_data_part->checksums.files["ttl.txt"].file_size = out_hashing.count(); @@ -646,7 +645,7 @@ void finalizeMutatedPart( if (!new_data_part->getSerializationInfos().empty()) { - auto out = data_part_storage_builder->writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, context->getWriteSettings()); + auto out = new_data_part->data_part_storage->writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, context->getWriteSettings()); HashingWriteBuffer out_hashing(*out); new_data_part->getSerializationInfos().writeJSON(out_hashing); new_data_part->checksums.files[IMergeTreeDataPart::SERIALIZATION_FILE_NAME].file_size = out_hashing.count(); @@ -655,18 +654,18 @@ void finalizeMutatedPart( { /// Write file with checksums. - auto out_checksums = data_part_storage_builder->writeFile("checksums.txt", 4096, context->getWriteSettings()); + auto out_checksums = new_data_part->data_part_storage->writeFile("checksums.txt", 4096, context->getWriteSettings()); new_data_part->checksums.write(*out_checksums); } /// close fd { - auto out = data_part_storage_builder->writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, context->getWriteSettings()); + auto out = new_data_part->data_part_storage->writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, context->getWriteSettings()); DB::writeText(queryToString(codec->getFullCodecDesc()), *out); } /// close fd { /// Write a file with a description of columns. - auto out_columns = data_part_storage_builder->writeFile("columns.txt", 4096, context->getWriteSettings()); + auto out_columns = new_data_part->data_part_storage->writeFile("columns.txt", 4096, context->getWriteSettings()); new_data_part->getColumns().writeText(*out_columns); } /// close fd @@ -734,8 +733,6 @@ struct MutationContext = MutationsInterpreter::MutationKind::MutationKindEnum::MUTATE_UNKNOWN; MergeTreeData::MutableDataPartPtr new_data_part; - DataPartStorageBuilderPtr data_part_storage_builder; - IMergedBlockOutputStreamPtr out{nullptr}; String mrk_extension; @@ -816,11 +813,9 @@ public: if (next_level_parts.empty()) { LOG_DEBUG(log, "Merged a projection part in level {}", current_level); - auto builder = selected_parts[0]->data_part_storage->getBuilder(); - selected_parts[0]->renameTo(projection.name + ".proj", true, builder); + selected_parts[0]->renameTo(projection.name + ".proj", true); selected_parts[0]->name = projection.name; selected_parts[0]->is_temp = false; - builder->commit(); ctx->new_data_part->addProjectionPart(name, std::move(selected_parts[0])); /// Task is finished @@ -865,7 +860,6 @@ public: projection_merging_params, NO_TRANSACTION_PTR, ctx->new_data_part.get(), - ctx->data_part_storage_builder.get(), ".tmp_proj"); next_level_parts.push_back(executeHere(tmp_part_merge_task)); @@ -1025,8 +1019,7 @@ bool PartMergerWriter::mutateOriginalPartAndPrepareProjections() if (projection_block) { auto tmp_part = MergeTreeDataWriter::writeTempProjectionPart( - *ctx->data, ctx->log, projection_block, projection, ctx->data_part_storage_builder, ctx->new_data_part.get(), ++block_num); - tmp_part.builder->commit(); + *ctx->data, ctx->log, projection_block, projection, ctx->new_data_part.get(), ++block_num); tmp_part.finalize(); projection_parts[projection.name].emplace_back(std::move(tmp_part.part)); } @@ -1048,8 +1041,7 @@ bool PartMergerWriter::mutateOriginalPartAndPrepareProjections() if (projection_block) { auto temp_part = MergeTreeDataWriter::writeTempProjectionPart( - *ctx->data, ctx->log, projection_block, projection, ctx->data_part_storage_builder, ctx->new_data_part.get(), ++block_num); - temp_part.builder->commit(); + *ctx->data, ctx->log, projection_block, projection, ctx->new_data_part.get(), ++block_num); temp_part.finalize(); projection_parts[projection.name].emplace_back(std::move(temp_part.part)); } @@ -1149,7 +1141,7 @@ private: void prepare() { - ctx->data_part_storage_builder->createDirectories(); + ctx->new_data_part->data_part_storage->createDirectories(); /// Note: this is done before creating input streams, because otherwise data.data_parts_mutex /// (which is locked in data.getTotalActiveSizeInBytes()) @@ -1184,7 +1176,6 @@ private: ctx->out = std::make_shared( ctx->new_data_part, - ctx->data_part_storage_builder, ctx->metadata_snapshot, ctx->new_data_part->getColumns(), skip_part_indices, @@ -1280,7 +1271,7 @@ private: if (ctx->execute_ttl_type != ExecuteTTLType::NONE) ctx->files_to_skip.insert("ttl.txt"); - ctx->data_part_storage_builder->createDirectories(); + ctx->new_data_part->data_part_storage->createDirectories(); /// We should write version metadata on part creation to distinguish it from parts that were created without transaction. TransactionID tid = ctx->txn ? ctx->txn->tid : Tx::PrehistoricTID; @@ -1317,22 +1308,22 @@ private: if (it->isFile()) { - ctx->data_part_storage_builder->createHardLinkFrom( + ctx->new_data_part->data_part_storage->createHardLinkFrom( *ctx->source_part->data_part_storage, it->name(), destination); hardlinked_files.insert(it->name()); } else if (!endsWith(it->name(), ".tmp_proj")) // ignore projection tmp merge dir { // it's a projection part directory - ctx->data_part_storage_builder->createProjection(destination); + ctx->new_data_part->data_part_storage->createProjection(destination); - auto projection_data_part_storage = ctx->source_part->data_part_storage->getProjection(destination); - auto projection_data_part_storage_builder = ctx->data_part_storage_builder->getProjection(destination); + auto projection_data_part_storage_src = ctx->source_part->data_part_storage->getProjection(destination); + auto projection_data_part_storage_dst = ctx->new_data_part->data_part_storage->getProjection(destination); - for (auto p_it = projection_data_part_storage->iterate(); p_it->isValid(); p_it->next()) + for (auto p_it = projection_data_part_storage_src->iterate(); p_it->isValid(); p_it->next()) { - projection_data_part_storage_builder->createHardLinkFrom( - *projection_data_part_storage, p_it->name(), p_it->name()); + projection_data_part_storage_dst->createHardLinkFrom( + *projection_data_part_storage_src, p_it->name(), p_it->name()); hardlinked_files.insert(p_it->name()); } } @@ -1362,7 +1353,6 @@ private: builder.addTransform(std::make_shared(builder.getHeader(), *ctx->data, ctx->metadata_snapshot, ctx->new_data_part, ctx->time_of_mutation, true)); ctx->out = std::make_shared( - ctx->data_part_storage_builder, ctx->new_data_part, ctx->metadata_snapshot, ctx->updated_header, @@ -1414,7 +1404,7 @@ private: } } - MutationHelpers::finalizeMutatedPart(ctx->source_part, ctx->data_part_storage_builder, ctx->new_data_part, ctx->execute_ttl_type, ctx->compression_codec, ctx->context); + MutationHelpers::finalizeMutatedPart(ctx->source_part, ctx->new_data_part, ctx->execute_ttl_type, ctx->compression_codec, ctx->context); } @@ -1584,10 +1574,7 @@ bool MutateTask::prepare() ctx->data->getRelativeDataPath(), tmp_part_dir_name); - ctx->data_part_storage_builder = std::make_shared( - single_disk_volume, - ctx->data->getRelativeDataPath(), - tmp_part_dir_name); + data_part_storage->beginTransaction(); ctx->new_data_part = ctx->data->createPart( ctx->future_part->name, ctx->future_part->type, ctx->future_part->part_info, data_part_storage); @@ -1690,9 +1677,4 @@ const MergeTreeData::HardlinkedFiles & MutateTask::getHardlinkedFiles() const return ctx->hardlinked_files; } -DataPartStorageBuilderPtr MutateTask::getBuilder() const -{ - return ctx->data_part_storage_builder; -} - } diff --git a/src/Storages/MergeTree/MutateTask.h b/src/Storages/MergeTree/MutateTask.h index 1f2e8a6fd20..3df30670d7f 100644 --- a/src/Storages/MergeTree/MutateTask.h +++ b/src/Storages/MergeTree/MutateTask.h @@ -46,7 +46,7 @@ public: const MergeTreeData::HardlinkedFiles & getHardlinkedFiles() const; - DataPartStorageBuilderPtr getBuilder() const; + MutableDataPartStoragePtr getBuilder() const; private: diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp index 158cbfca9fd..293c679dcd0 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp +++ b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp @@ -268,7 +268,7 @@ void ReplicatedMergeTreeSink::finishDelayedChunk(zkutil::ZooKeeperPtr & zookeepe try { - commitPart(zookeeper, part, partition.block_id, partition.temp_part.builder, delayed_chunk->replicas_num); + commitPart(zookeeper, part, partition.block_id, delayed_chunk->replicas_num); last_block_is_duplicate = last_block_is_duplicate || part->is_duplicate; @@ -301,7 +301,7 @@ void ReplicatedMergeTreeSink::writeExistingPart(MergeTreeData::MutableDataPartPt try { part->version.setCreationTID(Tx::PrehistoricTID, nullptr); - commitPart(zookeeper, part, "", part->data_part_storage->getBuilder(), replicas_num); + commitPart(zookeeper, part, "", replicas_num); PartLog::addNewPart(storage.getContext(), part, watch.elapsed()); } catch (...) @@ -315,7 +315,6 @@ void ReplicatedMergeTreeSink::commitPart( zkutil::ZooKeeperPtr & zookeeper, MergeTreeData::MutableDataPartPtr & part, const String & block_id, - DataPartStorageBuilderPtr builder, size_t replicas_num) { /// It is possible that we alter a part with different types of source columns. @@ -499,7 +498,7 @@ void ReplicatedMergeTreeSink::commitPart( try { auto lock = storage.lockParts(); - renamed = storage.renameTempPartAndAdd(part, transaction, builder, lock); + renamed = storage.renameTempPartAndAdd(part, transaction, lock); } catch (const Exception & e) { @@ -563,8 +562,7 @@ void ReplicatedMergeTreeSink::commitPart( transaction.rollbackPartsToTemporaryState(); part->is_temp = true; - part->renameTo(temporary_part_relative_path, false, builder); - builder->commit(); + part->renameTo(temporary_part_relative_path, false); /// If this part appeared on other replica than it's better to try to write it locally one more time. If it's our part /// than it will be ignored on the next itration. diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeSink.h b/src/Storages/MergeTree/ReplicatedMergeTreeSink.h index ab729e6edec..da87ddc0d63 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeSink.h +++ b/src/Storages/MergeTree/ReplicatedMergeTreeSink.h @@ -79,7 +79,6 @@ private: zkutil::ZooKeeperPtr & zookeeper, MergeTreeData::MutableDataPartPtr & part, const String & block_id, - DataPartStorageBuilderPtr part_builder, size_t replicas_num); /// Wait for quorum to be satisfied on path (quorum_path) form part (part_name) diff --git a/src/Storages/StorageMergeTree.cpp b/src/Storages/StorageMergeTree.cpp index 7cfce882e7a..1aeca1343c2 100644 --- a/src/Storages/StorageMergeTree.cpp +++ b/src/Storages/StorageMergeTree.cpp @@ -1519,9 +1519,8 @@ PartitionCommandsResultInfo StorageMergeTree::attachPartition( MergeTreeData::Transaction transaction(*this, local_context->getCurrentTransaction().get()); { auto lock = lockParts(); - auto builder = loaded_parts[i]->data_part_storage->getBuilder(); fillNewPartName(loaded_parts[i], lock); - renameTempPartAndAdd(loaded_parts[i], transaction, builder, lock); + renameTempPartAndAdd(loaded_parts[i], transaction, lock); transaction.commit(&lock); } @@ -1604,9 +1603,7 @@ void StorageMergeTree::replacePartitionFrom(const StoragePtr & source_table, con for (auto part : dst_parts) { fillNewPartName(part, data_parts_lock); - - auto builder = part->data_part_storage->getBuilder(); - renameTempPartAndReplaceUnlocked(part, transaction, builder, data_parts_lock); + renameTempPartAndReplaceUnlocked(part, transaction, data_parts_lock); } /// Populate transaction transaction.commit(&data_parts_lock); @@ -1685,9 +1682,8 @@ void StorageMergeTree::movePartitionToTable(const StoragePtr & dest_table, const for (auto & part : dst_parts) { - auto builder = part->data_part_storage->getBuilder(); dest_table_storage->fillNewPartName(part, dest_data_parts_lock); - dest_table_storage->renameTempPartAndReplaceUnlocked(part, transaction, builder, dest_data_parts_lock); + dest_table_storage->renameTempPartAndReplaceUnlocked(part, transaction, dest_data_parts_lock); } @@ -1810,17 +1806,15 @@ BackupEntries StorageMergeTree::backupMutations(UInt64 version, const String & d void StorageMergeTree::attachRestoredParts(MutableDataPartsVector && parts) { - for (auto part : parts) { /// It's important to create it outside of lock scope because /// otherwise it can lock parts in destructor and deadlock is possible. MergeTreeData::Transaction transaction(*this, NO_TRANSACTION_RAW); - auto builder = part->data_part_storage->getBuilder(); { auto lock = lockParts(); fillNewPartName(part, lock); - renameTempPartAndAdd(part, transaction, builder, lock); + renameTempPartAndAdd(part, transaction, lock); transaction.commit(&lock); } } diff --git a/src/Storages/StorageReplicatedMergeTree.cpp b/src/Storages/StorageReplicatedMergeTree.cpp index ef2d91fabfc..52a30dd37b3 100644 --- a/src/Storages/StorageReplicatedMergeTree.cpp +++ b/src/Storages/StorageReplicatedMergeTree.cpp @@ -1454,6 +1454,7 @@ MergeTreeData::MutableDataPartPtr StorageReplicatedMergeTree::attachPartHelperFo const String part_new_name = actual_part_info.getPartName(); for (const DiskPtr & disk : getStoragePolicy()->getDisks()) + { for (const auto it = disk->iterateDirectory(fs::path(relative_data_path) / "detached/"); it->isValid(); it->next()) { const auto part_info = MergeTreePartInfo::tryParsePartName(it->name(), format_version); @@ -1490,6 +1491,7 @@ MergeTreeData::MutableDataPartPtr StorageReplicatedMergeTree::attachPartHelperFo return part; } } + } return {}; } @@ -1540,8 +1542,7 @@ bool StorageReplicatedMergeTree::executeLogEntry(LogEntry & entry) Transaction transaction(*this, NO_TRANSACTION_RAW); part->version.setCreationTID(Tx::PrehistoricTID, nullptr); - auto builder = part->data_part_storage->getBuilder(); - renameTempPartAndReplace(part, transaction, builder); + renameTempPartAndReplace(part, transaction); checkPartChecksumsAndCommit(transaction, part); writePartLog(PartLogElement::Type::NEW_PART, {}, 0 /** log entry is fake so we don't measure the time */, @@ -1780,7 +1781,7 @@ bool StorageReplicatedMergeTree::executeFetch(LogEntry & entry, bool need_to_che } -DataPartStoragePtr StorageReplicatedMergeTree::executeFetchShared( +bool StorageReplicatedMergeTree::executeFetchShared( const String & source_replica, const String & new_part_name, const DiskPtr & disk, @@ -1789,7 +1790,7 @@ DataPartStoragePtr StorageReplicatedMergeTree::executeFetchShared( if (source_replica.empty()) { LOG_INFO(log, "No active replica has part {} on shared storage.", new_part_name); - return nullptr; + return false; } const auto storage_settings_ptr = getSettings(); @@ -2225,8 +2226,7 @@ bool StorageReplicatedMergeTree::executeReplaceRange(const LogEntry & entry) Coordination::Requests ops; for (PartDescriptionPtr & part_desc : final_parts) { - auto builder = part_desc->res_part->data_part_storage->getBuilder(); - renameTempPartAndReplace(part_desc->res_part, transaction, builder); + renameTempPartAndReplace(part_desc->res_part, transaction); getCommitPartOps(ops, part_desc->res_part); lockSharedData(*part_desc->res_part, false, part_desc->hardlinked_files); @@ -2323,9 +2323,7 @@ void StorageReplicatedMergeTree::executeClonePartFromShard(const LogEntry & entr part = get_part(); // The fetched part is valuable and should not be cleaned like a temp part. part->is_temp = false; - auto builder = part->data_part_storage->getBuilder(); - part->renameTo("detached/" + entry.new_part_name, true, builder); - builder->commit(); + part->renameTo("detached/" + entry.new_part_name, true); LOG_INFO(log, "Cloned part {} to detached directory", part->name); } @@ -3990,11 +3988,10 @@ bool StorageReplicatedMergeTree::fetchPart(const String & part_name, const Stora { part = get_part(); - auto builder = part->data_part_storage->getBuilder(); if (!to_detached) { Transaction transaction(*this, NO_TRANSACTION_RAW); - renameTempPartAndReplace(part, transaction, builder); + renameTempPartAndReplace(part, transaction); replaced_parts = checkPartChecksumsAndCommit(transaction, part, hardlinked_files); @@ -4036,8 +4033,7 @@ bool StorageReplicatedMergeTree::fetchPart(const String & part_name, const Stora { // The fetched part is valuable and should not be cleaned like a temp part. part->is_temp = false; - part->renameTo(fs::path("detached") / part_name, true, builder); - builder->commit(); + part->renameTo(fs::path("detached") / part_name, true); } } catch (const Exception & e) @@ -4071,7 +4067,7 @@ bool StorageReplicatedMergeTree::fetchPart(const String & part_name, const Stora } -DataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( +bool StorageReplicatedMergeTree::fetchExistsPart( const String & part_name, const StorageMetadataPtr & metadata_snapshot, const String & source_replica_path, @@ -4086,7 +4082,7 @@ DataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( LOG_DEBUG(log, "Part {} should be deleted after previous attempt before fetch", part->name); /// Force immediate parts cleanup to delete the part that was left from the previous fetch attempt. cleanup_thread.wakeup(); - return nullptr; + return false; } { @@ -4094,7 +4090,7 @@ DataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( if (!currently_fetching_parts.insert(part_name).second) { LOG_DEBUG(log, "Part {} is already fetching right now", part_name); - return nullptr; + return false; } } @@ -4150,10 +4146,7 @@ DataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( throw Exception("Part " + part->name + " fetched on wrong disk " + part->data_part_storage->getDiskName(), ErrorCodes::LOGICAL_ERROR); auto replaced_path = fs::path(replaced_part_path); - auto builder = part->data_part_storage->getBuilder(); - builder->rename(replaced_path.parent_path(), replaced_path.filename(), nullptr, true, false); - part->data_part_storage->onRename(replaced_path.parent_path(), replaced_path.filename()); - builder->commit(); + part->data_part_storage->rename(replaced_path.parent_path(), replaced_path.filename(), nullptr, true, false); } catch (const Exception & e) { @@ -4162,7 +4155,7 @@ DataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( if (e.code() == ErrorCodes::DIRECTORY_ALREADY_EXISTS) { LOG_TRACE(log, "Not fetching part: {}", e.message()); - return nullptr; + return false; } throw; @@ -4176,8 +4169,7 @@ DataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( ProfileEvents::increment(ProfileEvents::ReplicatedPartFetches); LOG_DEBUG(log, "Fetched part {} from {}", part_name, source_replica_path); - - return part->data_part_storage; + return true; } void StorageReplicatedMergeTree::startup() @@ -6592,10 +6584,7 @@ void StorageReplicatedMergeTree::replacePartitionFrom( { auto data_parts_lock = lockParts(); for (auto & part : dst_parts) - { - auto builder = part->data_part_storage->getBuilder(); - renameTempPartAndReplaceUnlocked(part, transaction, builder, data_parts_lock); - } + renameTempPartAndReplaceUnlocked(part, transaction, data_parts_lock); } for (size_t i = 0; i < dst_parts.size(); ++i) @@ -6831,10 +6820,7 @@ void StorageReplicatedMergeTree::movePartitionToTable(const StoragePtr & dest_ta auto dest_data_parts_lock = dest_table_storage->lockParts(); for (auto & part : dst_parts) - { - auto builder = part->data_part_storage->getBuilder(); - dest_table_storage->renameTempPartAndReplaceUnlocked(part, transaction, builder, dest_data_parts_lock); - } + dest_table_storage->renameTempPartAndReplaceUnlocked(part, transaction, dest_data_parts_lock); for (size_t i = 0; i < dst_parts.size(); ++i) dest_table_storage->lockSharedData(*dst_parts[i], false, hardlinked_files_for_parts[i]); @@ -7888,7 +7874,7 @@ std::pair StorageReplicatedMergeTree::unlockSharedDataByID( } -DataPartStoragePtr StorageReplicatedMergeTree::tryToFetchIfShared( +bool StorageReplicatedMergeTree::tryToFetchIfShared( const IMergeTreeDataPart & part, const DiskPtr & disk, const String & path) @@ -7896,13 +7882,13 @@ DataPartStoragePtr StorageReplicatedMergeTree::tryToFetchIfShared( const auto settings = getSettings(); auto data_source_description = disk->getDataSourceDescription(); if (!(disk->supportZeroCopyReplication() && settings->allow_remote_fs_zero_copy_replication)) - return nullptr; + return false; String replica = getSharedDataReplica(part, data_source_description.type); /// We can't fetch part when none replicas have this part on a same type remote disk if (replica.empty()) - return nullptr; + return false; return executeFetchShared(replica, part.name, disk, path); } @@ -8123,15 +8109,13 @@ bool StorageReplicatedMergeTree::createEmptyPartInsteadOfLost(zkutil::ZooKeeperP minmax_idx->update(block, getMinMaxColumnsNames(metadata_snapshot->getPartitionKey())); auto new_volume = createVolumeFromReservation(reservation, volume); + auto data_part_storage = std::make_shared( new_volume, relative_data_path, TMP_PREFIX + lost_part_name); - DataPartStorageBuilderPtr data_part_storage_builder = std::make_shared( - new_volume, - relative_data_path, - TMP_PREFIX + lost_part_name); + data_part_storage->beginTransaction(); auto new_data_part = createPart( lost_part_name, @@ -8174,16 +8158,16 @@ bool StorageReplicatedMergeTree::createEmptyPartInsteadOfLost(zkutil::ZooKeeperP if (new_data_part->isStoredOnDisk()) { /// The name could be non-unique in case of stale files from previous runs. - if (data_part_storage_builder->exists()) + if (data_part_storage->exists()) { LOG_WARNING(log, "Removing old temporary directory {}", new_data_part->data_part_storage->getFullPath()); - data_part_storage_builder->removeRecursive(); + data_part_storage->removeRecursive(); } - data_part_storage_builder->createDirectories(); + data_part_storage->createDirectories(); if (getSettings()->fsync_part_directory) - sync_guard = data_part_storage_builder->getDirectorySyncGuard(); + sync_guard = data_part_storage->getDirectorySyncGuard(); } /// This effectively chooses minimal compression method: @@ -8191,7 +8175,7 @@ bool StorageReplicatedMergeTree::createEmptyPartInsteadOfLost(zkutil::ZooKeeperP auto compression_codec = getContext()->chooseCompressionCodec(0, 0); const auto & index_factory = MergeTreeIndexFactory::instance(); - MergedBlockOutputStream out(new_data_part, data_part_storage_builder, metadata_snapshot, columns, + MergedBlockOutputStream out(new_data_part, metadata_snapshot, columns, index_factory.getMany(metadata_snapshot->getSecondaryIndices()), compression_codec, NO_TRANSACTION_PTR); bool sync_on_insert = settings->fsync_after_insert; @@ -8205,7 +8189,7 @@ bool StorageReplicatedMergeTree::createEmptyPartInsteadOfLost(zkutil::ZooKeeperP try { MergeTreeData::Transaction transaction(*this, NO_TRANSACTION_RAW); - auto replaced_parts = renameTempPartAndReplace(new_data_part, transaction, data_part_storage_builder); + auto replaced_parts = renameTempPartAndReplace(new_data_part, transaction); if (!replaced_parts.empty()) { diff --git a/src/Storages/StorageReplicatedMergeTree.h b/src/Storages/StorageReplicatedMergeTree.h index e10ffcce22c..c9af1ab5f93 100644 --- a/src/Storages/StorageReplicatedMergeTree.h +++ b/src/Storages/StorageReplicatedMergeTree.h @@ -263,7 +263,7 @@ public: bool canExecuteFetch(const ReplicatedMergeTreeLogEntry & entry, String & disable_reason) const; /// Fetch part only when it stored on shared storage like S3 - DataPartStoragePtr executeFetchShared(const String & source_replica, const String & new_part_name, const DiskPtr & disk, const String & path); + bool executeFetchShared(const String & source_replica, const String & new_part_name, const DiskPtr & disk, const String & path); /// Lock part in zookeeper for use shared data in several nodes void lockSharedData(const IMergeTreeDataPart & part, bool replace_existing_lock, std::optional hardlinked_files) const override; @@ -283,7 +283,7 @@ public: const String & zookeeper_path_old, MergeTreeDataFormatVersion data_format_version); /// Fetch part only if some replica has it on shared storage like S3 - DataPartStoragePtr tryToFetchIfShared(const IMergeTreeDataPart & part, const DiskPtr & disk, const String & path) override; + bool tryToFetchIfShared(const IMergeTreeDataPart & part, const DiskPtr & disk, const String & path) override; /// Get best replica having this partition on a same type remote disk String getSharedDataReplica(const IMergeTreeDataPart & part, DataSourceType data_source_type) const; @@ -682,7 +682,7 @@ private: * Used for replace local part on the same s3-shared part in hybrid storage. * Returns false if part is already fetching right now. */ - DataPartStoragePtr fetchExistsPart( + bool fetchExistsPart( const String & part_name, const StorageMetadataPtr & metadata_snapshot, const String & replica_path, From b40d9200d20c0ce92ef40c601087ddb576575ba7 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Sun, 23 Oct 2022 03:29:26 +0000 Subject: [PATCH 066/239] better semantic of constsness of DataPartStorage --- .../getCompressionCodecForFile.cpp | 4 +- src/Compression/getCompressionCodecForFile.h | 2 +- src/Interpreters/PartLog.cpp | 4 +- .../MergeTree/DataPartStorageOnDisk.cpp | 26 +--- .../MergeTree/DataPartStorageOnDisk.h | 11 +- src/Storages/MergeTree/DataPartsExchange.cpp | 24 +-- src/Storages/MergeTree/IDataPartStorage.h | 36 +++-- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 138 +++++++++--------- src/Storages/MergeTree/IMergeTreeDataPart.h | 18 +-- .../MergeTree/IMergedBlockOutputStream.cpp | 2 +- .../LoadedMergeTreeDataPartInfoForReader.h | 2 +- .../MergeTree/MergeFromLogEntryTask.cpp | 2 +- src/Storages/MergeTree/MergeList.cpp | 2 +- src/Storages/MergeTree/MergeTask.cpp | 2 +- src/Storages/MergeTree/MergeTask.h | 2 +- src/Storages/MergeTree/MergeTreeData.cpp | 131 +++++++++-------- src/Storages/MergeTree/MergeTreeData.h | 21 +-- .../MergeTree/MergeTreeDataMergerMutator.cpp | 2 +- .../MergeTree/MergeTreeDataMergerMutator.h | 2 +- .../MergeTree/MergeTreeDataPartCompact.cpp | 38 ++--- .../MergeTree/MergeTreeDataPartCompact.h | 2 +- .../MergeTree/MergeTreeDataPartInMemory.cpp | 20 ++- .../MergeTree/MergeTreeDataPartInMemory.h | 2 +- .../MergeTree/MergeTreeDataPartWide.cpp | 30 ++-- .../MergeTree/MergeTreeDataPartWide.h | 3 +- .../MergeTreeDataPartWriterCompact.cpp | 4 +- .../MergeTreeDataPartWriterOnDisk.cpp | 8 +- .../MergeTree/MergeTreeDataPartWriterWide.cpp | 10 +- .../MergeTree/MergeTreeDataSelectExecutor.cpp | 6 +- .../MergeTree/MergeTreeDataWriter.cpp | 10 +- .../MergeTreeIndexGranularityInfo.cpp | 8 +- .../MergeTree/MergeTreeIndexGranularityInfo.h | 4 +- .../MergeTree/MergeTreeIndexMinMax.cpp | 6 +- src/Storages/MergeTree/MergeTreeIndexMinMax.h | 2 +- .../MergeTree/MergeTreeIndexReader.cpp | 4 +- src/Storages/MergeTree/MergeTreeIndices.h | 4 +- src/Storages/MergeTree/MergeTreePartition.cpp | 6 +- src/Storages/MergeTree/MergeTreePartition.h | 4 +- .../MergeTree/MergeTreePartsMover.cpp | 28 ++-- src/Storages/MergeTree/MergeTreePartsMover.h | 4 +- src/Storages/MergeTree/MergeTreeReadPool.cpp | 2 +- .../MergeTree/MergeTreeReaderCompact.cpp | 10 +- src/Storages/MergeTree/MergeTreeSink.cpp | 2 +- .../MergeTree/MergedBlockOutputStream.cpp | 40 ++--- .../MergeTree/MergedBlockOutputStream.h | 20 +-- .../MergedColumnOnlyOutputStream.cpp | 2 +- .../MergeTree/MutateFromLogEntryTask.cpp | 2 +- src/Storages/MergeTree/MutateTask.cpp | 28 ++-- .../MergeTree/PartMetadataManagerOrdinary.cpp | 16 +- .../MergeTree/PartMetadataManagerOrdinary.h | 2 +- .../PartMetadataManagerWithCache.cpp | 28 ++-- .../MergeTree/ReplicatedMergeTreeSink.cpp | 2 +- src/Storages/MergeTree/checkDataPart.cpp | 35 ++--- src/Storages/StorageMergeTree.cpp | 12 +- src/Storages/StorageReplicatedMergeTree.cpp | 58 ++++---- src/Storages/StorageReplicatedMergeTree.h | 6 +- src/Storages/System/StorageSystemParts.cpp | 4 +- .../System/StorageSystemPartsColumns.cpp | 4 +- .../System/StorageSystemProjectionParts.cpp | 4 +- .../StorageSystemProjectionPartsColumns.cpp | 4 +- 60 files changed, 456 insertions(+), 459 deletions(-) diff --git a/src/Compression/getCompressionCodecForFile.cpp b/src/Compression/getCompressionCodecForFile.cpp index f9365862c5b..027ee0ac57a 100644 --- a/src/Compression/getCompressionCodecForFile.cpp +++ b/src/Compression/getCompressionCodecForFile.cpp @@ -13,9 +13,9 @@ namespace DB using Checksum = CityHash_v1_0_2::uint128; -CompressionCodecPtr getCompressionCodecForFile(const DataPartStoragePtr & data_part_storage, const String & relative_path) +CompressionCodecPtr getCompressionCodecForFile(const IDataPartStorage & data_part_storage, const String & relative_path) { - auto read_buffer = data_part_storage->readFile(relative_path, {}, std::nullopt, std::nullopt); + auto read_buffer = data_part_storage.readFile(relative_path, {}, std::nullopt, std::nullopt); read_buffer->ignore(sizeof(Checksum)); UInt8 header_size = ICompressionCodec::getHeaderSize(); diff --git a/src/Compression/getCompressionCodecForFile.h b/src/Compression/getCompressionCodecForFile.h index ad855684128..b6f22750e4d 100644 --- a/src/Compression/getCompressionCodecForFile.h +++ b/src/Compression/getCompressionCodecForFile.h @@ -11,6 +11,6 @@ namespace DB /// clickhouse fashion (with checksums, headers for each block, etc). This /// method should be used as fallback when we cannot deduce compression codec /// from metadata. -CompressionCodecPtr getCompressionCodecForFile(const DataPartStoragePtr & data_part_storage, const String & relative_path); +CompressionCodecPtr getCompressionCodecForFile(const IDataPartStorage & data_part_storage, const String & relative_path); } diff --git a/src/Interpreters/PartLog.cpp b/src/Interpreters/PartLog.cpp index 75e6d02d6e1..b35ee50b98e 100644 --- a/src/Interpreters/PartLog.cpp +++ b/src/Interpreters/PartLog.cpp @@ -207,8 +207,8 @@ bool PartLog::addNewParts( elem.table_name = table_id.table_name; elem.partition_id = part->info.partition_id; elem.part_name = part->name; - elem.disk_name = part->data_part_storage->getDiskName(); - elem.path_on_disk = part->data_part_storage->getFullPath(); + elem.disk_name = part->getDataPartStorage().getDiskName(); + elem.path_on_disk = part->getDataPartStorage().getFullPath(); elem.part_type = part->getType(); elem.bytes_compressed_on_disk = part->getBytesOnDisk(); diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index 765c62ba903..e28aa359c99 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -30,11 +30,6 @@ DataPartStorageOnDisk::DataPartStorageOnDisk(VolumePtr volume_, std::string root { } -std::shared_ptr DataPartStorageOnDisk::clone() const -{ - return std::make_shared(volume, root_path, part_dir); -} - std::string DataPartStorageOnDisk::getFullPath() const { return fs::path(volume->getDisk()->getPath()) / root_path / part_dir / ""; @@ -55,12 +50,7 @@ std::string DataPartStorageOnDisk::getFullRootPath() const return fs::path(volume->getDisk()->getPath()) / root_path / ""; } -DataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) const -{ - return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); -} - -MutableDataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) +MutableDataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) const { return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); } @@ -279,7 +269,7 @@ void DataPartStorageOnDisk::remove( try { disk->moveDirectory(from, to); - onRename(root_path, part_dir_without_slash); + part_dir = part_dir_without_slash; } catch (const Exception & e) { @@ -524,7 +514,7 @@ bool DataPartStorageOnDisk::isBroken() const return volume->getDisk()->isBroken(); } -void DataPartStorageOnDisk::syncRevision(UInt64 revision) +void DataPartStorageOnDisk::syncRevision(UInt64 revision) const { volume->getDisk()->syncRevision(revision); } @@ -549,7 +539,7 @@ DataPartStorageOnDisk::DisksSet::const_iterator DataPartStorageOnDisk::isStoredO return disks.find(volume->getDisk()); } -ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) +ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) const { auto res = volume->reserve(bytes); if (!res) @@ -558,7 +548,7 @@ ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) return res; } -ReservationPtr DataPartStorageOnDisk::tryReserve(UInt64 bytes) +ReservationPtr DataPartStorageOnDisk::tryReserve(UInt64 bytes) const { return volume->reserve(bytes); } @@ -845,12 +835,6 @@ MutableDataPartStoragePtr DataPartStorageOnDisk::clonePart( return std::make_shared(single_disk_volume, to, dir_path); } -void DataPartStorageOnDisk::onRename(const std::string & new_root_path, const std::string & new_part_dir) -{ - part_dir = new_part_dir; - root_path = new_root_path; -} - void DataPartStorageOnDisk::rename( const std::string & new_root_path, const std::string & new_part_dir, diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.h b/src/Storages/MergeTree/DataPartStorageOnDisk.h index 3ce063ca990..80946c37f79 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.h +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.h @@ -15,15 +15,13 @@ class DataPartStorageOnDisk final : public IDataPartStorage { public: DataPartStorageOnDisk(VolumePtr volume_, std::string root_path_, std::string part_dir_); - std::shared_ptr clone() const override; std::string getFullPath() const override; std::string getRelativePath() const override; std::string getPartDirectory() const override { return part_dir; } std::string getFullRootPath() const override; - DataPartStoragePtr getProjection(const std::string & name) const override; - MutableDataPartStoragePtr getProjection(const std::string & name) override; + MutableDataPartStoragePtr getProjection(const std::string & name) const override; bool exists() const override; bool exists(const std::string & name) const override; @@ -62,7 +60,6 @@ public: std::optional & original_files_list) const; void setRelativePath(const std::string & path) override; - void onRename(const std::string & new_root_path, const std::string & new_part_dir) override; std::string getDiskName() const override; std::string getDiskType() const override; @@ -70,15 +67,15 @@ public: bool supportZeroCopyReplication() const override; bool supportParallelWrite() const override; bool isBroken() const override; - void syncRevision(UInt64 revision) override; + void syncRevision(UInt64 revision) const override; UInt64 getRevision() const override; std::unordered_map getSerializedMetadata(const std::vector & paths) const override; std::string getDiskPath() const override; DisksSet::const_iterator isStoredOnDisk(const DisksSet & disks) const override; - ReservationPtr reserve(UInt64 bytes) override; - ReservationPtr tryReserve(UInt64 bytes) override; + ReservationPtr reserve(UInt64 bytes) const override; + ReservationPtr tryReserve(UInt64 bytes) const override; size_t getVolumeIndex(const IStoragePolicy &) const override; void writeChecksums(const MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const override; diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index 3398839131c..88a745820e9 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -147,12 +147,12 @@ void Service::processQuery(const HTMLForm & params, ReadBuffer & /*body*/, Write CurrentMetrics::Increment metric_increment{CurrentMetrics::ReplicatedSend}; - if (part->data_part_storage->isStoredOnRemoteDisk()) + if (part->getDataPartStorage().isStoredOnRemoteDisk()) { UInt64 revision = parse(params.get("disk_revision", "0")); if (revision) - part->data_part_storage->syncRevision(revision); - revision = part->data_part_storage->getRevision(); + part->getDataPartStorage().syncRevision(revision); + revision = part->getDataPartStorage().getRevision(); if (revision) response.addCookie({"disk_revision", toString(revision)}); } @@ -184,8 +184,8 @@ void Service::processQuery(const HTMLForm & params, ReadBuffer & /*body*/, Write !isInMemoryPart(part) && client_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_ZERO_COPY) { - auto disk_type = part->data_part_storage->getDiskType(); - if (part->data_part_storage->supportZeroCopyReplication() && std::find(capability.begin(), capability.end(), disk_type) != capability.end()) + auto disk_type = part->getDataPartStorage().getDiskType(); + if (part->getDataPartStorage().supportZeroCopyReplication() && std::find(capability.begin(), capability.end(), disk_type) != capability.end()) { /// Send metadata if the receiver's capability covers the source disk type. response.addCookie({"remote_fs_metadata", disk_type}); @@ -307,12 +307,12 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDisk( { String file_name = it.first; - UInt64 size = part->data_part_storage->getFileSize(file_name); + UInt64 size = part->getDataPartStorage().getFileSize(file_name); writeStringBinary(it.first, out); writeBinary(size, out); - auto file_in = part->data_part_storage->readFile(file_name, {}, std::nullopt, std::nullopt); + auto file_in = part->getDataPartStorage().readFile(file_name, {}, std::nullopt, std::nullopt); HashingWriteBuffer hashing_out(out); copyDataWithThrottler(*file_in, hashing_out, blocker.getCounter(), data.getSendsThrottler()); @@ -323,7 +323,7 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDisk( throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Unexpected size of file {}, expected {} got {}", - std::string(fs::path(part->data_part_storage->getRelativePath()) / file_name), + std::string(fs::path(part->getDataPartStorage().getRelativePath()) / file_name), hashing_out.count(), size); writePODBinary(hashing_out.getHash(), out); @@ -342,9 +342,9 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( bool send_part_id, const std::map> & projections) { - const auto * data_part_storage_on_disk = dynamic_cast(part->data_part_storage.get()); + const auto * data_part_storage_on_disk = dynamic_cast(&part->getDataPartStorage()); if (!data_part_storage_on_disk) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Storage '{}' doesn't support zero-copy replication", part->data_part_storage->getDiskName()); + throw Exception(ErrorCodes::LOGICAL_ERROR, "Storage '{}' doesn't support zero-copy replication", part->getDataPartStorage().getDiskName()); if (!data_part_storage_on_disk->supportZeroCopyReplication()) throw Exception(ErrorCodes::LOGICAL_ERROR, "Disk '{}' doesn't support zero-copy replication", data_part_storage_on_disk->getDiskName()); @@ -365,7 +365,7 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( std::vector paths; paths.reserve(checksums.files.size()); for (const auto & it : checksums.files) - paths.push_back(fs::path(part->data_part_storage->getRelativePath()) / it.first); + paths.push_back(fs::path(part->getDataPartStorage().getRelativePath()) / it.first); /// Serialized metadatadatas with zero ref counts. auto metadatas = data_part_storage_on_disk->getSerializedMetadata(paths); @@ -399,7 +399,7 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( for (const auto & it : checksums.files) { const String & file_name = it.first; - String file_path_prefix = fs::path(part->data_part_storage->getRelativePath()) / file_name; + String file_path_prefix = fs::path(part->getDataPartStorage().getRelativePath()) / file_name; /// Just some additional checks String metadata_file_path = fs::path(data_part_storage_on_disk->getDiskPath()) / file_path_prefix; diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index 3b005942c54..9e3fbe7d13b 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -4,6 +4,7 @@ #include #include #include +#include #include #include @@ -70,7 +71,6 @@ class IDataPartStorage : public boost::noncopyable { public: virtual ~IDataPartStorage() = default; - virtual std::shared_ptr clone() const = 0; /// Methods to get path components of a data part. virtual std::string getFullPath() const = 0; /// '/var/lib/clickhouse/data/database/table/moving/all_1_5_1' @@ -81,8 +81,7 @@ public: /// virtual std::string getRelativeRootPath() const = 0; /// Get a storage for projection. - virtual std::shared_ptr getProjection(const std::string & name) const = 0; - virtual std::shared_ptr getProjection(const std::string & name) = 0; + virtual std::shared_ptr getProjection(const std::string & name) const = 0; /// Part directory exists. virtual bool exists() const = 0; @@ -132,10 +131,9 @@ public: /// TODO: remove it. virtual std::optional getRelativePathForPrefix(Poco::Logger * log, const String & prefix, bool detached, bool broken) const = 0; - /// Reset part directory, used for im-memory parts. + /// Reset part directory, used for in-memory parts. /// TODO: remove it. virtual void setRelativePath(const std::string & path) = 0; - virtual void onRename(const std::string & new_root_path, const std::string & new_part_dir) = 0; /// Some methods from IDisk. Needed to avoid getting internal IDisk interface. virtual std::string getDiskName() const = 0; @@ -144,7 +142,8 @@ public: virtual bool supportZeroCopyReplication() const { return false; } virtual bool supportParallelWrite() const = 0; virtual bool isBroken() const = 0; - virtual void syncRevision(UInt64 revision) = 0; + /// TODO: remove or at least remove const. + virtual void syncRevision(UInt64 revision) const = 0; virtual UInt64 getRevision() const = 0; virtual std::unordered_map getSerializedMetadata(const std::vector & paths) const = 0; /// Get a path for internal disk if relevant. It is used mainly for logging. @@ -156,8 +155,9 @@ public: /// Reserve space on the same disk. /// Probably we should try to remove it later. - virtual ReservationPtr reserve(UInt64 /*bytes*/) { return nullptr; } - virtual ReservationPtr tryReserve(UInt64 /*bytes*/) { return nullptr; } + /// TODO: remove constness + virtual ReservationPtr reserve(UInt64 /*bytes*/) const { return nullptr; } + virtual ReservationPtr tryReserve(UInt64 /*bytes*/) const { return nullptr; } virtual size_t getVolumeIndex(const IStoragePolicy &) const { return 0; } /// Some methods which change data part internals possibly after creation. @@ -234,8 +234,6 @@ public: /// Ideally, new_root_path should be the same as current root (but it is not true). /// Examples are: 'all_1_2_1' -> 'detached/all_1_2_1' /// 'moving/tmp_all_1_2_1' -> 'all_1_2_1' - /// - /// To notify storage also call onRename for it with first two args virtual void rename( const std::string & new_root_path, const std::string & new_part_dir, @@ -251,4 +249,22 @@ public: using DataPartStoragePtr = std::shared_ptr; using MutableDataPartStoragePtr = std::shared_ptr; +class DataPartStorageHolder : public boost::noncopyable +{ +public: + explicit DataPartStorageHolder(MutableDataPartStoragePtr storage_) + : storage(std::move(storage_)) + { + } + + IDataPartStorage & getDataPartStorage() { return *storage; } + const IDataPartStorage & getDataPartStorage() const { return *storage; } + + MutableDataPartStoragePtr getDataPartStoragePtr() { return storage; } + DataPartStoragePtr getDataPartStoragePtr() const { return storage; } + +private: + MutableDataPartStoragePtr storage; +}; + } diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index 6a641f0c94e..fdaa4231e9c 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -1,4 +1,5 @@ #include "IMergeTreeDataPart.h" +#include "Storages/MergeTree/IDataPartStorage.h" #include #include @@ -101,7 +102,7 @@ void IMergeTreeDataPart::MinMaxIndex::load(const MergeTreeData & data, const Par } IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::store( - const MergeTreeData & data, const MutableDataPartStoragePtr & part_storage, Checksums & out_checksums) const + const MergeTreeData & data, IDataPartStorage & part_storage, Checksums & out_checksums) const { auto metadata_snapshot = data.getInMemoryMetadataPtr(); const auto & partition_key = metadata_snapshot->getPartitionKey(); @@ -115,14 +116,14 @@ IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::s IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::store( const Names & column_names, const DataTypes & data_types, - const MutableDataPartStoragePtr & part_storage, + IDataPartStorage & part_storage, Checksums & out_checksums) const { if (!initialized) throw Exception( ErrorCodes::LOGICAL_ERROR, "Attempt to store uninitialized MinMax index for part {}. This is a bug", - part_storage->getFullPath()); + part_storage.getFullPath()); WrittenFiles written_files; @@ -131,7 +132,7 @@ IMergeTreeDataPart::MinMaxIndex::WrittenFiles IMergeTreeDataPart::MinMaxIndex::s String file_name = "minmax_" + escapeForFileName(column_names[i]) + ".idx"; auto serialization = data_types.at(i)->getDefaultSerialization(); - auto out = part_storage->writeFile(file_name, DBMS_DEFAULT_BUFFER_SIZE, {}); + auto out = part_storage.writeFile(file_name, DBMS_DEFAULT_BUFFER_SIZE, {}); HashingWriteBuffer out_hashing(*out); serialization->serializeBinary(hyperrectangle[i].left, out_hashing); serialization->serializeBinary(hyperrectangle[i].right, out_hashing); @@ -304,10 +305,10 @@ IMergeTreeDataPart::IMergeTreeDataPart( const MutableDataPartStoragePtr & data_part_storage_, Type part_type_, const IMergeTreeDataPart * parent_part_) - : storage(storage_) + : DataPartStorageHolder(data_part_storage_) + , storage(storage_) , name(name_) , info(MergeTreePartInfo::fromPartName(name_, storage.format_version)) - , data_part_storage(parent_part_ ? parent_part_->data_part_storage : data_part_storage_) , index_granularity_info(storage_, part_type_) , part_type(part_type_) , parent_part(parent_part_) @@ -332,10 +333,10 @@ IMergeTreeDataPart::IMergeTreeDataPart( const MutableDataPartStoragePtr & data_part_storage_, Type part_type_, const IMergeTreeDataPart * parent_part_) - : storage(storage_) + : DataPartStorageHolder(data_part_storage_) + , storage(storage_) , name(name_) , info(info_) - , data_part_storage(data_part_storage_) , index_granularity_info(storage_, part_type_) , part_type(part_type_) , parent_part(parent_part_) @@ -343,6 +344,7 @@ IMergeTreeDataPart::IMergeTreeDataPart( { if (parent_part) state = MergeTreeDataPartState::Active; + incrementStateMetric(state); incrementTypeMetric(part_type); @@ -506,17 +508,17 @@ void IMergeTreeDataPart::removeIfNeeded() std::string path; try { - path = data_part_storage->getRelativePath(); + path = getDataPartStorage().getRelativePath(); - if (!data_part_storage->exists()) // path + if (!getDataPartStorage().exists()) // path return; if (is_temp) { - String file_name = fileName(data_part_storage->getPartDirectory()); + String file_name = fileName(getDataPartStorage().getPartDirectory()); if (file_name.empty()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "relative_path {} of part {} is invalid or not set", data_part_storage->getPartDirectory(), name); + throw Exception(ErrorCodes::LOGICAL_ERROR, "relative_path {} of part {} is invalid or not set", getDataPartStorage().getPartDirectory(), name); if (!startsWith(file_name, "tmp") && !endsWith(file_name, ".tmp_proj")) { @@ -621,7 +623,7 @@ String IMergeTreeDataPart::getColumnNameWithMinimumCompressedSize(bool with_subc } if (!minimum_size_column) - throw Exception("Could not find a column of minimum size in MergeTree, part " + data_part_storage->getFullPath(), ErrorCodes::LOGICAL_ERROR); + throw Exception("Could not find a column of minimum size in MergeTree, part " + getDataPartStorage().getFullPath(), ErrorCodes::LOGICAL_ERROR); return *minimum_size_column; } @@ -699,9 +701,9 @@ void IMergeTreeDataPart::loadProjections(bool require_columns_checksums, bool ch for (const auto & projection : metadata_snapshot->projections) { String path = /*getRelativePath() + */ projection.name + ".proj"; - if (data_part_storage->exists(path)) + if (getDataPartStorage().exists(path)) { - auto projection_part_storage = data_part_storage->getProjection(projection.name + ".proj"); + auto projection_part_storage = getDataPartStorage().getProjection(projection.name + ".proj"); auto part = storage.createPart(projection.name, {"all", 0, 0, 0}, projection_part_storage, this); part->loadColumnsChecksumsIndexes(require_columns_checksums, check_consistency); projection_parts.emplace(projection.name, std::move(part)); @@ -742,8 +744,8 @@ void IMergeTreeDataPart::loadIndex() loaded_index[i]->reserve(index_granularity.getMarksCount()); } - String index_name = "primary" + getIndexExtensionFromFilesystem(data_part_storage).value(); - String index_path = fs::path(data_part_storage->getRelativePath()) / index_name; + String index_name = "primary" + getIndexExtensionFromFilesystem(getDataPartStorage()).value(); + String index_path = fs::path(getDataPartStorage().getRelativePath()) / index_name; auto index_file = metadata_manager->read(index_name); size_t marks_count = index_granularity.getMarksCount(); @@ -782,7 +784,7 @@ void IMergeTreeDataPart::appendFilesOfIndex(Strings & files) const if (metadata_snapshot->hasPrimaryKey()) { - String index_name = "primary" + getIndexExtensionFromFilesystem(data_part_storage).value(); + String index_name = "primary" + getIndexExtensionFromFilesystem(getDataPartStorage()).value(); files.push_back(index_name); } } @@ -794,10 +796,10 @@ NameSet IMergeTreeDataPart::getFileNamesWithoutChecksums() const NameSet result = {"checksums.txt", "columns.txt"}; - if (data_part_storage->exists(DEFAULT_COMPRESSION_CODEC_FILE_NAME)) + if (getDataPartStorage().exists(DEFAULT_COMPRESSION_CODEC_FILE_NAME)) result.emplace(DEFAULT_COMPRESSION_CODEC_FILE_NAME); - if (data_part_storage->exists(TXN_VERSION_METADATA_FILE_NAME)) + if (getDataPartStorage().exists(TXN_VERSION_METADATA_FILE_NAME)) result.emplace(TXN_VERSION_METADATA_FILE_NAME); return result; @@ -812,7 +814,7 @@ void IMergeTreeDataPart::loadDefaultCompressionCodec() return; } - String path = fs::path(data_part_storage->getRelativePath()) / DEFAULT_COMPRESSION_CODEC_FILE_NAME; + String path = fs::path(getDataPartStorage().getRelativePath()) / DEFAULT_COMPRESSION_CODEC_FILE_NAME; bool exists = metadata_manager->exists(DEFAULT_COMPRESSION_CODEC_FILE_NAME); if (!exists) { @@ -881,7 +883,7 @@ CompressionCodecPtr IMergeTreeDataPart::detectDefaultCompressionCodec() const String candidate_path = /*fs::path(getRelativePath()) */ (ISerialization::getFileNameForStream(part_column, substream_path) + ".bin"); /// We can have existing, but empty .bin files. Example: LowCardinality(Nullable(...)) columns and column_name.dict.null.bin file. - if (data_part_storage->exists(candidate_path) && data_part_storage->getFileSize(candidate_path) != 0) + if (getDataPartStorage().exists(candidate_path) && getDataPartStorage().getFileSize(candidate_path) != 0) path_to_data_file = candidate_path; } }); @@ -892,7 +894,7 @@ CompressionCodecPtr IMergeTreeDataPart::detectDefaultCompressionCodec() const continue; } - result = getCompressionCodecForFile(data_part_storage, path_to_data_file); + result = getCompressionCodecForFile(getDataPartStorage(), path_to_data_file); break; } } @@ -937,7 +939,7 @@ void IMergeTreeDataPart::loadPartitionAndMinMaxIndex() String calculated_partition_id = partition.getID(metadata_snapshot->getPartitionKey().sample_block); if (calculated_partition_id != info.partition_id) throw Exception( - "While loading part " + data_part_storage->getFullPath() + ": calculated partition ID: " + calculated_partition_id + "While loading part " + getDataPartStorage().getFullPath() + ": calculated partition ID: " + calculated_partition_id + " differs from partition ID in part name: " + info.partition_id, ErrorCodes::CORRUPTED_DATA); } @@ -966,7 +968,7 @@ void IMergeTreeDataPart::loadChecksums(bool require) bytes_on_disk = checksums.getTotalSizeOnDisk(); } else - bytes_on_disk = data_part_storage->calculateTotalSizeOnDisk(); + bytes_on_disk = getDataPartStorage().calculateTotalSizeOnDisk(); } else { @@ -978,7 +980,7 @@ void IMergeTreeDataPart::loadChecksums(bool require) LOG_WARNING(storage.log, "Checksums for part {} not found. Will calculate them from data on disk.", name); checksums = checkDataPart(shared_from_this(), false); - data_part_storage->writeChecksums(checksums, {}); + getDataPartStorage().writeChecksums(checksums, {}); bytes_on_disk = checksums.getTotalSizeOnDisk(); } @@ -1063,7 +1065,7 @@ void IMergeTreeDataPart::loadRowsCount() } else { - if (data_part_storage->exists("count.txt")) + if (getDataPartStorage().exists("count.txt")) { read_rows_count(); return; @@ -1162,7 +1164,7 @@ void IMergeTreeDataPart::appendFilesOfUUID(Strings & files) void IMergeTreeDataPart::loadColumns(bool require) { - String path = fs::path(data_part_storage->getRelativePath()) / "columns.txt"; + String path = fs::path(getDataPartStorage().getRelativePath()) / "columns.txt"; auto metadata_snapshot = storage.getInMemoryMetadataPtr(); if (parent_part) metadata_snapshot = metadata_snapshot->projections.get(name).metadata; @@ -1173,18 +1175,18 @@ void IMergeTreeDataPart::loadColumns(bool require) { /// We can get list of columns only from columns.txt in compact parts. if (require || part_type == Type::Compact) - throw Exception("No columns.txt in part " + name + ", expected path " + path + " on drive " + data_part_storage->getDiskName(), + throw Exception("No columns.txt in part " + name + ", expected path " + path + " on drive " + getDataPartStorage().getDiskName(), ErrorCodes::NO_FILE_IN_DATA_PART); /// If there is no file with a list of columns, write it down. for (const NameAndTypePair & column : metadata_snapshot->getColumns().getAllPhysical()) - if (data_part_storage->exists(getFileNameForColumn(column) + ".bin")) + if (getDataPartStorage().exists(getFileNameForColumn(column) + ".bin")) loaded_columns.push_back(column); if (columns.empty()) throw Exception("No columns in part " + name, ErrorCodes::NO_FILE_IN_DATA_PART); - data_part_storage->writeColumns(loaded_columns, {}); + getDataPartStorage().writeColumns(loaded_columns, {}); } else { @@ -1228,7 +1230,7 @@ void IMergeTreeDataPart::assertHasVersionMetadata(MergeTreeTransaction * txn) co name, storage.getStorageID().getNameForLogs(), version.creation_tid, txn ? txn->dumpDescription() : ""); assert(!txn || storage.supportsTransactions()); - assert(!txn || data_part_storage->exists(TXN_VERSION_METADATA_FILE_NAME)); + assert(!txn || getDataPartStorage().exists(TXN_VERSION_METADATA_FILE_NAME)); } void IMergeTreeDataPart::storeVersionMetadata(bool force) const @@ -1243,7 +1245,7 @@ void IMergeTreeDataPart::storeVersionMetadata(bool force) const throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Transactions are not supported for in-memory parts (table: {}, part: {})", storage.getStorageID().getNameForLogs(), name); - data_part_storage->writeVersionMetadata(version, storage.getSettings()->fsync_part_directory); + getDataPartStorage().writeVersionMetadata(version, storage.getSettings()->fsync_part_directory); } void IMergeTreeDataPart::appendCSNToVersionMetadata(VersionMetadata::WhichCSN which_csn) const @@ -1255,7 +1257,7 @@ void IMergeTreeDataPart::appendCSNToVersionMetadata(VersionMetadata::WhichCSN wh chassert(!(which_csn == VersionMetadata::WhichCSN::REMOVAL && version.removal_csn == 0)); chassert(isStoredOnDisk()); - data_part_storage->appendCSNToVersionMetadata(version, which_csn); + getDataPartStorage().appendCSNToVersionMetadata(version, which_csn); } void IMergeTreeDataPart::appendRemovalTIDToVersionMetadata(bool clear) const @@ -1278,13 +1280,13 @@ void IMergeTreeDataPart::appendRemovalTIDToVersionMetadata(bool clear) const else LOG_TEST(storage.log, "Appending removal TID for {} (creation: {}, removal {})", name, version.creation_tid, version.removal_tid); - data_part_storage->appendRemovalTIDToVersionMetadata(version, clear); + getDataPartStorage().appendRemovalTIDToVersionMetadata(version, clear); } void IMergeTreeDataPart::loadVersionMetadata() const try { - data_part_storage->loadVersionMetadata(version, storage.log); + getDataPartStorage().loadVersionMetadata(version, storage.log); } catch (Exception & e) { @@ -1321,15 +1323,15 @@ bool IMergeTreeDataPart::assertHasValidVersionMetadata() const if (state == MergeTreeDataPartState::Temporary) return true; - if (!data_part_storage->exists()) + if (!getDataPartStorage().exists()) return true; String content; String version_file_name = TXN_VERSION_METADATA_FILE_NAME; try { - size_t file_size = data_part_storage->getFileSize(TXN_VERSION_METADATA_FILE_NAME); - auto buf = data_part_storage->readFile(TXN_VERSION_METADATA_FILE_NAME, ReadSettings().adjustBufferSize(file_size), file_size, std::nullopt); + size_t file_size = getDataPartStorage().getFileSize(TXN_VERSION_METADATA_FILE_NAME); + auto buf = getDataPartStorage().readFile(TXN_VERSION_METADATA_FILE_NAME, ReadSettings().adjustBufferSize(file_size), file_size, std::nullopt); readStringUntilEOF(content, *buf); ReadBufferFromString str_buf{content}; @@ -1363,10 +1365,10 @@ void IMergeTreeDataPart::appendFilesOfColumns(Strings & files) bool IMergeTreeDataPart::shallParticipateInMerges(const StoragePolicyPtr & storage_policy) const { - return data_part_storage->shallParticipateInMerges(*storage_policy); + return getDataPartStorage().shallParticipateInMerges(*storage_policy); } -void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) const +void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) try { assertOnDisk(); @@ -1377,20 +1379,21 @@ try if (parent_part) { /// For projections, move is only possible inside parent part dir. - relative_path = parent_part->data_part_storage->getRelativePath(); + relative_path = parent_part->getDataPartStorage().getRelativePath(); } - String from = data_part_storage->getRelativePath(); auto to = fs::path(relative_path) / new_relative_path; metadata_manager->deleteAll(true); metadata_manager->assertAllDeleted(true); - data_part_storage->rename(to.parent_path(), to.filename(), storage.log, remove_new_dir_if_exists, fsync_dir); - data_part_storage->onRename(to.parent_path(), to.filename()); + getDataPartStorage().rename(to.parent_path(), to.filename(), storage.log, remove_new_dir_if_exists, fsync_dir); metadata_manager->updateAll(true); - for (const auto & [p_name, part] : projection_parts) - part->data_part_storage = data_part_storage->getProjection(p_name + ".proj"); + auto old_projection_root_path = getDataPartStorage().getRelativePath(); + auto new_projection_root_path = to.string(); + + for (const auto & [_, part] : projection_parts) + part->getDataPartStorage().changeRootPath(old_projection_root_path, new_projection_root_path); } catch (...) { @@ -1431,14 +1434,14 @@ void IMergeTreeDataPart::initializePartMetadataManager() void IMergeTreeDataPart::initializeIndexGranularityInfo() { - auto mrk_ext = MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(data_part_storage); + auto mrk_ext = MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(getDataPartStorage()); if (mrk_ext) index_granularity_info = MergeTreeIndexGranularityInfo(storage, MarkType{*mrk_ext}); else index_granularity_info = MergeTreeIndexGranularityInfo(storage, part_type); } -void IMergeTreeDataPart::remove() const +void IMergeTreeDataPart::remove() { assert(assertHasValidVersionMetadata()); part_is_probably_removed_from_disk = true; @@ -1455,7 +1458,6 @@ void IMergeTreeDataPart::remove() const return CanRemoveDescription{.can_remove_anything = can_remove, .files_not_to_remove = files_not_to_remove }; }; - if (!isStoredOnDisk()) return; @@ -1474,7 +1476,7 @@ void IMergeTreeDataPart::remove() const projection_checksums.emplace_back(IDataPartStorage::ProjectionChecksums{.name = p_name, .checksums = projection_part->checksums}); } - data_part_storage->remove(std::move(can_remove_callback), checksums, projection_checksums, is_temp, getState(), storage.log); + getDataPartStorage().remove(std::move(can_remove_callback), checksums, projection_checksums, is_temp, getState(), storage.log); } std::optional IMergeTreeDataPart::getRelativePathForPrefix(const String & prefix, bool detached, bool broken) const @@ -1491,7 +1493,7 @@ std::optional IMergeTreeDataPart::getRelativePathForPrefix(const String if (detached && parent_part) throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot detach projection"); - return data_part_storage->getRelativePathForPrefix(storage.log, prefix, detached, broken); + return getDataPartStorage().getRelativePathForPrefix(storage.log, prefix, detached, broken); } std::optional IMergeTreeDataPart::getRelativePathForDetachedPart(const String & prefix, bool broken) const @@ -1506,7 +1508,7 @@ std::optional IMergeTreeDataPart::getRelativePathForDetachedPart(const S return {}; } -void IMergeTreeDataPart::renameToDetached(const String & prefix) const +void IMergeTreeDataPart::renameToDetached(const String & prefix) { auto path_to_detach = getRelativePathForDetachedPart(prefix, /* broken */ false); assert(path_to_detach); @@ -1529,7 +1531,7 @@ void IMergeTreeDataPart::makeCloneInDetached(const String & prefix, const Storag if (!maybe_path_in_detached) return; - data_part_storage->freeze( + getDataPartStorage().freeze( storage.relative_data_path, *maybe_path_in_detached, /*make_source_readonly*/ true, @@ -1542,13 +1544,13 @@ MutableDataPartStoragePtr IMergeTreeDataPart::makeCloneOnDisk(const DiskPtr & di { assertOnDisk(); - if (disk->getName() == data_part_storage->getDiskName()) - throw Exception("Can not clone data part " + name + " to same disk " + data_part_storage->getDiskName(), ErrorCodes::LOGICAL_ERROR); + if (disk->getName() == getDataPartStorage().getDiskName()) + throw Exception("Can not clone data part " + name + " to same disk " + getDataPartStorage().getDiskName(), ErrorCodes::LOGICAL_ERROR); if (directory_name.empty()) throw Exception("Can not clone data part " + name + " to empty directory.", ErrorCodes::LOGICAL_ERROR); String path_to_clone = fs::path(storage.relative_data_path) / directory_name / ""; - return data_part_storage->clonePart(path_to_clone, data_part_storage->getPartDirectory(), disk, storage.log); + return getDataPartStorage().clonePart(path_to_clone, getDataPartStorage().getPartDirectory(), disk, storage.log); } void IMergeTreeDataPart::checkConsistencyBase() const @@ -1589,26 +1591,26 @@ void IMergeTreeDataPart::checkConsistencyBase() const } } - data_part_storage->checkConsistency(checksums); + getDataPartStorage().checkConsistency(checksums); } else { auto check_file_not_empty = [this](const String & file_path) { UInt64 file_size; - if (!data_part_storage->exists(file_path) || (file_size = data_part_storage->getFileSize(file_path)) == 0) + if (!getDataPartStorage().exists(file_path) || (file_size = getDataPartStorage().getFileSize(file_path)) == 0) throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Part {} is broken: {} is empty", - data_part_storage->getFullPath(), - std::string(fs::path(data_part_storage->getFullPath()) / file_path)); + getDataPartStorage().getFullPath(), + std::string(fs::path(getDataPartStorage().getFullPath()) / file_path)); return file_size; }; /// Check that the primary key index is not empty. if (!pk.column_names.empty()) { - String index_name = "primary" + getIndexExtensionFromFilesystem(data_part_storage).value(); + String index_name = "primary" + getIndexExtensionFromFilesystem(getDataPartStorage()).value(); check_file_not_empty(index_name); } @@ -1752,7 +1754,7 @@ bool IMergeTreeDataPart::checkAllTTLCalculated(const StorageMetadataPtr & metada String IMergeTreeDataPart::getUniqueId() const { - return data_part_storage->getUniqueId(); + return getDataPartStorage().getUniqueId(); } String IMergeTreeDataPart::getZeroLevelPartBlockID(std::string_view token) const @@ -1791,11 +1793,11 @@ IMergeTreeDataPart::uint128 IMergeTreeDataPart::getActualChecksumByFile(const St return it->second.file_hash; } - if (!data_part_storage->exists(file_name)) + if (!getDataPartStorage().exists(file_name)) { return {}; } - std::unique_ptr in_file = data_part_storage->readFile(file_name, {}, std::nullopt, std::nullopt); + std::unique_ptr in_file = getDataPartStorage().readFile(file_name, {}, std::nullopt, std::nullopt); HashingReadBuffer in_hash(*in_file); String value; @@ -1823,11 +1825,11 @@ bool isInMemoryPart(const MergeTreeDataPartPtr & data_part) return (data_part && data_part->getType() == MergeTreeDataPartType::InMemory); } -std::optional getIndexExtensionFromFilesystem(const DataPartStoragePtr & data_part_storage) +std::optional getIndexExtensionFromFilesystem(const IDataPartStorage & data_part_storage) { - if (data_part_storage->exists()) + if (data_part_storage.exists()) { - for (auto it = data_part_storage->iterate(); it->isValid(); it->next()) + for (auto it = data_part_storage.iterate(); it->isValid(); it->next()) { const auto & extension = fs::path(it->name()).extension(); if (extension == getIndexExtension(false) diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.h b/src/Storages/MergeTree/IMergeTreeDataPart.h index 0fe94b666b6..fbe4f992de4 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.h +++ b/src/Storages/MergeTree/IMergeTreeDataPart.h @@ -46,7 +46,7 @@ class UncompressedCache; class MergeTreeTransaction; /// Description of the data part. -class IMergeTreeDataPart : public std::enable_shared_from_this +class IMergeTreeDataPart : public std::enable_shared_from_this, public DataPartStorageHolder { public: static constexpr auto DATA_FILE_EXTENSION = ".bin"; @@ -150,7 +150,7 @@ public: /// Throws an exception if part is not stored in on-disk format. void assertOnDisk() const; - void remove() const; + void remove(); /// Initialize columns (from columns.txt if exists, or create from column files if not). /// Load checksums from checksums.txt if exists. Load index if required. @@ -198,10 +198,6 @@ public: /// processed by multiple shards. UUID uuid = UUIDHelpers::Nil; - /// This is an object which encapsulates all the operations with disk. - /// Contains a path to stored data. - MutableDataPartStoragePtr data_part_storage; - MergeTreeIndexGranularityInfo index_granularity_info; size_t rows_count = 0; @@ -287,8 +283,8 @@ public: using WrittenFiles = std::vector>; - [[nodiscard]] WrittenFiles store(const MergeTreeData & data, const MutableDataPartStoragePtr & part_storage, Checksums & checksums) const; - [[nodiscard]] WrittenFiles store(const Names & column_names, const DataTypes & data_types, const MutableDataPartStoragePtr & part_storage, Checksums & checksums) const; + [[nodiscard]] WrittenFiles store(const MergeTreeData & data, IDataPartStorage & part_storage, Checksums & checksums) const; + [[nodiscard]] WrittenFiles store(const Names & column_names, const DataTypes & data_types, IDataPartStorage & part_storage, Checksums & checksums) const; void update(const Block & block, const Names & column_names); void merge(const MinMaxIndex & other); @@ -319,11 +315,11 @@ public: size_t getFileSizeOrZero(const String & file_name) const; /// Moves a part to detached/ directory and adds prefix to its name - void renameToDetached(const String & prefix) const; + void renameToDetached(const String & prefix); /// Makes checks and move part to new directory /// Changes only relative_dir_name, you need to update other metadata (name, is_temp) explicitly - virtual void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) const; + virtual void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists); /// Makes clone of a part in detached/ directory via hard links virtual void makeCloneInDetached(const String & prefix, const StorageMetadataPtr & metadata_snapshot) const; @@ -583,7 +579,7 @@ bool isCompactPart(const MergeTreeDataPartPtr & data_part); bool isWidePart(const MergeTreeDataPartPtr & data_part); bool isInMemoryPart(const MergeTreeDataPartPtr & data_part); inline String getIndexExtension(bool is_compressed_primary_key) { return is_compressed_primary_key ? ".cidx" : ".idx"; } -std::optional getIndexExtensionFromFilesystem(const DataPartStoragePtr & data_part_storage); +std::optional getIndexExtensionFromFilesystem(const IDataPartStorage & data_part_storage); bool isCompressedFromIndexExtension(const String & index_extension); } diff --git a/src/Storages/MergeTree/IMergedBlockOutputStream.cpp b/src/Storages/MergeTree/IMergedBlockOutputStream.cpp index 54f393a65a2..37da6014d1b 100644 --- a/src/Storages/MergeTree/IMergedBlockOutputStream.cpp +++ b/src/Storages/MergeTree/IMergedBlockOutputStream.cpp @@ -12,7 +12,7 @@ IMergedBlockOutputStream::IMergedBlockOutputStream( bool reset_columns_) : storage(data_part->storage) , metadata_snapshot(metadata_snapshot_) - , data_part_storage(data_part->data_part_storage) + , data_part_storage(data_part->getDataPartStoragePtr()) , reset_columns(reset_columns_) { if (reset_columns) diff --git a/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h b/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h index ee265ee6fb1..bc786ec0428 100644 --- a/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h +++ b/src/Storages/MergeTree/LoadedMergeTreeDataPartInfoForReader.h @@ -23,7 +23,7 @@ public: bool isProjectionPart() const override { return data_part->isProjectionPart(); } - DataPartStoragePtr getDataPartStorage() const override { return data_part->data_part_storage; } + DataPartStoragePtr getDataPartStorage() const override { return data_part->getDataPartStoragePtr(); } const NamesAndTypesList & getColumns() const override { return data_part->getColumns(); } diff --git a/src/Storages/MergeTree/MergeFromLogEntryTask.cpp b/src/Storages/MergeTree/MergeFromLogEntryTask.cpp index 182d5df4960..f3b81a4793e 100644 --- a/src/Storages/MergeTree/MergeFromLogEntryTask.cpp +++ b/src/Storages/MergeTree/MergeFromLogEntryTask.cpp @@ -160,7 +160,7 @@ ReplicatedMergeMutateTaskBase::PrepareResult MergeFromLogEntryTask::prepare() for (auto & part_ptr : parts) { ttl_infos.update(part_ptr->ttl_infos); - max_volume_index = std::max(max_volume_index, part_ptr->data_part_storage->getVolumeIndex(*storage.getStoragePolicy())); + max_volume_index = std::max(max_volume_index, part_ptr->getDataPartStorage().getVolumeIndex(*storage.getStoragePolicy())); } /// It will live until the whole task is being destroyed diff --git a/src/Storages/MergeTree/MergeList.cpp b/src/Storages/MergeTree/MergeList.cpp index ebe826531d2..02e61a70eb6 100644 --- a/src/Storages/MergeTree/MergeList.cpp +++ b/src/Storages/MergeTree/MergeList.cpp @@ -65,7 +65,7 @@ MergeListElement::MergeListElement( for (const auto & source_part : future_part->parts) { source_part_names.emplace_back(source_part->name); - source_part_paths.emplace_back(source_part->data_part_storage->getFullPath()); + source_part_paths.emplace_back(source_part->getDataPartStorage().getFullPath()); total_size_bytes_compressed += source_part->getBytesOnDisk(); total_size_marks += source_part->getMarksCount(); diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp index 130d156e53c..27a71345a5e 100644 --- a/src/Storages/MergeTree/MergeTask.cpp +++ b/src/Storages/MergeTree/MergeTask.cpp @@ -130,7 +130,7 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() if (global_ctx->parent_part) { - data_part_storage = global_ctx->parent_part->data_part_storage->getProjection(local_tmp_part_basename); + data_part_storage = global_ctx->parent_part->getDataPartStorage().getProjection(local_tmp_part_basename); } else { diff --git a/src/Storages/MergeTree/MergeTask.h b/src/Storages/MergeTree/MergeTask.h index 07d46460423..f21d542c7a0 100644 --- a/src/Storages/MergeTree/MergeTask.h +++ b/src/Storages/MergeTree/MergeTask.h @@ -59,7 +59,7 @@ public: bool deduplicate_, Names deduplicate_by_columns_, MergeTreeData::MergingParams merging_params_, - const IMergeTreeDataPart * parent_part_, + IMergeTreeDataPart * parent_part_, String suffix_, MergeTreeTransactionPtr txn, MergeTreeData * data_, diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 9e9d90b6b1e..83f3a167fa7 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -943,8 +943,8 @@ Int64 MergeTreeData::getMaxBlockNumber() const } void MergeTreeData::loadDataPartsFromDisk( - DataPartsVector & broken_parts_to_detach, - DataPartsVector & duplicate_parts_to_remove, + MutableDataPartsVector & broken_parts_to_detach, + MutableDataPartsVector & duplicate_parts_to_remove, ThreadPool & pool, size_t num_parts, std::queue>> & parts_queue, @@ -1199,8 +1199,7 @@ void MergeTreeData::loadDataPartsFromDisk( void MergeTreeData::loadDataPartsFromWAL( - DataPartsVector & /* broken_parts_to_detach */, - DataPartsVector & duplicate_parts_to_remove, + MutableDataPartsVector & duplicate_parts_to_remove, MutableDataPartsVector & parts_from_wal) { for (auto & part : parts_from_wal) @@ -1214,7 +1213,7 @@ void MergeTreeData::loadDataPartsFromWAL( { if ((*it)->checksums.getTotalChecksumHex() == part->checksums.getTotalChecksumHex()) { - LOG_ERROR(log, "Remove duplicate part {}", part->data_part_storage->getFullPath()); + LOG_ERROR(log, "Remove duplicate part {}", part->getDataPartStorage().getFullPath()); duplicate_parts_to_remove.push_back(part); } else @@ -1328,8 +1327,8 @@ void MergeTreeData::loadDataParts(bool skip_sanity_checks) auto part_lock = lockParts(); data_parts_indexes.clear(); - DataPartsVector broken_parts_to_detach; - DataPartsVector duplicate_parts_to_remove; + MutableDataPartsVector broken_parts_to_detach; + MutableDataPartsVector duplicate_parts_to_remove; if (num_parts > 0) loadDataPartsFromDisk( @@ -1383,7 +1382,7 @@ void MergeTreeData::loadDataParts(bool skip_sanity_checks) parts_from_wal.insert( parts_from_wal.end(), std::make_move_iterator(disk_wal_parts.begin()), std::make_move_iterator(disk_wal_parts.end())); - loadDataPartsFromWAL(broken_parts_to_detach, duplicate_parts_to_remove, parts_from_wal); + loadDataPartsFromWAL(duplicate_parts_to_remove, parts_from_wal); num_parts += parts_from_wal.size(); } @@ -1684,6 +1683,14 @@ scope_guard MergeTreeData::getTemporaryPartDirectoryHolder(const String & part_d return [this, part_dir_name]() { temporary_parts.remove(part_dir_name); }; } +MergeTreeData::MutableDataPartPtr MergeTreeData::preparePartForRemoval(const DataPartPtr & part) +{ + if (part->getState() != DataPartState::Deleting) + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Cannot remove part {}, because it has state: {}", part->name, magic_enum::enum_name(part->getState())); + + return std::const_pointer_cast(part); +} MergeTreeData::DataPartsVector MergeTreeData::grabOldParts(bool force) { @@ -1859,7 +1866,7 @@ void MergeTreeData::flushAllInMemoryPartsIfNeeded() { if (auto part_in_memory = asInMemoryPart(part)) { - part_in_memory->flushToDisk(part_in_memory->data_part_storage->getPartDirectory(), metadata_snapshot); + part_in_memory->flushToDisk(part_in_memory->getDataPartStorage().getPartDirectory(), metadata_snapshot); } } } @@ -1943,7 +1950,7 @@ void MergeTreeData::clearPartsFromFilesystemImpl(const DataPartsVector & parts_t if (thread_group) CurrentThread::attachToIfDetached(thread_group); - part->remove(); + preparePartForRemoval(part)->remove(); if (part_names_succeed) { std::lock_guard lock(part_names_mutex); @@ -1959,7 +1966,7 @@ void MergeTreeData::clearPartsFromFilesystemImpl(const DataPartsVector & parts_t LOG_DEBUG(log, "Removing {} parts from filesystem: {}", parts_to_remove.size(), fmt::join(parts_to_remove, ", ")); for (const DataPartPtr & part : parts_to_remove) { - part->remove(); + preparePartForRemoval(part)->remove(); if (part_names_succeed) part_names_succeed->insert(part->name); } @@ -2139,11 +2146,14 @@ void MergeTreeData::rename(const String & new_table_path, const StorageID & new_ if (!getStorageID().hasUUID()) getContext()->dropCaches(); + /// TODO: remove const_cast for (const auto & part : data_parts_by_info) - part->data_part_storage->changeRootPath(relative_data_path, new_table_path); + { + auto & part_mutable = const_cast(*part); + part_mutable.getDataPartStorage().changeRootPath(relative_data_path, new_table_path); + } relative_data_path = new_table_path; - renameInMemory(new_table_id); } @@ -2744,7 +2754,7 @@ MergeTreeData::MutableDataPartPtr MergeTreeData::createPart( const MutableDataPartStoragePtr & data_part_storage, const IMergeTreeDataPart * parent_part) const { MergeTreeDataPartType type; - auto mrk_ext = MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(data_part_storage); + auto mrk_ext = MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(*data_part_storage); if (mrk_ext) { @@ -2983,7 +2993,7 @@ void MergeTreeData::preparePartForCommit(MutableDataPartPtr & part, Transaction assert([&]() { - String dir_name = fs::path(part->data_part_storage->getRelativePath()).filename(); + String dir_name = fs::path(part->getDataPartStorage().getRelativePath()).filename(); bool may_be_cleaned_up = dir_name.starts_with("tmp_") || dir_name.starts_with("tmp-fetch_"); return !may_be_cleaned_up || temporary_parts.contains(dir_name); }()); @@ -3000,7 +3010,7 @@ bool MergeTreeData::renameTempPartAndReplaceImpl( DataPartsLock & lock, DataPartsVector * out_covered_parts) { - LOG_TRACE(log, "Renaming temporary part {} to {}.", part->data_part_storage->getPartDirectory(), part->name); + LOG_TRACE(log, "Renaming temporary part {} to {}.", part->getDataPartStorage().getPartDirectory(), part->name); if (&out_transaction.data != this) throw Exception("MergeTreeData::Transaction for one table cannot be used with another. It is a bug.", @@ -3231,9 +3241,9 @@ void MergeTreeData::outdateBrokenPartAndCloneToDetached(const DataPartPtr & part { auto metadata_snapshot = getInMemoryMetadataPtr(); if (prefix.empty()) - LOG_INFO(log, "Cloning part {} to {} and making it obsolete.", part_to_detach->data_part_storage->getPartDirectory(), part_to_detach->name); + LOG_INFO(log, "Cloning part {} to {} and making it obsolete.", part_to_detach->getDataPartStorage().getPartDirectory(), part_to_detach->name); else - LOG_INFO(log, "Cloning part {} to {}_{} and making it obsolete.", part_to_detach->data_part_storage->getPartDirectory(), prefix, part_to_detach->name); + LOG_INFO(log, "Cloning part {} to {}_{} and making it obsolete.", part_to_detach->getDataPartStorage().getPartDirectory(), prefix, part_to_detach->name); part_to_detach->makeCloneInDetached(prefix, metadata_snapshot); @@ -3245,9 +3255,9 @@ void MergeTreeData::outdateBrokenPartAndCloneToDetached(const DataPartPtr & part void MergeTreeData::forcefullyMovePartToDetachedAndRemoveFromMemory(const MergeTreeData::DataPartPtr & part_to_detach, const String & prefix, bool restore_covered) { if (prefix.empty()) - LOG_INFO(log, "Renaming {} to {} and forgetting it.", part_to_detach->data_part_storage->getPartDirectory(), part_to_detach->name); + LOG_INFO(log, "Renaming {} to {} and forgetting it.", part_to_detach->getDataPartStorage().getPartDirectory(), part_to_detach->name); else - LOG_INFO(log, "Renaming {} to {}_{} and forgetting it.", part_to_detach->data_part_storage->getPartDirectory(), prefix, part_to_detach->name); + LOG_INFO(log, "Renaming {} to {}_{} and forgetting it.", part_to_detach->getDataPartStorage().getPartDirectory(), prefix, part_to_detach->name); auto lock = lockParts(); bool removed_active_part = false; @@ -3270,9 +3280,7 @@ void MergeTreeData::forcefullyMovePartToDetachedAndRemoveFromMemory(const MergeT } modifyPartState(it_part, DataPartState::Deleting); - - part->renameToDetached(prefix); - + preparePartForRemoval(part)->renameToDetached(prefix); data_parts_indexes.erase(it_part); if (restore_covered && part->info.level == 0) @@ -3426,7 +3434,7 @@ void MergeTreeData::tryRemovePartImmediately(DataPartPtr && part) try { - part_to_delete->remove(); + preparePartForRemoval(part_to_delete)->remove(); } catch (...) { @@ -3636,9 +3644,9 @@ void MergeTreeData::swapActivePart(MergeTreeData::DataPartPtr part_copy) /// when allow_remote_fs_zero_copy_replication turned on and off again original_active_part->force_keep_shared_data = false; - if (original_active_part->data_part_storage->supportZeroCopyReplication() && - part_copy->data_part_storage->supportZeroCopyReplication() && - original_active_part->data_part_storage->getUniqueId() == part_copy->data_part_storage->getUniqueId()) + if (original_active_part->getDataPartStorage().supportZeroCopyReplication() && + part_copy->getDataPartStorage().supportZeroCopyReplication() && + original_active_part->getDataPartStorage().getUniqueId() == part_copy->getDataPartStorage().getUniqueId()) { /// May be when several volumes use the same S3/HDFS storage original_active_part->force_keep_shared_data = true; @@ -3658,7 +3666,7 @@ void MergeTreeData::swapActivePart(MergeTreeData::DataPartPtr part_copy) /// All other locks are taken in StorageReplicatedMergeTree lockSharedData(*part_copy); - original_active_part->data_part_storage->writeDeleteOnDestroyMarker(log); + original_active_part->getDataPartStorage().writeDeleteOnDestroyMarker(log); return; } } @@ -3792,9 +3800,9 @@ MergeTreeData::DataPartPtr MergeTreeData::getPartIfExists(const String & part_na static void loadPartAndFixMetadataImpl(MergeTreeData::MutableDataPartPtr part) { part->loadColumnsChecksumsIndexes(false, true); - part->modification_time = part->data_part_storage->getLastModified().epochTime(); - part->data_part_storage->removeDeleteOnDestroyMarker(); - part->data_part_storage->removeVersionMetadata(); + part->modification_time = part->getDataPartStorage().getLastModified().epochTime(); + part->getDataPartStorage().removeDeleteOnDestroyMarker(); + part->getDataPartStorage().removeVersionMetadata(); } void MergeTreeData::calculateColumnAndSecondaryIndexSizesImpl() @@ -3954,7 +3962,7 @@ void MergeTreeData::movePartitionToDisk(const ASTPtr & partition, const String & auto disk = getStoragePolicy()->getDiskByName(name); std::erase_if(parts, [&](auto part_ptr) { - return part_ptr->data_part_storage->getDiskName() == disk->getName(); + return part_ptr->getDataPartStorage().getDiskName() == disk->getName(); }); if (parts.empty()) @@ -4004,7 +4012,7 @@ void MergeTreeData::movePartitionToVolume(const ASTPtr & partition, const String { for (const auto & disk : volume->getDisks()) { - if (part_ptr->data_part_storage->getDiskName() == disk->getName()) + if (part_ptr->getDataPartStorage().getDiskName() == disk->getName()) { return true; } @@ -4201,7 +4209,7 @@ BackupEntries MergeTreeData::backupParts(const DataPartsVector & data_parts, con make_temporary_hard_links = false; hold_storage_and_part_ptrs = true; } - else if (supportsReplication() && part->data_part_storage->supportZeroCopyReplication() && getSettings()->allow_remote_fs_zero_copy_replication) + else if (supportsReplication() && part->getDataPartStorage().supportZeroCopyReplication() && getSettings()->allow_remote_fs_zero_copy_replication) { /// Hard links don't work correctly with zero copy replication. make_temporary_hard_links = false; @@ -4213,7 +4221,7 @@ BackupEntries MergeTreeData::backupParts(const DataPartsVector & data_parts, con table_lock = lockForShare(local_context->getCurrentQueryId(), local_context->getSettingsRef().lock_acquire_timeout); BackupEntries backup_entries_from_part; - part->data_part_storage->backup( + part->getDataPartStorage().backup( part->checksums, part->getFileNamesWithoutChecksums(), data_path_in_backup, @@ -4224,7 +4232,7 @@ BackupEntries MergeTreeData::backupParts(const DataPartsVector & data_parts, con auto projection_parts = part->getProjectionParts(); for (const auto & [projection_name, projection_part] : projection_parts) { - projection_part->data_part_storage->backup( + projection_part->getDataPartStorage().backup( projection_part->checksums, projection_part->getFileNamesWithoutChecksums(), fs::path{data_path_in_backup} / part->name, @@ -4900,16 +4908,16 @@ ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, SpacePtr space) return checkAndReturnReservation(expected_size, std::move(reservation)); } -ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage) +ReservationPtr MergeTreeData::reserveSpace(UInt64 expected_size, const IDataPartStorage & data_part_storage) { expected_size = std::max(RESERVATION_MIN_ESTIMATION_SIZE, expected_size); - return data_part_storage->reserve(expected_size); + return data_part_storage.reserve(expected_size); } -ReservationPtr MergeTreeData::tryReserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage) +ReservationPtr MergeTreeData::tryReserveSpace(UInt64 expected_size, const IDataPartStorage & data_part_storage) { expected_size = std::max(RESERVATION_MIN_ESTIMATION_SIZE, expected_size); - return data_part_storage->tryReserve(expected_size); + return data_part_storage.tryReserve(expected_size); } ReservationPtr MergeTreeData::tryReserveSpace(UInt64 expected_size, SpacePtr space) @@ -5058,11 +5066,11 @@ bool MergeTreeData::isPartInTTLDestination(const TTLDescription & ttl, const IMe if (ttl.destination_type == DataDestinationType::VOLUME) { for (const auto & disk : policy->getVolumeByName(ttl.destination_name)->getDisks()) - if (disk->getName() == part.data_part_storage->getDiskName()) + if (disk->getName() == part.getDataPartStorage().getDiskName()) return true; } else if (ttl.destination_type == DataDestinationType::DISK) - return policy->getDiskByName(ttl.destination_name)->getName() == part.data_part_storage->getDiskName(); + return policy->getDiskByName(ttl.destination_name)->getName() == part.getDataPartStorage().getDiskName(); return false; } @@ -5134,7 +5142,7 @@ void MergeTreeData::Transaction::rollbackPartsToTemporaryState() WriteBufferFromOwnString buf; buf << " Rollbacking parts state to temporary and removing from working set:"; for (const auto & part : precommitted_parts) - buf << " " << part->data_part_storage->getPartDirectory(); + buf << " " << part->getDataPartStorage().getPartDirectory(); buf << "."; LOG_DEBUG(data.log, "Undoing transaction.{}", buf.str()); @@ -5159,7 +5167,7 @@ void MergeTreeData::Transaction::rollback() WriteBufferFromOwnString buf; buf << " Removing parts:"; for (const auto & part : precommitted_parts) - buf << " " << part->data_part_storage->getPartDirectory(); + buf << " " << part->getDataPartStorage().getPartDirectory(); buf << "."; LOG_DEBUG(data.log, "Undoing transaction.{}", buf.str()); @@ -5188,8 +5196,8 @@ MergeTreeData::DataPartsVector MergeTreeData::Transaction::commit(MergeTreeData: auto * owing_parts_lock = acquired_parts_lock ? acquired_parts_lock : &parts_lock; for (const auto & part : precommitted_parts) - if (part->data_part_storage->hasActiveTransaction()) - part->data_part_storage->commitTransaction(); + if (part->getDataPartStorage().hasActiveTransaction()) + part->getDataPartStorage().commitTransaction(); bool commit_to_wal = has_in_memory_parts && settings->in_memory_parts_enable_wal; if (txn || commit_to_wal) @@ -6210,7 +6218,7 @@ std::pair MergeTreeData::cloneAn bool does_storage_policy_allow_same_disk = false; for (const DiskPtr & disk : getStoragePolicy()->getDisks()) { - if (disk->getName() == src_part->data_part_storage->getDiskName()) + if (disk->getName() == src_part->getDataPartStorage().getDiskName()) { does_storage_policy_allow_same_disk = true; break; @@ -6220,7 +6228,7 @@ std::pair MergeTreeData::cloneAn throw Exception( ErrorCodes::BAD_ARGUMENTS, "Could not clone and load part {} because disk does not belong to storage policy", - quoteString(src_part->data_part_storage->getFullPath())); + quoteString(src_part->getDataPartStorage().getFullPath())); String dst_part_name = src_part->getNewName(dst_part_info); assert(!tmp_part_prefix.empty()); @@ -6228,9 +6236,8 @@ std::pair MergeTreeData::cloneAn auto temporary_directory_lock = getTemporaryPartDirectoryHolder(tmp_dst_part_name); /// Why it is needed if we only hardlink files? - auto reservation = src_part->data_part_storage->reserve(src_part->getBytesOnDisk()); - - auto src_part_storage = src_part->data_part_storage; + auto reservation = src_part->getDataPartStorage().reserve(src_part->getBytesOnDisk()); + auto src_part_storage = src_part->getDataPartStoragePtr(); /// If source part is in memory, flush it to disk and clone it already in on-disk format if (auto src_part_in_memory = asInMemoryPart(src_part)) @@ -6257,7 +6264,7 @@ std::pair MergeTreeData::cloneAn hardlinked_files->source_part_name = src_part->name; hardlinked_files->source_table_shared_id = src_part->storage.getTableSharedID(); - for (auto it = src_part->data_part_storage->iterate(); it->isValid(); it->next()) + for (auto it = src_part->getDataPartStorage().iterate(); it->isValid(); it->next()) { if (!files_to_copy_instead_of_hardlinks.contains(it->name()) && it->name() != IMergeTreeDataPart::DELETE_ON_DESTROY_MARKER_FILE_NAME @@ -6316,14 +6323,14 @@ Strings MergeTreeData::getDataPaths() const void MergeTreeData::reportBrokenPart(MergeTreeData::DataPartPtr & data_part) const { - if (data_part->data_part_storage && data_part->data_part_storage->isBroken()) + if (data_part->getDataPartStorage().isBroken()) { auto parts = getDataPartsForInternalUsage(); - LOG_WARNING(log, "Scanning parts to recover on broken disk {}@{}.", data_part->data_part_storage->getDiskName(), data_part->data_part_storage->getDiskPath()); + LOG_WARNING(log, "Scanning parts to recover on broken disk {}@{}.", data_part->getDataPartStorage().getDiskName(), data_part->getDataPartStorage().getDiskPath()); for (const auto & part : parts) { - if (part->data_part_storage && part->data_part_storage->getDiskName() == data_part->data_part_storage->getDiskName()) + if (part->getDataPartStorage().getDiskName() == data_part->getDataPartStorage().getDiskName()) broken_part_callback(part->name); } } @@ -6414,7 +6421,7 @@ PartitionCommandsResultInfo MergeTreeData::freezePartitionsByMatcher( LOG_DEBUG(log, "Freezing part {} snapshot will be placed at {}", part->name, backup_path); - auto data_part_storage = part->data_part_storage; + auto data_part_storage = part->getDataPartStoragePtr(); String src_part_path = data_part_storage->getRelativePath(); String backup_part_path = fs::path(backup_path) / relative_data_path; if (auto part_in_memory = asInMemoryPart(part)) @@ -6428,12 +6435,12 @@ PartitionCommandsResultInfo MergeTreeData::freezePartitionsByMatcher( // Store metadata for replicated table. // Do nothing for non-replicated. - createAndStoreFreezeMetadata(disk, part, fs::path(backup_part_path) / part->data_part_storage->getPartDirectory()); + createAndStoreFreezeMetadata(disk, part, fs::path(backup_part_path) / part->getDataPartStorage().getPartDirectory()); }; auto new_storage = data_part_storage->freeze( backup_part_path, - part->data_part_storage->getPartDirectory(), + part->getDataPartStorage().getPartDirectory(), /*make_source_readonly*/ true, callback, /*copy_instead_of_hardlink*/ false, @@ -6555,8 +6562,8 @@ try if (result_part) { - part_log_elem.disk_name = result_part->data_part_storage->getDiskName(); - part_log_elem.path_on_disk = result_part->data_part_storage->getFullPath(); + part_log_elem.disk_name = result_part->getDataPartStorage().getDiskName(); + part_log_elem.path_on_disk = result_part->getDataPartStorage().getFullPath(); part_log_elem.bytes_compressed_on_disk = result_part->getBytesOnDisk(); part_log_elem.rows = result_part->rows_count; part_log_elem.part_type = result_part->getType(); @@ -6712,7 +6719,7 @@ bool MergeTreeData::moveParts(const CurrentlyMovingPartsTaggerPtr & moving_tagge for (const auto & moving_part : moving_tagger->parts_to_move) { Stopwatch stopwatch; - DataPartPtr cloned_part; + MutableDataPartPtr cloned_part; auto write_part_log = [&](const ExecutionStatus & execution_status) { @@ -6975,7 +6982,7 @@ ReservationPtr MergeTreeData::balancedReservation( if (part->isStoredOnDisk() && part->getBytesOnDisk() >= min_bytes_to_rebalance_partition_over_jbod && part_info.partition_id == part->info.partition_id) { - auto name = part->data_part_storage->getDiskName(); + auto name = part->getDataPartStorage().getDiskName(); auto it = disk_occupation.find(name); if (it != disk_occupation.end()) { diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 327718d15ed..2b67face570 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -214,6 +214,7 @@ public: }; using DataParts = std::set; + using MutableDataParts = std::set; using DataPartsVector = std::vector; using DataPartsLock = std::unique_lock; @@ -275,8 +276,8 @@ public: MergeTreeData & data; MergeTreeTransaction * txn; - DataParts precommitted_parts; - DataParts locked_parts; + MutableDataParts precommitted_parts; + MutableDataParts locked_parts; bool has_in_memory_parts = false; void clear(); @@ -413,8 +414,8 @@ public: SelectQueryInfo & info) const override; ReservationPtr reserveSpace(UInt64 expected_size, VolumePtr & volume) const; - static ReservationPtr tryReserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage); - static ReservationPtr reserveSpace(UInt64 expected_size, const MutableDataPartStoragePtr & data_part_storage); + static ReservationPtr tryReserveSpace(UInt64 expected_size, const IDataPartStorage & data_part_storage); + static ReservationPtr reserveSpace(UInt64 expected_size, const IDataPartStorage & data_part_storage); static bool partsContainSameProjections(const DataPartPtr & left, const DataPartPtr & right); @@ -974,7 +975,7 @@ public: /// Fetch part only if some replica has it on shared storage like S3 /// Overridden in StorageReplicatedMergeTree - virtual bool tryToFetchIfShared(const IMergeTreeDataPart &, const DiskPtr &, const String &) { return false; } + virtual MutableDataPartStoragePtr tryToFetchIfShared(const IMergeTreeDataPart &, const DiskPtr &, const String &) { return nullptr; } /// Check shared data usage on other replicas for detached/freezed part /// Remove local files and remote files if needed @@ -1259,7 +1260,6 @@ protected: static void incrementMergedPartsProfileEvent(MergeTreeDataPartType type); private: - /// Checking that candidate part doesn't break invariants: correct partition and doesn't exist already void checkPartCanBeAddedToTable(MutableDataPartPtr & part, DataPartsLock & lock) const; @@ -1328,8 +1328,8 @@ private: virtual std::unique_ptr getDefaultSettings() const = 0; void loadDataPartsFromDisk( - DataPartsVector & broken_parts_to_detach, - DataPartsVector & duplicate_parts_to_remove, + MutableDataPartsVector & broken_parts_to_detach, + MutableDataPartsVector & duplicate_parts_to_remove, ThreadPool & pool, size_t num_parts, std::queue>> & parts_queue, @@ -1337,8 +1337,7 @@ private: const MergeTreeSettingsPtr & settings); void loadDataPartsFromWAL( - DataPartsVector & broken_parts_to_detach, - DataPartsVector & duplicate_parts_to_remove, + MutableDataPartsVector & duplicate_parts_to_remove, MutableDataPartsVector & parts_from_wal); /// Create zero-copy exclusive lock for part and disk. Useful for coordination of @@ -1350,6 +1349,8 @@ private: /// Otherwise, in non-parallel case will break and return. void clearPartsFromFilesystemImpl(const DataPartsVector & parts, NameSet * part_names_succeed); + static MutableDataPartPtr preparePartForRemoval(const DataPartPtr & part); + TemporaryParts temporary_parts; }; diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp index c96003c8938..4d5dea94d44 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.cpp @@ -482,7 +482,7 @@ MergeTaskPtr MergeTreeDataMergerMutator::mergePartsToTemporaryPart( const Names & deduplicate_by_columns, const MergeTreeData::MergingParams & merging_params, const MergeTreeTransactionPtr & txn, - const IMergeTreeDataPart * parent_part, + IMergeTreeDataPart * parent_part, const String & suffix) { return std::make_shared( diff --git a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h index d1b9d3c99e7..5d98f526325 100644 --- a/src/Storages/MergeTree/MergeTreeDataMergerMutator.h +++ b/src/Storages/MergeTree/MergeTreeDataMergerMutator.h @@ -113,7 +113,7 @@ public: const Names & deduplicate_by_columns, const MergeTreeData::MergingParams & merging_params, const MergeTreeTransactionPtr & txn, - const IMergeTreeDataPart * parent_part = nullptr, + IMergeTreeDataPart * parent_part = nullptr, const String & suffix = ""); /// Mutate a single data part with the specified commands. Will create and return a temporary part. diff --git a/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp index 4fedacee13c..a537b44d9ea 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartCompact.cpp @@ -96,21 +96,21 @@ void MergeTreeDataPartCompact::calculateEachColumnSizes(ColumnSizeByName & /*eac void MergeTreeDataPartCompact::loadIndexGranularityImpl( MergeTreeIndexGranularity & index_granularity_, const MergeTreeIndexGranularityInfo & index_granularity_info_, - size_t columns_count, const DataPartStoragePtr & data_part_storage_) + size_t columns_count, const IDataPartStorage & data_part_storage_) { if (!index_granularity_info_.mark_type.adaptive) throw Exception("MergeTreeDataPartCompact cannot be created with non-adaptive granulary.", ErrorCodes::NOT_IMPLEMENTED); auto marks_file_path = index_granularity_info_.getMarksFilePath("data"); - if (!data_part_storage_->exists(marks_file_path)) + if (!data_part_storage_.exists(marks_file_path)) throw Exception( ErrorCodes::NO_FILE_IN_DATA_PART, "Marks file '{}' doesn't exist", - std::string(fs::path(data_part_storage_->getFullPath()) / marks_file_path)); + std::string(fs::path(data_part_storage_.getFullPath()) / marks_file_path)); - size_t marks_file_size = data_part_storage_->getFileSize(marks_file_path); + size_t marks_file_size = data_part_storage_.getFileSize(marks_file_path); - std::unique_ptr buffer = data_part_storage_->readFile( + std::unique_ptr buffer = data_part_storage_.readFile( marks_file_path, ReadSettings().adjustBufferSize(marks_file_size), marks_file_size, std::nullopt); std::unique_ptr marks_reader; @@ -139,7 +139,7 @@ void MergeTreeDataPartCompact::loadIndexGranularity() if (columns.empty()) throw Exception("No columns in part " + name, ErrorCodes::NO_FILE_IN_DATA_PART); - loadIndexGranularityImpl(index_granularity, index_granularity_info, columns.size(), data_part_storage); + loadIndexGranularityImpl(index_granularity, index_granularity_info, columns.size(), getDataPartStorage()); } bool MergeTreeDataPartCompact::hasColumnFiles(const NameAndTypePair & column) const @@ -170,12 +170,12 @@ void MergeTreeDataPartCompact::checkConsistency(bool require_part_metadata) cons throw Exception( ErrorCodes::NO_FILE_IN_DATA_PART, "No marks file checksum for column in part {}", - data_part_storage->getFullPath()); + getDataPartStorage().getFullPath()); if (!checksums.files.contains(DATA_FILE_NAME_WITH_EXTENSION)) throw Exception( ErrorCodes::NO_FILE_IN_DATA_PART, "No data file checksum for in part {}", - data_part_storage->getFullPath()); + getDataPartStorage().getFullPath()); } } else @@ -183,33 +183,33 @@ void MergeTreeDataPartCompact::checkConsistency(bool require_part_metadata) cons { /// count.txt should be present even in non custom-partitioned parts std::string file_path = "count.txt"; - if (!data_part_storage->exists(file_path) || data_part_storage->getFileSize(file_path) == 0) + if (!getDataPartStorage().exists(file_path) || getDataPartStorage().getFileSize(file_path) == 0) throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Part {} is broken: {} is empty", - data_part_storage->getRelativePath(), - std::string(fs::path(data_part_storage->getFullPath()) / file_path)); + getDataPartStorage().getRelativePath(), + std::string(fs::path(getDataPartStorage().getFullPath()) / file_path)); } /// Check that marks are nonempty and have the consistent size with columns number. - if (data_part_storage->exists(mrk_file_name)) + if (getDataPartStorage().exists(mrk_file_name)) { - UInt64 file_size = data_part_storage->getFileSize(mrk_file_name); + UInt64 file_size = getDataPartStorage().getFileSize(mrk_file_name); if (!file_size) throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Part {} is broken: {} is empty.", - data_part_storage->getRelativePath(), - std::string(fs::path(data_part_storage->getFullPath()) / mrk_file_name)); + getDataPartStorage().getRelativePath(), + std::string(fs::path(getDataPartStorage().getFullPath()) / mrk_file_name)); UInt64 expected_file_size = index_granularity_info.getMarkSizeInBytes(columns.size()) * index_granularity.getMarksCount(); if (expected_file_size != file_size) throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Part {} is broken: bad size of marks file '{}': {}, must be: {}", - data_part_storage->getRelativePath(), - std::string(fs::path(data_part_storage->getFullPath()) / mrk_file_name), + getDataPartStorage().getRelativePath(), + std::string(fs::path(getDataPartStorage().getFullPath()) / mrk_file_name), std::to_string(file_size), std::to_string(expected_file_size)); } } @@ -217,12 +217,12 @@ void MergeTreeDataPartCompact::checkConsistency(bool require_part_metadata) cons bool MergeTreeDataPartCompact::isStoredOnRemoteDisk() const { - return data_part_storage->isStoredOnRemoteDisk(); + return getDataPartStorage().isStoredOnRemoteDisk(); } bool MergeTreeDataPartCompact::isStoredOnRemoteDiskWithZeroCopySupport() const { - return data_part_storage->supportZeroCopyReplication(); + return getDataPartStorage().supportZeroCopyReplication(); } MergeTreeDataPartCompact::~MergeTreeDataPartCompact() diff --git a/src/Storages/MergeTree/MergeTreeDataPartCompact.h b/src/Storages/MergeTree/MergeTreeDataPartCompact.h index 7c3fe012616..e275c586cb9 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartCompact.h +++ b/src/Storages/MergeTree/MergeTreeDataPartCompact.h @@ -67,7 +67,7 @@ public: protected: static void loadIndexGranularityImpl( MergeTreeIndexGranularity & index_granularity_, const MergeTreeIndexGranularityInfo & index_granularity_info_, - size_t columns_count, const DataPartStoragePtr & data_part_storage_); + size_t columns_count, const IDataPartStorage & data_part_storage_); private: void checkConsistency(bool require_part_metadata) const override; diff --git a/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp b/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp index c30efbc7969..48b1b6bab60 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartInMemory.cpp @@ -6,6 +6,7 @@ #include #include #include +#include #include #include #include @@ -71,12 +72,18 @@ IMergeTreeDataPart::MergeTreeWriterPtr MergeTreeDataPartInMemory::getWriter( MutableDataPartStoragePtr MergeTreeDataPartInMemory::flushToDisk(const String & new_relative_path, const StorageMetadataPtr & metadata_snapshot) const { - auto current_full_path = data_part_storage->getFullPath(); - auto new_data_part_storage = data_part_storage->clone(); + auto reservation = storage.reserveSpace(block.bytes(), getDataPartStorage()); + VolumePtr volume = storage.getStoragePolicy()->getVolume(0); + VolumePtr data_part_volume = createVolumeFromReservation(reservation, volume); + + auto new_data_part_storage = std::make_shared( + data_part_volume, + storage.getRelativeDataPath(), + new_relative_path); - new_data_part_storage->setRelativePath(new_relative_path); new_data_part_storage->beginTransaction(); + auto current_full_path = getDataPartStorage().getFullPath(); auto new_type = storage.choosePartTypeOnDisk(block.bytes(), rows_count); auto new_data_part = storage.createPart(name, new_type, info, new_data_part_storage); @@ -148,12 +155,9 @@ void MergeTreeDataPartInMemory::makeCloneInDetached(const String & prefix, const flushToDisk(detached_path, metadata_snapshot); } -void MergeTreeDataPartInMemory::renameTo(const String & new_relative_path, bool /* remove_new_dir_if_exists */) const +void MergeTreeDataPartInMemory::renameTo(const String & new_relative_path, bool /* remove_new_dir_if_exists */) { - data_part_storage->setRelativePath(new_relative_path); - - if (data_part_storage) - data_part_storage->setRelativePath(new_relative_path); + getDataPartStorage().setRelativePath(new_relative_path); } void MergeTreeDataPartInMemory::calculateEachColumnSizes(ColumnSizeByName & each_columns_size, ColumnSize & total_size) const diff --git a/src/Storages/MergeTree/MergeTreeDataPartInMemory.h b/src/Storages/MergeTree/MergeTreeDataPartInMemory.h index 49bc5eff1ea..e58701b04a1 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartInMemory.h +++ b/src/Storages/MergeTree/MergeTreeDataPartInMemory.h @@ -46,7 +46,7 @@ public: bool isStoredOnRemoteDiskWithZeroCopySupport() const override { return false; } bool hasColumnFiles(const NameAndTypePair & column) const override { return !!getColumnPosition(column.getNameInStorage()); } String getFileNameForColumn(const NameAndTypePair & /* column */) const override { return ""; } - void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) const override; + void renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) override; void makeCloneInDetached(const String & prefix, const StorageMetadataPtr & metadata_snapshot) const override; MutableDataPartStoragePtr flushToDisk(const String & new_relative_path, const StorageMetadataPtr & metadata_snapshot) const; diff --git a/src/Storages/MergeTree/MergeTreeDataPartWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWide.cpp index 18467f2cef7..2418960f992 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWide.cpp @@ -104,18 +104,18 @@ ColumnSize MergeTreeDataPartWide::getColumnSizeImpl( void MergeTreeDataPartWide::loadIndexGranularityImpl( MergeTreeIndexGranularity & index_granularity_, MergeTreeIndexGranularityInfo & index_granularity_info_, - const DataPartStoragePtr & data_part_storage_, const std::string & any_column_file_name) + const IDataPartStorage & data_part_storage_, const std::string & any_column_file_name) { index_granularity_info_.changeGranularityIfRequired(data_part_storage_); /// We can use any column, it doesn't matter std::string marks_file_path = index_granularity_info_.getMarksFilePath(any_column_file_name); - if (!data_part_storage_->exists(marks_file_path)) + if (!data_part_storage_.exists(marks_file_path)) throw Exception( ErrorCodes::NO_FILE_IN_DATA_PART, "Marks file '{}' doesn't exist", - std::string(fs::path(data_part_storage_->getFullPath()) / marks_file_path)); + std::string(fs::path(data_part_storage_.getFullPath()) / marks_file_path)); - size_t marks_file_size = data_part_storage_->getFileSize(marks_file_path); + size_t marks_file_size = data_part_storage_.getFileSize(marks_file_path); if (!index_granularity_info_.mark_type.adaptive && !index_granularity_info_.mark_type.compressed) { @@ -125,7 +125,7 @@ void MergeTreeDataPartWide::loadIndexGranularityImpl( } else { - auto marks_file = data_part_storage_->readFile(marks_file_path, ReadSettings().adjustBufferSize(marks_file_size), marks_file_size, std::nullopt); + auto marks_file = data_part_storage_.readFile(marks_file_path, ReadSettings().adjustBufferSize(marks_file_size), marks_file_size, std::nullopt); std::unique_ptr marks_reader; if (!index_granularity_info_.mark_type.compressed) @@ -162,18 +162,18 @@ void MergeTreeDataPartWide::loadIndexGranularity() if (columns.empty()) throw Exception("No columns in part " + name, ErrorCodes::NO_FILE_IN_DATA_PART); - loadIndexGranularityImpl(index_granularity, index_granularity_info, data_part_storage, getFileNameForColumn(columns.front())); + loadIndexGranularityImpl(index_granularity, index_granularity_info, getDataPartStorage(), getFileNameForColumn(columns.front())); } bool MergeTreeDataPartWide::isStoredOnRemoteDisk() const { - return data_part_storage->isStoredOnRemoteDisk(); + return getDataPartStorage().isStoredOnRemoteDisk(); } bool MergeTreeDataPartWide::isStoredOnRemoteDiskWithZeroCopySupport() const { - return data_part_storage->supportZeroCopyReplication(); + return getDataPartStorage().supportZeroCopyReplication(); } MergeTreeDataPartWide::~MergeTreeDataPartWide() @@ -202,13 +202,13 @@ void MergeTreeDataPartWide::checkConsistency(bool require_part_metadata) const throw Exception( ErrorCodes::NO_FILE_IN_DATA_PART, "No {} file checksum for column {} in part {} ", - mrk_file_name, name_type.name, data_part_storage->getFullPath()); + mrk_file_name, name_type.name, getDataPartStorage().getFullPath()); if (!checksums.files.contains(bin_file_name)) throw Exception( ErrorCodes::NO_FILE_IN_DATA_PART, "No {} file checksum for column {} in part ", - bin_file_name, name_type.name, data_part_storage->getFullPath()); + bin_file_name, name_type.name, getDataPartStorage().getFullPath()); }); } } @@ -224,23 +224,23 @@ void MergeTreeDataPartWide::checkConsistency(bool require_part_metadata) const auto file_path = ISerialization::getFileNameForStream(name_type, substream_path) + marks_file_extension; /// Missing file is Ok for case when new column was added. - if (data_part_storage->exists(file_path)) + if (getDataPartStorage().exists(file_path)) { - UInt64 file_size = data_part_storage->getFileSize(file_path); + UInt64 file_size = getDataPartStorage().getFileSize(file_path); if (!file_size) throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Part {} is broken: {} is empty.", - data_part_storage->getFullPath(), - std::string(fs::path(data_part_storage->getFullPath()) / file_path)); + getDataPartStorage().getFullPath(), + std::string(fs::path(getDataPartStorage().getFullPath()) / file_path)); if (!marks_size) marks_size = file_size; else if (file_size != *marks_size) throw Exception( ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, - "Part {} is broken: marks have different sizes.", data_part_storage->getFullPath()); + "Part {} is broken: marks have different sizes.", getDataPartStorage().getFullPath()); } }); } diff --git a/src/Storages/MergeTree/MergeTreeDataPartWide.h b/src/Storages/MergeTree/MergeTreeDataPartWide.h index 4343148b175..601bdff51a1 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWide.h +++ b/src/Storages/MergeTree/MergeTreeDataPartWide.h @@ -1,5 +1,6 @@ #pragma once +#include "Storages/MergeTree/IDataPartStorage.h" #include namespace DB @@ -63,7 +64,7 @@ public: protected: static void loadIndexGranularityImpl( MergeTreeIndexGranularity & index_granularity_, MergeTreeIndexGranularityInfo & index_granularity_info_, - const DataPartStoragePtr & data_part_storage_, const std::string & any_column_file_name); + const IDataPartStorage & data_part_storage_, const std::string & any_column_file_name); private: void checkConsistency(bool require_part_metadata) const override; diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp index 6821c52f0d2..c9f3c3b5101 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterCompact.cpp @@ -21,13 +21,13 @@ MergeTreeDataPartWriterCompact::MergeTreeDataPartWriterCompact( : MergeTreeDataPartWriterOnDisk(data_part_, columns_list_, metadata_snapshot_, indices_to_recalc_, marks_file_extension_, default_codec_, settings_, index_granularity_) - , plain_file(data_part_->data_part_storage->writeFile( + , plain_file(data_part_->getDataPartStorage().writeFile( MergeTreeDataPartCompact::DATA_FILE_NAME_WITH_EXTENSION, settings.max_compress_block_size, settings_.query_write_settings)) , plain_hashing(*plain_file) { - marks_file = data_part_->data_part_storage->writeFile( + marks_file = data_part_->getDataPartStorage().writeFile( MergeTreeDataPartCompact::DATA_FILE_NAME + marks_file_extension_, 4096, settings_.query_write_settings); diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp index 383cd79734f..d085bb29b20 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp @@ -114,8 +114,8 @@ MergeTreeDataPartWriterOnDisk::MergeTreeDataPartWriterOnDisk( if (settings.blocks_are_granules_size && !index_granularity.empty()) throw Exception("Can't take information about index granularity from blocks, when non empty index_granularity array specified", ErrorCodes::LOGICAL_ERROR); - if (!data_part->data_part_storage->exists()) - data_part->data_part_storage->createDirectories(); + if (!data_part->getDataPartStorage().exists()) + data_part->getDataPartStorage().createDirectories(); if (settings.rewrite_primary_key) initPrimaryIndex(); @@ -176,7 +176,7 @@ void MergeTreeDataPartWriterOnDisk::initPrimaryIndex() if (metadata_snapshot->hasPrimaryKey()) { String index_name = "primary" + getIndexExtension(compress_primary_key); - index_file_stream = data_part->data_part_storage->writeFile(index_name, DBMS_DEFAULT_BUFFER_SIZE, settings.query_write_settings); + index_file_stream = data_part->getDataPartStorage().writeFile(index_name, DBMS_DEFAULT_BUFFER_SIZE, settings.query_write_settings); index_file_hashing_stream = std::make_unique(*index_file_stream); if (compress_primary_key) @@ -202,7 +202,7 @@ void MergeTreeDataPartWriterOnDisk::initSkipIndices() skip_indices_streams.emplace_back( std::make_unique( stream_name, - data_part->data_part_storage, + data_part->getDataPartStoragePtr(), stream_name, index_helper->getSerializedFileExtension(), stream_name, marks_file_extension, default_codec, settings.max_compress_block_size, diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp index 5c6e69abe4d..f48b350a981 100644 --- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp +++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp @@ -116,7 +116,7 @@ void MergeTreeDataPartWriterWide::addStreams( column_streams[stream_name] = std::make_unique( stream_name, - data_part->data_part_storage, + data_part->getDataPartStoragePtr(), stream_name, DATA_FILE_EXTENSION, stream_name, marks_file_extension, compression_codec, @@ -421,17 +421,17 @@ void MergeTreeDataPartWriterWide::validateColumnOfFixedSize(const NameAndTypePai String bin_path = escaped_name + DATA_FILE_EXTENSION; /// Some columns may be removed because of ttl. Skip them. - if (!data_part->data_part_storage->exists(mrk_path)) + if (!data_part->getDataPartStorage().exists(mrk_path)) return; - auto mrk_file_in = data_part->data_part_storage->readFile(mrk_path, {}, std::nullopt, std::nullopt); + auto mrk_file_in = data_part->getDataPartStorage().readFile(mrk_path, {}, std::nullopt, std::nullopt); std::unique_ptr mrk_in; if (data_part->index_granularity_info.mark_type.compressed) mrk_in = std::make_unique(std::move(mrk_file_in)); else mrk_in = std::move(mrk_file_in); - DB::CompressedReadBufferFromFile bin_in(data_part->data_part_storage->readFile(bin_path, {}, std::nullopt, std::nullopt)); + DB::CompressedReadBufferFromFile bin_in(data_part->getDataPartStorage().readFile(bin_path, {}, std::nullopt, std::nullopt)); bool must_be_last = false; UInt64 offset_in_compressed_file = 0; UInt64 offset_in_decompressed_block = 0; @@ -482,7 +482,7 @@ void MergeTreeDataPartWriterWide::validateColumnOfFixedSize(const NameAndTypePai if (index_granularity_rows != index_granularity.getMarkRows(mark_num)) throw Exception( ErrorCodes::LOGICAL_ERROR, "Incorrect mark rows for part {} for mark #{} (compressed offset {}, decompressed offset {}), in-memory {}, on disk {}, total marks {}", - data_part->data_part_storage->getFullPath(), mark_num, offset_in_compressed_file, offset_in_decompressed_block, index_granularity.getMarkRows(mark_num), index_granularity_rows, index_granularity.getMarksCount()); + data_part->getDataPartStorage().getFullPath(), mark_num, offset_in_compressed_file, offset_in_decompressed_block, index_granularity.getMarkRows(mark_num), index_granularity_rows, index_granularity.getMarksCount()); auto column = type->createColumn(); diff --git a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp index 91ecb3a37a0..b99b9047308 100644 --- a/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp +++ b/src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp @@ -1618,10 +1618,10 @@ MarkRanges MergeTreeDataSelectExecutor::filterMarksUsingIndex( UncompressedCache * uncompressed_cache, Poco::Logger * log) { - if (!index_helper->getDeserializedFormat(part->data_part_storage, index_helper->getFileName())) + if (!index_helper->getDeserializedFormat(part->getDataPartStorage(), index_helper->getFileName())) { LOG_DEBUG(log, "File for index {} does not exist ({}.*). Skipping it.", backQuote(index_helper->index.name), - (fs::path(part->data_part_storage->getFullPath()) / index_helper->getFileName()).string()); + (fs::path(part->getDataPartStorage().getFullPath()) / index_helper->getFileName()).string()); return ranges; } @@ -1736,7 +1736,7 @@ MarkRanges MergeTreeDataSelectExecutor::filterMarksUsingMergedIndex( { for (const auto & index_helper : indices) { - if (!part->data_part_storage->exists(index_helper->getFileName() + ".idx")) + if (!part->getDataPartStorage().exists(index_helper->getFileName() + ".idx")) { LOG_DEBUG(log, "File for index {} does not exist. Skipping it.", backQuote(index_helper->index.name)); return ranges; diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 62d01a2d555..856a684d18d 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -405,9 +405,9 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( if (new_data_part->isStoredOnDisk()) { /// The name could be non-unique in case of stale files from previous runs. - String full_path = new_data_part->data_part_storage->getFullPath(); + String full_path = new_data_part->getDataPartStorage().getFullPath(); - if (new_data_part->data_part_storage->exists()) + if (new_data_part->getDataPartStorage().exists()) { LOG_WARNING(log, "Removing old temporary directory {}", full_path); data_part_storage->removeRecursive(); @@ -493,7 +493,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( const StorageMetadataPtr & metadata_snapshot = projection.metadata; MergeTreePartInfo new_part_info("all", 0, 0, 0); - auto projection_part_storage = parent_part->data_part_storage->getProjection(relative_path); + auto projection_part_storage = parent_part->getDataPartStorage().getProjection(relative_path); auto new_data_part = data.createPart( part_name, part_type, @@ -600,7 +600,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPart( /// Size of part would not be greater than block.bytes() + epsilon size_t expected_size = block.bytes(); // just check if there is enough space on parent volume - data.reserveSpace(expected_size, parent_part->data_part_storage); + data.reserveSpace(expected_size, parent_part->getDataPartStorage()); part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); } @@ -637,7 +637,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( /// Size of part would not be greater than block.bytes() + epsilon size_t expected_size = block.bytes(); // just check if there is enough space on parent volume - data.reserveSpace(expected_size, parent_part->data_part_storage); + data.reserveSpace(expected_size, parent_part->getDataPartStorage()); part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); } diff --git a/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp b/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp index 9c154f786f7..11e1f9efcc2 100644 --- a/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.cpp @@ -89,10 +89,10 @@ std::string MarkType::getFileExtension() const } -std::optional MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(const DataPartStoragePtr & data_part_storage) +std::optional MergeTreeIndexGranularityInfo::getMarksExtensionFromFilesystem(const IDataPartStorage & data_part_storage) { - if (data_part_storage->exists()) - for (auto it = data_part_storage->iterate(); it->isValid(); it->next()) + if (data_part_storage.exists()) + for (auto it = data_part_storage.iterate(); it->isValid(); it->next()) if (it->isFile()) if (std::string ext = fs::path(it->name()).extension(); MarkType::isMarkFileExtension(ext)) return ext; @@ -110,7 +110,7 @@ MergeTreeIndexGranularityInfo::MergeTreeIndexGranularityInfo(const MergeTreeData fixed_index_granularity = storage.getSettings()->index_granularity; } -void MergeTreeIndexGranularityInfo::changeGranularityIfRequired(const DataPartStoragePtr & data_part_storage) +void MergeTreeIndexGranularityInfo::changeGranularityIfRequired(const IDataPartStorage & data_part_storage) { auto mrk_ext = getMarksExtensionFromFilesystem(data_part_storage); if (mrk_ext && !MarkType(*mrk_ext).adaptive) diff --git a/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.h b/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.h index 883fe3c899e..aed3081d3d0 100644 --- a/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.h +++ b/src/Storages/MergeTree/MergeTreeIndexGranularityInfo.h @@ -48,7 +48,7 @@ public: MergeTreeIndexGranularityInfo(MergeTreeDataPartType type_, bool is_adaptive_, size_t index_granularity_, size_t index_granularity_bytes_); - void changeGranularityIfRequired(const DataPartStoragePtr & data_part_storage); + void changeGranularityIfRequired(const IDataPartStorage & data_part_storage); String getMarksFilePath(const String & path_prefix) const { @@ -57,7 +57,7 @@ public: size_t getMarkSizeInBytes(size_t columns_num = 1) const; - static std::optional getMarksExtensionFromFilesystem(const DataPartStoragePtr & data_part_storage); + static std::optional getMarksExtensionFromFilesystem(const IDataPartStorage & data_part_storage); }; constexpr inline auto getNonAdaptiveMrkSizeWide() { return sizeof(UInt64) * 2; } diff --git a/src/Storages/MergeTree/MergeTreeIndexMinMax.cpp b/src/Storages/MergeTree/MergeTreeIndexMinMax.cpp index b190ac2b2fd..43e655a4ee5 100644 --- a/src/Storages/MergeTree/MergeTreeIndexMinMax.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexMinMax.cpp @@ -211,11 +211,11 @@ bool MergeTreeIndexMinMax::mayBenefitFromIndexForIn(const ASTPtr & node) const return false; } -MergeTreeIndexFormat MergeTreeIndexMinMax::getDeserializedFormat(const DataPartStoragePtr & data_part_storage, const std::string & relative_path_prefix) const +MergeTreeIndexFormat MergeTreeIndexMinMax::getDeserializedFormat(const IDataPartStorage & data_part_storage, const std::string & relative_path_prefix) const { - if (data_part_storage->exists(relative_path_prefix + ".idx2")) + if (data_part_storage.exists(relative_path_prefix + ".idx2")) return {2, ".idx2"}; - else if (data_part_storage->exists(relative_path_prefix + ".idx")) + else if (data_part_storage.exists(relative_path_prefix + ".idx")) return {1, ".idx"}; return {0 /* unknown */, ""}; } diff --git a/src/Storages/MergeTree/MergeTreeIndexMinMax.h b/src/Storages/MergeTree/MergeTreeIndexMinMax.h index 0566a15d535..af420613855 100644 --- a/src/Storages/MergeTree/MergeTreeIndexMinMax.h +++ b/src/Storages/MergeTree/MergeTreeIndexMinMax.h @@ -83,7 +83,7 @@ public: bool mayBenefitFromIndexForIn(const ASTPtr & node) const override; const char* getSerializedFileExtension() const override { return ".idx2"; } - MergeTreeIndexFormat getDeserializedFormat(const DataPartStoragePtr & data_part_storage, const std::string & path_prefix) const override; /// NOLINT + MergeTreeIndexFormat getDeserializedFormat(const IDataPartStorage & data_part_storage, const std::string & path_prefix) const override; /// NOLINT }; } diff --git a/src/Storages/MergeTree/MergeTreeIndexReader.cpp b/src/Storages/MergeTree/MergeTreeIndexReader.cpp index 33106f7ab64..7d7024a8ac2 100644 --- a/src/Storages/MergeTree/MergeTreeIndexReader.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexReader.cpp @@ -20,7 +20,7 @@ std::unique_ptr makeIndexReader( auto * load_marks_threadpool = settings.read_settings.load_marks_asynchronously ? &context->getLoadMarksThreadpool() : nullptr; return std::make_unique( - part->data_part_storage, + part->getDataPartStoragePtr(), index->getFileName(), extension, marks_count, all_mark_ranges, std::move(settings), mark_cache, uncompressed_cache, @@ -44,7 +44,7 @@ MergeTreeIndexReader::MergeTreeIndexReader( MergeTreeReaderSettings settings) : index(index_) { - auto index_format = index->getDeserializedFormat(part_->data_part_storage, index->getFileName()); + auto index_format = index->getDeserializedFormat(part_->getDataPartStorage(), index->getFileName()); stream = makeIndexReader( index_format.extension, diff --git a/src/Storages/MergeTree/MergeTreeIndices.h b/src/Storages/MergeTree/MergeTreeIndices.h index 14002534c94..6a671c31944 100644 --- a/src/Storages/MergeTree/MergeTreeIndices.h +++ b/src/Storages/MergeTree/MergeTreeIndices.h @@ -148,9 +148,9 @@ struct IMergeTreeIndex /// Returns extension for deserialization. /// /// Return pair. - virtual MergeTreeIndexFormat getDeserializedFormat(const DataPartStoragePtr & data_part_storage, const std::string & relative_path_prefix) const + virtual MergeTreeIndexFormat getDeserializedFormat(const IDataPartStorage & data_part_storage, const std::string & relative_path_prefix) const { - if (data_part_storage->exists(relative_path_prefix + ".idx")) + if (data_part_storage.exists(relative_path_prefix + ".idx")) return {1, ".idx"}; return {0 /*unknown*/, ""}; } diff --git a/src/Storages/MergeTree/MergeTreePartition.cpp b/src/Storages/MergeTree/MergeTreePartition.cpp index 47f07e7bb08..10f5cc95baf 100644 --- a/src/Storages/MergeTree/MergeTreePartition.cpp +++ b/src/Storages/MergeTree/MergeTreePartition.cpp @@ -382,7 +382,7 @@ void MergeTreePartition::load(const MergeTreeData & storage, const PartMetadataM partition_key_sample.getByPosition(i).type->getDefaultSerialization()->deserializeBinary(value[i], *file); } -std::unique_ptr MergeTreePartition::store(const MergeTreeData & storage, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums) const +std::unique_ptr MergeTreePartition::store(const MergeTreeData & storage, IDataPartStorage & data_part_storage, MergeTreeDataPartChecksums & checksums) const { auto metadata_snapshot = storage.getInMemoryMetadataPtr(); const auto & context = storage.getContext(); @@ -390,12 +390,12 @@ std::unique_ptr MergeTreePartition::store(const MergeTr return store(partition_key_sample, data_part_storage, checksums, context->getWriteSettings()); } -std::unique_ptr MergeTreePartition::store(const Block & partition_key_sample, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const +std::unique_ptr MergeTreePartition::store(const Block & partition_key_sample, IDataPartStorage & data_part_storage, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const { if (!partition_key_sample) return nullptr; - auto out = data_part_storage->writeFile("partition.dat", DBMS_DEFAULT_BUFFER_SIZE, settings); + auto out = data_part_storage.writeFile("partition.dat", DBMS_DEFAULT_BUFFER_SIZE, settings); HashingWriteBuffer out_hashing(*out); for (size_t i = 0; i < value.size(); ++i) { diff --git a/src/Storages/MergeTree/MergeTreePartition.h b/src/Storages/MergeTree/MergeTreePartition.h index c9c6723df5a..78b141f26ec 100644 --- a/src/Storages/MergeTree/MergeTreePartition.h +++ b/src/Storages/MergeTree/MergeTreePartition.h @@ -44,8 +44,8 @@ public: /// Store functions return write buffer with written but not finalized data. /// User must call finish() for returned object. - [[nodiscard]] std::unique_ptr store(const MergeTreeData & storage, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums) const; - [[nodiscard]] std::unique_ptr store(const Block & partition_key_sample, const MutableDataPartStoragePtr & data_part_storage, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const; + [[nodiscard]] std::unique_ptr store(const MergeTreeData & storage, IDataPartStorage & data_part_storage, MergeTreeDataPartChecksums & checksums) const; + [[nodiscard]] std::unique_ptr store(const Block & partition_key_sample, IDataPartStorage & data_part_storage, MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const; void assign(const MergeTreePartition & other) { value = other.value; } diff --git a/src/Storages/MergeTree/MergeTreePartsMover.cpp b/src/Storages/MergeTree/MergeTreePartsMover.cpp index 5a3f138d727..bd277ca4374 100644 --- a/src/Storages/MergeTree/MergeTreePartsMover.cpp +++ b/src/Storages/MergeTree/MergeTreePartsMover.cpp @@ -140,7 +140,7 @@ bool MergeTreePartsMover::selectPartsForMove( auto ttl_entry = selectTTLDescriptionForTTLInfos(metadata_snapshot->getMoveTTLs(), part->ttl_infos.moves_ttl, time_of_move, true); auto to_insert = need_to_move.end(); - if (auto disk_it = part->data_part_storage->isStoredOnDisk(need_to_move_disks); disk_it != need_to_move_disks.end()) + if (auto disk_it = part->getDataPartStorage().isStoredOnDisk(need_to_move_disks); disk_it != need_to_move_disks.end()) to_insert = need_to_move.find(*disk_it); ReservationPtr reservation; @@ -199,7 +199,7 @@ bool MergeTreePartsMover::selectPartsForMove( return false; } -MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEntry & moving_part) const +MergeTreeMutableDataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEntry & moving_part) const { if (moves_blocker.isCancelled()) throw Exception("Cancelled moving parts.", ErrorCodes::ABORTED); @@ -207,7 +207,7 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt auto settings = data->getSettings(); auto part = moving_part.part; auto disk = moving_part.reserved_space->getDisk(); - LOG_DEBUG(log, "Cloning part {} from '{}' to '{}'", part->name, part->data_part_storage->getDiskName(), disk->getName()); + LOG_DEBUG(log, "Cloning part {} from '{}' to '{}'", part->name, part->getDataPartStorage().getDiskName(), disk->getName()); MutableDataPartStoragePtr cloned_part_storage; if (disk->supportZeroCopyReplication() && settings->allow_remote_fs_zero_copy_replication) @@ -215,7 +215,7 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt /// Try zero-copy replication and fallback to default copy if it's not possible moving_part.part->assertOnDisk(); String path_to_clone = fs::path(data->getRelativeDataPath()) / MergeTreeData::MOVING_DIR_NAME / ""; - String relative_path = part->data_part_storage->getPartDirectory(); + String relative_path = part->getDataPartStorage().getPartDirectory(); if (disk->exists(path_to_clone + relative_path)) { LOG_WARNING(log, "Path {} already exists. Will remove it and clone again.", fullPath(disk, path_to_clone + relative_path)); @@ -224,16 +224,12 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt disk->createDirectories(path_to_clone); - bool is_fetched = data->tryToFetchIfShared(*part, disk, fs::path(path_to_clone) / part->name); + cloned_part_storage = data->tryToFetchIfShared(*part, disk, fs::path(path_to_clone) / part->name); - if (!is_fetched) + if (!cloned_part_storage) { LOG_INFO(log, "Part {} was not fetched, we are the first who move it to another disk, so we will copy it", part->name); - cloned_part_storage = part->data_part_storage->clonePart(path_to_clone, part->data_part_storage->getPartDirectory(), disk, log); - } - else - { - cloned_part_storage = part->data_part_storage->clone(); + cloned_part_storage = part->getDataPartStorage().clonePart(path_to_clone, part->getDataPartStorage().getPartDirectory(), disk, log); } } else @@ -242,16 +238,16 @@ MergeTreeData::DataPartPtr MergeTreePartsMover::clonePart(const MergeTreeMoveEnt } auto cloned_part = data->createPart(part->name, cloned_part_storage); - LOG_TRACE(log, "Part {} was cloned to {}", part->name, cloned_part->data_part_storage->getFullPath()); + LOG_TRACE(log, "Part {} was cloned to {}", part->name, cloned_part->getDataPartStorage().getFullPath()); cloned_part->loadColumnsChecksumsIndexes(true, true); cloned_part->loadVersionMetadata(); - cloned_part->modification_time = cloned_part->data_part_storage->getLastModified().epochTime(); + cloned_part->modification_time = cloned_part->getDataPartStorage().getLastModified().epochTime(); return cloned_part; } -void MergeTreePartsMover::swapClonedPart(const MergeTreeData::DataPartPtr & cloned_part) const +void MergeTreePartsMover::swapClonedPart(const MergeTreeMutableDataPartPtr & cloned_part) const { if (moves_blocker.isCancelled()) throw Exception("Cancelled moving parts.", ErrorCodes::ABORTED); @@ -261,7 +257,7 @@ void MergeTreePartsMover::swapClonedPart(const MergeTreeData::DataPartPtr & clon /// It's ok, because we don't block moving parts for merges or mutations if (!active_part || active_part->name != cloned_part->name) { - LOG_INFO(log, "Failed to swap {}. Active part doesn't exist. Possible it was merged or mutated. Will remove copy on path '{}'.", cloned_part->name, cloned_part->data_part_storage->getFullPath()); + LOG_INFO(log, "Failed to swap {}. Active part doesn't exist. Possible it was merged or mutated. Will remove copy on path '{}'.", cloned_part->name, cloned_part->getDataPartStorage().getFullPath()); return; } @@ -271,7 +267,7 @@ void MergeTreePartsMover::swapClonedPart(const MergeTreeData::DataPartPtr & clon /// TODO what happen if server goes down here? data->swapActivePart(cloned_part); - LOG_TRACE(log, "Part {} was moved to {}", cloned_part->name, cloned_part->data_part_storage->getFullPath()); + LOG_TRACE(log, "Part {} was moved to {}", cloned_part->name, cloned_part->getDataPartStorage().getFullPath()); } } diff --git a/src/Storages/MergeTree/MergeTreePartsMover.h b/src/Storages/MergeTree/MergeTreePartsMover.h index 6ad658c2cb3..0266b2daa46 100644 --- a/src/Storages/MergeTree/MergeTreePartsMover.h +++ b/src/Storages/MergeTree/MergeTreePartsMover.h @@ -50,14 +50,14 @@ public: const std::lock_guard & moving_parts_lock); /// Copies part to selected reservation in detached folder. Throws exception if part already exists. - MergeTreeDataPartPtr clonePart(const MergeTreeMoveEntry & moving_part) const; + MergeTreeMutableDataPartPtr clonePart(const MergeTreeMoveEntry & moving_part) const; /// Replaces cloned part from detached directory into active data parts set. /// Replacing part changes state to DeleteOnDestroy and will be removed from disk after destructor of ///IMergeTreeDataPart called. If replacing part doesn't exists or not active (committed) than /// cloned part will be removed and log message will be reported. It may happen in case of concurrent /// merge or mutation. - void swapClonedPart(const MergeTreeDataPartPtr & cloned_parts) const; + void swapClonedPart(const MergeTreeMutableDataPartPtr & cloned_parts) const; /// Can stop background moves and moves from queries ActionBlocker moves_blocker; diff --git a/src/Storages/MergeTree/MergeTreeReadPool.cpp b/src/Storages/MergeTree/MergeTreeReadPool.cpp index 3f51673a6b1..ca9cde0ae61 100644 --- a/src/Storages/MergeTree/MergeTreeReadPool.cpp +++ b/src/Storages/MergeTree/MergeTreeReadPool.cpp @@ -263,7 +263,7 @@ void MergeTreeReadPool::fillPerThreadInfo( { PartInfo part_info{parts[i], per_part_sum_marks[i], i}; if (parts[i].data_part->isStoredOnDisk()) - parts_per_disk[parts[i].data_part->data_part_storage->getDiskName()].push_back(std::move(part_info)); + parts_per_disk[parts[i].data_part->getDataPartStorage().getDiskName()].push_back(std::move(part_info)); else parts_per_disk[""].push_back(std::move(part_info)); } diff --git a/src/Storages/MergeTree/MergeTreeReaderCompact.cpp b/src/Storages/MergeTree/MergeTreeReaderCompact.cpp index 4801c9a4058..b0488d29f8e 100644 --- a/src/Storages/MergeTree/MergeTreeReaderCompact.cpp +++ b/src/Storages/MergeTree/MergeTreeReaderCompact.cpp @@ -59,13 +59,15 @@ MergeTreeReaderCompact::MergeTreeReaderCompact( throw Exception(ErrorCodes::CANNOT_READ_ALL_DATA, "Cannot read to empty buffer."); const String path = MergeTreeDataPartCompact::DATA_FILE_NAME_WITH_EXTENSION; + auto data_part_storage = data_part_info_for_read->getDataPartStorage(); + if (uncompressed_cache) { auto buffer = std::make_unique( - std::string(fs::path(data_part_info_for_read->getDataPartStorage()->getFullPath()) / path), - [this, path]() + std::string(fs::path(data_part_storage->getFullPath()) / path), + [this, path, data_part_storage]() { - return data_part_info_for_read->getDataPartStorage()->readFile( + return data_part_storage->readFile( path, settings.read_settings, std::nullopt, std::nullopt); @@ -87,7 +89,7 @@ MergeTreeReaderCompact::MergeTreeReaderCompact( { auto buffer = std::make_unique( - data_part_info_for_read->getDataPartStorage()->readFile( + data_part_storage->readFile( path, settings.read_settings, std::nullopt, std::nullopt), diff --git a/src/Storages/MergeTree/MergeTreeSink.cpp b/src/Storages/MergeTree/MergeTreeSink.cpp index 6178af83913..a7ddb9d9e1d 100644 --- a/src/Storages/MergeTree/MergeTreeSink.cpp +++ b/src/Storages/MergeTree/MergeTreeSink.cpp @@ -81,7 +81,7 @@ void MergeTreeSink::consume(Chunk chunk) if (!temp_part.part) continue; - if (!support_parallel_write && temp_part.part->data_part_storage->supportParallelWrite()) + if (!support_parallel_write && temp_part.part->getDataPartStorage().supportParallelWrite()) support_parallel_write = true; if (storage.getDeduplicationLog()) diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.cpp b/src/Storages/MergeTree/MergedBlockOutputStream.cpp index 1d0743a0429..991a8d359a8 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.cpp +++ b/src/Storages/MergeTree/MergedBlockOutputStream.cpp @@ -94,7 +94,7 @@ void MergedBlockOutputStream::Finalizer::Impl::finish() writer.finish(sync); for (const auto & file_name : files_to_remove_after_finish) - part->data_part_storage->removeFile(file_name); + part->getDataPartStorage().removeFile(file_name); for (auto & file : written_files) { @@ -121,19 +121,19 @@ MergedBlockOutputStream::Finalizer & MergedBlockOutputStream::Finalizer::operato MergedBlockOutputStream::Finalizer::Finalizer(std::unique_ptr impl_) : impl(std::move(impl_)) {} void MergedBlockOutputStream::finalizePart( - MergeTreeData::MutableDataPartPtr & new_part, - bool sync, - const NamesAndTypesList * total_columns_list, - MergeTreeData::DataPart::Checksums * additional_column_checksums) + const MergeTreeMutableDataPartPtr & new_part, + bool sync, + const NamesAndTypesList * total_columns_list, + MergeTreeData::DataPart::Checksums * additional_column_checksums) { finalizePartAsync(new_part, sync, total_columns_list, additional_column_checksums).finish(); } MergedBlockOutputStream::Finalizer MergedBlockOutputStream::finalizePartAsync( - MergeTreeData::MutableDataPartPtr & new_part, - bool sync, - const NamesAndTypesList * total_columns_list, - MergeTreeData::DataPart::Checksums * additional_column_checksums) + const MergeTreeMutableDataPartPtr & new_part, + bool sync, + const NamesAndTypesList * total_columns_list, + MergeTreeData::DataPart::Checksums * additional_column_checksums) { /// Finish write and get checksums. MergeTreeData::DataPart::Checksums checksums; @@ -183,7 +183,7 @@ MergedBlockOutputStream::Finalizer MergedBlockOutputStream::finalizePartAsync( } MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDisk( - const MergeTreeData::DataPartPtr & new_part, + const MergeTreeMutableDataPartPtr & new_part, MergeTreeData::DataPart::Checksums & checksums) { WrittenFiles written_files; @@ -191,7 +191,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { if (storage.format_version >= MERGE_TREE_DATA_MIN_FORMAT_VERSION_WITH_CUSTOM_PARTITIONING || isCompactPart(new_part)) { - auto count_out = new_part->data_part_storage->writeFile("count.txt", 4096, write_settings); + auto count_out = new_part->getDataPartStorage().writeFile("count.txt", 4096, write_settings); HashingWriteBuffer count_out_hashing(*count_out); writeIntText(rows_count, count_out_hashing); count_out_hashing.next(); @@ -205,7 +205,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { if (new_part->uuid != UUIDHelpers::Nil) { - auto out = new_part->data_part_storage->writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, write_settings); + auto out = new_part->getDataPartStorage().writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, write_settings); HashingWriteBuffer out_hashing(*out); writeUUIDText(new_part->uuid, out_hashing); checksums.files[IMergeTreeDataPart::UUID_FILE_NAME].file_size = out_hashing.count(); @@ -216,12 +216,12 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (storage.format_version >= MERGE_TREE_DATA_MIN_FORMAT_VERSION_WITH_CUSTOM_PARTITIONING) { - if (auto file = new_part->partition.store(storage, new_part->data_part_storage, checksums)) + if (auto file = new_part->partition.store(storage, new_part->getDataPartStorage(), checksums)) written_files.emplace_back(std::move(file)); if (new_part->minmax_idx->initialized) { - auto files = new_part->minmax_idx->store(storage, new_part->data_part_storage, checksums); + auto files = new_part->minmax_idx->store(storage, new_part->getDataPartStorage(), checksums); for (auto & file : files) written_files.emplace_back(std::move(file)); } @@ -231,7 +231,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis } { - auto count_out = new_part->data_part_storage->writeFile("count.txt", 4096, write_settings); + auto count_out = new_part->getDataPartStorage().writeFile("count.txt", 4096, write_settings); HashingWriteBuffer count_out_hashing(*count_out); writeIntText(rows_count, count_out_hashing); count_out_hashing.next(); @@ -245,7 +245,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (!new_part->ttl_infos.empty()) { /// Write a file with ttl infos in json format. - auto out = new_part->data_part_storage->writeFile("ttl.txt", 4096, write_settings); + auto out = new_part->getDataPartStorage().writeFile("ttl.txt", 4096, write_settings); HashingWriteBuffer out_hashing(*out); new_part->ttl_infos.write(out_hashing); checksums.files["ttl.txt"].file_size = out_hashing.count(); @@ -256,7 +256,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (!new_part->getSerializationInfos().empty()) { - auto out = new_part->data_part_storage->writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, write_settings); + auto out = new_part->getDataPartStorage().writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, write_settings); HashingWriteBuffer out_hashing(*out); new_part->getSerializationInfos().writeJSON(out_hashing); checksums.files[IMergeTreeDataPart::SERIALIZATION_FILE_NAME].file_size = out_hashing.count(); @@ -267,7 +267,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { /// Write a file with a description of columns. - auto out = new_part->data_part_storage->writeFile("columns.txt", 4096, write_settings); + auto out = new_part->getDataPartStorage().writeFile("columns.txt", 4096, write_settings); new_part->getColumns().writeText(*out); out->preFinalize(); written_files.emplace_back(std::move(out)); @@ -275,7 +275,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis if (default_codec != nullptr) { - auto out = new_part->data_part_storage->writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, write_settings); + auto out = new_part->getDataPartStorage().writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, write_settings); DB::writeText(queryToString(default_codec->getFullCodecDesc()), *out); out->preFinalize(); written_files.emplace_back(std::move(out)); @@ -288,7 +288,7 @@ MergedBlockOutputStream::WrittenFiles MergedBlockOutputStream::finalizePartOnDis { /// Write file with checksums. - auto out = new_part->data_part_storage->writeFile("checksums.txt", 4096, write_settings); + auto out = new_part->getDataPartStorage().writeFile("checksums.txt", 4096, write_settings); checksums.write(*out); out->preFinalize(); written_files.emplace_back(std::move(out)); diff --git a/src/Storages/MergeTree/MergedBlockOutputStream.h b/src/Storages/MergeTree/MergedBlockOutputStream.h index 799bae8e94b..ad1bb584788 100644 --- a/src/Storages/MergeTree/MergedBlockOutputStream.h +++ b/src/Storages/MergeTree/MergedBlockOutputStream.h @@ -54,16 +54,16 @@ public: /// Finalize writing part and fill inner structures /// If part is new and contains projections, they should be added before invoking this method. Finalizer finalizePartAsync( - MergeTreeData::MutableDataPartPtr & new_part, - bool sync, - const NamesAndTypesList * total_columns_list = nullptr, - MergeTreeData::DataPart::Checksums * additional_column_checksums = nullptr); + const MergeTreeMutableDataPartPtr & new_part, + bool sync, + const NamesAndTypesList * total_columns_list = nullptr, + MergeTreeData::DataPart::Checksums * additional_column_checksums = nullptr); void finalizePart( - MergeTreeData::MutableDataPartPtr & new_part, - bool sync, - const NamesAndTypesList * total_columns_list = nullptr, - MergeTreeData::DataPart::Checksums * additional_column_checksums = nullptr); + const MergeTreeMutableDataPartPtr & new_part, + bool sync, + const NamesAndTypesList * total_columns_list = nullptr, + MergeTreeData::DataPart::Checksums * additional_column_checksums = nullptr); private: /** If `permutation` is given, it rearranges the values in the columns when writing. @@ -73,8 +73,8 @@ private: using WrittenFiles = std::vector>; WrittenFiles finalizePartOnDisk( - const MergeTreeData::DataPartPtr & new_part, - MergeTreeData::DataPart::Checksums & checksums); + const MergeTreeMutableDataPartPtr & new_part, + MergeTreeData::DataPart::Checksums & checksums); NamesAndTypesList columns_list; IMergeTreeDataPart::MinMaxIndex minmax_idx; diff --git a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp index 310bc849ffe..e4a5a0bc3ba 100644 --- a/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp +++ b/src/Storages/MergeTree/MergedColumnOnlyOutputStream.cpp @@ -79,7 +79,7 @@ MergedColumnOnlyOutputStream::fillChecksums( for (const String & removed_file : removed_files) { - new_part->data_part_storage->removeFileIfExists(removed_file); + new_part->getDataPartStorage().removeFileIfExists(removed_file); if (all_checksums.files.contains(removed_file)) all_checksums.files.erase(removed_file); diff --git a/src/Storages/MergeTree/MutateFromLogEntryTask.cpp b/src/Storages/MergeTree/MutateFromLogEntryTask.cpp index a4a75a637a4..9e3cbb0640b 100644 --- a/src/Storages/MergeTree/MutateFromLogEntryTask.cpp +++ b/src/Storages/MergeTree/MutateFromLogEntryTask.cpp @@ -92,7 +92,7 @@ ReplicatedMergeMutateTaskBase::PrepareResult MutateFromLogEntryTask::prepare() /// Once we mutate part, we must reserve space on the same disk, because mutations can possibly create hardlinks. /// Can throw an exception. - reserved_space = storage.reserveSpace(estimated_space_for_result, source_part->data_part_storage); + reserved_space = storage.reserveSpace(estimated_space_for_result, source_part->getDataPartStorage()); table_lock_holder = storage.lockForShare( RWLockImpl::NO_QUERY, storage_settings_ptr->lock_acquire_timeout_for_background_operations); diff --git a/src/Storages/MergeTree/MutateTask.cpp b/src/Storages/MergeTree/MutateTask.cpp index 80bf7e3c7d9..e5ba771a198 100644 --- a/src/Storages/MergeTree/MutateTask.cpp +++ b/src/Storages/MergeTree/MutateTask.cpp @@ -626,7 +626,7 @@ void finalizeMutatedPart( { if (new_data_part->uuid != UUIDHelpers::Nil) { - auto out = new_data_part->data_part_storage->writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, context->getWriteSettings()); + auto out = new_data_part->getDataPartStorage().writeFile(IMergeTreeDataPart::UUID_FILE_NAME, 4096, context->getWriteSettings()); HashingWriteBuffer out_hashing(*out); writeUUIDText(new_data_part->uuid, out_hashing); new_data_part->checksums.files[IMergeTreeDataPart::UUID_FILE_NAME].file_size = out_hashing.count(); @@ -636,7 +636,7 @@ void finalizeMutatedPart( if (execute_ttl_type != ExecuteTTLType::NONE) { /// Write a file with ttl infos in json format. - auto out_ttl = new_data_part->data_part_storage->writeFile("ttl.txt", 4096, context->getWriteSettings()); + auto out_ttl = new_data_part->getDataPartStorage().writeFile("ttl.txt", 4096, context->getWriteSettings()); HashingWriteBuffer out_hashing(*out_ttl); new_data_part->ttl_infos.write(out_hashing); new_data_part->checksums.files["ttl.txt"].file_size = out_hashing.count(); @@ -645,7 +645,7 @@ void finalizeMutatedPart( if (!new_data_part->getSerializationInfos().empty()) { - auto out = new_data_part->data_part_storage->writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, context->getWriteSettings()); + auto out = new_data_part->getDataPartStorage().writeFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, 4096, context->getWriteSettings()); HashingWriteBuffer out_hashing(*out); new_data_part->getSerializationInfos().writeJSON(out_hashing); new_data_part->checksums.files[IMergeTreeDataPart::SERIALIZATION_FILE_NAME].file_size = out_hashing.count(); @@ -654,18 +654,18 @@ void finalizeMutatedPart( { /// Write file with checksums. - auto out_checksums = new_data_part->data_part_storage->writeFile("checksums.txt", 4096, context->getWriteSettings()); + auto out_checksums = new_data_part->getDataPartStorage().writeFile("checksums.txt", 4096, context->getWriteSettings()); new_data_part->checksums.write(*out_checksums); } /// close fd { - auto out = new_data_part->data_part_storage->writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, context->getWriteSettings()); + auto out = new_data_part->getDataPartStorage().writeFile(IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME, 4096, context->getWriteSettings()); DB::writeText(queryToString(codec->getFullCodecDesc()), *out); } /// close fd { /// Write a file with a description of columns. - auto out_columns = new_data_part->data_part_storage->writeFile("columns.txt", 4096, context->getWriteSettings()); + auto out_columns = new_data_part->getDataPartStorage().writeFile("columns.txt", 4096, context->getWriteSettings()); new_data_part->getColumns().writeText(*out_columns); } /// close fd @@ -1141,7 +1141,7 @@ private: void prepare() { - ctx->new_data_part->data_part_storage->createDirectories(); + ctx->new_data_part->getDataPartStorage().createDirectories(); /// Note: this is done before creating input streams, because otherwise data.data_parts_mutex /// (which is locked in data.getTotalActiveSizeInBytes()) @@ -1271,7 +1271,7 @@ private: if (ctx->execute_ttl_type != ExecuteTTLType::NONE) ctx->files_to_skip.insert("ttl.txt"); - ctx->new_data_part->data_part_storage->createDirectories(); + ctx->new_data_part->getDataPartStorage().createDirectories(); /// We should write version metadata on part creation to distinguish it from parts that were created without transaction. TransactionID tid = ctx->txn ? ctx->txn->tid : Tx::PrehistoricTID; @@ -1282,7 +1282,7 @@ private: NameSet hardlinked_files; /// Create hardlinks for unchanged files - for (auto it = ctx->source_part->data_part_storage->iterate(); it->isValid(); it->next()) + for (auto it = ctx->source_part->getDataPartStorage().iterate(); it->isValid(); it->next()) { if (ctx->files_to_skip.contains(it->name())) continue; @@ -1308,17 +1308,17 @@ private: if (it->isFile()) { - ctx->new_data_part->data_part_storage->createHardLinkFrom( - *ctx->source_part->data_part_storage, it->name(), destination); + ctx->new_data_part->getDataPartStorage().createHardLinkFrom( + ctx->source_part->getDataPartStorage(), it->name(), destination); hardlinked_files.insert(it->name()); } else if (!endsWith(it->name(), ".tmp_proj")) // ignore projection tmp merge dir { // it's a projection part directory - ctx->new_data_part->data_part_storage->createProjection(destination); + ctx->new_data_part->getDataPartStorage().createProjection(destination); - auto projection_data_part_storage_src = ctx->source_part->data_part_storage->getProjection(destination); - auto projection_data_part_storage_dst = ctx->new_data_part->data_part_storage->getProjection(destination); + auto projection_data_part_storage_src = ctx->source_part->getDataPartStorage().getProjection(destination); + auto projection_data_part_storage_dst = ctx->new_data_part->getDataPartStorage().getProjection(destination); for (auto p_it = projection_data_part_storage_src->iterate(); p_it->isValid(); p_it->next()) { diff --git a/src/Storages/MergeTree/PartMetadataManagerOrdinary.cpp b/src/Storages/MergeTree/PartMetadataManagerOrdinary.cpp index 7eb868f7754..30823d593a2 100644 --- a/src/Storages/MergeTree/PartMetadataManagerOrdinary.cpp +++ b/src/Storages/MergeTree/PartMetadataManagerOrdinary.cpp @@ -8,20 +8,10 @@ namespace DB { -static std::unique_ptr openForReading(const DataPartStoragePtr & data_part_storage, const String & path) -{ - size_t file_size = data_part_storage->getFileSize(path); - return data_part_storage->readFile(path, ReadSettings().adjustBufferSize(file_size), file_size, std::nullopt); -} - -PartMetadataManagerOrdinary::PartMetadataManagerOrdinary(const IMergeTreeDataPart * part_) : IPartMetadataManager(part_) -{ -} - - std::unique_ptr PartMetadataManagerOrdinary::read(const String & file_name) const { - auto res = openForReading(part->data_part_storage, file_name); + size_t file_size = part->getDataPartStorage().getFileSize(file_name); + auto res = part->getDataPartStorage().readFile(file_name, ReadSettings().adjustBufferSize(file_size), file_size, std::nullopt); if (isCompressedFromFileName(file_name)) return std::make_unique(std::move(res)); @@ -31,7 +21,7 @@ std::unique_ptr PartMetadataManagerOrdinary::read(const String & fil bool PartMetadataManagerOrdinary::exists(const String & file_name) const { - return part->data_part_storage->exists(file_name); + return part->getDataPartStorage().exists(file_name); } diff --git a/src/Storages/MergeTree/PartMetadataManagerOrdinary.h b/src/Storages/MergeTree/PartMetadataManagerOrdinary.h index d86d5c54c00..428b6d4710a 100644 --- a/src/Storages/MergeTree/PartMetadataManagerOrdinary.h +++ b/src/Storages/MergeTree/PartMetadataManagerOrdinary.h @@ -8,7 +8,7 @@ namespace DB class PartMetadataManagerOrdinary : public IPartMetadataManager { public: - explicit PartMetadataManagerOrdinary(const IMergeTreeDataPart * part_); + explicit PartMetadataManagerOrdinary(const IMergeTreeDataPart * part_) : IPartMetadataManager(part_) {} ~PartMetadataManagerOrdinary() override = default; diff --git a/src/Storages/MergeTree/PartMetadataManagerWithCache.cpp b/src/Storages/MergeTree/PartMetadataManagerWithCache.cpp index ee0970984f9..90fd25bc4e7 100644 --- a/src/Storages/MergeTree/PartMetadataManagerWithCache.cpp +++ b/src/Storages/MergeTree/PartMetadataManagerWithCache.cpp @@ -31,24 +31,24 @@ PartMetadataManagerWithCache::PartMetadataManagerWithCache(const IMergeTreeDataP String PartMetadataManagerWithCache::getKeyFromFilePath(const String & file_path) const { - return part->data_part_storage->getDiskName() + ":" + file_path; + return part->getDataPartStorage().getDiskName() + ":" + file_path; } String PartMetadataManagerWithCache::getFilePathFromKey(const String & key) const { - return key.substr(part->data_part_storage->getDiskName().size() + 1); + return key.substr(part->getDataPartStorage().getDiskName().size() + 1); } std::unique_ptr PartMetadataManagerWithCache::read(const String & file_name) const { - String file_path = fs::path(part->data_part_storage->getRelativePath()) / file_name; + String file_path = fs::path(part->getDataPartStorage().getRelativePath()) / file_name; String key = getKeyFromFilePath(file_path); String value; auto status = cache->get(key, value); if (!status.ok()) { ProfileEvents::increment(ProfileEvents::MergeTreeMetadataCacheMiss); - auto in = part->data_part_storage->readFile(file_name, {}, std::nullopt, std::nullopt); + auto in = part->getDataPartStorage().readFile(file_name, {}, std::nullopt, std::nullopt); std::unique_ptr reader; if (!isCompressedFromFileName(file_name)) reader = std::move(in); @@ -67,7 +67,7 @@ std::unique_ptr PartMetadataManagerWithCache::read(const String & fi bool PartMetadataManagerWithCache::exists(const String & file_name) const { - String file_path = fs::path(part->data_part_storage->getRelativePath()) / file_name; + String file_path = fs::path(part->getDataPartStorage().getRelativePath()) / file_name; String key = getKeyFromFilePath(file_path); String value; auto status = cache->get(key, value); @@ -79,7 +79,7 @@ bool PartMetadataManagerWithCache::exists(const String & file_name) const else { ProfileEvents::increment(ProfileEvents::MergeTreeMetadataCacheMiss); - return part->data_part_storage->exists(file_name); + return part->getDataPartStorage().exists(file_name); } } @@ -91,7 +91,7 @@ void PartMetadataManagerWithCache::deleteAll(bool include_projection) String value; for (const auto & file_name : file_names) { - String file_path = fs::path(part->data_part_storage->getRelativePath()) / file_name; + String file_path = fs::path(part->getDataPartStorage().getRelativePath()) / file_name; String key = getKeyFromFilePath(file_path); auto status = cache->del(key); if (!status.ok()) @@ -119,10 +119,10 @@ void PartMetadataManagerWithCache::updateAll(bool include_projection) String read_value; for (const auto & file_name : file_names) { - String file_path = fs::path(part->data_part_storage->getRelativePath()) / file_name; - if (!part->data_part_storage->exists(file_name)) + String file_path = fs::path(part->getDataPartStorage().getRelativePath()) / file_name; + if (!part->getDataPartStorage().exists(file_name)) continue; - auto in = part->data_part_storage->readFile(file_name, {}, std::nullopt, std::nullopt); + auto in = part->getDataPartStorage().readFile(file_name, {}, std::nullopt, std::nullopt); readStringUntilEOF(value, *in); String key = getKeyFromFilePath(file_path); @@ -159,7 +159,7 @@ void PartMetadataManagerWithCache::assertAllDeleted(bool include_projection) con file_name = fs::path(file_path).filename(); /// Metadata file belongs to current part - if (fs::path(part->data_part_storage->getRelativePath()) / file_name == file_path) + if (fs::path(part->getDataPartStorage().getRelativePath()) / file_name == file_path) throw Exception( ErrorCodes::LOGICAL_ERROR, "Data part {} with type {} with meta file {} still in cache", @@ -173,7 +173,7 @@ void PartMetadataManagerWithCache::assertAllDeleted(bool include_projection) con const auto & projection_parts = part->getProjectionParts(); for (const auto & [projection_name, projection_part] : projection_parts) { - if (fs::path(part->data_part_storage->getRelativePath()) / (projection_name + ".proj") / file_name == file_path) + if (fs::path(part->getDataPartStorage().getRelativePath()) / (projection_name + ".proj") / file_name == file_path) { throw Exception( ErrorCodes::LOGICAL_ERROR, @@ -190,7 +190,7 @@ void PartMetadataManagerWithCache::assertAllDeleted(bool include_projection) con void PartMetadataManagerWithCache::getKeysAndCheckSums(Strings & keys, std::vector & checksums) const { - String prefix = getKeyFromFilePath(fs::path(part->data_part_storage->getRelativePath()) / ""); + String prefix = getKeyFromFilePath(fs::path(part->getDataPartStorage().getRelativePath()) / ""); Strings values; cache->getByPrefix(prefix, keys, values); size_t size = keys.size(); @@ -225,7 +225,7 @@ std::unordered_map PartMetadataManagerWit results.emplace(file_name, cache_checksums[i]); /// File belongs to normal part - if (fs::path(part->data_part_storage->getRelativePath()) / file_name == file_path) + if (fs::path(part->getDataPartStorage().getRelativePath()) / file_name == file_path) { auto disk_checksum = part->getActualChecksumByFile(file_name); if (disk_checksum != cache_checksums[i]) diff --git a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp index 293c679dcd0..082228d7ebf 100644 --- a/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp +++ b/src/Storages/MergeTree/ReplicatedMergeTreeSink.cpp @@ -325,7 +325,7 @@ void ReplicatedMergeTreeSink::commitPart( assertSessionIsNotExpired(zookeeper); - String temporary_part_relative_path = part->data_part_storage->getPartDirectory(); + String temporary_part_relative_path = part->getDataPartStorage().getPartDirectory(); /// There is one case when we need to retry transaction in a loop. /// But don't do it too many times - just as defensive measure. diff --git a/src/Storages/MergeTree/checkDataPart.cpp b/src/Storages/MergeTree/checkDataPart.cpp index d5a838668d2..4758ccb201a 100644 --- a/src/Storages/MergeTree/checkDataPart.cpp +++ b/src/Storages/MergeTree/checkDataPart.cpp @@ -1,3 +1,4 @@ +#include "Storages/MergeTree/IDataPartStorage.h" #include #include @@ -46,7 +47,7 @@ bool isNotEnoughMemoryErrorCode(int code) IMergeTreeDataPart::Checksums checkDataPart( MergeTreeData::DataPartPtr data_part, - const DataPartStoragePtr & data_part_storage, + const IDataPartStorage & data_part_storage, const NamesAndTypesList & columns_list, const MergeTreeDataPartType & part_type, const NameSet & files_without_checksums, @@ -64,13 +65,13 @@ IMergeTreeDataPart::Checksums checkDataPart( NamesAndTypesList columns_txt; { - auto buf = data_part_storage->readFile("columns.txt", {}, std::nullopt, std::nullopt); + auto buf = data_part_storage.readFile("columns.txt", {}, std::nullopt, std::nullopt); columns_txt.readText(*buf); assertEOF(*buf); } if (columns_txt != columns_list) - throw Exception("Columns doesn't match in part " + data_part_storage->getFullPath() + throw Exception("Columns doesn't match in part " + data_part_storage.getFullPath() + ". Expected: " + columns_list.toString() + ". Found: " + columns_txt.toString(), ErrorCodes::CORRUPTED_DATA); @@ -78,9 +79,9 @@ IMergeTreeDataPart::Checksums checkDataPart( IMergeTreeDataPart::Checksums checksums_data; /// This function calculates checksum for both compressed and decompressed contents of compressed file. - auto checksum_compressed_file = [](const DataPartStoragePtr & data_part_storage_, const String & file_path) + auto checksum_compressed_file = [](const IDataPartStorage & data_part_storage_, const String & file_path) { - auto file_buf = data_part_storage_->readFile(file_path, {}, std::nullopt, std::nullopt); + auto file_buf = data_part_storage_.readFile(file_path, {}, std::nullopt, std::nullopt); HashingReadBuffer compressed_hashing_buf(*file_buf); CompressedReadBuffer uncompressing_buf(compressed_hashing_buf); HashingReadBuffer uncompressed_hashing_buf(uncompressing_buf); @@ -96,9 +97,9 @@ IMergeTreeDataPart::Checksums checkDataPart( auto ratio_of_defaults = data_part->storage.getSettings()->ratio_of_defaults_for_sparse_serialization; SerializationInfoByName serialization_infos(columns_txt, SerializationInfo::Settings{ratio_of_defaults, false}); - if (data_part_storage->exists(IMergeTreeDataPart::SERIALIZATION_FILE_NAME)) + if (data_part_storage.exists(IMergeTreeDataPart::SERIALIZATION_FILE_NAME)) { - auto serialization_file = data_part_storage->readFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, {}, std::nullopt, std::nullopt); + auto serialization_file = data_part_storage.readFile(IMergeTreeDataPart::SERIALIZATION_FILE_NAME, {}, std::nullopt, std::nullopt); serialization_infos.readJSON(*serialization_file); } @@ -114,7 +115,7 @@ IMergeTreeDataPart::Checksums checkDataPart( /// It also calculates checksum of projections. auto checksum_file = [&](const String & file_name) { - if (data_part_storage->isDirectory(file_name) && endsWith(file_name, ".proj")) + if (data_part_storage.isDirectory(file_name) && endsWith(file_name, ".proj")) { auto projection_name = file_name.substr(0, file_name.size() - sizeof(".proj") + 1); auto pit = data_part->getProjectionParts().find(projection_name); @@ -129,7 +130,7 @@ IMergeTreeDataPart::Checksums checkDataPart( const auto & projection = pit->second; IMergeTreeDataPart::Checksums projection_checksums_data; - auto projection_part_storage = data_part_storage->getProjection(file_name); + auto projection_part_storage = data_part_storage.getProjection(file_name); if (projection->getType() == MergeTreeDataPartType::Compact) { @@ -148,7 +149,7 @@ IMergeTreeDataPart::Checksums checkDataPart( [&](const ISerialization::SubstreamPath & substream_path) { String projection_file_name = ISerialization::getFileNameForStream(projection_column, substream_path) + ".bin"; - projection_checksums_data.files[projection_file_name] = checksum_compressed_file(projection_part_storage, projection_file_name); + projection_checksums_data.files[projection_file_name] = checksum_compressed_file(*projection_part_storage, projection_file_name); }); } } @@ -183,7 +184,7 @@ IMergeTreeDataPart::Checksums checkDataPart( } else { - projection_checksums_data.files[projection_file_name] = checksum_compressed_file(projection_part_storage, projection_file_name); + projection_checksums_data.files[projection_file_name] = checksum_compressed_file(*projection_part_storage, projection_file_name); } } } @@ -195,7 +196,7 @@ IMergeTreeDataPart::Checksums checkDataPart( } else { - auto file_buf = data_part_storage->readFile(file_name, {}, std::nullopt, std::nullopt); + auto file_buf = data_part_storage.readFile(file_name, {}, std::nullopt, std::nullopt); HashingReadBuffer hashing_buf(*file_buf); hashing_buf.ignoreAll(); checksums_data.files[file_name] = IMergeTreeDataPart::Checksums::Checksum(hashing_buf.count(), hashing_buf.getHash()); @@ -224,21 +225,21 @@ IMergeTreeDataPart::Checksums checkDataPart( } else { - throw Exception("Unknown type in part " + data_part_storage->getFullPath(), ErrorCodes::UNKNOWN_PART_TYPE); + throw Exception("Unknown type in part " + data_part_storage.getFullPath(), ErrorCodes::UNKNOWN_PART_TYPE); } /// Checksums from the rest files listed in checksums.txt. May be absent. If present, they are subsequently compared with the actual data checksums. IMergeTreeDataPart::Checksums checksums_txt; - if (require_checksums || data_part_storage->exists("checksums.txt")) + if (require_checksums || data_part_storage.exists("checksums.txt")) { - auto buf = data_part_storage->readFile("checksums.txt", {}, std::nullopt, std::nullopt); + auto buf = data_part_storage.readFile("checksums.txt", {}, std::nullopt, std::nullopt); checksums_txt.read(*buf); assertEOF(*buf); } const auto & checksum_files_txt = checksums_txt.files; - for (auto it = data_part_storage->iterate(); it->isValid(); it->next()) + for (auto it = data_part_storage.iterate(); it->isValid(); it->next()) { const String & file_name = it->name(); auto checksum_it = checksums_data.files.find(file_name); @@ -285,7 +286,7 @@ IMergeTreeDataPart::Checksums checkDataPart( return checkDataPart( data_part, - data_part->data_part_storage, + data_part->getDataPartStorage(), data_part->getColumns(), data_part->getType(), data_part->getFileNamesWithoutChecksums(), diff --git a/src/Storages/StorageMergeTree.cpp b/src/Storages/StorageMergeTree.cpp index 1aeca1343c2..3b456ab5360 100644 --- a/src/Storages/StorageMergeTree.cpp +++ b/src/Storages/StorageMergeTree.cpp @@ -378,7 +378,9 @@ CurrentlyMergingPartsTagger::CurrentlyMergingPartsTagger( /// if we mutate part, than we should reserve space on the same disk, because mutations possible can create hardlinks if (is_mutation) - reserved_space = storage.tryReserveSpace(total_size, future_part->parts[0]->data_part_storage); + { + reserved_space = storage.tryReserveSpace(total_size, future_part->parts[0]->getDataPartStorage()); + } else { IMergeTreeDataPart::TTLInfos ttl_infos; @@ -386,7 +388,7 @@ CurrentlyMergingPartsTagger::CurrentlyMergingPartsTagger( for (auto & part_ptr : future_part->parts) { ttl_infos.update(part_ptr->ttl_infos); - max_volume_index = std::max(max_volume_index, part_ptr->data_part_storage->getVolumeIndex(*storage.getStoragePolicy())); + max_volume_index = std::max(max_volume_index, part_ptr->getDataPartStorage().getVolumeIndex(*storage.getStoragePolicy())); } reserved_space = storage.balancedReservation( @@ -1474,7 +1476,7 @@ void StorageMergeTree::dropPartsImpl(DataPartsVector && parts_to_remove, bool de /// NOTE: no race with background cleanup until we hold pointers to parts for (const auto & part : parts_to_remove) { - LOG_INFO(log, "Detaching {}", part->data_part_storage->getPartDirectory()); + LOG_INFO(log, "Detaching {}", part->getDataPartStorage().getPartDirectory()); part->makeCloneInDetached("", metadata_snapshot); } } @@ -1739,14 +1741,14 @@ CheckResults StorageMergeTree::checkData(const ASTPtr & query, ContextPtr local_ /// If the checksums file is not present, calculate the checksums and write them to disk. String checksums_path = "checksums.txt"; String tmp_checksums_path = "checksums.txt.tmp"; - if (part->isStoredOnDisk() && !part->data_part_storage->exists(checksums_path)) + if (part->isStoredOnDisk() && !part->getDataPartStorage().exists(checksums_path)) { try { auto calculated_checksums = checkDataPart(part, false); calculated_checksums.checkEqual(part->checksums, true); - part->data_part_storage->writeChecksums(part->checksums, local_context->getWriteSettings()); + part->getDataPartStorage().writeChecksums(part->checksums, local_context->getWriteSettings()); part->checkMetadata(); results.emplace_back(part->name, true, "Checksums recounted and written to disk."); diff --git a/src/Storages/StorageReplicatedMergeTree.cpp b/src/Storages/StorageReplicatedMergeTree.cpp index 52a30dd37b3..286a9d6bdd8 100644 --- a/src/Storages/StorageReplicatedMergeTree.cpp +++ b/src/Storages/StorageReplicatedMergeTree.cpp @@ -1,5 +1,6 @@ #include +#include #include #include "Common/hex.h" #include @@ -1781,7 +1782,7 @@ bool StorageReplicatedMergeTree::executeFetch(LogEntry & entry, bool need_to_che } -bool StorageReplicatedMergeTree::executeFetchShared( +MutableDataPartStoragePtr StorageReplicatedMergeTree::executeFetchShared( const String & source_replica, const String & new_part_name, const DiskPtr & disk, @@ -1790,7 +1791,7 @@ bool StorageReplicatedMergeTree::executeFetchShared( if (source_replica.empty()) { LOG_INFO(log, "No active replica has part {} on shared storage.", new_part_name); - return false; + return nullptr; } const auto storage_settings_ptr = getSettings(); @@ -1847,7 +1848,7 @@ void StorageReplicatedMergeTree::executeDropRange(const LogEntry & entry) /// If DETACH clone parts to detached/ directory for (const auto & part : parts_to_remove) { - LOG_INFO(log, "Detaching {}", part->data_part_storage->getPartDirectory()); + LOG_INFO(log, "Detaching {}", part->getDataPartStorage().getPartDirectory()); part->makeCloneInDetached("", metadata_snapshot); } } @@ -2538,7 +2539,7 @@ void StorageReplicatedMergeTree::cloneReplica(const String & source_replica, Coo for (const auto & part : parts_to_remove_from_working_set) { - LOG_INFO(log, "Detaching {}", part->data_part_storage->getPartDirectory()); + LOG_INFO(log, "Detaching {}", part->getDataPartStorage().getPartDirectory()); part->makeCloneInDetached("clone", metadata_snapshot); } } @@ -3890,7 +3891,7 @@ bool StorageReplicatedMergeTree::fetchPart(const String & part_name, const Stora auto source_part = getActiveContainingPart(covered_part_info); /// Fetch for zero-copy replication is cheap and straightforward, so we don't use local clone here - if (source_part && (!settings_ptr->allow_remote_fs_zero_copy_replication || !source_part->data_part_storage->supportZeroCopyReplication())) + if (source_part && (!settings_ptr->allow_remote_fs_zero_copy_replication || !source_part->getDataPartStorage().supportZeroCopyReplication())) { auto source_part_header = ReplicatedMergeTreePartHeader::fromColumnsAndChecksums( source_part->getColumns(), source_part->checksums); @@ -4067,7 +4068,7 @@ bool StorageReplicatedMergeTree::fetchPart(const String & part_name, const Stora } -bool StorageReplicatedMergeTree::fetchExistsPart( +MutableDataPartStoragePtr StorageReplicatedMergeTree::fetchExistsPart( const String & part_name, const StorageMetadataPtr & metadata_snapshot, const String & source_replica_path, @@ -4082,7 +4083,7 @@ bool StorageReplicatedMergeTree::fetchExistsPart( LOG_DEBUG(log, "Part {} should be deleted after previous attempt before fetch", part->name); /// Force immediate parts cleanup to delete the part that was left from the previous fetch attempt. cleanup_thread.wakeup(); - return false; + return nullptr; } { @@ -4090,7 +4091,7 @@ bool StorageReplicatedMergeTree::fetchExistsPart( if (!currently_fetching_parts.insert(part_name).second) { LOG_DEBUG(log, "Part {} is already fetching right now", part_name); - return false; + return nullptr; } } @@ -4142,11 +4143,11 @@ bool StorageReplicatedMergeTree::fetchExistsPart( { part = get_part(); - if (part->data_part_storage->getDiskName() != replaced_disk->getName()) - throw Exception("Part " + part->name + " fetched on wrong disk " + part->data_part_storage->getDiskName(), ErrorCodes::LOGICAL_ERROR); + if (part->getDataPartStorage().getDiskName() != replaced_disk->getName()) + throw Exception("Part " + part->name + " fetched on wrong disk " + part->getDataPartStorage().getDiskName(), ErrorCodes::LOGICAL_ERROR); auto replaced_path = fs::path(replaced_part_path); - part->data_part_storage->rename(replaced_path.parent_path(), replaced_path.filename(), nullptr, true, false); + part->getDataPartStorage().rename(replaced_path.parent_path(), replaced_path.filename(), nullptr, true, false); } catch (const Exception & e) { @@ -4155,7 +4156,7 @@ bool StorageReplicatedMergeTree::fetchExistsPart( if (e.code() == ErrorCodes::DIRECTORY_ALREADY_EXISTS) { LOG_TRACE(log, "Not fetching part: {}", e.message()); - return false; + return nullptr; } throw; @@ -4169,7 +4170,7 @@ bool StorageReplicatedMergeTree::fetchExistsPart( ProfileEvents::increment(ProfileEvents::ReplicatedPartFetches); LOG_DEBUG(log, "Fetched part {} from {}", part_name, source_replica_path); - return true; + return part->getDataPartStoragePtr(); } void StorageReplicatedMergeTree::startup() @@ -7409,7 +7410,7 @@ void StorageReplicatedMergeTree::checkBrokenDisks() for (auto & part : *parts) { - if (part->data_part_storage && part->data_part_storage->getDiskName() == disk_ptr->getName()) + if (part->getDataPartStorage().getDiskName() == disk_ptr->getName()) broken_part_callback(part->name); } continue; @@ -7572,10 +7573,10 @@ void StorageReplicatedMergeTree::lockSharedData(const IMergeTreeDataPart & part, { auto settings = getSettings(); - if (!part.data_part_storage || !part.isStoredOnDisk() || !settings->allow_remote_fs_zero_copy_replication) + if (!part.isStoredOnDisk() || !settings->allow_remote_fs_zero_copy_replication) return; - if (!part.data_part_storage->supportZeroCopyReplication()) + if (!part.getDataPartStorage().supportZeroCopyReplication()) return; zkutil::ZooKeeperPtr zookeeper = tryGetZooKeeper(); @@ -7586,7 +7587,7 @@ void StorageReplicatedMergeTree::lockSharedData(const IMergeTreeDataPart & part, boost::replace_all(id, "/", "_"); Strings zc_zookeeper_paths = getZeroCopyPartPath( - *getSettings(), part.data_part_storage->getDiskType(), getTableSharedID(), + *getSettings(), part.getDataPartStorage().getDiskType(), getTableSharedID(), part.name, zookeeper_path); String path_to_set_hardlinked_files; @@ -7595,7 +7596,7 @@ void StorageReplicatedMergeTree::lockSharedData(const IMergeTreeDataPart & part, if (hardlinked_files.has_value() && !hardlinked_files->hardlinks_from_source_part.empty()) { path_to_set_hardlinked_files = getZeroCopyPartPath( - *getSettings(), part.data_part_storage->getDiskType(), hardlinked_files->source_table_shared_id, + *getSettings(), part.getDataPartStorage().getDiskType(), hardlinked_files->source_table_shared_id, hardlinked_files->source_part_name, zookeeper_path)[0]; hardlinks = hardlinked_files->hardlinks_from_source_part; @@ -7619,25 +7620,22 @@ std::pair StorageReplicatedMergeTree::unlockSharedData(const IMer if (!settings->allow_remote_fs_zero_copy_replication) return std::make_pair(true, NameSet{}); - if (!part.data_part_storage) - LOG_WARNING(log, "Datapart storage for part {} (temp: {}) is not initialzied", part.name, part.is_temp); - - if (!part.data_part_storage || !part.isStoredOnDisk()) + if (!part.isStoredOnDisk()) { LOG_TRACE(log, "Part {} is not stored on disk, blobs can be removed", part.name); return std::make_pair(true, NameSet{}); } - if (!part.data_part_storage || !part.data_part_storage->supportZeroCopyReplication()) + if (!part.getDataPartStorage().supportZeroCopyReplication()) { LOG_TRACE(log, "Part {} is not stored on zero-copy replicated disk, blobs can be removed", part.name); return std::make_pair(true, NameSet{}); } /// If part is temporary refcount file may be absent - if (part.data_part_storage->exists(IMergeTreeDataPart::FILE_FOR_REFERENCES_CHECK)) + if (part.getDataPartStorage().exists(IMergeTreeDataPart::FILE_FOR_REFERENCES_CHECK)) { - auto ref_count = part.data_part_storage->getRefCount(IMergeTreeDataPart::FILE_FOR_REFERENCES_CHECK); + auto ref_count = part.getDataPartStorage().getRefCount(IMergeTreeDataPart::FILE_FOR_REFERENCES_CHECK); if (ref_count > 0) /// Keep part shard info for frozen backups { LOG_TRACE(log, "Part {} has more than zero local references ({}), blobs cannot be removed", part.name, ref_count); @@ -7675,7 +7673,7 @@ std::pair StorageReplicatedMergeTree::unlockSharedData(const IMer return unlockSharedDataByID( part.getUniqueId(), getTableSharedID(), part.name, replica_name, - part.data_part_storage->getDiskType(), zookeeper, *getSettings(), log, zookeeper_path, format_version); + part.getDataPartStorage().getDiskType(), zookeeper, *getSettings(), log, zookeeper_path, format_version); } namespace @@ -7874,7 +7872,7 @@ std::pair StorageReplicatedMergeTree::unlockSharedDataByID( } -bool StorageReplicatedMergeTree::tryToFetchIfShared( +MutableDataPartStoragePtr StorageReplicatedMergeTree::tryToFetchIfShared( const IMergeTreeDataPart & part, const DiskPtr & disk, const String & path) @@ -7882,13 +7880,13 @@ bool StorageReplicatedMergeTree::tryToFetchIfShared( const auto settings = getSettings(); auto data_source_description = disk->getDataSourceDescription(); if (!(disk->supportZeroCopyReplication() && settings->allow_remote_fs_zero_copy_replication)) - return false; + return nullptr; String replica = getSharedDataReplica(part, data_source_description.type); /// We can't fetch part when none replicas have this part on a same type remote disk if (replica.empty()) - return false; + return nullptr; return executeFetchShared(replica, part.name, disk, path); } @@ -8160,7 +8158,7 @@ bool StorageReplicatedMergeTree::createEmptyPartInsteadOfLost(zkutil::ZooKeeperP /// The name could be non-unique in case of stale files from previous runs. if (data_part_storage->exists()) { - LOG_WARNING(log, "Removing old temporary directory {}", new_data_part->data_part_storage->getFullPath()); + LOG_WARNING(log, "Removing old temporary directory {}", new_data_part->getDataPartStorage().getFullPath()); data_part_storage->removeRecursive(); } diff --git a/src/Storages/StorageReplicatedMergeTree.h b/src/Storages/StorageReplicatedMergeTree.h index c9af1ab5f93..c7399d46ce3 100644 --- a/src/Storages/StorageReplicatedMergeTree.h +++ b/src/Storages/StorageReplicatedMergeTree.h @@ -263,7 +263,7 @@ public: bool canExecuteFetch(const ReplicatedMergeTreeLogEntry & entry, String & disable_reason) const; /// Fetch part only when it stored on shared storage like S3 - bool executeFetchShared(const String & source_replica, const String & new_part_name, const DiskPtr & disk, const String & path); + MutableDataPartStoragePtr executeFetchShared(const String & source_replica, const String & new_part_name, const DiskPtr & disk, const String & path); /// Lock part in zookeeper for use shared data in several nodes void lockSharedData(const IMergeTreeDataPart & part, bool replace_existing_lock, std::optional hardlinked_files) const override; @@ -283,7 +283,7 @@ public: const String & zookeeper_path_old, MergeTreeDataFormatVersion data_format_version); /// Fetch part only if some replica has it on shared storage like S3 - bool tryToFetchIfShared(const IMergeTreeDataPart & part, const DiskPtr & disk, const String & path) override; + MutableDataPartStoragePtr tryToFetchIfShared(const IMergeTreeDataPart & part, const DiskPtr & disk, const String & path) override; /// Get best replica having this partition on a same type remote disk String getSharedDataReplica(const IMergeTreeDataPart & part, DataSourceType data_source_type) const; @@ -682,7 +682,7 @@ private: * Used for replace local part on the same s3-shared part in hybrid storage. * Returns false if part is already fetching right now. */ - bool fetchExistsPart( + MutableDataPartStoragePtr fetchExistsPart( const String & part_name, const StorageMetadataPtr & metadata_snapshot, const String & replica_path, diff --git a/src/Storages/System/StorageSystemParts.cpp b/src/Storages/System/StorageSystemParts.cpp index d788efd8860..fa1c26b623d 100644 --- a/src/Storages/System/StorageSystemParts.cpp +++ b/src/Storages/System/StorageSystemParts.cpp @@ -198,9 +198,9 @@ void StorageSystemParts::processNextStorage( if (part->isStoredOnDisk()) { if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getDiskName()); + columns[res_index++]->insert(part->getDataPartStorage().getDiskName()); if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getFullPath()); + columns[res_index++]->insert(part->getDataPartStorage().getFullPath()); } else { diff --git a/src/Storages/System/StorageSystemPartsColumns.cpp b/src/Storages/System/StorageSystemPartsColumns.cpp index cc6e69b160f..cd51c767eae 100644 --- a/src/Storages/System/StorageSystemPartsColumns.cpp +++ b/src/Storages/System/StorageSystemPartsColumns.cpp @@ -190,9 +190,9 @@ void StorageSystemPartsColumns::processNextStorage( if (columns_mask[src_index++]) columns[res_index++]->insert(info.engine); if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getDiskName()); + columns[res_index++]->insert(part->getDataPartStorage().getDiskName()); if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getFullPath()); + columns[res_index++]->insert(part->getDataPartStorage().getFullPath()); if (columns_mask[src_index++]) columns[res_index++]->insert(column.name); diff --git a/src/Storages/System/StorageSystemProjectionParts.cpp b/src/Storages/System/StorageSystemProjectionParts.cpp index 3934e7c9623..37c62ba5eb0 100644 --- a/src/Storages/System/StorageSystemProjectionParts.cpp +++ b/src/Storages/System/StorageSystemProjectionParts.cpp @@ -200,9 +200,9 @@ void StorageSystemProjectionParts::processNextStorage( if (part->isStoredOnDisk()) { if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getDiskName()); + columns[res_index++]->insert(part->getDataPartStorage().getDiskName()); if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getFullPath()); + columns[res_index++]->insert(part->getDataPartStorage().getFullPath()); } else { diff --git a/src/Storages/System/StorageSystemProjectionPartsColumns.cpp b/src/Storages/System/StorageSystemProjectionPartsColumns.cpp index 0847010faaa..a5968597885 100644 --- a/src/Storages/System/StorageSystemProjectionPartsColumns.cpp +++ b/src/Storages/System/StorageSystemProjectionPartsColumns.cpp @@ -211,9 +211,9 @@ void StorageSystemProjectionPartsColumns::processNextStorage( if (columns_mask[src_index++]) columns[res_index++]->insert(info.engine); if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getDiskName()); + columns[res_index++]->insert(part->getDataPartStorage().getDiskName()); if (columns_mask[src_index++]) - columns[res_index++]->insert(part->data_part_storage->getFullPath()); + columns[res_index++]->insert(part->getDataPartStorage().getFullPath()); if (columns_mask[src_index++]) columns[res_index++]->insert(column.name); From 4a8326ff0c33758570aee25702ed99143e43d853 Mon Sep 17 00:00:00 2001 From: Han Fei Date: Sun, 23 Oct 2022 23:22:52 +0200 Subject: [PATCH 067/239] fix typo --- src/Storages/SelectQueryInfo.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Storages/SelectQueryInfo.h b/src/Storages/SelectQueryInfo.h index 94a8c1143f3..565594569ce 100644 --- a/src/Storages/SelectQueryInfo.h +++ b/src/Storages/SelectQueryInfo.h @@ -220,7 +220,7 @@ struct SelectQueryInfo Block minmax_count_projection_block; MergeTreeDataSelectAnalysisResultPtr merge_tree_select_result_ptr; - // If limit is not 0, that means it's a trival limit query. + // If limit is not 0, that means it's a trivial limit query. UInt64 limit = 0; InputOrderInfoPtr getInputOrderInfo() const From cf375c9732d4380a203a4094574b098ebe9a885b Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Sun, 23 Oct 2022 22:29:24 +0000 Subject: [PATCH 068/239] better interface --- .../MergeTree/DataPartStorageOnDisk.cpp | 176 +++--------------- .../MergeTree/DataPartStorageOnDisk.h | 15 +- src/Storages/MergeTree/IDataPartStorage.h | 24 +-- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 139 +++++++++++++- src/Storages/MergeTree/IMergeTreeDataPart.h | 13 ++ src/Storages/MergeTree/MergeTreeData.cpp | 9 +- src/Storages/StorageMergeTree.cpp | 7 +- 7 files changed, 200 insertions(+), 183 deletions(-) diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index e28aa359c99..cdbd01efab4 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -558,154 +558,6 @@ size_t DataPartStorageOnDisk::getVolumeIndex(const IStoragePolicy & storage_poli return storage_policy.getVolumeIndexByDisk(volume->getDisk()); } -void DataPartStorageOnDisk::writeChecksums(const MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const -{ - std::string path = fs::path(root_path) / part_dir / "checksums.txt"; - - try - { - { - auto out = volume->getDisk()->writeFile(path + ".tmp", 4096, WriteMode::Rewrite, settings); - checksums.write(*out); - } - - volume->getDisk()->moveFile(path + ".tmp", path); - } - catch (...) - { - try - { - if (volume->getDisk()->exists(path + ".tmp")) - volume->getDisk()->removeFile(path + ".tmp"); - } - catch (...) - { - tryLogCurrentException("DataPartStorageOnDisk"); - } - - throw; - } -} - -void DataPartStorageOnDisk::writeColumns(const NamesAndTypesList & columns, const WriteSettings & settings) const -{ - std::string path = fs::path(root_path) / part_dir / "columns.txt"; - - try - { - auto buf = volume->getDisk()->writeFile(path + ".tmp", 4096, WriteMode::Rewrite, settings); - columns.writeText(*buf); - buf->finalize(); - - volume->getDisk()->moveFile(path + ".tmp", path); - } - catch (...) - { - try - { - if (volume->getDisk()->exists(path + ".tmp")) - volume->getDisk()->removeFile(path + ".tmp"); - } - catch (...) - { - tryLogCurrentException("DataPartStorageOnDisk"); - } - - throw; - } -} - -void DataPartStorageOnDisk::writeVersionMetadata(const VersionMetadata & version, bool fsync_part_dir) const -{ - std::string path = fs::path(root_path) / part_dir / "txn_version.txt"; - try - { - { - /// TODO IDisk interface does not allow to open file with O_EXCL flag (for DiskLocal), - /// so we create empty file at first (expecting that createFile throws if file already exists) - /// and then overwrite it. - volume->getDisk()->createFile(path + ".tmp"); - auto buf = volume->getDisk()->writeFile(path + ".tmp", 256); - version.write(*buf); - buf->finalize(); - buf->sync(); - } - - SyncGuardPtr sync_guard; - if (fsync_part_dir) - sync_guard = volume->getDisk()->getDirectorySyncGuard(getRelativePath()); - volume->getDisk()->replaceFile(path + ".tmp", path); - - } - catch (...) - { - try - { - if (volume->getDisk()->exists(path + ".tmp")) - volume->getDisk()->removeFile(path + ".tmp"); - } - catch (...) - { - tryLogCurrentException("DataPartStorageOnDisk"); - } - - throw; - } -} - -void DataPartStorageOnDisk::appendCSNToVersionMetadata(const VersionMetadata & version, VersionMetadata::WhichCSN which_csn) const -{ - /// Small enough appends to file are usually atomic, - /// so we append new metadata instead of rewriting file to reduce number of fsyncs. - /// We don't need to do fsync when writing CSN, because in case of hard restart - /// we will be able to restore CSN from transaction log in Keeper. - - std::string version_file_name = fs::path(root_path) / part_dir / "txn_version.txt"; - DiskPtr disk = volume->getDisk(); - auto out = disk->writeFile(version_file_name, 256, WriteMode::Append); - version.writeCSN(*out, which_csn); - out->finalize(); -} - -void DataPartStorageOnDisk::appendRemovalTIDToVersionMetadata(const VersionMetadata & version, bool clear) const -{ - String version_file_name = fs::path(root_path) / part_dir / "txn_version.txt"; - DiskPtr disk = volume->getDisk(); - auto out = disk->writeFile(version_file_name, 256, WriteMode::Append); - version.writeRemovalTID(*out, clear); - out->finalize(); - - /// fsync is not required when we clearing removal TID, because after hard restart we will fix metadata - if (!clear) - out->sync(); -} - -void DataPartStorageOnDisk::writeDeleteOnDestroyMarker(Poco::Logger * log) const -{ - String marker_path = fs::path(root_path) / part_dir / "delete-on-destroy.txt"; - auto disk = volume->getDisk(); - try - { - volume->getDisk()->createFile(marker_path); - } - catch (Poco::Exception & e) - { - LOG_ERROR(log, "{} (while creating DeleteOnDestroy marker: {})", e.what(), backQuote(fullPath(disk, marker_path))); - } -} - -void DataPartStorageOnDisk::removeDeleteOnDestroyMarker() const -{ - std::string delete_on_destroy_file_name = fs::path(root_path) / part_dir / "delete-on-destroy.txt"; - volume->getDisk()->removeFileIfExists(delete_on_destroy_file_name); -} - -void DataPartStorageOnDisk::removeVersionMetadata() const -{ - std::string version_file_name = fs::path(root_path) / part_dir / "txn_version.txt"; - volume->getDisk()->removeFileIfExists(version_file_name); -} - String DataPartStorageOnDisk::getUniqueId() const { auto disk = volume->getDisk(); @@ -935,6 +787,34 @@ std::unique_ptr DataPartStorageOnDisk::writeFile( return volume->getDisk()->writeFile(fs::path(root_path) / part_dir / name, buf_size, WriteMode::Rewrite, settings); } +std::unique_ptr DataPartStorageOnDisk::writeTransactionFile(WriteMode mode) const +{ + return volume->getDisk()->writeFile(fs::path(root_path) / part_dir / "txn_version.txt", 256, mode); +} + +void DataPartStorageOnDisk::createFile(const String & name) +{ + executeOperation([&](auto & disk) { disk.createFile(fs::path(root_path) / part_dir / name); }); +} + +void DataPartStorageOnDisk::moveFile(const String & from_name, const String & to_name) +{ + executeOperation([&](auto & disk) + { + auto relative_path = fs::path(root_path) / part_dir; + disk.moveFile(relative_path / from_name, relative_path / to_name); + }); +} + +void DataPartStorageOnDisk::replaceFile(const String & from_name, const String & to_name) +{ + executeOperation([&](auto & disk) + { + auto relative_path = fs::path(root_path) / part_dir; + disk.replaceFile(relative_path / from_name, relative_path / to_name); + }); +} + void DataPartStorageOnDisk::removeFile(const String & name) { executeOperation([&](auto & disk) { disk.removeFile(fs::path(root_path) / part_dir / name); }); diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.h b/src/Storages/MergeTree/DataPartStorageOnDisk.h index 80946c37f79..fda901d0204 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.h +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.h @@ -78,15 +78,6 @@ public: ReservationPtr tryReserve(UInt64 bytes) const override; size_t getVolumeIndex(const IStoragePolicy &) const override; - void writeChecksums(const MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const override; - void writeColumns(const NamesAndTypesList & columns, const WriteSettings & settings) const override; - void writeVersionMetadata(const VersionMetadata & version, bool fsync_part_dir) const override; - void appendCSNToVersionMetadata(const VersionMetadata & version, VersionMetadata::WhichCSN which_csn) const override; - void appendRemovalTIDToVersionMetadata(const VersionMetadata & version, bool clear) const override; - void writeDeleteOnDestroyMarker(Poco::Logger * log) const override; - void removeDeleteOnDestroyMarker() const override; - void removeVersionMetadata() const override; - String getUniqueId() const override; bool shallParticipateInMerges(const IStoragePolicy &) const override; @@ -123,6 +114,12 @@ public: size_t buf_size, const WriteSettings & settings) override; + std::unique_ptr writeTransactionFile(WriteMode mode) const override; + + void createFile(const String & name) override; + void moveFile(const String & from_name, const String & to_name) override; + void replaceFile(const String & from_name, const String & to_name) override; + void removeFile(const String & name) override; void removeFileIfExists(const String & name) override; void removeRecursive() override; diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index 9e3fbe7d13b..3f73199305d 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -4,6 +4,7 @@ #include #include #include +#include #include #include #include @@ -142,6 +143,7 @@ public: virtual bool supportZeroCopyReplication() const { return false; } virtual bool supportParallelWrite() const = 0; virtual bool isBroken() const = 0; + /// TODO: remove or at least remove const. virtual void syncRevision(UInt64 revision) const = 0; virtual UInt64 getRevision() const = 0; @@ -160,17 +162,6 @@ public: virtual ReservationPtr tryReserve(UInt64 /*bytes*/) const { return nullptr; } virtual size_t getVolumeIndex(const IStoragePolicy &) const { return 0; } - /// Some methods which change data part internals possibly after creation. - /// Probably we should try to remove it later. - virtual void writeChecksums(const MergeTreeDataPartChecksums & checksums, const WriteSettings & settings) const = 0; - virtual void writeColumns(const NamesAndTypesList & columns, const WriteSettings & settings) const = 0; - virtual void writeVersionMetadata(const VersionMetadata & version, bool fsync_part_dir) const = 0; - virtual void appendCSNToVersionMetadata(const VersionMetadata & version, VersionMetadata::WhichCSN which_csn) const = 0; - virtual void appendRemovalTIDToVersionMetadata(const VersionMetadata & version, bool clear) const = 0; - virtual void writeDeleteOnDestroyMarker(Poco::Logger * log) const = 0; - virtual void removeDeleteOnDestroyMarker() const = 0; - virtual void removeVersionMetadata() const = 0; - /// A leak of abstraction. /// Return some uniq string for file. /// Required for distinguish different copies of the same part on remote FS. @@ -219,7 +210,16 @@ public: virtual void createDirectories() = 0; virtual void createProjection(const std::string & name) = 0; - virtual std::unique_ptr writeFile(const String & name, size_t buf_size, const WriteSettings & settings) = 0; + virtual std::unique_ptr writeFile( + const String & name, + size_t buf_size, + const WriteSettings & settings) = 0; + + virtual std::unique_ptr writeTransactionFile(WriteMode mode) const = 0; + + virtual void createFile(const String & name) = 0; + virtual void moveFile(const String & from_name, const String & to_name) = 0; + virtual void replaceFile(const String & from_name, const String & to_name) = 0; virtual void removeFile(const String & name) = 0; virtual void removeFileIfExists(const String & name) = 0; diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index fdaa4231e9c..f5348311468 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -854,6 +854,120 @@ void IMergeTreeDataPart::loadDefaultCompressionCodec() } } +template +void IMergeTreeDataPart::writeMetadata(const String & filename, const WriteSettings & settings, Writer && writer) +{ + auto & data_part_storage = getDataPartStorage(); + auto tmp_filename = filename + ".tmp"; + + try + { + { + auto out = data_part_storage.writeFile(tmp_filename, 4096, settings); + writer(*out); + out->finalize(); + } + + data_part_storage.moveFile(tmp_filename, filename); + } + catch (...) + { + try + { + if (data_part_storage.exists(tmp_filename)) + data_part_storage.removeFile(tmp_filename); + } + catch (...) + { + tryLogCurrentException("DataPartStorageOnDisk"); + } + + throw; + } +} + +void IMergeTreeDataPart::writeChecksums(const MergeTreeDataPartChecksums & checksums_, const WriteSettings & settings) +{ + writeMetadata("checksums.txt", settings, [&checksums_](auto & buffer) + { + checksums_.write(buffer); + }); +} + +void IMergeTreeDataPart::writeColumns(const NamesAndTypesList & columns_, const WriteSettings & settings) +{ + writeMetadata("columns.txt", settings, [&columns_](auto & buffer) + { + columns_.writeText(buffer); + }); +} + +void IMergeTreeDataPart::writeVersionMetadata(const VersionMetadata & version_, bool fsync_part_dir) const +{ + auto & data_part_storage = const_cast(getDataPartStorage()); + static constexpr auto filename = "txn_version.txt"; + static constexpr auto tmp_filename = "txn_version.txt.tmp"; + + try + { + { + /// TODO IDisk interface does not allow to open file with O_EXCL flag (for DiskLocal), + /// so we create empty file at first (expecting that createFile throws if file already exists) + /// and then overwrite it. + data_part_storage.createFile(tmp_filename); + auto write_settings = storage.getContext()->getWriteSettings(); + auto buf = data_part_storage.writeFile(tmp_filename, 256, write_settings); + version_.write(*buf); + buf->finalize(); + buf->sync(); + } + + SyncGuardPtr sync_guard; + if (fsync_part_dir) + sync_guard = data_part_storage.getDirectorySyncGuard(); + data_part_storage.replaceFile(tmp_filename, filename); + } + catch (...) + { + try + { + if (data_part_storage.exists(tmp_filename)) + data_part_storage.removeFile(tmp_filename); + } + catch (...) + { + tryLogCurrentException("DataPartStorageOnDisk"); + } + + throw; + } +} + +void IMergeTreeDataPart::writeDeleteOnDestroyMarker() +{ + static constexpr auto marker_path = "delete-on-destroy.txt"; + + try + { + getDataPartStorage().createFile(marker_path); + } + catch (Poco::Exception & e) + { + LOG_ERROR(storage.log, "{} (while creating DeleteOnDestroy marker: {})", + e.what(), (fs::path(getDataPartStorage().getFullPath()) / marker_path).string()); + } +} + +void IMergeTreeDataPart::removeDeleteOnDestroyMarker() +{ + getDataPartStorage().removeFileIfExists("delete-on-destroy.txt"); +} + +void IMergeTreeDataPart::removeVersionMetadata() +{ + getDataPartStorage().removeFileIfExists("txn_version.txt"); +} + void IMergeTreeDataPart::appendFilesOfDefaultCompressionCodec(Strings & files) { files.push_back(DEFAULT_COMPRESSION_CODEC_FILE_NAME); @@ -980,7 +1094,7 @@ void IMergeTreeDataPart::loadChecksums(bool require) LOG_WARNING(storage.log, "Checksums for part {} not found. Will calculate them from data on disk.", name); checksums = checkDataPart(shared_from_this(), false); - getDataPartStorage().writeChecksums(checksums, {}); + writeChecksums(checksums, {}); bytes_on_disk = checksums.getTotalSizeOnDisk(); } @@ -993,8 +1107,6 @@ void IMergeTreeDataPart::appendFilesOfChecksums(Strings & files) void IMergeTreeDataPart::loadRowsCount() { - //String path = fs::path(getRelativePath()) / "count.txt"; - auto read_rows_count = [&]() { auto buf = metadata_manager->read("count.txt"); @@ -1186,7 +1298,7 @@ void IMergeTreeDataPart::loadColumns(bool require) if (columns.empty()) throw Exception("No columns in part " + name, ErrorCodes::NO_FILE_IN_DATA_PART); - getDataPartStorage().writeColumns(loaded_columns, {}); + writeColumns(loaded_columns, {}); } else { @@ -1245,7 +1357,7 @@ void IMergeTreeDataPart::storeVersionMetadata(bool force) const throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Transactions are not supported for in-memory parts (table: {}, part: {})", storage.getStorageID().getNameForLogs(), name); - getDataPartStorage().writeVersionMetadata(version, storage.getSettings()->fsync_part_directory); + writeVersionMetadata(version, storage.getSettings()->fsync_part_directory); } void IMergeTreeDataPart::appendCSNToVersionMetadata(VersionMetadata::WhichCSN which_csn) const @@ -1257,7 +1369,14 @@ void IMergeTreeDataPart::appendCSNToVersionMetadata(VersionMetadata::WhichCSN wh chassert(!(which_csn == VersionMetadata::WhichCSN::REMOVAL && version.removal_csn == 0)); chassert(isStoredOnDisk()); - getDataPartStorage().appendCSNToVersionMetadata(version, which_csn); + /// Small enough appends to file are usually atomic, + /// so we append new metadata instead of rewriting file to reduce number of fsyncs. + /// We don't need to do fsync when writing CSN, because in case of hard restart + /// we will be able to restore CSN from transaction log in Keeper. + + auto out = getDataPartStorage().writeTransactionFile(WriteMode::Append); + version.writeCSN(*out, which_csn); + out->finalize(); } void IMergeTreeDataPart::appendRemovalTIDToVersionMetadata(bool clear) const @@ -1280,7 +1399,13 @@ void IMergeTreeDataPart::appendRemovalTIDToVersionMetadata(bool clear) const else LOG_TEST(storage.log, "Appending removal TID for {} (creation: {}, removal {})", name, version.creation_tid, version.removal_tid); - getDataPartStorage().appendRemovalTIDToVersionMetadata(version, clear); + auto out = getDataPartStorage().writeTransactionFile(WriteMode::Append); + version.writeRemovalTID(*out, clear); + out->finalize(); + + /// fsync is not required when we clearing removal TID, because after hard restart we will fix metadata + if (!clear) + out->sync(); } void IMergeTreeDataPart::loadVersionMetadata() const diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.h b/src/Storages/MergeTree/IMergeTreeDataPart.h index fbe4f992de4..6515eb1a65c 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.h +++ b/src/Storages/MergeTree/IMergeTreeDataPart.h @@ -1,5 +1,6 @@ #pragma once +#include "IO/WriteSettings.h" #include #include #include @@ -439,6 +440,12 @@ public: /// True if here is lightweight deleted mask file in part. bool hasLightweightDelete() const { return columns.contains(LightweightDeleteDescription::FILTER_COLUMN.name); } + void writeChecksums(const MergeTreeDataPartChecksums & checksums_, const WriteSettings & settings); + + void writeDeleteOnDestroyMarker(); + void removeDeleteOnDestroyMarker(); + void removeVersionMetadata(); + protected: /// Total size of all columns, calculated once in calcuateColumnSizesOnDisk @@ -560,6 +567,12 @@ private: /// any specifial compression. void loadDefaultCompressionCodec(); + void writeColumns(const NamesAndTypesList & columns_, const WriteSettings & settings); + void writeVersionMetadata(const VersionMetadata & version_, bool fsync_part_dir) const; + + template + void writeMetadata(const String & filename, const WriteSettings & settings, Writer && writer); + static void appendFilesOfDefaultCompressionCodec(Strings & files); /// Found column without specific compression and return codec diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 83f3a167fa7..04e5432f239 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -1685,7 +1685,8 @@ scope_guard MergeTreeData::getTemporaryPartDirectoryHolder(const String & part_d MergeTreeData::MutableDataPartPtr MergeTreeData::preparePartForRemoval(const DataPartPtr & part) { - if (part->getState() != DataPartState::Deleting) + auto state = part->getState(); + if (state != DataPartState::Deleting && state != DataPartState::DeleteOnDestroy) throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot remove part {}, because it has state: {}", part->name, magic_enum::enum_name(part->getState())); @@ -3666,7 +3667,7 @@ void MergeTreeData::swapActivePart(MergeTreeData::DataPartPtr part_copy) /// All other locks are taken in StorageReplicatedMergeTree lockSharedData(*part_copy); - original_active_part->getDataPartStorage().writeDeleteOnDestroyMarker(log); + preparePartForRemoval(original_active_part)->writeDeleteOnDestroyMarker(); return; } } @@ -3801,8 +3802,8 @@ static void loadPartAndFixMetadataImpl(MergeTreeData::MutableDataPartPtr part) { part->loadColumnsChecksumsIndexes(false, true); part->modification_time = part->getDataPartStorage().getLastModified().epochTime(); - part->getDataPartStorage().removeDeleteOnDestroyMarker(); - part->getDataPartStorage().removeVersionMetadata(); + part->removeDeleteOnDestroyMarker(); + part->removeVersionMetadata(); } void MergeTreeData::calculateColumnAndSecondaryIndexSizesImpl() diff --git a/src/Storages/StorageMergeTree.cpp b/src/Storages/StorageMergeTree.cpp index 3b456ab5360..90340bb01a7 100644 --- a/src/Storages/StorageMergeTree.cpp +++ b/src/Storages/StorageMergeTree.cpp @@ -1,4 +1,5 @@ #include "StorageMergeTree.h" +#include "Storages/MergeTree/IMergeTreeDataPart.h" #include @@ -1739,8 +1740,7 @@ CheckResults StorageMergeTree::checkData(const ASTPtr & query, ContextPtr local_ for (auto & part : data_parts) { /// If the checksums file is not present, calculate the checksums and write them to disk. - String checksums_path = "checksums.txt"; - String tmp_checksums_path = "checksums.txt.tmp"; + static constexpr auto checksums_path = "checksums.txt"; if (part->isStoredOnDisk() && !part->getDataPartStorage().exists(checksums_path)) { try @@ -1748,7 +1748,8 @@ CheckResults StorageMergeTree::checkData(const ASTPtr & query, ContextPtr local_ auto calculated_checksums = checkDataPart(part, false); calculated_checksums.checkEqual(part->checksums, true); - part->getDataPartStorage().writeChecksums(part->checksums, local_context->getWriteSettings()); + auto & part_mutable = const_cast(*part); + part_mutable.writeChecksums(part->checksums, local_context->getWriteSettings()); part->checkMetadata(); results.emplace_back(part->name, true, "Checksums recounted and written to disk."); From 7243e12ef88b6c762561e523ce8bdd8f1599988e Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Mon, 24 Oct 2022 07:41:12 +0000 Subject: [PATCH 069/239] Fix build --- contrib/libcxxabi-cmake/CMakeLists.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/libcxxabi-cmake/CMakeLists.txt b/contrib/libcxxabi-cmake/CMakeLists.txt index 221a18de6e5..a59452eee9a 100644 --- a/contrib/libcxxabi-cmake/CMakeLists.txt +++ b/contrib/libcxxabi-cmake/CMakeLists.txt @@ -9,7 +9,7 @@ set(SRCS "${LIBCXXABI_SOURCE_DIR}/src/cxa_exception_storage.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_guard.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_handlers.cpp" -"${LIBCXXABI_SOURCE_DIR}/src/cxa_noexception.cpp" +# "${LIBCXXABI_SOURCE_DIR}/src/cxa_noexception.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_personality.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_thread_atexit.cpp" "${LIBCXXABI_SOURCE_DIR}/src/cxa_vector.cpp" From 5579d139eba0cb533a2722eaa11371b4212e12c0 Mon Sep 17 00:00:00 2001 From: Antonio Andelic Date: Mon, 24 Oct 2022 08:37:52 +0000 Subject: [PATCH 070/239] Avoid crash with big int in prewhere --- .../MergeTree/MergeTreeBaseSelectProcessor.cpp | 4 ++-- .../02473_prewhere_with_bigint.reference | 0 .../0_stateless/02473_prewhere_with_bigint.sql | 15 +++++++++++++++ 3 files changed, 17 insertions(+), 2 deletions(-) create mode 100644 tests/queries/0_stateless/02473_prewhere_with_bigint.reference create mode 100644 tests/queries/0_stateless/02473_prewhere_with_bigint.sql diff --git a/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp b/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp index 475407a402b..710b8aef50c 100644 --- a/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp +++ b/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp @@ -628,12 +628,12 @@ Block MergeTreeBaseSelectProcessor::transformHeader( else { WhichDataType which(removeNullable(recursiveRemoveLowCardinality(prewhere_column.type))); - if (which.isInt() || which.isUInt()) + if (which.isNativeInt() || which.isNativeUInt()) prewhere_column.column = prewhere_column.type->createColumnConst(block.rows(), 1u)->convertToFullColumnIfConst(); else if (which.isFloat()) prewhere_column.column = prewhere_column.type->createColumnConst(block.rows(), 1.0f)->convertToFullColumnIfConst(); else - throw Exception("Illegal type " + prewhere_column.type->getName() + " of column for filter.", + throw Exception("Illegal type " + prewhere_column.type->getName() + " of column for filter", ErrorCodes::ILLEGAL_TYPE_OF_COLUMN_FOR_FILTER); } } diff --git a/tests/queries/0_stateless/02473_prewhere_with_bigint.reference b/tests/queries/0_stateless/02473_prewhere_with_bigint.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02473_prewhere_with_bigint.sql b/tests/queries/0_stateless/02473_prewhere_with_bigint.sql new file mode 100644 index 00000000000..852da729648 --- /dev/null +++ b/tests/queries/0_stateless/02473_prewhere_with_bigint.sql @@ -0,0 +1,15 @@ +CREATE TABLE prewhere_int128 (a Int128) ENGINE=MergeTree ORDER BY a; +SELECT a FROM prewhere_int128 WHERE a; -- { serverError 59 } +DROP TABLE prewhere_int128; + +CREATE TABLE prewhere_int256 (a Int256) ENGINE=MergeTree ORDER BY a; +SELECT a FROM prewhere_int256 WHERE a; -- { serverError 59 } +DROP TABLE prewhere_int256; + +CREATE TABLE prewhere_uint128 (a UInt128) ENGINE=MergeTree ORDER BY a; +SELECT a FROM prewhere_uint128 WHERE a; -- { serverError 59 } +DROP TABLE prewhere_uint128; + +CREATE TABLE prewhere_uint256 (a UInt256) ENGINE=MergeTree ORDER BY a; +SELECT a FROM prewhere_uint256 WHERE a; -- { serverError 59 } +DROP TABLE prewhere_uint256; From 8f00d0d1cc709733b2571b23cba6bead445fed72 Mon Sep 17 00:00:00 2001 From: Han Fei Date: Mon, 24 Oct 2022 12:56:36 +0200 Subject: [PATCH 071/239] fix tidy --- src/Processors/QueryPlan/ReadFromMergeTree.cpp | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.cpp b/src/Processors/QueryPlan/ReadFromMergeTree.cpp index 164ec8777de..ba7b0e963eb 100644 --- a/src/Processors/QueryPlan/ReadFromMergeTree.cpp +++ b/src/Processors/QueryPlan/ReadFromMergeTree.cpp @@ -253,17 +253,19 @@ ProcessorPtr ReadFromMergeTree::createSource( if (query_info.limit > 0 && query_info.limit < total_rows) total_rows = query_info.limit; - auto source = std::make_shared( - data, storage_snapshot, part.data_part, max_block_size, preferred_block_size_bytes, - preferred_max_column_in_block_size_bytes, required_columns, part.ranges, use_uncompressed_cache, prewhere_info, - actions_settings, reader_settings, virt_column_names, part.part_index_in_query, has_limit_below_one_block, std::move(extension)); - /// Actually it means that parallel reading from replicas enabled /// and we have to collaborate with initiator. /// In this case we won't set approximate rows, because it will be accounted multiple times. /// Also do not count amount of read rows if we read in order of sorting key, /// because we don't know actual amount of read rows in case when limit is set. - if (!extension.has_value() && !reader_settings.read_in_order) + bool set_rows_approx = !extension.has_value() && !reader_settings.read_in_order; + + auto source = std::make_shared( + data, storage_snapshot, part.data_part, max_block_size, preferred_block_size_bytes, + preferred_max_column_in_block_size_bytes, required_columns, part.ranges, use_uncompressed_cache, prewhere_info, + actions_settings, reader_settings, virt_column_names, part.part_index_in_query, has_limit_below_one_block, std::move(extension)); + + if (set_rows_approx) source -> addTotalRowsApprox(total_rows); return source; From 14e3bb6354f4acf4341731b67ba6a626752c69e3 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Mon, 24 Oct 2022 13:58:26 +0000 Subject: [PATCH 072/239] Another case. --- src/Functions/FunctionBinaryArithmetic.h | 6 ++-- ...461_mullable_pk_monotonicity_bug.reference | 32 +++++++++++++++++++ .../02461_mullable_pk_monotonicity_bug.sql | 32 ++++++++++++++++--- 3 files changed, 63 insertions(+), 7 deletions(-) diff --git a/src/Functions/FunctionBinaryArithmetic.h b/src/Functions/FunctionBinaryArithmetic.h index 3ffe054a439..1c1d0453230 100644 --- a/src/Functions/FunctionBinaryArithmetic.h +++ b/src/Functions/FunctionBinaryArithmetic.h @@ -1781,6 +1781,7 @@ public: { auto left_type = removeNullable(removeLowCardinality(left.type)); auto right_type = removeNullable(removeLowCardinality(right.type)); + auto ret_type = removeNullable(removeLowCardinality(return_type)); auto transform = [&](const Field & point) { @@ -1791,7 +1792,7 @@ public: /// This is a bit dangerous to call Base::executeImpl cause it ignores `use Default Implementation For XXX` flags. /// It was possible to check monotonicity for nullable right type which result to exception. /// Adding removeNullable above fixes the issue, but some other inconsistency may left. - auto col = Base::executeImpl(columns_with_constant, return_type, 1); + auto col = Base::executeImpl(columns_with_constant, ret_type, 1); Field point_transformed; col->get(0, point_transformed); return point_transformed; @@ -1822,6 +1823,7 @@ public: { auto left_type = removeNullable(removeLowCardinality(left.type)); auto right_type = removeNullable(removeLowCardinality(right.type)); + auto ret_type = removeNullable(removeLowCardinality(return_type)); auto transform = [&](const Field & point) { @@ -1829,7 +1831,7 @@ public: = {{left_type->createColumnConst(1, point), left_type, left.name}, {right_type->createColumnConst(1, (*right.column)[0]), right_type, right.name}}; - auto col = Base::executeImpl(columns_with_constant, return_type, 1); + auto col = Base::executeImpl(columns_with_constant, ret_type, 1); Field point_transformed; col->get(0, point_transformed); return point_transformed; diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference index 5ee6e6c67c4..29da9bc651b 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference @@ -30,3 +30,35 @@ 2 1 2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1 +2 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 +1970-01-01 00:00:01 +1970-01-01 00:00:02 diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql index ce190694514..af2d5b859b3 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql @@ -1,6 +1,5 @@ create table tab (x Nullable(UInt8)) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; insert into tab select number from numbers(4); - set allow_suspicious_low_cardinality_types=1; set max_rows_to_read = 2; @@ -14,9 +13,7 @@ SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; drop table tab; - set max_rows_to_read = 100; - create table tab (x LowCardinality(UInt8)) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; insert into tab select number from numbers(4); @@ -31,10 +28,35 @@ SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; drop table tab; - set max_rows_to_read = 100; - create table tab (x UInt128) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; insert into tab select number from numbers(4); +set max_rows_to_read = 2; +SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(Nullable(UInt8))) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; + +set max_rows_to_read = 100; SELECT x + 1 FROM tab WHERE (x + 1::LowCardinality(UInt8)) <= -9223372036854775808 order by x; + +drop table tab; +create table tab (x DateTime) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; +insert into tab select number from numbers(4); + +set max_rows_to_read = 2; +SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(Nullable(UInt8))) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; + +SELECT x + 1 FROM tab WHERE (x + CAST('1', 'Nullable(UInt8)')) <= -2147483647 ORDER BY x ASC NULLS FIRST; From 128fdd785e772bcdd1c8b7b3c558363cf58a44d8 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 24 Oct 2022 14:44:22 +0000 Subject: [PATCH 073/239] better interface --- .../MergeTree/DataPartStorageOnDisk.cpp | 67 ++----------------- .../MergeTree/DataPartStorageOnDisk.h | 4 +- src/Storages/MergeTree/IDataPartStorage.h | 9 ++- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 59 +++++++++++++++- src/Storages/MergeTree/MergeTask.h | 2 +- .../MergeTree/MergeTreeDataWriter.cpp | 8 +-- src/Storages/MergeTree/MergeTreeDataWriter.h | 8 +-- 7 files changed, 81 insertions(+), 76 deletions(-) diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index cdbd01efab4..a3df17c57de 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -1,3 +1,4 @@ +#include "Storages/MergeTree/IDataPartStorage.h" #include #include #include @@ -50,7 +51,12 @@ std::string DataPartStorageOnDisk::getFullRootPath() const return fs::path(volume->getDisk()->getPath()) / root_path / ""; } -MutableDataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) const +MutableDataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) +{ + return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); +} + +DataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) const { return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); } @@ -136,65 +142,6 @@ std::unique_ptr DataPartStorageOnDisk::readFile( return volume->getDisk()->readFile(fs::path(root_path) / part_dir / name, settings, read_hint, file_size); } -static std::unique_ptr openForReading(const DiskPtr & disk, const String & path) -{ - size_t file_size = disk->getFileSize(path); - return disk->readFile(path, ReadSettings().adjustBufferSize(file_size), file_size); -} - -void DataPartStorageOnDisk::loadVersionMetadata(VersionMetadata & version, Poco::Logger * log) const -{ - std::string version_file_name = fs::path(root_path) / part_dir / "txn_version.txt"; - String tmp_version_file_name = version_file_name + ".tmp"; - DiskPtr disk = volume->getDisk(); - - auto remove_tmp_file = [&]() - { - auto last_modified = disk->getLastModified(tmp_version_file_name); - auto buf = openForReading(disk, tmp_version_file_name); - String content; - readStringUntilEOF(content, *buf); - LOG_WARNING(log, "Found file {} that was last modified on {}, has size {} and the following content: {}", - tmp_version_file_name, last_modified.epochTime(), content.size(), content); - disk->removeFile(tmp_version_file_name); - }; - - if (disk->exists(version_file_name)) - { - auto buf = openForReading(disk, version_file_name); - version.read(*buf); - if (disk->exists(tmp_version_file_name)) - remove_tmp_file(); - return; - } - - /// Four (?) cases are possible: - /// 1. Part was created without transactions. - /// 2. Version metadata file was not renamed from *.tmp on part creation. - /// 3. Version metadata were written to *.tmp file, but hard restart happened before fsync. - /// 4. Fsyncs in storeVersionMetadata() work incorrectly. - - if (!disk->exists(tmp_version_file_name)) - { - /// Case 1. - /// We do not have version metadata and transactions history for old parts, - /// so let's consider that such parts were created by some ancient transaction - /// and were committed with some prehistoric CSN. - /// NOTE It might be Case 3, but version metadata file is written on part creation before other files, - /// so it's not Case 3 if part is not broken. - version.setCreationTID(Tx::PrehistoricTID, nullptr); - version.creation_csn = Tx::PrehistoricCSN; - return; - } - - /// Case 2. - /// Content of *.tmp file may be broken, just use fake TID. - /// Transaction was not committed if *.tmp file was not renamed, so we should complete rollback by removing part. - version.setCreationTID(Tx::DummyTID, nullptr); - version.creation_csn = Tx::RolledBackCSN; - remove_tmp_file(); -} - void DataPartStorageOnDisk::checkConsistency(const MergeTreeDataPartChecksums & checksums) const { checksums.checkSizes(volume->getDisk(), getRelativePath()); diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.h b/src/Storages/MergeTree/DataPartStorageOnDisk.h index fda901d0204..f93c6a235d6 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.h +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.h @@ -21,7 +21,8 @@ public: std::string getPartDirectory() const override { return part_dir; } std::string getFullRootPath() const override; - MutableDataPartStoragePtr getProjection(const std::string & name) const override; + MutableDataPartStoragePtr getProjection(const std::string & name) override; + DataPartStoragePtr getProjection(const std::string & name) const override; bool exists() const override; bool exists(const std::string & name) const override; @@ -41,7 +42,6 @@ public: std::optional read_hint, std::optional file_size) const override; - void loadVersionMetadata(VersionMetadata & version, Poco::Logger * log) const override; void checkConsistency(const MergeTreeDataPartChecksums & checksums) const override; void remove( diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index 3f73199305d..8f478b3c688 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -82,16 +82,19 @@ public: /// virtual std::string getRelativeRootPath() const = 0; /// Get a storage for projection. - virtual std::shared_ptr getProjection(const std::string & name) const = 0; + virtual std::shared_ptr getProjection(const std::string & name) = 0; + virtual std::shared_ptr getProjection(const std::string & name) const = 0; /// Part directory exists. virtual bool exists() const = 0; + /// File inside part directory exists. Specified path is relative to the part path. virtual bool exists(const std::string & name) const = 0; virtual bool isDirectory(const std::string & name) const = 0; /// Modification time for part directory. virtual Poco::Timestamp getLastModified() const = 0; + /// Iterate part directory. Iteration in subdirectory is not needed yet. virtual DataPartStorageIteratorPtr iterate() const = 0; @@ -108,7 +111,6 @@ public: std::optional read_hint, std::optional file_size) const = 0; - virtual void loadVersionMetadata(VersionMetadata & version, Poco::Logger * log) const = 0; virtual void checkConsistency(const MergeTreeDataPartChecksums & checksums) const = 0; struct ProjectionChecksums @@ -130,7 +132,8 @@ public: /// Get a name like 'prefix_partdir_tryN' which does not exist in a root dir. /// TODO: remove it. - virtual std::optional getRelativePathForPrefix(Poco::Logger * log, const String & prefix, bool detached, bool broken) const = 0; + virtual std::optional getRelativePathForPrefix( + Poco::Logger * log, const String & prefix, bool detached, bool broken) const = 0; /// Reset part directory, used for in-memory parts. /// TODO: remove it. diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index f5348311468..24fabc7cff1 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -904,9 +904,9 @@ void IMergeTreeDataPart::writeColumns(const NamesAndTypesList & columns_, const void IMergeTreeDataPart::writeVersionMetadata(const VersionMetadata & version_, bool fsync_part_dir) const { - auto & data_part_storage = const_cast(getDataPartStorage()); static constexpr auto filename = "txn_version.txt"; static constexpr auto tmp_filename = "txn_version.txt.tmp"; + auto & data_part_storage = const_cast(getDataPartStorage()); try { @@ -1408,10 +1408,65 @@ void IMergeTreeDataPart::appendRemovalTIDToVersionMetadata(bool clear) const out->sync(); } +static std::unique_ptr openForReading(const IDataPartStorage & part_storage, const String & filename) +{ + size_t file_size = part_storage.getFileSize(filename); + return part_storage.readFile(filename, ReadSettings().adjustBufferSize(file_size), file_size, file_size); +} + void IMergeTreeDataPart::loadVersionMetadata() const try { - getDataPartStorage().loadVersionMetadata(version, storage.log); + static constexpr auto version_file_name = "txn_version.txt"; + static constexpr auto tmp_version_file_name = "txn_version.txt.tmp"; + auto & data_part_storage = const_cast(getDataPartStorage()); + + auto remove_tmp_file = [&]() + { + auto last_modified = data_part_storage.getLastModified(); + auto buf = openForReading(data_part_storage, tmp_version_file_name); + + String content; + readStringUntilEOF(content, *buf); + LOG_WARNING(storage.log, "Found file {} that was last modified on {}, has size {} and the following content: {}", + tmp_version_file_name, last_modified.epochTime(), content.size(), content); + data_part_storage.removeFile(tmp_version_file_name); + }; + + if (data_part_storage.exists(version_file_name)) + { + auto buf = openForReading(data_part_storage, version_file_name); + version.read(*buf); + if (data_part_storage.exists(tmp_version_file_name)) + remove_tmp_file(); + return; + } + + /// Four (?) cases are possible: + /// 1. Part was created without transactions. + /// 2. Version metadata file was not renamed from *.tmp on part creation. + /// 3. Version metadata were written to *.tmp file, but hard restart happened before fsync. + /// 4. Fsyncs in storeVersionMetadata() work incorrectly. + + if (!data_part_storage.exists(tmp_version_file_name)) + { + /// Case 1. + /// We do not have version metadata and transactions history for old parts, + /// so let's consider that such parts were created by some ancient transaction + /// and were committed with some prehistoric CSN. + /// NOTE It might be Case 3, but version metadata file is written on part creation before other files, + /// so it's not Case 3 if part is not broken. + version.setCreationTID(Tx::PrehistoricTID, nullptr); + version.creation_csn = Tx::PrehistoricCSN; + return; + } + + /// Case 2. + /// Content of *.tmp file may be broken, just use fake TID. + /// Transaction was not committed if *.tmp file was not renamed, so we should complete rollback by removing part. + version.setCreationTID(Tx::DummyTID, nullptr); + version.creation_csn = Tx::RolledBackCSN; + remove_tmp_file(); } catch (Exception & e) { diff --git a/src/Storages/MergeTree/MergeTask.h b/src/Storages/MergeTree/MergeTask.h index f21d542c7a0..6a29cdbb5ca 100644 --- a/src/Storages/MergeTree/MergeTask.h +++ b/src/Storages/MergeTree/MergeTask.h @@ -134,7 +134,7 @@ private: StorageMetadataPtr metadata_snapshot{nullptr}; FutureMergedMutatedPartPtr future_part{nullptr}; /// This will be either nullptr or new_data_part, so raw pointer is ok. - const IMergeTreeDataPart * parent_part{nullptr}; + IMergeTreeDataPart * parent_part{nullptr}; ContextPtr context{nullptr}; time_t time_of_merge{0}; ReservationSharedPtr space_reservation{nullptr}; diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 856a684d18d..55404324b75 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -483,7 +483,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( MergeTreeDataPartType part_type, const String & relative_path, bool is_temp, - const IMergeTreeDataPart * parent_part, + IMergeTreeDataPart * parent_part, const MergeTreeData & data, Poco::Logger * log, Block block, @@ -587,7 +587,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPart( Poco::Logger * log, Block block, const ProjectionDescription & projection, - const IMergeTreeDataPart * parent_part) + IMergeTreeDataPart * parent_part) { String part_name = projection.name; MergeTreeDataPartType part_type; @@ -623,7 +623,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( Poco::Logger * log, Block block, const ProjectionDescription & projection, - const IMergeTreeDataPart * parent_part, + IMergeTreeDataPart * parent_part, size_t block_num) { String part_name = fmt::format("{}_{}", projection.name, block_num); @@ -658,7 +658,7 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeInMemoryProjectionP Poco::Logger * log, Block block, const ProjectionDescription & projection, - const IMergeTreeDataPart * parent_part) + IMergeTreeDataPart * parent_part) { return writeProjectionPartImpl( projection.name, diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.h b/src/Storages/MergeTree/MergeTreeDataWriter.h index 72ceb8b38e3..8e405016cde 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.h +++ b/src/Storages/MergeTree/MergeTreeDataWriter.h @@ -77,7 +77,7 @@ public: Poco::Logger * log, Block block, const ProjectionDescription & projection, - const IMergeTreeDataPart * parent_part); + IMergeTreeDataPart * parent_part); /// For mutation: MATERIALIZE PROJECTION. static TemporaryPart writeTempProjectionPart( @@ -85,7 +85,7 @@ public: Poco::Logger * log, Block block, const ProjectionDescription & projection, - const IMergeTreeDataPart * parent_part, + IMergeTreeDataPart * parent_part, size_t block_num); /// For WriteAheadLog AddPart. @@ -94,7 +94,7 @@ public: Poco::Logger * log, Block block, const ProjectionDescription & projection, - const IMergeTreeDataPart * parent_part); + IMergeTreeDataPart * parent_part); static Block mergeBlock( const Block & block, @@ -109,7 +109,7 @@ private: MergeTreeDataPartType part_type, const String & relative_path, bool is_temp, - const IMergeTreeDataPart * parent_part, + IMergeTreeDataPart * parent_part, const MergeTreeData & data, Poco::Logger * log, Block block, From 724f8335294603ed5fd1724f2f01d299c57f1175 Mon Sep 17 00:00:00 2001 From: Antonio Andelic Date: Mon, 24 Oct 2022 15:08:41 +0000 Subject: [PATCH 074/239] Use PREWHERE in tests --- tests/queries/0_stateless/02473_prewhere_with_bigint.sql | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/queries/0_stateless/02473_prewhere_with_bigint.sql b/tests/queries/0_stateless/02473_prewhere_with_bigint.sql index 852da729648..f11a1a279de 100644 --- a/tests/queries/0_stateless/02473_prewhere_with_bigint.sql +++ b/tests/queries/0_stateless/02473_prewhere_with_bigint.sql @@ -1,15 +1,15 @@ CREATE TABLE prewhere_int128 (a Int128) ENGINE=MergeTree ORDER BY a; -SELECT a FROM prewhere_int128 WHERE a; -- { serverError 59 } +SELECT a FROM prewhere_int128 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_int128; CREATE TABLE prewhere_int256 (a Int256) ENGINE=MergeTree ORDER BY a; -SELECT a FROM prewhere_int256 WHERE a; -- { serverError 59 } +SELECT a FROM prewhere_int256 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_int256; CREATE TABLE prewhere_uint128 (a UInt128) ENGINE=MergeTree ORDER BY a; -SELECT a FROM prewhere_uint128 WHERE a; -- { serverError 59 } +SELECT a FROM prewhere_uint128 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_uint128; CREATE TABLE prewhere_uint256 (a UInt256) ENGINE=MergeTree ORDER BY a; -SELECT a FROM prewhere_uint256 WHERE a; -- { serverError 59 } +SELECT a FROM prewhere_uint256 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_uint256; From dcb7667808373cf760c2eed6ddba58e119e9d846 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 24 Oct 2022 15:08:51 +0000 Subject: [PATCH 075/239] share transaction for projections --- src/Storages/MergeTree/DataPartStorageOnDisk.cpp | 16 ++++++++++++++-- src/Storages/MergeTree/DataPartStorageOnDisk.h | 3 +++ 2 files changed, 17 insertions(+), 2 deletions(-) diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index a3df17c57de..f1690a2cc59 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -1,4 +1,3 @@ -#include "Storages/MergeTree/IDataPartStorage.h" #include #include #include @@ -31,6 +30,16 @@ DataPartStorageOnDisk::DataPartStorageOnDisk(VolumePtr volume_, std::string root { } +DataPartStorageOnDisk::DataPartStorageOnDisk( + VolumePtr volume_, std::string root_path_, std::string part_dir_, DiskTransactionPtr transaction_) + : volume(std::move(volume_)) + , root_path(std::move(root_path_)) + , part_dir(std::move(part_dir_)) + , transaction(std::move(transaction_)) + , has_shared_transaction(transaction != nullptr) +{ +} + std::string DataPartStorageOnDisk::getFullPath() const { return fs::path(volume->getDisk()->getPath()) / root_path / part_dir / ""; @@ -53,7 +62,7 @@ std::string DataPartStorageOnDisk::getFullRootPath() const MutableDataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) { - return std::make_shared(volume, std::string(fs::path(root_path) / part_dir), name); + return std::shared_ptr(new DataPartStorageOnDisk(volume, std::string(fs::path(root_path) / part_dir), name, transaction)); } DataPartStoragePtr DataPartStorageOnDisk::getProjection(const std::string & name) const @@ -822,6 +831,9 @@ void DataPartStorageOnDisk::commitTransaction() if (!transaction) throw Exception(ErrorCodes::LOGICAL_ERROR, "There is no uncommitted transaction"); + if (has_shared_transaction) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Cannot commit shared transaction"); + transaction->commit(); transaction.reset(); } diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.h b/src/Storages/MergeTree/DataPartStorageOnDisk.h index f93c6a235d6..b5030f484ae 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.h +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.h @@ -145,6 +145,9 @@ private: std::string root_path; std::string part_dir; DiskTransactionPtr transaction; + bool has_shared_transaction = false; + + DataPartStorageOnDisk(VolumePtr volume_, std::string root_path_, std::string part_dir_, DiskTransactionPtr transaction_); template void executeOperation(Op && op); From 0c5011419009f39be9ca7aeec6adca30be8511a6 Mon Sep 17 00:00:00 2001 From: Dan Roscigno Date: Mon, 24 Oct 2022 14:41:38 -0400 Subject: [PATCH 076/239] Update Fedora build info I am using Fedora 36. yum install of `clang-c++` failed for me, `clang` worked. And I had an error message telling me to install `ccache` also. --- docs/en/development/build.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/development/build.md b/docs/en/development/build.md index f397dc0d037..8982a3bc0a4 100644 --- a/docs/en/development/build.md +++ b/docs/en/development/build.md @@ -105,7 +105,7 @@ ninja Example for Fedora Rawhide: ``` bash sudo yum update -yum --nogpg install git cmake make clang-c++ python3 +sudo yum --nogpg install git cmake make clang python3 ccache git clone --recursive https://github.com/ClickHouse/ClickHouse.git mkdir build && cd build cmake ../ClickHouse From f865b28da496efd8d85c77812b2eacb1c17056ee Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 24 Oct 2022 22:38:53 +0000 Subject: [PATCH 077/239] fix drop of MergeTree tables --- src/Storages/MergeTree/DataPartStorageOnDisk.cpp | 2 +- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 2 +- src/Storages/MergeTree/MergeTreeData.cpp | 8 ++++++-- 3 files changed, 8 insertions(+), 4 deletions(-) diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index f1690a2cc59..250ee4792fc 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -705,7 +705,7 @@ void DataPartStorageOnDisk::changeRootPath(const std::string & from_root, const --prefix_size; if (prefix_size > root_path.size() - || std::string_view(from_root).substr(0, prefix_size) != std::string_view(root_path).substr(0, prefix_size)) + || std::string_view(from_root).substr(0, prefix_size) != std::string_view(root_path).substr(0, prefix_size)) throw Exception( ErrorCodes::LOGICAL_ERROR, "Cannot change part root to {} because it is not a prefix of current root {}", diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index 24fabc7cff1..e9b6a28d8a1 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -1562,6 +1562,7 @@ try relative_path = parent_part->getDataPartStorage().getRelativePath(); } + auto old_projection_root_path = getDataPartStorage().getRelativePath(); auto to = fs::path(relative_path) / new_relative_path; metadata_manager->deleteAll(true); @@ -1569,7 +1570,6 @@ try getDataPartStorage().rename(to.parent_path(), to.filename(), storage.log, remove_new_dir_if_exists, fsync_dir); metadata_manager->updateAll(true); - auto old_projection_root_path = getDataPartStorage().getRelativePath(); auto new_projection_root_path = to.string(); for (const auto & [_, part] : projection_parts) diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 04e5432f239..47fcdfa6907 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -2172,7 +2172,12 @@ void MergeTreeData::dropAllData() auto lock = lockParts(); - DataPartsVector all_parts(data_parts_by_info.begin(), data_parts_by_info.end()); + DataPartsVector all_parts; + for (auto it = data_parts_by_info.begin(); it != data_parts_by_info.end(); ++it) + { + modifyPartState(it, DataPartState::Deleting); + all_parts.push_back(*it); + } { std::lock_guard wal_lock(write_ahead_log_mutex); @@ -2185,7 +2190,6 @@ void MergeTreeData::dropAllData() if (!getStorageID().hasUUID()) getContext()->dropCaches(); - /// Removing of each data part before recursive removal of directory is to speed-up removal, because there will be less number of syscalls. NameSet part_names_failed; try From cbfa887e52546a94a852c71d760a2293bba26282 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Mon, 24 Oct 2022 23:51:57 +0000 Subject: [PATCH 078/239] fix merge of projections --- src/Storages/MergeTree/MergeTask.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp index 27a71345a5e..98f05ade18d 100644 --- a/src/Storages/MergeTree/MergeTask.cpp +++ b/src/Storages/MergeTree/MergeTask.cpp @@ -140,9 +140,9 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() local_single_disk_volume, global_ctx->data->relative_data_path, local_tmp_part_basename); - } - data_part_storage->beginTransaction(); + data_part_storage->beginTransaction(); + } if (data_part_storage->exists()) throw Exception("Directory " + data_part_storage->getFullPath() + " already exists", ErrorCodes::DIRECTORY_ALREADY_EXISTS); From dd4121a2906883d23c65e0eeb96bbc0c2c3c338e Mon Sep 17 00:00:00 2001 From: peter279k Date: Tue, 25 Oct 2022 15:27:52 +0800 Subject: [PATCH 079/239] Migration method for installing the deb-packages --- docs/en/getting-started/install.md | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/docs/en/getting-started/install.md b/docs/en/getting-started/install.md index 61303eddab9..e88e9e06a68 100644 --- a/docs/en/getting-started/install.md +++ b/docs/en/getting-started/install.md @@ -128,6 +128,24 @@ clickhouse-client # or "clickhouse-client --password" if you set up a password. +
+Migration Method for installing the deb-packages + +```bash +sudo apt-key del E0C56BD4 +sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv 8919F6BD2B48D754 +echo "deb https://packages.clickhouse.com/deb stable main" | sudo tee \ + /etc/apt/sources.list.d/clickhouse.list +sudo apt-get update + +sudo apt-get install -y clickhouse-server clickhouse-client + +sudo service clickhouse-server start +clickhouse-client # or "clickhouse-client --password" if you set up a password. +``` + +
+ You can replace `stable` with `lts` to use different [release kinds](/docs/en/faq/operations/production.md) based on your needs. You can also download and install packages manually from [here](https://packages.clickhouse.com/deb/pool/main/c/). From e6bfff3de2e719af596cda9265ffebb3c16cd177 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Tue, 25 Oct 2022 12:35:25 +0200 Subject: [PATCH 080/239] Analyzer change setting into allow_experimental_analyzer --- src/Core/Settings.h | 2 +- src/Interpreters/InterpreterDescribeQuery.cpp | 2 +- src/Interpreters/InterpreterExplainQuery.cpp | 4 ++-- src/Interpreters/InterpreterFactory.cpp | 4 ++-- .../Optimizations/reuseStorageOrderingForWindowFunctions.cpp | 2 +- tests/queries/0_stateless/02337_analyzer_columns_basic.sql | 2 +- tests/queries/0_stateless/02338_analyzer_constants_basic.sql | 2 +- tests/queries/0_stateless/02339_analyzer_matcher_basic.sql | 2 +- tests/queries/0_stateless/02340_analyzer_functions.sql | 2 +- tests/queries/0_stateless/02341_analyzer_aliases_basics.sql | 2 +- tests/queries/0_stateless/02342_analyzer_compound_types.sql | 2 +- .../0_stateless/02343_analyzer_column_transformers_strict.sql | 2 +- tests/queries/0_stateless/02343_analyzer_lambdas.sql | 2 +- .../02344_analyzer_multiple_aliases_for_expression.sql | 2 +- tests/queries/0_stateless/02345_analyzer_subqueries.sql | 4 ++-- tests/queries/0_stateless/02366_explain_query_tree.sql | 2 +- .../0_stateless/02367_analyzer_table_alias_columns.sql | 2 +- tests/queries/0_stateless/02368_analyzer_table_functions.sql | 2 +- .../0_stateless/02369_analyzer_array_join_function.sql | 2 +- tests/queries/0_stateless/02370_analyzer_in_function.sql | 2 +- tests/queries/0_stateless/02371_analyzer_join_cross.sql | 2 +- tests/queries/0_stateless/02372_analyzer_join.sql.j2 | 2 +- tests/queries/0_stateless/02373_analyzer_join_use_nulls.sql | 2 +- tests/queries/0_stateless/02374_analyzer_array_join.sql | 2 +- tests/queries/0_stateless/02374_analyzer_join_using.sql.j2 | 2 +- tests/queries/0_stateless/02375_analyzer_union.sql | 2 +- .../0_stateless/02376_analyzer_in_function_subquery.sql | 2 +- tests/queries/0_stateless/02377_analyzer_in_function_set.sql | 2 +- tests/queries/0_stateless/02378_analyzer_projection_names.sql | 2 +- tests/queries/0_stateless/02379_analyzer_subquery_depth.sql | 2 +- tests/queries/0_stateless/02380_analyzer_join_sample.sql | 2 +- tests/queries/0_stateless/02381_analyzer_join_final.sql | 2 +- .../queries/0_stateless/02382_analyzer_matcher_join_using.sql | 2 +- .../0_stateless/02383_analyzer_merge_tree_self_join.sql | 2 +- .../queries/0_stateless/02384_analyzer_dict_get_join_get.sql | 2 +- .../02385_analyzer_aliases_compound_expression.sql | 2 +- .../02386_analyzer_in_function_nested_subqueries.sql | 2 +- tests/queries/0_stateless/02387_analyzer_cte.sql | 2 +- tests/queries/0_stateless/02388_analyzer_recursive_lambda.sql | 2 +- tests/queries/0_stateless/02389_analyzer_nested_lambda.sql | 2 +- 40 files changed, 43 insertions(+), 43 deletions(-) diff --git a/src/Core/Settings.h b/src/Core/Settings.h index bf117163431..148dc56a084 100644 --- a/src/Core/Settings.h +++ b/src/Core/Settings.h @@ -302,7 +302,7 @@ static constexpr UInt64 operator""_GiB(unsigned long long value) M(Float, opentelemetry_start_trace_probability, 0., "Probability to start an OpenTelemetry trace for an incoming query.", 0) \ M(Bool, opentelemetry_trace_processors, false, "Collect OpenTelemetry spans for processors.", 0) \ M(Bool, prefer_column_name_to_alias, false, "Prefer using column names instead of aliases if possible.", 0) \ - M(Bool, use_analyzer, false, "Use analyzer", 0) \ + M(Bool, allow_experimental_analyzer, false, "Allow experimental analyzer", 0) \ M(Bool, prefer_global_in_and_join, false, "If enabled, all IN/JOIN operators will be rewritten as GLOBAL IN/JOIN. It's useful when the to-be-joined tables are only available on the initiator and we need to always scatter their data on-the-fly during distributed processing with the GLOBAL keyword. It's also useful to reduce the need to access the external sources joining external tables.", 0) \ \ \ diff --git a/src/Interpreters/InterpreterDescribeQuery.cpp b/src/Interpreters/InterpreterDescribeQuery.cpp index 60894d331a1..512f9be6fa1 100644 --- a/src/Interpreters/InterpreterDescribeQuery.cpp +++ b/src/Interpreters/InterpreterDescribeQuery.cpp @@ -75,7 +75,7 @@ BlockIO InterpreterDescribeQuery::execute() auto select_query = table_expression.subquery->children.at(0); auto current_context = getContext(); - if (settings.use_analyzer) + if (settings.allow_experimental_analyzer) { SelectQueryOptions select_query_options; names_and_types = InterpreterSelectQueryAnalyzer(select_query, select_query_options, current_context).getSampleBlock().getNamesAndTypesList(); diff --git a/src/Interpreters/InterpreterExplainQuery.cpp b/src/Interpreters/InterpreterExplainQuery.cpp index 6ea1b6a62b3..fb8d3c6049f 100644 --- a/src/Interpreters/InterpreterExplainQuery.cpp +++ b/src/Interpreters/InterpreterExplainQuery.cpp @@ -419,7 +419,7 @@ QueryPipeline InterpreterExplainQuery::executeImpl() auto settings = checkAndGetSettings(ast.getSettings()); QueryPlan plan; - if (getContext()->getSettingsRef().use_analyzer) + if (getContext()->getSettingsRef().allow_experimental_analyzer) { InterpreterSelectQueryAnalyzer interpreter(ast.getExplainedQuery(), options, getContext()); plan = std::move(interpreter).extractQueryPlan(); @@ -462,7 +462,7 @@ QueryPipeline InterpreterExplainQuery::executeImpl() auto settings = checkAndGetSettings(ast.getSettings()); QueryPlan plan; - if (getContext()->getSettingsRef().use_analyzer) + if (getContext()->getSettingsRef().allow_experimental_analyzer) { InterpreterSelectQueryAnalyzer interpreter(ast.getExplainedQuery(), options, getContext()); plan = std::move(interpreter).extractQueryPlan(); diff --git a/src/Interpreters/InterpreterFactory.cpp b/src/Interpreters/InterpreterFactory.cpp index 3e074ec8713..06d5746af59 100644 --- a/src/Interpreters/InterpreterFactory.cpp +++ b/src/Interpreters/InterpreterFactory.cpp @@ -119,7 +119,7 @@ std::unique_ptr InterpreterFactory::get(ASTPtr & query, ContextMut if (query->as()) { - if (context->getSettingsRef().use_analyzer) + if (context->getSettingsRef().allow_experimental_analyzer) return std::make_unique(query, options, context); /// This is internal part of ASTSelectWithUnionQuery. @@ -130,7 +130,7 @@ std::unique_ptr InterpreterFactory::get(ASTPtr & query, ContextMut { ProfileEvents::increment(ProfileEvents::SelectQuery); - if (context->getSettingsRef().use_analyzer) + if (context->getSettingsRef().allow_experimental_analyzer) return std::make_unique(query, options, context); return std::make_unique(query, context, options); diff --git a/src/Processors/QueryPlan/Optimizations/reuseStorageOrderingForWindowFunctions.cpp b/src/Processors/QueryPlan/Optimizations/reuseStorageOrderingForWindowFunctions.cpp index f5e7a438877..c74b5ed915b 100644 --- a/src/Processors/QueryPlan/Optimizations/reuseStorageOrderingForWindowFunctions.cpp +++ b/src/Processors/QueryPlan/Optimizations/reuseStorageOrderingForWindowFunctions.cpp @@ -62,7 +62,7 @@ size_t tryReuseStorageOrderingForWindowFunctions(QueryPlan::Node * parent_node, } auto context = read_from_merge_tree->getContext(); - if (!context->getSettings().optimize_read_in_window_order || context->getSettingsRef().use_analyzer) + if (!context->getSettings().optimize_read_in_window_order || context->getSettingsRef().allow_experimental_analyzer) { return 0; } diff --git a/tests/queries/0_stateless/02337_analyzer_columns_basic.sql b/tests/queries/0_stateless/02337_analyzer_columns_basic.sql index 00dc6c1351e..76f9f8b25e4 100644 --- a/tests/queries/0_stateless/02337_analyzer_columns_basic.sql +++ b/tests/queries/0_stateless/02337_analyzer_columns_basic.sql @@ -1,6 +1,6 @@ -- Tags: no-parallel -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; -- Empty from section diff --git a/tests/queries/0_stateless/02338_analyzer_constants_basic.sql b/tests/queries/0_stateless/02338_analyzer_constants_basic.sql index 56f82763f1b..6d6249538a4 100644 --- a/tests/queries/0_stateless/02338_analyzer_constants_basic.sql +++ b/tests/queries/0_stateless/02338_analyzer_constants_basic.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DESCRIBE (SELECT 1); SELECT 1; diff --git a/tests/queries/0_stateless/02339_analyzer_matcher_basic.sql b/tests/queries/0_stateless/02339_analyzer_matcher_basic.sql index 862cb2f88ff..9d7c486b28a 100644 --- a/tests/queries/0_stateless/02339_analyzer_matcher_basic.sql +++ b/tests/queries/0_stateless/02339_analyzer_matcher_basic.sql @@ -1,6 +1,6 @@ -- Tags: no-parallel -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT 'Matchers without FROM section'; diff --git a/tests/queries/0_stateless/02340_analyzer_functions.sql b/tests/queries/0_stateless/02340_analyzer_functions.sql index b1bfeabd836..101a5bfcc86 100644 --- a/tests/queries/0_stateless/02340_analyzer_functions.sql +++ b/tests/queries/0_stateless/02340_analyzer_functions.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DESCRIBE (SELECT 1 + 1); SELECT 1 + 1; diff --git a/tests/queries/0_stateless/02341_analyzer_aliases_basics.sql b/tests/queries/0_stateless/02341_analyzer_aliases_basics.sql index 636936d875a..52a1cd1dae8 100644 --- a/tests/queries/0_stateless/02341_analyzer_aliases_basics.sql +++ b/tests/queries/0_stateless/02341_analyzer_aliases_basics.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT 'Aliases to constants'; diff --git a/tests/queries/0_stateless/02342_analyzer_compound_types.sql b/tests/queries/0_stateless/02342_analyzer_compound_types.sql index 6b4f323fd66..0fd96928496 100644 --- a/tests/queries/0_stateless/02342_analyzer_compound_types.sql +++ b/tests/queries/0_stateless/02342_analyzer_compound_types.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT 'Constant tuple'; diff --git a/tests/queries/0_stateless/02343_analyzer_column_transformers_strict.sql b/tests/queries/0_stateless/02343_analyzer_column_transformers_strict.sql index 657e2098b4a..98ee7bc8f58 100644 --- a/tests/queries/0_stateless/02343_analyzer_column_transformers_strict.sql +++ b/tests/queries/0_stateless/02343_analyzer_column_transformers_strict.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02343_analyzer_lambdas.sql b/tests/queries/0_stateless/02343_analyzer_lambdas.sql index 1c6200e5eb6..b90f7b32b57 100644 --- a/tests/queries/0_stateless/02343_analyzer_lambdas.sql +++ b/tests/queries/0_stateless/02343_analyzer_lambdas.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02344_analyzer_multiple_aliases_for_expression.sql b/tests/queries/0_stateless/02344_analyzer_multiple_aliases_for_expression.sql index 2569dc8ba9c..cd1bca8285b 100644 --- a/tests/queries/0_stateless/02344_analyzer_multiple_aliases_for_expression.sql +++ b/tests/queries/0_stateless/02344_analyzer_multiple_aliases_for_expression.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02345_analyzer_subqueries.sql b/tests/queries/0_stateless/02345_analyzer_subqueries.sql index 3694028e5d2..c0cc242b57b 100644 --- a/tests/queries/0_stateless/02345_analyzer_subqueries.sql +++ b/tests/queries/0_stateless/02345_analyzer_subqueries.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table @@ -32,7 +32,7 @@ WITH 1 AS global_a SELECT b.a FROM (SELECT global_a AS a) AS b; SELECT * FROM (SELECT * FROM (SELECT * FROM test_table)); SELECT * FROM (SELECT id, value FROM (SELECT * FROM test_table)); -WITH 1 AS a SELECT (SELECT * FROM (SELECT * FROM (SELECT a + 1))) SETTINGS use_analyzer=1; +WITH 1 AS a SELECT (SELECT * FROM (SELECT * FROM (SELECT a + 1))); SELECT 'Subqueries CTE'; diff --git a/tests/queries/0_stateless/02366_explain_query_tree.sql b/tests/queries/0_stateless/02366_explain_query_tree.sql index 23df6d12121..c942f0cac37 100644 --- a/tests/queries/0_stateless/02366_explain_query_tree.sql +++ b/tests/queries/0_stateless/02366_explain_query_tree.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; EXPLAIN QUERY TREE SELECT 1; diff --git a/tests/queries/0_stateless/02367_analyzer_table_alias_columns.sql b/tests/queries/0_stateless/02367_analyzer_table_alias_columns.sql index 769ba769f4d..f41680cd9f4 100644 --- a/tests/queries/0_stateless/02367_analyzer_table_alias_columns.sql +++ b/tests/queries/0_stateless/02367_analyzer_table_alias_columns.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02368_analyzer_table_functions.sql b/tests/queries/0_stateless/02368_analyzer_table_functions.sql index 6d9accef30c..456e095c6c1 100644 --- a/tests/queries/0_stateless/02368_analyzer_table_functions.sql +++ b/tests/queries/0_stateless/02368_analyzer_table_functions.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT c1, c2, c3, c4 FROM format('CSV', '1,2,"[1,2,3]","[[\'abc\'], [], [\'d\', \'e\']]"'); SELECT f.c1, f.c2, f.c3, f.c4 FROM format('CSV', '1,2,"[1,2,3]","[[\'abc\'], [], [\'d\', \'e\']]"') AS f; diff --git a/tests/queries/0_stateless/02369_analyzer_array_join_function.sql b/tests/queries/0_stateless/02369_analyzer_array_join_function.sql index db9884f9cdd..9a9939d2a2f 100644 --- a/tests/queries/0_stateless/02369_analyzer_array_join_function.sql +++ b/tests/queries/0_stateless/02369_analyzer_array_join_function.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT arrayJoin([1, 2, 3]); diff --git a/tests/queries/0_stateless/02370_analyzer_in_function.sql b/tests/queries/0_stateless/02370_analyzer_in_function.sql index 01e5cf498cd..a7128ced449 100644 --- a/tests/queries/0_stateless/02370_analyzer_in_function.sql +++ b/tests/queries/0_stateless/02370_analyzer_in_function.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT 1 IN 1; SELECT 1 IN (1); diff --git a/tests/queries/0_stateless/02371_analyzer_join_cross.sql b/tests/queries/0_stateless/02371_analyzer_join_cross.sql index 0c0709fb951..8261572cdf2 100644 --- a/tests/queries/0_stateless/02371_analyzer_join_cross.sql +++ b/tests/queries/0_stateless/02371_analyzer_join_cross.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02372_analyzer_join.sql.j2 b/tests/queries/0_stateless/02372_analyzer_join.sql.j2 index 970c0d3da17..9b3c212562b 100644 --- a/tests/queries/0_stateless/02372_analyzer_join.sql.j2 +++ b/tests/queries/0_stateless/02372_analyzer_join.sql.j2 @@ -1,6 +1,6 @@ -- Tags: long -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02373_analyzer_join_use_nulls.sql b/tests/queries/0_stateless/02373_analyzer_join_use_nulls.sql index a67a11bf8b0..db7895084e8 100644 --- a/tests/queries/0_stateless/02373_analyzer_join_use_nulls.sql +++ b/tests/queries/0_stateless/02373_analyzer_join_use_nulls.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SET join_use_nulls = 1; DROP TABLE IF EXISTS test_table_join_1; diff --git a/tests/queries/0_stateless/02374_analyzer_array_join.sql b/tests/queries/0_stateless/02374_analyzer_array_join.sql index b7e19bceadf..8ebfdef349c 100644 --- a/tests/queries/0_stateless/02374_analyzer_array_join.sql +++ b/tests/queries/0_stateless/02374_analyzer_array_join.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02374_analyzer_join_using.sql.j2 b/tests/queries/0_stateless/02374_analyzer_join_using.sql.j2 index a8e9cc9eefc..26fb52716ff 100644 --- a/tests/queries/0_stateless/02374_analyzer_join_using.sql.j2 +++ b/tests/queries/0_stateless/02374_analyzer_join_using.sql.j2 @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02375_analyzer_union.sql b/tests/queries/0_stateless/02375_analyzer_union.sql index 4983040bd23..5e41f07d217 100644 --- a/tests/queries/0_stateless/02375_analyzer_union.sql +++ b/tests/queries/0_stateless/02375_analyzer_union.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 0; +SET allow_experimental_analyzer = 0; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02376_analyzer_in_function_subquery.sql b/tests/queries/0_stateless/02376_analyzer_in_function_subquery.sql index 90b2f75b6c6..72a4edb8567 100644 --- a/tests/queries/0_stateless/02376_analyzer_in_function_subquery.sql +++ b/tests/queries/0_stateless/02376_analyzer_in_function_subquery.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02377_analyzer_in_function_set.sql b/tests/queries/0_stateless/02377_analyzer_in_function_set.sql index e5c27e72ea1..e3cbcf75a9c 100644 --- a/tests/queries/0_stateless/02377_analyzer_in_function_set.sql +++ b/tests/queries/0_stateless/02377_analyzer_in_function_set.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02378_analyzer_projection_names.sql b/tests/queries/0_stateless/02378_analyzer_projection_names.sql index a31f5afd8fb..907cc79dcec 100644 --- a/tests/queries/0_stateless/02378_analyzer_projection_names.sql +++ b/tests/queries/0_stateless/02378_analyzer_projection_names.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02379_analyzer_subquery_depth.sql b/tests/queries/0_stateless/02379_analyzer_subquery_depth.sql index 3fea94e1f1c..c2109f543eb 100644 --- a/tests/queries/0_stateless/02379_analyzer_subquery_depth.sql +++ b/tests/queries/0_stateless/02379_analyzer_subquery_depth.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT (SELECT a FROM (SELECT 1 AS a)) SETTINGS max_subquery_depth = 1; -- { serverError 162 } SELECT (SELECT a FROM (SELECT 1 AS a)) SETTINGS max_subquery_depth = 2; diff --git a/tests/queries/0_stateless/02380_analyzer_join_sample.sql b/tests/queries/0_stateless/02380_analyzer_join_sample.sql index 1cf5a4a04df..e417f47d173 100644 --- a/tests/queries/0_stateless/02380_analyzer_join_sample.sql +++ b/tests/queries/0_stateless/02380_analyzer_join_sample.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02381_analyzer_join_final.sql b/tests/queries/0_stateless/02381_analyzer_join_final.sql index 6870b6e5f98..57fc3aedd8f 100644 --- a/tests/queries/0_stateless/02381_analyzer_join_final.sql +++ b/tests/queries/0_stateless/02381_analyzer_join_final.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02382_analyzer_matcher_join_using.sql b/tests/queries/0_stateless/02382_analyzer_matcher_join_using.sql index 8b6d9832086..25d493dc422 100644 --- a/tests/queries/0_stateless/02382_analyzer_matcher_join_using.sql +++ b/tests/queries/0_stateless/02382_analyzer_matcher_join_using.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02383_analyzer_merge_tree_self_join.sql b/tests/queries/0_stateless/02383_analyzer_merge_tree_self_join.sql index adb858c217f..c22a0f4244b 100644 --- a/tests/queries/0_stateless/02383_analyzer_merge_tree_self_join.sql +++ b/tests/queries/0_stateless/02383_analyzer_merge_tree_self_join.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table_join_1; CREATE TABLE test_table_join_1 diff --git a/tests/queries/0_stateless/02384_analyzer_dict_get_join_get.sql b/tests/queries/0_stateless/02384_analyzer_dict_get_join_get.sql index 79aa0703aaa..ff6e417d756 100644 --- a/tests/queries/0_stateless/02384_analyzer_dict_get_join_get.sql +++ b/tests/queries/0_stateless/02384_analyzer_dict_get_join_get.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02385_analyzer_aliases_compound_expression.sql b/tests/queries/0_stateless/02385_analyzer_aliases_compound_expression.sql index ec2155d241b..1a195bbfffe 100644 --- a/tests/queries/0_stateless/02385_analyzer_aliases_compound_expression.sql +++ b/tests/queries/0_stateless/02385_analyzer_aliases_compound_expression.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT cast(tuple(1, 'Value'), 'Tuple(first UInt64, second String)') AS value, value.first, value.second; diff --git a/tests/queries/0_stateless/02386_analyzer_in_function_nested_subqueries.sql b/tests/queries/0_stateless/02386_analyzer_in_function_nested_subqueries.sql index bffbc46809a..c8ca3ff21d4 100644 --- a/tests/queries/0_stateless/02386_analyzer_in_function_nested_subqueries.sql +++ b/tests/queries/0_stateless/02386_analyzer_in_function_nested_subqueries.sql @@ -1,3 +1,3 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; SELECT (NULL IN (SELECT 9223372036854775806 IN (SELECT 65536), inf, NULL IN (NULL))) IN (SELECT NULL IN (NULL)); diff --git a/tests/queries/0_stateless/02387_analyzer_cte.sql b/tests/queries/0_stateless/02387_analyzer_cte.sql index 725e18af315..1f10ac10438 100644 --- a/tests/queries/0_stateless/02387_analyzer_cte.sql +++ b/tests/queries/0_stateless/02387_analyzer_cte.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; DROP TABLE IF EXISTS test_table; CREATE TABLE test_table diff --git a/tests/queries/0_stateless/02388_analyzer_recursive_lambda.sql b/tests/queries/0_stateless/02388_analyzer_recursive_lambda.sql index dd342b32d1f..6fc8ff2aae0 100644 --- a/tests/queries/0_stateless/02388_analyzer_recursive_lambda.sql +++ b/tests/queries/0_stateless/02388_analyzer_recursive_lambda.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; WITH x -> plus(lambda(1), x) AS lambda SELECT lambda(1048576); -- { serverError 1 }; diff --git a/tests/queries/0_stateless/02389_analyzer_nested_lambda.sql b/tests/queries/0_stateless/02389_analyzer_nested_lambda.sql index 3be07a35425..8f8b5537da9 100644 --- a/tests/queries/0_stateless/02389_analyzer_nested_lambda.sql +++ b/tests/queries/0_stateless/02389_analyzer_nested_lambda.sql @@ -1,4 +1,4 @@ -SET use_analyzer = 1; +SET allow_experimental_analyzer = 1; -- { echoOn } From a28d1f01fd23e67eef8ab57f245b50f200d96c70 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Tue, 25 Oct 2022 12:54:17 +0200 Subject: [PATCH 081/239] Analyzer IQueryTreeNode remove getName method --- src/Analyzer/ColumnNode.h | 5 -- src/Analyzer/ConstantNode.h | 5 -- src/Analyzer/FunctionNode.cpp | 21 ----- src/Analyzer/FunctionNode.h | 2 - src/Analyzer/IQueryTreeNode.h | 8 -- src/Analyzer/IdentifierNode.h | 5 -- src/Analyzer/InterpolateNode.cpp | 9 --- src/Analyzer/InterpolateNode.h | 2 - src/Analyzer/LambdaNode.cpp | 5 -- src/Analyzer/LambdaNode.h | 2 - src/Analyzer/ListNode.cpp | 18 ----- src/Analyzer/ListNode.h | 2 - src/Analyzer/MatcherNode.cpp | 49 ------------ src/Analyzer/MatcherNode.h | 2 - src/Analyzer/Passes/QueryAnalysisPass.cpp | 2 +- src/Analyzer/QueryNode.cpp | 94 ----------------------- src/Analyzer/QueryNode.h | 2 - src/Analyzer/SortNode.cpp | 32 -------- src/Analyzer/SortNode.h | 3 - src/Analyzer/TableFunctionNode.cpp | 12 --- src/Analyzer/TableFunctionNode.h | 2 - src/Analyzer/TableNode.cpp | 5 -- src/Analyzer/TableNode.h | 2 - src/Analyzer/UnionNode.cpp | 40 ---------- src/Analyzer/UnionNode.h | 2 - src/Analyzer/WindowNode.cpp | 69 ----------------- src/Analyzer/WindowNode.h | 2 - src/Planner/PlannerActionsVisitor.cpp | 11 ++- 28 files changed, 11 insertions(+), 402 deletions(-) diff --git a/src/Analyzer/ColumnNode.h b/src/Analyzer/ColumnNode.h index fbd788ae6fd..e378bc5f3d0 100644 --- a/src/Analyzer/ColumnNode.h +++ b/src/Analyzer/ColumnNode.h @@ -112,11 +112,6 @@ public: return QueryTreeNodeType::COLUMN; } - String getName() const override - { - return column.name; - } - DataTypePtr getResultType() const override { return column.type; diff --git a/src/Analyzer/ConstantNode.h b/src/Analyzer/ConstantNode.h index 29e8cd25532..0a6868b72d4 100644 --- a/src/Analyzer/ConstantNode.h +++ b/src/Analyzer/ConstantNode.h @@ -51,11 +51,6 @@ public: return QueryTreeNodeType::CONSTANT; } - String getName() const override - { - return value_string; - } - DataTypePtr getResultType() const override { return constant_value->getType(); diff --git a/src/Analyzer/FunctionNode.cpp b/src/Analyzer/FunctionNode.cpp index 7468141b3d5..55e097cc5f4 100644 --- a/src/Analyzer/FunctionNode.cpp +++ b/src/Analyzer/FunctionNode.cpp @@ -93,27 +93,6 @@ void FunctionNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state } } -String FunctionNode::getName() const -{ - String name = function_name; - - const auto & parameters = getParameters(); - const auto & parameters_nodes = parameters.getNodes(); - if (!parameters_nodes.empty()) - { - name += '('; - name += parameters.getName(); - name += ')'; - } - - const auto & arguments = getArguments(); - name += '('; - name += arguments.getName(); - name += ')'; - - return name; -} - bool FunctionNode::isEqualImpl(const IQueryTreeNode & rhs) const { const auto & rhs_typed = assert_cast(rhs); diff --git a/src/Analyzer/FunctionNode.h b/src/Analyzer/FunctionNode.h index 18b4c6d445c..cb1bf0d8cc5 100644 --- a/src/Analyzer/FunctionNode.h +++ b/src/Analyzer/FunctionNode.h @@ -203,8 +203,6 @@ public: return result_type; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/IQueryTreeNode.h b/src/Analyzer/IQueryTreeNode.h index 2ba96d27575..c3b067ab005 100644 --- a/src/Analyzer/IQueryTreeNode.h +++ b/src/Analyzer/IQueryTreeNode.h @@ -82,14 +82,6 @@ public: return toString(getNodeType()); } - /** Get name of query tree node that can be used as part of expression. - * TODO: Projection name, expression name must be refactored in better interface. - */ - virtual String getName() const - { - throw Exception(ErrorCodes::UNSUPPORTED_METHOD, "Method getName is not supported for {} query node", getNodeTypeName()); - } - /** Get result type of query tree node that can be used as part of expression. * If node does not support this method exception is thrown. * TODO: Maybe this can be a part of ExpressionQueryTreeNode. diff --git a/src/Analyzer/IdentifierNode.h b/src/Analyzer/IdentifierNode.h index 7a2351f0ece..358511d1f90 100644 --- a/src/Analyzer/IdentifierNode.h +++ b/src/Analyzer/IdentifierNode.h @@ -50,11 +50,6 @@ public: return QueryTreeNodeType::IDENTIFIER; } - String getName() const override - { - return identifier.getFullName(); - } - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/InterpolateNode.cpp b/src/Analyzer/InterpolateNode.cpp index dcc14d6b6d5..c8c61b05853 100644 --- a/src/Analyzer/InterpolateNode.cpp +++ b/src/Analyzer/InterpolateNode.cpp @@ -17,15 +17,6 @@ InterpolateNode::InterpolateNode(QueryTreeNodePtr expression_, QueryTreeNodePtr children[interpolate_expression_child_index] = std::move(interpolate_expression_); } -String InterpolateNode::getName() const -{ - String result = getExpression()->getName(); - result += " AS "; - result += getInterpolateExpression()->getName(); - - return result; -} - void InterpolateNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const { buffer << std::string(indent, ' ') << "INTERPOLATE id: " << format_state.getNodeId(this); diff --git a/src/Analyzer/InterpolateNode.h b/src/Analyzer/InterpolateNode.h index 5bc8eded0bb..5764ea561c0 100644 --- a/src/Analyzer/InterpolateNode.h +++ b/src/Analyzer/InterpolateNode.h @@ -50,8 +50,6 @@ public: return QueryTreeNodeType::INTERPOLATE; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/LambdaNode.cpp b/src/Analyzer/LambdaNode.cpp index ccc43e75644..809f73072d2 100644 --- a/src/Analyzer/LambdaNode.cpp +++ b/src/Analyzer/LambdaNode.cpp @@ -44,11 +44,6 @@ void LambdaNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, getExpression()->dumpTreeImpl(buffer, format_state, indent + 4); } -String LambdaNode::getName() const -{ - return "lambda(" + children[arguments_child_index]->getName() + ") -> " + children[expression_child_index]->getName(); -} - bool LambdaNode::isEqualImpl(const IQueryTreeNode & rhs) const { const auto & rhs_typed = assert_cast(rhs); diff --git a/src/Analyzer/LambdaNode.h b/src/Analyzer/LambdaNode.h index 6061e854ab0..ad80c645656 100644 --- a/src/Analyzer/LambdaNode.h +++ b/src/Analyzer/LambdaNode.h @@ -84,8 +84,6 @@ public: return QueryTreeNodeType::LAMBDA; } - String getName() const override; - DataTypePtr getResultType() const override { return getExpression()->getResultType(); diff --git a/src/Analyzer/ListNode.cpp b/src/Analyzer/ListNode.cpp index b98e29deecd..7bbb884fa7f 100644 --- a/src/Analyzer/ListNode.cpp +++ b/src/Analyzer/ListNode.cpp @@ -38,24 +38,6 @@ void ListNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, si } } -String ListNode::getName() const -{ - if (children.empty()) - return ""; - - std::string result; - for (const auto & node : children) - { - result += node->getName(); - result += ", "; - } - - result.pop_back(); - result.pop_back(); - - return result; -} - bool ListNode::isEqualImpl(const IQueryTreeNode &) const { /// No state diff --git a/src/Analyzer/ListNode.h b/src/Analyzer/ListNode.h index 7d941731b25..5b328d0f870 100644 --- a/src/Analyzer/ListNode.h +++ b/src/Analyzer/ListNode.h @@ -39,8 +39,6 @@ public: return QueryTreeNodeType::LIST; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/MatcherNode.cpp b/src/Analyzer/MatcherNode.cpp index 0bed9110d25..9d822771087 100644 --- a/src/Analyzer/MatcherNode.cpp +++ b/src/Analyzer/MatcherNode.cpp @@ -146,55 +146,6 @@ void MatcherNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, } } -String MatcherNode::getName() const -{ - WriteBufferFromOwnString buffer; - - if (!qualified_identifier.empty()) - buffer << qualified_identifier.getFullName() << '.'; - - if (matcher_type == MatcherNodeType::ASTERISK) - { - buffer << '*'; - } - else - { - buffer << "COLUMNS("; - - if (columns_matcher) - { - buffer << ' ' << columns_matcher->pattern(); - } - else if (matcher_type == MatcherNodeType::COLUMNS_LIST) - { - size_t columns_identifiers_size = columns_identifiers.size(); - for (size_t i = 0; i < columns_identifiers_size; ++i) - { - buffer << columns_identifiers[i].getFullName(); - - if (i + 1 != columns_identifiers_size) - buffer << ", "; - } - } - } - - buffer << ')'; - - const auto & column_transformers = getColumnTransformers().getNodes(); - size_t column_transformers_size = column_transformers.size(); - - for (size_t i = 0; i < column_transformers_size; ++i) - { - const auto & column_transformer = column_transformers[i]; - buffer << column_transformer->getName(); - - if (i + 1 != column_transformers_size) - buffer << ' '; - } - - return buffer.str(); -} - bool MatcherNode::isEqualImpl(const IQueryTreeNode & rhs) const { const auto & rhs_typed = assert_cast(rhs); diff --git a/src/Analyzer/MatcherNode.h b/src/Analyzer/MatcherNode.h index 3635d038549..e79c1cb4bf2 100644 --- a/src/Analyzer/MatcherNode.h +++ b/src/Analyzer/MatcherNode.h @@ -139,8 +139,6 @@ public: return QueryTreeNodeType::MATCHER; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/Passes/QueryAnalysisPass.cpp b/src/Analyzer/Passes/QueryAnalysisPass.cpp index e55f296e1b7..6b91d0f8053 100644 --- a/src/Analyzer/Passes/QueryAnalysisPass.cpp +++ b/src/Analyzer/Passes/QueryAnalysisPass.cpp @@ -2255,7 +2255,7 @@ QueryTreeNodePtr QueryAnalyzer::tryResolveIdentifierFromJoin(const IdentifierLoo for (auto & join_using_node : join_using_list.getNodes()) { auto & column_node = join_using_node->as(); - join_using_column_name_to_column_node.emplace(column_node.getName(), std::static_pointer_cast(join_using_node)); + join_using_column_name_to_column_node.emplace(column_node.getColumnName(), std::static_pointer_cast(join_using_node)); } } diff --git a/src/Analyzer/QueryNode.cpp b/src/Analyzer/QueryNode.cpp index 378bd7259bf..557baad2654 100644 --- a/src/Analyzer/QueryNode.cpp +++ b/src/Analyzer/QueryNode.cpp @@ -32,100 +32,6 @@ QueryNode::QueryNode() children[limit_by_child_index] = std::make_shared(); } -String QueryNode::getName() const -{ - WriteBufferFromOwnString buffer; - - if (hasWith()) - { - buffer << getWith().getName(); - buffer << ' '; - } - - buffer << "SELECT "; - buffer << getProjection().getName(); - - if (getJoinTree()) - { - buffer << " FROM "; - buffer << getJoinTree()->getName(); - } - - if (getPrewhere()) - { - buffer << " PREWHERE "; - buffer << getPrewhere()->getName(); - } - - if (getWhere()) - { - buffer << " WHERE "; - buffer << getWhere()->getName(); - } - - if (hasGroupBy()) - { - buffer << " GROUP BY "; - buffer << getGroupBy().getName(); - } - - if (hasHaving()) - { - buffer << " HAVING "; - buffer << getHaving()->getName(); - } - - if (hasWindow()) - { - buffer << " WINDOW "; - buffer << getWindow().getName(); - } - - if (hasOrderBy()) - { - buffer << " ORDER BY "; - buffer << getOrderByNode()->getName(); - } - - if (hasInterpolate()) - { - buffer << " INTERPOLATE "; - buffer << getInterpolate()->getName(); - } - - if (hasLimitByLimit()) - { - buffer << "LIMIT "; - buffer << getLimitByLimit()->getName(); - } - - if (hasLimitByOffset()) - { - buffer << "OFFSET "; - buffer << getLimitByOffset()->getName(); - } - - if (hasLimitBy()) - { - buffer << " BY "; - buffer << getLimitBy().getName(); - } - - if (hasLimit()) - { - buffer << " LIMIT "; - buffer << getLimit()->getName(); - } - - if (hasOffset()) - { - buffer << " OFFSET "; - buffer << getOffset()->getName(); - } - - return buffer.str(); -} - void QueryNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const { buffer << std::string(indent, ' ') << "QUERY id: " << format_state.getNodeId(this); diff --git a/src/Analyzer/QueryNode.h b/src/Analyzer/QueryNode.h index 05d393b4212..6bb6613fc2b 100644 --- a/src/Analyzer/QueryNode.h +++ b/src/Analyzer/QueryNode.h @@ -559,8 +559,6 @@ public: return QueryTreeNodeType::QUERY; } - String getName() const override; - DataTypePtr getResultType() const override { if (constant_value) diff --git a/src/Analyzer/SortNode.cpp b/src/Analyzer/SortNode.cpp index f6faccdb7c5..3f91724e9b7 100644 --- a/src/Analyzer/SortNode.cpp +++ b/src/Analyzer/SortNode.cpp @@ -35,38 +35,6 @@ SortNode::SortNode(QueryTreeNodePtr expression_, children[sort_expression_child_index] = std::move(expression_); } -String SortNode::getName() const -{ - String result = getExpression()->getName(); - - if (sort_direction == SortDirection::ASCENDING) - result += " ASC"; - else - result += " DESC"; - - if (nulls_sort_direction) - { - if (*nulls_sort_direction == SortDirection::ASCENDING) - result += " NULLS FIRST"; - else - result += " NULLS LAST"; - } - - if (with_fill) - result += " WITH FILL"; - - if (hasFillFrom()) - result += " FROM " + getFillFrom()->getName(); - - if (hasFillStep()) - result += " STEP " + getFillStep()->getName(); - - if (hasFillTo()) - result += " TO " + getFillTo()->getName(); - - return result; -} - void SortNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const { buffer << std::string(indent, ' ') << "SORT id: " << format_state.getNodeId(this); diff --git a/src/Analyzer/SortNode.h b/src/Analyzer/SortNode.h index b0fe835cf45..04f9fe798e1 100644 --- a/src/Analyzer/SortNode.h +++ b/src/Analyzer/SortNode.h @@ -128,11 +128,8 @@ public: return QueryTreeNodeType::SORT; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; - protected: bool isEqualImpl(const IQueryTreeNode & rhs) const override; diff --git a/src/Analyzer/TableFunctionNode.cpp b/src/Analyzer/TableFunctionNode.cpp index 2056e10ff6f..c8cd05cf685 100644 --- a/src/Analyzer/TableFunctionNode.cpp +++ b/src/Analyzer/TableFunctionNode.cpp @@ -50,18 +50,6 @@ const StorageSnapshotPtr & TableFunctionNode::getStorageSnapshot() const return storage_snapshot; } -String TableFunctionNode::getName() const -{ - String name = table_function_name; - - const auto & arguments = getArguments(); - name += '('; - name += arguments.getName(); - name += ')'; - - return name; -} - void TableFunctionNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const { buffer << std::string(indent, ' ') << "TABLE_FUNCTION id: " << format_state.getNodeId(this); diff --git a/src/Analyzer/TableFunctionNode.h b/src/Analyzer/TableFunctionNode.h index a9f08ed4d1a..292ab740c5b 100644 --- a/src/Analyzer/TableFunctionNode.h +++ b/src/Analyzer/TableFunctionNode.h @@ -127,8 +127,6 @@ public: return QueryTreeNodeType::TABLE_FUNCTION; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/TableNode.cpp b/src/Analyzer/TableNode.cpp index 3075bce238a..0d9a351e9a2 100644 --- a/src/Analyzer/TableNode.cpp +++ b/src/Analyzer/TableNode.cpp @@ -66,11 +66,6 @@ void TableNode::updateTreeHashImpl(HashState & state) const table_expression_modifiers->updateTreeHash(state); } -String TableNode::getName() const -{ - return storage->getStorageID().getFullNameNotQuoted(); -} - QueryTreeNodePtr TableNode::cloneImpl() const { auto result_table_node = std::make_shared(storage, storage_id, storage_lock, storage_snapshot); diff --git a/src/Analyzer/TableNode.h b/src/Analyzer/TableNode.h index f8e897378d6..c7feedd908f 100644 --- a/src/Analyzer/TableNode.h +++ b/src/Analyzer/TableNode.h @@ -76,8 +76,6 @@ public: return QueryTreeNodeType::TABLE; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Analyzer/UnionNode.cpp b/src/Analyzer/UnionNode.cpp index fe913373981..b8ed46c645e 100644 --- a/src/Analyzer/UnionNode.cpp +++ b/src/Analyzer/UnionNode.cpp @@ -79,46 +79,6 @@ NamesAndTypes UnionNode::computeProjectionColumns() const return result_columns; } -String UnionNode::getName() const -{ - WriteBufferFromOwnString buffer; - - auto query_nodes = getQueries().getNodes(); - size_t query_nodes_size = query_nodes.size(); - - for (size_t i = 0; i < query_nodes_size; ++i) - { - const auto & query_node = query_nodes[i]; - buffer << query_node->getName(); - - if (i == 0) - continue; - - auto query_union_mode = union_modes.at(i - 1); - - if (query_union_mode == SelectUnionMode::UNION_DEFAULT) - buffer << "UNION"; - else if (query_union_mode == SelectUnionMode::UNION_ALL) - buffer << "UNION ALL"; - else if (query_union_mode == SelectUnionMode::UNION_DISTINCT) - buffer << "UNION DISTINCT"; - else if (query_union_mode == SelectUnionMode::EXCEPT_DEFAULT) - buffer << "EXCEPT"; - else if (query_union_mode == SelectUnionMode::EXCEPT_ALL) - buffer << "EXCEPT ALL"; - else if (query_union_mode == SelectUnionMode::EXCEPT_DISTINCT) - buffer << "EXCEPT DISTINCT"; - else if (query_union_mode == SelectUnionMode::INTERSECT_DEFAULT) - buffer << "INTERSECT"; - else if (query_union_mode == SelectUnionMode::INTERSECT_ALL) - buffer << "INTERSECT ALL"; - else if (query_union_mode == SelectUnionMode::INTERSECT_DISTINCT) - buffer << "INTERSECT DISTINCT"; - } - - return buffer.str(); -} - void UnionNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const { buffer << std::string(indent, ' ') << "UNION id: " << format_state.getNodeId(this); diff --git a/src/Analyzer/UnionNode.h b/src/Analyzer/UnionNode.h index 9c502c8fe5b..05e70b87a27 100644 --- a/src/Analyzer/UnionNode.h +++ b/src/Analyzer/UnionNode.h @@ -154,8 +154,6 @@ public: return QueryTreeNodeType::UNION; } - String getName() const override; - DataTypePtr getResultType() const override { if (constant_value) diff --git a/src/Analyzer/WindowNode.cpp b/src/Analyzer/WindowNode.cpp index ccf49bc37d8..3e8537302e5 100644 --- a/src/Analyzer/WindowNode.cpp +++ b/src/Analyzer/WindowNode.cpp @@ -18,75 +18,6 @@ WindowNode::WindowNode(WindowFrame window_frame_) children[order_by_child_index] = std::make_shared(); } -String WindowNode::getName() const -{ - String result; - - if (hasPartitionBy()) - { - result += "PARTITION BY"; - result += getPartitionBy().getName(); - } - - if (hasOrderBy()) - { - result += "ORDER BY"; - result += getOrderBy().getName(); - } - - if (!window_frame.is_default) - { - if (hasPartitionBy() || hasOrderBy()) - result += ' '; - - if (window_frame.type == WindowFrame::FrameType::ROWS) - result += "ROWS"; - else if (window_frame.type == WindowFrame::FrameType::GROUPS) - result += "GROUPS"; - else if (window_frame.type == WindowFrame::FrameType::RANGE) - result += "RANGE"; - - result += " BETWEEN "; - if (window_frame.begin_type == WindowFrame::BoundaryType::Current) - { - result += "CURRENT ROW"; - } - else if (window_frame.begin_type == WindowFrame::BoundaryType::Unbounded) - { - result += "UNBOUNDED"; - result += " "; - result += (window_frame.begin_preceding ? "PRECEDING" : "FOLLOWING"); - } - else - { - result += getFrameBeginOffsetNode()->getName(); - result += " "; - result += (window_frame.begin_preceding ? "PRECEDING" : "FOLLOWING"); - } - - result += " AND "; - - if (window_frame.end_type == WindowFrame::BoundaryType::Current) - { - result += "CURRENT ROW"; - } - else if (window_frame.end_type == WindowFrame::BoundaryType::Unbounded) - { - result += "UNBOUNDED"; - result += " "; - result += (window_frame.end_preceding ? "PRECEDING" : "FOLLOWING"); - } - else - { - result += getFrameEndOffsetNode()->getName(); - result += " "; - result += (window_frame.end_preceding ? "PRECEDING" : "FOLLOWING"); - } - } - - return result; -} - void WindowNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const { buffer << std::string(indent, ' ') << "WINDOW id: " << format_state.getNodeId(this); diff --git a/src/Analyzer/WindowNode.h b/src/Analyzer/WindowNode.h index 2f99524eebd..9dfb3e6ef2a 100644 --- a/src/Analyzer/WindowNode.h +++ b/src/Analyzer/WindowNode.h @@ -166,8 +166,6 @@ public: return QueryTreeNodeType::WINDOW; } - String getName() const override; - void dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, size_t indent) const override; protected: diff --git a/src/Planner/PlannerActionsVisitor.cpp b/src/Planner/PlannerActionsVisitor.cpp index 7bc53127101..a6f1a74f251 100644 --- a/src/Planner/PlannerActionsVisitor.cpp +++ b/src/Planner/PlannerActionsVisitor.cpp @@ -502,7 +502,16 @@ String calculateActionNodeName(const QueryTreeNodePtr & node, const PlannerConte case QueryTreeNodeType::COLUMN: { const auto * column_identifier = planner_context.getColumnNodeIdentifierOrNull(node); - result = column_identifier ? *column_identifier : node->getName(); + + if (column_identifier) + { + result = *column_identifier; + } + else + { + const auto & column_node = node->as(); + result = column_node.getColumnName(); + } break; } From af0c54e0557ff0867c01806b4999d038d6b6e9a4 Mon Sep 17 00:00:00 2001 From: Antonio Andelic Date: Tue, 25 Oct 2022 11:13:41 +0000 Subject: [PATCH 082/239] Small fix --- src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp | 4 ++-- tests/queries/0_stateless/02473_prewhere_with_bigint.sql | 9 +++++++++ 2 files changed, 11 insertions(+), 2 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp b/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp index 710b8aef50c..93b314cfd3f 100644 --- a/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp +++ b/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp @@ -633,8 +633,8 @@ Block MergeTreeBaseSelectProcessor::transformHeader( else if (which.isFloat()) prewhere_column.column = prewhere_column.type->createColumnConst(block.rows(), 1.0f)->convertToFullColumnIfConst(); else - throw Exception("Illegal type " + prewhere_column.type->getName() + " of column for filter", - ErrorCodes::ILLEGAL_TYPE_OF_COLUMN_FOR_FILTER); + throw Exception( + ErrorCodes::ILLEGAL_TYPE_OF_COLUMN_FOR_FILTER, "Illegal type {} of column for filter", prewhere_column.type->getName()); } } diff --git a/tests/queries/0_stateless/02473_prewhere_with_bigint.sql b/tests/queries/0_stateless/02473_prewhere_with_bigint.sql index f11a1a279de..29c6f0da2a1 100644 --- a/tests/queries/0_stateless/02473_prewhere_with_bigint.sql +++ b/tests/queries/0_stateless/02473_prewhere_with_bigint.sql @@ -1,15 +1,24 @@ +DROP TABLE IF EXISTS prewhere_int128; +DROP TABLE IF EXISTS prewhere_int256; +DROP TABLE IF EXISTS prewhere_uint128; +DROP TABLE IF EXISTS prewhere_uint256; + CREATE TABLE prewhere_int128 (a Int128) ENGINE=MergeTree ORDER BY a; +INSERT INTO prewhere_int128 VALUES (1); SELECT a FROM prewhere_int128 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_int128; CREATE TABLE prewhere_int256 (a Int256) ENGINE=MergeTree ORDER BY a; +INSERT INTO prewhere_int256 VALUES (1); SELECT a FROM prewhere_int256 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_int256; CREATE TABLE prewhere_uint128 (a UInt128) ENGINE=MergeTree ORDER BY a; +INSERT INTO prewhere_uint128 VALUES (1); SELECT a FROM prewhere_uint128 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_uint128; CREATE TABLE prewhere_uint256 (a UInt256) ENGINE=MergeTree ORDER BY a; +INSERT INTO prewhere_uint256 VALUES (1); SELECT a FROM prewhere_uint256 PREWHERE a; -- { serverError 59 } DROP TABLE prewhere_uint256; From e124920d02c9e591207e86b59b3269fb4a471012 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Tue, 25 Oct 2022 12:08:51 +0000 Subject: [PATCH 083/239] Remove flaky queries from 02457_datediff_via_unix_epoch test --- .../0_stateless/02457_datediff_via_unix_epoch.reference | 4 ---- tests/queries/0_stateless/02457_datediff_via_unix_epoch.sql | 5 ----- 2 files changed, 9 deletions(-) diff --git a/tests/queries/0_stateless/02457_datediff_via_unix_epoch.reference b/tests/queries/0_stateless/02457_datediff_via_unix_epoch.reference index ba12c868037..8b2de7a3eec 100644 --- a/tests/queries/0_stateless/02457_datediff_via_unix_epoch.reference +++ b/tests/queries/0_stateless/02457_datediff_via_unix_epoch.reference @@ -8,9 +8,5 @@ week 1 week 1 day 11 day 11 -hour 264 -hour 264 minute 1440 -minute 20 second 86400 -second 1200 diff --git a/tests/queries/0_stateless/02457_datediff_via_unix_epoch.sql b/tests/queries/0_stateless/02457_datediff_via_unix_epoch.sql index 796b4cc6e8f..ce977712943 100644 --- a/tests/queries/0_stateless/02457_datediff_via_unix_epoch.sql +++ b/tests/queries/0_stateless/02457_datediff_via_unix_epoch.sql @@ -13,11 +13,6 @@ select 'week', date_diff('week', toDateTime64('1969-12-25 10:00:00.000', 3), toD select 'day', date_diff('day', toDate32('1969-12-25'), toDate32('1970-01-05')); select 'day', date_diff('day', toDateTime64('1969-12-25 10:00:00.000', 3), toDateTime64('1970-01-05 10:00:00.000', 3)); -select 'hour', date_diff('hour', toDate32('1969-12-25'), toDate32('1970-01-05')); -select 'hour', date_diff('hour', toDateTime64('1969-12-25 10:00:00.000', 3), toDateTime64('1970-01-05 10:00:00.000', 3)); - select 'minute', date_diff('minute', toDate32('1969-12-31'), toDate32('1970-01-01')); -select 'minute', date_diff('minute', toDateTime64('1969-12-31 23:50:00.000', 3), toDateTime64('1970-01-01 00:10:00.000', 3)); select 'second', date_diff('second', toDate32('1969-12-31'), toDate32('1970-01-01')); -select 'second', date_diff('second', toDateTime64('1969-12-31 23:50:00.000', 3), toDateTime64('1970-01-01 00:10:00.000', 3)); From 9c9b8f4400bf88c26d2cc8ed0d50e4ef1bff533e Mon Sep 17 00:00:00 2001 From: Jordi Villar Date: Tue, 25 Oct 2022 15:23:28 +0200 Subject: [PATCH 084/239] Minor fix iotest_nonblock build --- utils/iotest/iotest_nonblock.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/utils/iotest/iotest_nonblock.cpp b/utils/iotest/iotest_nonblock.cpp index 33fab4d04e6..6db00045e03 100644 --- a/utils/iotest/iotest_nonblock.cpp +++ b/utils/iotest/iotest_nonblock.cpp @@ -101,7 +101,7 @@ int mainImpl(int argc, char ** argv) size_t ops = 0; while (ops < count) { - if (poll(polls.data(), descriptors, -1) <= 0) + if (poll(polls.data(), static_cast(descriptors), -1) <= 0) throwFromErrno("poll failed", ErrorCodes::SYSTEM_ERROR); for (size_t i = 0; i < descriptors; ++i) { From 40ac6fda7420073c1ec717e0ebab199195c98c78 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Tue, 25 Oct 2022 13:43:14 +0000 Subject: [PATCH 085/239] Fixing test. --- ...461_mullable_pk_monotonicity_bug.reference | 32 +++++++++---------- .../02461_mullable_pk_monotonicity_bug.sql | 18 +++++------ 2 files changed, 25 insertions(+), 25 deletions(-) diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference index 29da9bc651b..c0d3de1806a 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.reference @@ -46,19 +46,19 @@ 2 1 2 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 -1970-01-01 00:00:01 -1970-01-01 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 +2022-02-02 00:00:01 +2022-02-02 00:00:02 diff --git a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql index af2d5b859b3..75c8cb2b7e7 100644 --- a/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql +++ b/tests/queries/0_stateless/02461_mullable_pk_monotonicity_bug.sql @@ -47,16 +47,16 @@ SELECT x + 1 FROM tab WHERE (x + 1::LowCardinality(UInt8)) <= -92233720368547758 drop table tab; create table tab (x DateTime) engine = MergeTree order by x settings allow_nullable_key = 1, index_granularity = 2; -insert into tab select number from numbers(4); +insert into tab select toDateTime('2022-02-02') + number from numbers(4); set max_rows_to_read = 2; -SELECT x + 1 FROM tab where plus(x, 1) <= 2 order by x; -SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= 2 order by x; -SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= 2 order by x; -SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(Nullable(UInt8))) <= 2 order by x; -SELECT 1 + x FROM tab where plus(1, x) <= 2 order by x; -SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= 2 order by x; -SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= 2 order by x; -SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1) <= toDateTime('2022-02-02') + 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::Nullable(UInt8)) <= toDateTime('2022-02-02') + 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(UInt8)) <= toDateTime('2022-02-02') + 2 order by x; +SELECT x + 1 FROM tab where plus(x, 1::LowCardinality(Nullable(UInt8))) <= toDateTime('2022-02-02') + 2 order by x; +SELECT 1 + x FROM tab where plus(1, x) <= toDateTime('2022-02-02') + 2 order by x; +SELECT 1 + x FROM tab where plus(1::Nullable(UInt8), x) <= toDateTime('2022-02-02') + 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(UInt8), x) <= toDateTime('2022-02-02') + 2 order by x; +SELECT 1 + x FROM tab where plus(1::LowCardinality(Nullable(UInt8)), x) <= toDateTime('2022-02-02') + 2 order by x; SELECT x + 1 FROM tab WHERE (x + CAST('1', 'Nullable(UInt8)')) <= -2147483647 ORDER BY x ASC NULLS FIRST; From bfb5e6eae4c9a97fa7bffdb5445c33bad97e611d Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Tue, 25 Oct 2022 13:58:29 +0000 Subject: [PATCH 086/239] Fix logical error for invalid prewhere type in StorageMerge. --- src/Storages/StorageMerge.cpp | 10 ++++++++-- .../0_stateless/01902_table_function_merge_db_repr.sql | 2 +- 2 files changed, 9 insertions(+), 3 deletions(-) diff --git a/src/Storages/StorageMerge.cpp b/src/Storages/StorageMerge.cpp index 9891340a0d0..c68e9103704 100644 --- a/src/Storages/StorageMerge.cpp +++ b/src/Storages/StorageMerge.cpp @@ -520,6 +520,8 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( modified_select.setFinal(); } + modified_select.replaceDatabaseAndTable(database_name, table_name); + auto storage_stage = storage->getQueryProcessingStage(modified_context, QueryProcessingStage::Complete, storage_snapshot, modified_query_info); if (processed_stage <= storage_stage) @@ -545,6 +547,12 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( } else { + { + /// Analyze query to check that types are valid (e.g. in PREWHERE). + InterpreterSelectQuery interpreter + (modified_query_info.query, modified_context, SelectQueryOptions(processed_stage).ignoreProjections()); + } + storage->read( plan, real_column_names, @@ -569,8 +577,6 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( } else if (processed_stage > storage_stage) { - modified_select.replaceDatabaseAndTable(database_name, table_name); - /// Maximum permissible parallelism is streams_num modified_context->setSetting("max_threads", streams_num); modified_context->setSetting("max_streams_to_max_threads_ratio", 1); diff --git a/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql b/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql index 460ce16ccad..ee6f052d694 100644 --- a/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql +++ b/tests/queries/0_stateless/01902_table_function_merge_db_repr.sql @@ -65,7 +65,7 @@ SELECT _database, _table, n FROM merge(currentDatabase(), '^t') ORDER BY _databa --fuzzed LOGICAL_ERROR CREATE TABLE 01902_db.t4 (n Date) ENGINE=MergeTree ORDER BY n; INSERT INTO 01902_db.t4 SELECT * FROM numbers(10); -SELECT NULL FROM 01902_db.t_merge WHERE n ORDER BY _table DESC; +SELECT NULL FROM 01902_db.t_merge WHERE n ORDER BY _table DESC; -- {serverError ILLEGAL_TYPE_OF_COLUMN_FOR_FILTER} DROP DATABASE 01902_db; DROP DATABASE 01902_db1; From 728342ec7c57b8567aa4b93cce8bda7f03edd32a Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Tue, 25 Oct 2022 21:41:08 +0000 Subject: [PATCH 087/239] Un-inline stuff --- .../Algorithms/AggregatingSortedAlgorithm.cpp | 33 +++++++++++++++++++ .../Algorithms/AggregatingSortedAlgorithm.h | 31 +++-------------- 2 files changed, 38 insertions(+), 26 deletions(-) diff --git a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp index 7a0847bcbc5..db08f3ffbd3 100644 --- a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp +++ b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.cpp @@ -125,6 +125,39 @@ static void postprocessChunk(Chunk & chunk, const AggregatingSortedAlgorithm::Co } +AggregatingSortedAlgorithm::SimpleAggregateDescription::SimpleAggregateDescription( + AggregateFunctionPtr function_, const size_t column_number_, + DataTypePtr nested_type_, DataTypePtr real_type_) + : function(std::move(function_)), column_number(column_number_) + , nested_type(std::move(nested_type_)), real_type(std::move(real_type_)) +{ + add_function = function->getAddressOfAddFunction(); + state.reset(function->sizeOfData(), function->alignOfData()); +} + +void AggregatingSortedAlgorithm::SimpleAggregateDescription::createState() +{ + if (created) + return; + function->create(state.data()); + created = true; +} + +void AggregatingSortedAlgorithm::SimpleAggregateDescription::destroyState() +{ + if (!created) + return; + function->destroy(state.data()); + created = false; +} + +/// Explicitly destroy aggregation state if the stream is terminated +AggregatingSortedAlgorithm::SimpleAggregateDescription::~SimpleAggregateDescription() +{ + destroyState(); +} + + AggregatingSortedAlgorithm::AggregatingMergedData::AggregatingMergedData( MutableColumns columns_, UInt64 max_block_size_, ColumnsDefinition & def_) : MergedData(std::move(columns_), false, max_block_size_), def(def_) diff --git a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h index 0ffbd1262d3..d670242ed81 100644 --- a/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h +++ b/src/Processors/Merges/Algorithms/AggregatingSortedAlgorithm.h @@ -43,36 +43,15 @@ public: bool created = false; SimpleAggregateDescription( - AggregateFunctionPtr function_, const size_t column_number_, - DataTypePtr nested_type_, DataTypePtr real_type_) - : function(std::move(function_)), column_number(column_number_) - , nested_type(std::move(nested_type_)), real_type(std::move(real_type_)) - { - add_function = function->getAddressOfAddFunction(); - state.reset(function->sizeOfData(), function->alignOfData()); - } + AggregateFunctionPtr function_, const size_t column_number_, + DataTypePtr nested_type_, DataTypePtr real_type_); - void createState() - { - if (created) - return; - function->create(state.data()); - created = true; - } + void createState(); - void destroyState() - { - if (!created) - return; - function->destroy(state.data()); - created = false; - } + void destroyState(); /// Explicitly destroy aggregation state if the stream is terminated - ~SimpleAggregateDescription() - { - destroyState(); - } + ~SimpleAggregateDescription(); SimpleAggregateDescription() = default; SimpleAggregateDescription(SimpleAggregateDescription &&) = default; From c4d4f2dbedde1802a880fa1d2783ffd8fca20f2c Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 25 Oct 2022 22:14:06 +0000 Subject: [PATCH 088/239] better interface --- src/Disks/IStoragePolicy.cpp | 29 +++++++++++++++++++ src/Disks/IStoragePolicy.h | 10 +++++-- src/Disks/StoragePolicy.cpp | 16 ++-------- src/Disks/StoragePolicy.h | 5 +--- .../MergeTree/DataPartStorageOnDisk.cpp | 25 ++-------------- .../MergeTree/DataPartStorageOnDisk.h | 7 ----- src/Storages/MergeTree/DataPartsExchange.cpp | 1 + src/Storages/MergeTree/IDataPartStorage.h | 9 +----- src/Storages/MergeTree/IMergeTreeDataPart.cpp | 3 +- .../MergeTree/MergeFromLogEntryTask.cpp | 4 ++- src/Storages/MergeTree/MergeTreeData.cpp | 2 +- .../MergeTree/MergeTreePartsMover.cpp | 21 ++++++++------ src/Storages/StorageMergeTree.cpp | 4 ++- 13 files changed, 66 insertions(+), 70 deletions(-) diff --git a/src/Disks/IStoragePolicy.cpp b/src/Disks/IStoragePolicy.cpp index 2ba6df4be8f..c843ee11563 100644 --- a/src/Disks/IStoragePolicy.cpp +++ b/src/Disks/IStoragePolicy.cpp @@ -31,4 +31,33 @@ VolumePtr IStoragePolicy::getVolumeByName(const String & volume_name) const return volume; } +size_t IStoragePolicy::getVolumeIndexByDiskName(const String & disk_name) const +{ + auto index = tryGetVolumeIndexByDiskName(disk_name); + if (!index) + throw Exception(ErrorCodes::UNKNOWN_DISK, + "No disk {} in policy {}", backQuote(disk_name), backQuote(getName())); + + return *index; +} + +VolumePtr IStoragePolicy::tryGetVolumeByDiskName(const String & disk_name) const +{ + auto index = tryGetVolumeIndexByDiskName(disk_name); + if (!index) + return nullptr; + + return getVolume(*index); +} + +VolumePtr IStoragePolicy::getVolumeByDiskName(const String & disk_name) const +{ + auto volume = tryGetVolumeByDiskName(disk_name); + if (!volume) + throw Exception(ErrorCodes::UNKNOWN_DISK, + "No disk {} in policy {}", backQuote(disk_name), backQuote(getName())); + + return volume; +} + } diff --git a/src/Disks/IStoragePolicy.h b/src/Disks/IStoragePolicy.h index 8d14a26691b..a6a5fe5f692 100644 --- a/src/Disks/IStoragePolicy.h +++ b/src/Disks/IStoragePolicy.h @@ -4,6 +4,7 @@ #include #include +#include #include namespace DB @@ -55,12 +56,15 @@ public: /// Get volume by index. virtual VolumePtr getVolume(size_t index) const = 0; virtual VolumePtr tryGetVolumeByName(const String & volume_name) const = 0; - virtual VolumePtr tryGetVolumeByDisk(const DiskPtr & disk_ptr) const = 0; VolumePtr getVolumeByName(const String & volume_name) const; /// Checks if storage policy can be replaced by another one. virtual void checkCompatibleWith(const StoragePolicyPtr & new_storage_policy) const = 0; - /// Find volume index, which contains disk - virtual size_t getVolumeIndexByDisk(const DiskPtr & disk_ptr) const = 0; + /// Finds a volume index, which contains disk + virtual std::optional tryGetVolumeIndexByDiskName(const String & disk_name) const = 0; + size_t getVolumeIndexByDiskName(const String & disk_name) const; + /// Finds a volume which contains a specified disk. + VolumePtr tryGetVolumeByDiskName(const String & disk_name) const; + VolumePtr getVolumeByDiskName(const String & disk_name) const; /// Check if we have any volume with stopped merges virtual bool hasAnyVolumeWithDisabledMerges() const = 0; virtual bool containsVolume(const String & volume_name) const = 0; diff --git a/src/Disks/StoragePolicy.cpp b/src/Disks/StoragePolicy.cpp index 10513c6beae..1e5b5825b8c 100644 --- a/src/Disks/StoragePolicy.cpp +++ b/src/Disks/StoragePolicy.cpp @@ -311,22 +311,12 @@ void StoragePolicy::checkCompatibleWith(const StoragePolicyPtr & new_storage_pol } -size_t StoragePolicy::getVolumeIndexByDisk(const DiskPtr & disk_ptr) const +std::optional StoragePolicy::tryGetVolumeIndexByDiskName(const String & disk_name) const { - auto it = volume_index_by_disk_name.find(disk_ptr->getName()); + auto it = volume_index_by_disk_name.find(disk_name); if (it != volume_index_by_disk_name.end()) return it->second; - else - throw Exception("No disk " + backQuote(disk_ptr->getName()) + " in policy " + backQuote(name), ErrorCodes::UNKNOWN_DISK); -} - - -VolumePtr StoragePolicy::tryGetVolumeByDisk(const DiskPtr & disk_ptr) const -{ - auto it = volume_index_by_disk_name.find(disk_ptr->getName()); - if (it == volume_index_by_disk_name.end()) - return nullptr; - return getVolume(it->second); + return {}; } diff --git a/src/Disks/StoragePolicy.h b/src/Disks/StoragePolicy.h index fd0169a6ebe..9631f1c2e52 100644 --- a/src/Disks/StoragePolicy.h +++ b/src/Disks/StoragePolicy.h @@ -68,7 +68,7 @@ public: ReservationPtr reserve(UInt64 bytes, size_t min_volume_index) const override; /// Find volume index, which contains disk - size_t getVolumeIndexByDisk(const DiskPtr & disk_ptr) const override; + std::optional tryGetVolumeIndexByDiskName(const String & disk_name) const override; /// Reserves 0 bytes on disk with max available space /// Do not use this function when it is possible to predict size. @@ -85,9 +85,6 @@ public: VolumePtr tryGetVolumeByName(const String & volume_name) const override; - /// Finds a volume which contains a specified disk. - VolumePtr tryGetVolumeByDisk(const DiskPtr & disk_ptr) const override; - /// Checks if storage policy can be replaced by another one. void checkCompatibleWith(const StoragePolicyPtr & new_storage_policy) const override; diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp index 250ee4792fc..7b36a9873e4 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.cpp +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.cpp @@ -6,7 +6,6 @@ #include #include #include -#include #include #include #include @@ -129,6 +128,7 @@ static UInt64 calculateTotalSizeOnDiskImpl(const DiskPtr & disk, const String & { if (disk->isFile(from)) return disk->getFileSize(from); + std::vector files; disk->listFiles(from, files); UInt64 res = 0; @@ -490,11 +490,6 @@ std::string DataPartStorageOnDisk::getDiskPath() const return volume->getDisk()->getPath(); } -DataPartStorageOnDisk::DisksSet::const_iterator DataPartStorageOnDisk::isStoredOnDisk(const DisksSet & disks) const -{ - return disks.find(volume->getDisk()); -} - ReservationPtr DataPartStorageOnDisk::reserve(UInt64 bytes) const { auto res = volume->reserve(bytes); @@ -509,11 +504,6 @@ ReservationPtr DataPartStorageOnDisk::tryReserve(UInt64 bytes) const return volume->reserve(bytes); } -size_t DataPartStorageOnDisk::getVolumeIndex(const IStoragePolicy & storage_policy) const -{ - return storage_policy.getVolumeIndexByDisk(volume->getDisk()); -} - String DataPartStorageOnDisk::getUniqueId() const { auto disk = volume->getDisk(); @@ -523,16 +513,6 @@ String DataPartStorageOnDisk::getUniqueId() const return disk->getUniqueId(fs::path(getRelativePath()) / "checksums.txt"); } -bool DataPartStorageOnDisk::shallParticipateInMerges(const IStoragePolicy & storage_policy) const -{ - /// `IMergeTreeDataPart::volume` describes space where current part belongs, and holds - /// `SingleDiskVolume` object which does not contain up-to-date settings of corresponding volume. - /// Therefore we shall obtain volume from storage policy. - auto volume_ptr = storage_policy.getVolume(storage_policy.getVolumeIndexByDisk(volume->getDisk())); - - return !volume_ptr->areMergesAvoided(); -} - void DataPartStorageOnDisk::backup( const MergeTreeDataPartChecksums & checksums, const NameSet & files_without_checksums, @@ -821,7 +801,8 @@ void DataPartStorageOnDisk::createProjection(const std::string & name) void DataPartStorageOnDisk::beginTransaction() { if (transaction) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Uncommitted transaction already exists"); + throw Exception(ErrorCodes::LOGICAL_ERROR, + "Uncommitted {}transaction already exists", has_shared_transaction ? "shared " : ""); transaction = volume->getDisk()->createTransaction(); } diff --git a/src/Storages/MergeTree/DataPartStorageOnDisk.h b/src/Storages/MergeTree/DataPartStorageOnDisk.h index b5030f484ae..bea1596e1f7 100644 --- a/src/Storages/MergeTree/DataPartStorageOnDisk.h +++ b/src/Storages/MergeTree/DataPartStorageOnDisk.h @@ -71,17 +71,10 @@ public: UInt64 getRevision() const override; std::unordered_map getSerializedMetadata(const std::vector & paths) const override; std::string getDiskPath() const override; - - DisksSet::const_iterator isStoredOnDisk(const DisksSet & disks) const override; - ReservationPtr reserve(UInt64 bytes) const override; ReservationPtr tryReserve(UInt64 bytes) const override; - size_t getVolumeIndex(const IStoragePolicy &) const override; - String getUniqueId() const override; - bool shallParticipateInMerges(const IStoragePolicy &) const override; - void backup( const MergeTreeDataPartChecksums & checksums, const NameSet & files_without_checksums, diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index 88a745820e9..afef354a2e6 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -152,6 +152,7 @@ void Service::processQuery(const HTMLForm & params, ReadBuffer & /*body*/, Write UInt64 revision = parse(params.get("disk_revision", "0")); if (revision) part->getDataPartStorage().syncRevision(revision); + revision = part->getDataPartStorage().getRevision(); if (revision) response.addCookie({"disk_revision", toString(revision)}); diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index 8f478b3c688..0e4ebe68157 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -150,29 +150,22 @@ public: /// TODO: remove or at least remove const. virtual void syncRevision(UInt64 revision) const = 0; virtual UInt64 getRevision() const = 0; + virtual std::unordered_map getSerializedMetadata(const std::vector & paths) const = 0; /// Get a path for internal disk if relevant. It is used mainly for logging. virtual std::string getDiskPath() const = 0; - /// Check if data part is stored on one of the specified disk in set. - using DisksSet = std::unordered_set; - virtual DisksSet::const_iterator isStoredOnDisk(const DisksSet & disks) const { return disks.end(); } - /// Reserve space on the same disk. /// Probably we should try to remove it later. /// TODO: remove constness virtual ReservationPtr reserve(UInt64 /*bytes*/) const { return nullptr; } virtual ReservationPtr tryReserve(UInt64 /*bytes*/) const { return nullptr; } - virtual size_t getVolumeIndex(const IStoragePolicy &) const { return 0; } /// A leak of abstraction. /// Return some uniq string for file. /// Required for distinguish different copies of the same part on remote FS. virtual String getUniqueId() const = 0; - /// A leak of abstraction - virtual bool shallParticipateInMerges(const IStoragePolicy &) const { return true; } - /// Create a backup of a data part. /// This method adds a new entry to backup_entries. /// Also creates a new tmp_dir for internal disk (if disk is mentioned the first time). diff --git a/src/Storages/MergeTree/IMergeTreeDataPart.cpp b/src/Storages/MergeTree/IMergeTreeDataPart.cpp index e9b6a28d8a1..368af55aa15 100644 --- a/src/Storages/MergeTree/IMergeTreeDataPart.cpp +++ b/src/Storages/MergeTree/IMergeTreeDataPart.cpp @@ -1545,7 +1545,8 @@ void IMergeTreeDataPart::appendFilesOfColumns(Strings & files) bool IMergeTreeDataPart::shallParticipateInMerges(const StoragePolicyPtr & storage_policy) const { - return getDataPartStorage().shallParticipateInMerges(*storage_policy); + auto disk_name = getDataPartStorage().getDiskName(); + return !storage_policy->getVolumeByDiskName(disk_name)->areMergesAvoided(); } void IMergeTreeDataPart::renameTo(const String & new_relative_path, bool remove_new_dir_if_exists) diff --git a/src/Storages/MergeTree/MergeFromLogEntryTask.cpp b/src/Storages/MergeTree/MergeFromLogEntryTask.cpp index f3b81a4793e..9a9b8a4a6bb 100644 --- a/src/Storages/MergeTree/MergeFromLogEntryTask.cpp +++ b/src/Storages/MergeTree/MergeFromLogEntryTask.cpp @@ -160,7 +160,9 @@ ReplicatedMergeMutateTaskBase::PrepareResult MergeFromLogEntryTask::prepare() for (auto & part_ptr : parts) { ttl_infos.update(part_ptr->ttl_infos); - max_volume_index = std::max(max_volume_index, part_ptr->getDataPartStorage().getVolumeIndex(*storage.getStoragePolicy())); + auto disk_name = part_ptr->getDataPartStorage().getDiskName(); + size_t volume_index = storage.getStoragePolicy()->getVolumeIndexByDiskName(disk_name); + max_volume_index = std::max(max_volume_index, volume_index); } /// It will live until the whole task is being destroyed diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index 9dc8e147e76..bb589161b57 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -5059,7 +5059,7 @@ bool MergeTreeData::shouldPerformTTLMoveOnInsert(const SpacePtr & move_destinati if (move_destination->isDisk()) { auto disk = std::static_pointer_cast(move_destination); - if (auto volume = getStoragePolicy()->tryGetVolumeByDisk(disk)) + if (auto volume = getStoragePolicy()->tryGetVolumeByDiskName(disk->getName())) return volume->perform_ttl_move_on_insert; } return false; diff --git a/src/Storages/MergeTree/MergeTreePartsMover.cpp b/src/Storages/MergeTree/MergeTreePartsMover.cpp index bd277ca4374..b618b068769 100644 --- a/src/Storages/MergeTree/MergeTreePartsMover.cpp +++ b/src/Storages/MergeTree/MergeTreePartsMover.cpp @@ -100,7 +100,6 @@ bool MergeTreePartsMover::selectPartsForMove( return false; std::unordered_map need_to_move; - std::unordered_set need_to_move_disks; const auto policy = data->getStoragePolicy(); const auto & volumes = policy->getVolumes(); @@ -115,10 +114,7 @@ bool MergeTreePartsMover::selectPartsForMove( UInt64 unreserved_space = disk->getUnreservedSpace(); if (unreserved_space < required_maximum_available_space && !disk->isBroken()) - { need_to_move.emplace(disk, required_maximum_available_space - unreserved_space); - need_to_move_disks.emplace(disk); - } } } } @@ -140,8 +136,16 @@ bool MergeTreePartsMover::selectPartsForMove( auto ttl_entry = selectTTLDescriptionForTTLInfos(metadata_snapshot->getMoveTTLs(), part->ttl_infos.moves_ttl, time_of_move, true); auto to_insert = need_to_move.end(); - if (auto disk_it = part->getDataPartStorage().isStoredOnDisk(need_to_move_disks); disk_it != need_to_move_disks.end()) - to_insert = need_to_move.find(*disk_it); + auto part_disk_name = part->getDataPartStorage().getDiskName(); + + for (auto it = need_to_move.begin(); it != need_to_move.end(); ++it) + { + if (it->first->getName() == part_disk_name) + { + to_insert = it; + break; + } + } ReservationPtr reservation; if (ttl_entry) @@ -158,9 +162,8 @@ bool MergeTreePartsMover::selectPartsForMove( /// In order to not over-move, we need to "release" required space on this disk, /// possibly to zero. if (to_insert != need_to_move.end()) - { to_insert->second.decreaseRequiredSizeAndRemoveRedundantParts(part->getBytesOnDisk()); - } + ++parts_to_move_by_ttl_rules; parts_to_move_total_size_bytes += part->getBytesOnDisk(); } @@ -173,7 +176,7 @@ bool MergeTreePartsMover::selectPartsForMove( for (auto && move : need_to_move) { - auto min_volume_index = policy->getVolumeIndexByDisk(move.first) + 1; + auto min_volume_index = policy->getVolumeIndexByDiskName(move.first->getName()) + 1; for (auto && part : move.second.getAccumulatedParts()) { auto reservation = policy->reserve(part->getBytesOnDisk(), min_volume_index); diff --git a/src/Storages/StorageMergeTree.cpp b/src/Storages/StorageMergeTree.cpp index ee17dca567f..a450a9ef3a9 100644 --- a/src/Storages/StorageMergeTree.cpp +++ b/src/Storages/StorageMergeTree.cpp @@ -389,7 +389,9 @@ CurrentlyMergingPartsTagger::CurrentlyMergingPartsTagger( for (auto & part_ptr : future_part->parts) { ttl_infos.update(part_ptr->ttl_infos); - max_volume_index = std::max(max_volume_index, part_ptr->getDataPartStorage().getVolumeIndex(*storage.getStoragePolicy())); + auto disk_name = part_ptr->getDataPartStorage().getDiskName(); + size_t volume_index = storage.getStoragePolicy()->getVolumeIndexByDiskName(disk_name); + max_volume_index = std::max(max_volume_index, volume_index); } reserved_space = storage.balancedReservation( From d8b09b430f1e956d05d6a60b659b221c45dbbd06 Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 25 Oct 2022 22:27:11 +0000 Subject: [PATCH 089/239] add comments --- src/Storages/MergeTree/IDataPartStorage.h | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index 0e4ebe68157..baad9748b10 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -211,6 +211,9 @@ public: size_t buf_size, const WriteSettings & settings) = 0; + /// A special const method to write transaction file. + /// It's const, because file with transaction metadata + /// can be modifed after part creation. virtual std::unique_ptr writeTransactionFile(WriteMode mode) const = 0; virtual void createFile(const String & name) = 0; @@ -237,7 +240,9 @@ public: bool remove_new_dir_if_exists, bool fsync_part_dir) = 0; + /// Starts a transaction of mutable operations. virtual void beginTransaction() = 0; + /// Commits a transaction of mutable operations. virtual void commitTransaction() = 0; virtual bool hasActiveTransaction() const = 0; }; @@ -245,6 +250,9 @@ public: using DataPartStoragePtr = std::shared_ptr; using MutableDataPartStoragePtr = std::shared_ptr; +/// A holder that encapsulates data part storage and +/// gives access to const storage from const methods +/// and to mutable storage from non-const methods. class DataPartStorageHolder : public boost::noncopyable { public: From 79889198b1356a9d85fe9ca7dd76d2b95894f03a Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Tue, 25 Oct 2022 23:03:37 +0000 Subject: [PATCH 090/239] fix style check --- src/Disks/StoragePolicy.cpp | 1 - src/Storages/MergeTree/IDataPartStorage.h | 2 +- 2 files changed, 1 insertion(+), 2 deletions(-) diff --git a/src/Disks/StoragePolicy.cpp b/src/Disks/StoragePolicy.cpp index 1e5b5825b8c..10524ffcc0f 100644 --- a/src/Disks/StoragePolicy.cpp +++ b/src/Disks/StoragePolicy.cpp @@ -26,7 +26,6 @@ namespace ErrorCodes extern const int BAD_ARGUMENTS; extern const int EXCESSIVE_ELEMENT_IN_CONFIG; extern const int NO_ELEMENTS_IN_CONFIG; - extern const int UNKNOWN_DISK; extern const int UNKNOWN_POLICY; extern const int UNKNOWN_VOLUME; extern const int LOGICAL_ERROR; diff --git a/src/Storages/MergeTree/IDataPartStorage.h b/src/Storages/MergeTree/IDataPartStorage.h index baad9748b10..c6669908db4 100644 --- a/src/Storages/MergeTree/IDataPartStorage.h +++ b/src/Storages/MergeTree/IDataPartStorage.h @@ -213,7 +213,7 @@ public: /// A special const method to write transaction file. /// It's const, because file with transaction metadata - /// can be modifed after part creation. + /// can be modified after part creation. virtual std::unique_ptr writeTransactionFile(WriteMode mode) const = 0; virtual void createFile(const String & name) = 0; From fc223aa6731e15928e94a17f96b87644dc4017ee Mon Sep 17 00:00:00 2001 From: serxa Date: Tue, 25 Oct 2022 23:52:31 +0000 Subject: [PATCH 091/239] replace throttler algorithm by token bucket --- src/Common/Throttler.cpp | 80 ++++++++++++++++++---------------------- src/Common/Throttler.h | 37 ++++++++++--------- 2 files changed, 55 insertions(+), 62 deletions(-) diff --git a/src/Common/Throttler.cpp b/src/Common/Throttler.cpp index 2c9279e21e1..5a0f7e227a6 100644 --- a/src/Common/Throttler.cpp +++ b/src/Common/Throttler.cpp @@ -3,7 +3,6 @@ #include #include #include -#include namespace ProfileEvents { @@ -21,63 +20,56 @@ namespace ErrorCodes /// Just 10^9. static constexpr auto NS = 1000000000UL; -/// Tracking window. Actually the size is not really important. We just want to avoid -/// throttles when there are no actions for a long period time. -static const double window_ns = 1ULL * NS; +static const size_t default_burst_seconds = 10; + +Throttler::Throttler(size_t max_speed_, const std::shared_ptr & parent_) + : max_speed(max_speed_) + , max_burst(max_speed_ * default_burst_seconds) + , limit_exceeded_exception_message("") + , tokens(max_burst) + , parent(parent_) +{} + +Throttler::Throttler(size_t max_speed_, size_t limit_, const char * limit_exceeded_exception_message_, + const std::shared_ptr & parent_) + : max_speed(max_speed_) + , max_burst(max_speed_ * default_burst_seconds) + , limit(limit_) + , limit_exceeded_exception_message(limit_exceeded_exception_message_) + , tokens(max_burst) + , parent(parent_) +{} void Throttler::add(size_t amount) { - size_t new_count; - /// This outer variable is always equal to smoothed_speed. - /// We use to avoid race condition. - double current_speed = 0; - + // Values obtained under lock to be checked after release + size_t count_value; + double tokens_value; { std::lock_guard lock(mutex); - auto now = clock_gettime_ns_adjusted(prev_ns); - /// If prev_ns is equal to zero (first `add` call) we known nothing about speed - /// and don't track anything. if (max_speed && prev_ns != 0) { - /// Time spent to process the amount of bytes - double time_spent = now - prev_ns; - - /// The speed in bytes per second is equal to amount / time_spent in seconds - auto new_speed = amount / (time_spent / NS); - - /// We want to make old values of speed less important for our smoothed value - /// so we decay it's value with coef. - auto decay_coeff = std::pow(0.5, time_spent / window_ns); - - /// Weighted average between previous and new speed - smoothed_speed = smoothed_speed * decay_coeff + (1 - decay_coeff) * new_speed; - current_speed = smoothed_speed; + double delta_seconds = static_cast(now - prev_ns) / NS; + tokens = std::max(tokens + max_speed * delta_seconds - amount, max_burst); } - count += amount; - new_count = count; + count_value = count; + tokens_value = tokens; prev_ns = now; } - if (limit && new_count > limit) + if (limit && count_value > limit) throw Exception(limit_exceeded_exception_message + std::string(" Maximum: ") + toString(limit), ErrorCodes::LIMIT_EXCEEDED); - if (max_speed && current_speed > max_speed) + /// Wait unless there is positive amount of tokens - throttling + if (max_speed && tokens_value < 0) { - /// If we was too fast then we have to sleep until our smoothed speed became <= max_speed - int64_t sleep_time = static_cast(-window_ns * std::log2(max_speed / current_speed)); - - if (sleep_time > 0) - { - accumulated_sleep += sleep_time; - - sleepForNanoseconds(sleep_time); - - accumulated_sleep -= sleep_time; - - ProfileEvents::increment(ProfileEvents::ThrottlerSleepMicroseconds, sleep_time / 1000UL); - } + int64_t sleep_time = static_cast(-tokens_value * max_speed * NS); + accumulated_sleep += sleep_time; + sleepForNanoseconds(sleep_time); + accumulated_sleep -= sleep_time; + ProfileEvents::increment(ProfileEvents::ThrottlerSleepMicroseconds, sleep_time / 1000UL); } if (parent) @@ -89,9 +81,9 @@ void Throttler::reset() std::lock_guard lock(mutex); count = 0; - accumulated_sleep = 0; - smoothed_speed = 0; + tokens = max_burst; prev_ns = 0; + // NOTE: do not zero `accumulated_sleep` to avoid races } bool Throttler::isThrottling() const diff --git a/src/Common/Throttler.h b/src/Common/Throttler.h index 6d44ad6ca5f..a33637783e7 100644 --- a/src/Common/Throttler.h +++ b/src/Common/Throttler.h @@ -10,24 +10,26 @@ namespace DB { -/** Allows you to limit the speed of something (in entities per second) using sleep. - * Specifics of work: - * Tracks exponentially (pow of 1/2) smoothed speed with hardcoded window. - * See more comments in .cpp file. - * - * Also allows you to set a limit on the maximum number of entities. If exceeded, an exception will be thrown. +/** Allows you to limit the speed of something (in tokens per second) using sleep. + * Implemented using Token Bucket Throttling algorithm. + * Also allows you to set a limit on the maximum number of tokens. If exceeded, an exception will be thrown. */ class Throttler { public: - explicit Throttler(size_t max_speed_, const std::shared_ptr & parent_ = nullptr) - : max_speed(max_speed_), limit_exceeded_exception_message(""), parent(parent_) {} + Throttler(size_t max_speed_, size_t max_burst_, const std::shared_ptr & parent_ = nullptr) + : max_speed(max_speed_), max_burst(max_burst_), limit_exceeded_exception_message(""), tokens(max_burst), parent(parent_) {} + + explicit Throttler(size_t max_speed_, const std::shared_ptr & parent_ = nullptr); + + Throttler(size_t max_speed_, size_t max_burst_, size_t limit_, const char * limit_exceeded_exception_message_, + const std::shared_ptr & parent_ = nullptr) + : max_speed(max_speed_), max_burst(max_burst_), limit(limit_), limit_exceeded_exception_message(limit_exceeded_exception_message_), tokens(max_burst), parent(parent_) {} Throttler(size_t max_speed_, size_t limit_, const char * limit_exceeded_exception_message_, - const std::shared_ptr & parent_ = nullptr) - : max_speed(max_speed_), limit(limit_), limit_exceeded_exception_message(limit_exceeded_exception_message_), parent(parent_) {} + const std::shared_ptr & parent_ = nullptr); - /// Calculates the smoothed speed, sleeps if required and throws exception on + /// Use `amount` ent, sleeps if required and throws exception on /// limit overflow. void add(size_t amount); @@ -45,15 +47,14 @@ public: private: size_t count{0}; - const size_t max_speed{0}; - const uint64_t limit{0}; /// 0 - not limited. + const size_t max_speed{0}; /// in tokens per second. + const size_t max_burst{0}; /// in tokens. + const uint64_t limit{0}; /// 0 - not limited. const char * limit_exceeded_exception_message = nullptr; std::mutex mutex; - std::atomic accumulated_sleep{0}; - /// Smoothed value of current speed. Updated in `add` method. - double smoothed_speed{0}; - /// previous `add` call time (in nanoseconds) - uint64_t prev_ns{0}; + std::atomic accumulated_sleep{0}; // Accumulated sleep time over all waiting threads + double tokens{0}; /// Amount of tokens available in token bucket. Updated in `add` method. + uint64_t prev_ns{0}; /// Previous `add` call time (in nanoseconds). /// Used to implement a hierarchy of throttlers std::shared_ptr parent; From 27599ab70cf9c3b5b19efe497a554b1266b6210e Mon Sep 17 00:00:00 2001 From: serxa Date: Tue, 25 Oct 2022 23:58:25 +0000 Subject: [PATCH 092/239] fix comment --- src/Common/Throttler.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/Common/Throttler.h b/src/Common/Throttler.h index a33637783e7..9b6eff13506 100644 --- a/src/Common/Throttler.h +++ b/src/Common/Throttler.h @@ -29,8 +29,7 @@ public: Throttler(size_t max_speed_, size_t limit_, const char * limit_exceeded_exception_message_, const std::shared_ptr & parent_ = nullptr); - /// Use `amount` ent, sleeps if required and throws exception on - /// limit overflow. + /// Use `amount` tokens, sleeps if required or throws exception on limit overflow. void add(size_t amount); /// Not thread safe From 471b391ab2b92ebd18ca019ea00dca2aa6f99c1c Mon Sep 17 00:00:00 2001 From: serxa Date: Wed, 26 Oct 2022 00:00:40 +0000 Subject: [PATCH 093/239] fix --- src/Common/Throttler.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/Throttler.cpp b/src/Common/Throttler.cpp index 5a0f7e227a6..a3ae966c49b 100644 --- a/src/Common/Throttler.cpp +++ b/src/Common/Throttler.cpp @@ -51,7 +51,7 @@ void Throttler::add(size_t amount) if (max_speed && prev_ns != 0) { double delta_seconds = static_cast(now - prev_ns) / NS; - tokens = std::max(tokens + max_speed * delta_seconds - amount, max_burst); + tokens = std::min(tokens + max_speed * delta_seconds - amount, max_burst); } count += amount; count_value = count; From 91c3744cfe9e9d7aaf03bbdda2e45a960dc56e73 Mon Sep 17 00:00:00 2001 From: DanRoscigno Date: Tue, 25 Oct 2022 21:33:09 -0400 Subject: [PATCH 094/239] add backup to S3 endpoint --- docs/en/operations/{backup.md => _backup.md} | 140 ++++++++++++++++-- .../statements/alter/partition.md | 2 +- 2 files changed, 128 insertions(+), 14 deletions(-) rename docs/en/operations/{backup.md => _backup.md} (68%) diff --git a/docs/en/operations/backup.md b/docs/en/operations/_backup.md similarity index 68% rename from docs/en/operations/backup.md rename to docs/en/operations/_backup.md index a755e3ef9a6..c543c49a083 100644 --- a/docs/en/operations/backup.md +++ b/docs/en/operations/_backup.md @@ -5,6 +5,13 @@ sidebar_label: Data backup and restore title: Data backup and restore --- +- [Backup to a local disk](#backup-to-a-local-disk) +- [Configuring backup/restore to use an S3 endpoint](#configuring-backuprestore-to-use-an-s3-endpoint) +- [Backup/restore using an S3 disk](#backuprestore-using-an-s3-disk) +- [Alternatives](#alternatives) + +## Background + While [replication](../engines/table-engines/mergetree-family/replication.md) provides protection from hardware failures, it does not protect against human errors: accidental deletion of data, deletion of the wrong table or a table on the wrong cluster, and software bugs that result in incorrect data processing or data corruption. In many cases mistakes like these will affect all replicas. ClickHouse has built-in safeguards to prevent some types of mistakes — for example, by default [you can’t just drop tables with a MergeTree-like engine containing more than 50 Gb of data](server-configuration-parameters/settings.md#max-table-size-to-drop). However, these safeguards do not cover all possible cases and can be circumvented. In order to effectively mitigate possible human errors, you should carefully prepare a strategy for backing up and restoring your data **in advance**. @@ -15,7 +22,9 @@ Each company has different resources available and business requirements, so the Keep in mind that if you backed something up and never tried to restore it, chances are that restore will not work properly when you actually need it (or at least it will take longer than business can tolerate). So whatever backup approach you choose, make sure to automate the restore process as well, and practice it on a spare ClickHouse cluster regularly. ::: -## Configure a backup destination +## Backup to a local disk + +### Configure a backup destination In the examples below you will see the backup destination specified like `Disk('backups', '1.zip')`. To prepare the destination add a file to `/etc/clickhouse-server/config.d/backup_disk.xml` specifying the backup destination. For example, this file defines disk named `backups` and then adds that disk to the **backups > allowed_disk** list: @@ -39,7 +48,7 @@ In the examples below you will see the backup destination specified like `Disk(' ``` -## Parameters +### Parameters Backups can be either full or incremental, and can include tables (including materialized views, projections, and dictionaries), and databases. Backups can be synchronous (default) or asynchronous. They can be compressed. Backups can be password protected. @@ -52,7 +61,7 @@ The BACKUP and RESTORE statements take a list of DATABASE and TABLE names, a des - `password` for the file on disk - `base_backup`: the destination of the previous backup of this source. For example, `Disk('backups', '1.zip')` -## Usage examples +### Usage examples Backup and then restore a table: ``` @@ -81,7 +90,7 @@ RESTORE TABLE test.table AS test.table2 FROM Disk('backups', '1.zip') BACKUP TABLE test.table3 AS test.table4 TO Disk('backups', '2.zip') ``` -## Incremental backups +### Incremental backups Incremental backups can be taken by specifying the `base_backup`. :::note @@ -100,7 +109,7 @@ RESTORE TABLE test.table AS test.table2 FROM Disk('backups', 'incremental-a.zip'); ``` -## Assign a password to the backup +### Assign a password to the backup Backups written to disk can have a password applied to the file: ``` @@ -116,7 +125,7 @@ RESTORE TABLE test.table SETTINGS password='qwerty' ``` -## Compression settings +### Compression settings If you would like to specify the compression method or level: ``` @@ -125,14 +134,14 @@ BACKUP TABLE test.table SETTINGS compression_method='lzma', compression_level=3 ``` -## Restore specific partitions +### Restore specific partitions If specific partitions associated with a table need to be restored these can be specified. To restore partitions 1 and 4 from backup: ``` RESTORE TABLE test.table PARTITIONS '2', '3' FROM Disk('backups', 'filename.zip') ``` -## Check the status of backups +### Check the status of backups The backup command returns an `id` and `status`, and that `id` can be used to get the status of the backup. This is very useful to check the progress of long ASYNC backups. The example below shows a failure that happened when trying to overwrite an existing backup file: ```sql @@ -171,13 +180,118 @@ end_time: 2022-08-30 09:21:46 1 row in set. Elapsed: 0.002 sec. ``` -## Backup to S3 +## Configuring BACKUP/RESTORE to use an S3 Endpoint -It is possible to `BACKUP`/`RESTORE` to S3, but this disk should be configured -in a proper way, since by default you will need to backup metadata from local -disk to make backup full. +To write backups to an S3 bucket you need three pieces of information: +- S3 endpoint, + for example `https://mars-doc-test.s3.amazonaws.com/backup-S3/` +- Access key ID, + for example `ABC123` +- Secret access key, + for example `Abc+123` -First of all, you need to configure S3 disk in a special way: +:::note +Creating an S3 bucket is covered in [Use S3 Object Storage as a ClickHouse disk](/docs/en/integrations/data-ingestion/s3/configuring-s3-for-clickhouse-use.md), just come back to this doc after saving the policy, there is no need to configure ClickHouse to use the S3 bucket. +::: + +The destination for a backup will be specified like this: +``` +S3('/', '', ') +``` + +```sql +CREATE TABLE data +( + `key` Int, + `value` String, + `array` Array(String) +) +ENGINE = MergeTree +ORDER BY tuple() +``` + +```sql +INSERT INTO data SELECT * +FROM generateRandom('key Int, value String, array Array(String)') +LIMIT 1000 +``` + +### Create a base (initial) backup + +Incremental backups require a _base_ backup to start from, this example will be used +later as the base backup. The first parameter of the S3 destination is the S3 endpoint followed by the directory within the bucket to use for this backup. In this example the directory is named `my_backup`. + +```sql +BACKUP TABLE data TO S3('https://mars-doc-test.s3.amazonaws.com/backup-S3/my_backup', 'ABC123', 'Abc+123') +``` + +```response +┌─id───────────────────────────────────┬─status─────────┐ +│ de442b75-a66c-4a3c-a193-f76f278c70f3 │ BACKUP_CREATED │ +└──────────────────────────────────────┴────────────────┘ +``` + +### Add more data + +Incremental backups are populated with the difference between the base backup and the current content of the table being backed up. Add more data before taking the incremental backup: + +```sql +INSERT INTO data SELECT * +FROM generateRandom('key Int, value String, array Array(String)') +LIMIT 100 +``` +### Take an incremental backup + +This backup command is similar to the base backup, but adds `SETTINGS base_backup` and the location of the base backup. Note that the destination for the incremental backup is not the same directory as the base, it is the same endpoint with a different target directory within the bucket. The base backup is in `my_backup`, and the incremental will be written to `my_incremental`: +```sql +BACKUP TABLE data TO S3('https://mars-doc-test.s3.amazonaws.com/backup-S3/my_incremental', 'ABC123', 'Abc+123') SETTINGS base_backup = S3('https://mars-doc-test.s3.amazonaws.com/backup-S3/my_backup', 'ABC123', 'Abc+123') +``` + +```response +┌─id───────────────────────────────────┬─status─────────┐ +│ f6cd3900-850f-41c9-94f1-0c4df33ea528 │ BACKUP_CREATED │ +└──────────────────────────────────────┴────────────────┘ +``` +### Restore from the incremental backup + +This command restores the incremental backup into a new table, `data3`. Note that when an incremental backup is restored, the base backup is also included. Specify only the incremental backup when restoring: +```sql +RESTORE TABLE data AS data3 FROM S3('https://mars-doc-test.s3.amazonaws.com/backup-S3/my_incremental', 'ABC123', 'Abc+123') +``` + +```response +┌─id───────────────────────────────────┬─status───┐ +│ ff0c8c39-7dff-4324-a241-000796de11ca │ RESTORED │ +└──────────────────────────────────────┴──────────┘ +``` + +### Verify the count + +There were two inserts into the original table `data`, one with 1,000 rows and one with 100 rows, for a total of 1,100. Verify that the restored table has 1,100 rows: +```sql +SELECT count() +FROM data3 +``` +```response +┌─count()─┐ +│ 1100 │ +└─────────┘ +``` + +### Verify the content +This compares the content of the original table, `data` with the restored table `data3`: +```sql +SELECT throwIf(( + SELECT groupArray(tuple(*)) + FROM data + ) != ( + SELECT groupArray(tuple(*)) + FROM data3 + ), 'Data does not match after BACKUP/RESTORE') +``` +## BACKUP/RESTORE Using an S3 Disk + +It is also possible to `BACKUP`/`RESTORE` to S3 by configuring an S3 disk in the ClickHouse storage configuration. Configure the disk like this by adding a file to `/etc/clickhouse-server/config.d`: ```xml diff --git a/docs/en/sql-reference/statements/alter/partition.md b/docs/en/sql-reference/statements/alter/partition.md index a216de85cfc..da99c52538f 100644 --- a/docs/en/sql-reference/statements/alter/partition.md +++ b/docs/en/sql-reference/statements/alter/partition.md @@ -194,7 +194,7 @@ To restore data from a backup, do the following: Restoring from a backup does not require stopping the server. -For more information about backups and restoring data, see the [Data Backup](../../../operations/backup.md) section. +For more information about backups and restoring data, see the [Data Backup](/docs/en/manage/backups.mdx) section. ## UNFREEZE PARTITION From e996e2ff1560fecaf801c7cb42c3870839d6d4bc Mon Sep 17 00:00:00 2001 From: Robert Schulze Date: Wed, 26 Oct 2022 07:48:37 +0000 Subject: [PATCH 095/239] Fix clang-tidy --- programs/copier/TaskCluster.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/programs/copier/TaskCluster.cpp b/programs/copier/TaskCluster.cpp index 6b7911f56f2..957c7d2120d 100644 --- a/programs/copier/TaskCluster.cpp +++ b/programs/copier/TaskCluster.cpp @@ -40,7 +40,7 @@ void DB::TaskCluster::reloadSettings(const Poco::Util::AbstractConfiguration & c if (config.has(prefix + "settings")) settings_common.loadSettingsFromConfig(prefix + "settings", config); - settings_common.prefer_localhost_replica = 0; + settings_common.prefer_localhost_replica = false; settings_pull = settings_common; if (config.has(prefix + "settings_pull")) From 3eca9ada3b25ba789c049753138b799280c71fb8 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Wed, 26 Oct 2022 08:00:12 +0000 Subject: [PATCH 096/239] Add Date32 to formatDateTIme --- src/Functions/formatDateTime.cpp | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/src/Functions/formatDateTime.cpp b/src/Functions/formatDateTime.cpp index a10c059b342..4f28dae7a66 100644 --- a/src/Functions/formatDateTime.cpp +++ b/src/Functions/formatDateTime.cpp @@ -1,5 +1,6 @@ #include #include +#include #include #include #include @@ -45,6 +46,7 @@ template <> struct ActionValueTypeMap { using ActionValueTyp template <> struct ActionValueTypeMap { using ActionValueType = UInt32; }; template <> struct ActionValueTypeMap { using ActionValueType = UInt32; }; template <> struct ActionValueTypeMap { using ActionValueType = UInt16; }; +template <> struct ActionValueTypeMap { using ActionValueType = Int32; }; template <> struct ActionValueTypeMap { using ActionValueType = UInt32; }; // TODO(vnemkov): to add sub-second format instruction, make that DateTime64 and do some math in Action. template <> struct ActionValueTypeMap { using ActionValueType = Int64; }; @@ -324,7 +326,7 @@ public: "Illegal type " + arguments[0].type->getName() + " of 1 argument of function " + getName() + " when arguments size is 1. Should be integer", ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); - if (arguments.size() > 1 && !(isInteger(arguments[0].type) || isDate(arguments[0].type) || isDateTime(arguments[0].type) || isDateTime64(arguments[0].type))) + if (arguments.size() > 1 && !(isInteger(arguments[0].type) || isDate(arguments[0].type) || isDateTime(arguments[0].type) || isDate32(arguments[0].type) || isDateTime64(arguments[0].type))) throw Exception( "Illegal type " + arguments[0].type->getName() + " of 1 argument of function " + getName() + " when arguments size is 2 or 3. Should be a integer or a date with time", @@ -337,7 +339,7 @@ public: "Number of arguments for function " + getName() + " doesn't match: passed " + toString(arguments.size()) + ", should be 2 or 3", ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH); - if (!isDate(arguments[0].type) && !isDateTime(arguments[0].type) && !isDateTime64(arguments[0].type)) + if (!isDate(arguments[0].type) && !isDateTime(arguments[0].type) && !isDate32(arguments[0].type) && !isDateTime64(arguments[0].type)) throw Exception( "Illegal type " + arguments[0].type->getName() + " of 1 argument of function " + getName() + ". Should be a date or a date with time", @@ -393,6 +395,7 @@ public: })) { if (!((res = executeType(arguments, result_type)) + || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)))) throw Exception( @@ -405,6 +408,7 @@ public: else { if (!((res = executeType(arguments, result_type)) + || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)))) throw Exception( @@ -496,6 +500,13 @@ public: instruction.perform(pos, static_cast(c.whole), time_zone); } } + else if constexpr (std::is_same_v) + { + for (auto & instruction : instructions) + { + instruction.perform(pos, static_cast(vec[i]), time_zone); + } + } else { for (auto & instruction : instructions) From 117674c0cef594c125b472020a9e8339538d92ca Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Wed, 26 Oct 2022 08:48:12 +0000 Subject: [PATCH 097/239] Update version_date.tsv and changelogs after v22.10.1.1875-stable --- docker/server/Dockerfile.alpine | 2 +- docker/server/Dockerfile.ubuntu | 2 +- docs/changelogs/v22.10.1.1875-stable.md | 351 ++++++++++++++++++++++++ utils/list-versions/version_date.tsv | 1 + 4 files changed, 354 insertions(+), 2 deletions(-) create mode 100644 docs/changelogs/v22.10.1.1875-stable.md diff --git a/docker/server/Dockerfile.alpine b/docker/server/Dockerfile.alpine index 5b597f927a2..9b633b66188 100644 --- a/docker/server/Dockerfile.alpine +++ b/docker/server/Dockerfile.alpine @@ -33,7 +33,7 @@ RUN arch=${TARGETARCH:-amd64} \ # lts / testing / prestable / etc ARG REPO_CHANNEL="stable" ARG REPOSITORY="https://packages.clickhouse.com/tgz/${REPO_CHANNEL}" -ARG VERSION="22.9.3.18" +ARG VERSION="22.10.1.1875" ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static" # user/group precreated explicitly with fixed uid/gid on purpose. diff --git a/docker/server/Dockerfile.ubuntu b/docker/server/Dockerfile.ubuntu index c6254b898ed..35c78763a31 100644 --- a/docker/server/Dockerfile.ubuntu +++ b/docker/server/Dockerfile.ubuntu @@ -21,7 +21,7 @@ RUN sed -i "s|http://archive.ubuntu.com|${apt_archive}|g" /etc/apt/sources.list ARG REPO_CHANNEL="stable" ARG REPOSITORY="deb https://packages.clickhouse.com/deb ${REPO_CHANNEL} main" -ARG VERSION="22.9.3.18" +ARG VERSION="22.10.1.1875" ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static" # set non-empty deb_location_url url to create a docker image diff --git a/docs/changelogs/v22.10.1.1875-stable.md b/docs/changelogs/v22.10.1.1875-stable.md new file mode 100644 index 00000000000..49f93168a00 --- /dev/null +++ b/docs/changelogs/v22.10.1.1875-stable.md @@ -0,0 +1,351 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.10.1.1875-stable (011ae8675a2) FIXME as compared to v22.9.1.2603-stable (3030d4c7ff0) + +#### Backward Incompatible Change +* Rename cache commands: `show caches` -> `show filesystem caches`, `describe cache` -> `describe filesystem cache`. [#41508](https://github.com/ClickHouse/ClickHouse/pull/41508) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Remove support for the `WITH TIMEOUT` section for `LIVE VIEW`. This closes [#40557](https://github.com/ClickHouse/ClickHouse/issues/40557). [#42173](https://github.com/ClickHouse/ClickHouse/pull/42173) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### New Feature +* Add Rust code support into ClickHouse with BLAKE3 hash-function library as an example. [#33435](https://github.com/ClickHouse/ClickHouse/pull/33435) ([BoloniniD](https://github.com/BoloniniD)). +* This is the initial implement of Kusto Query Language. (MVP). [#37961](https://github.com/ClickHouse/ClickHouse/pull/37961) ([Yong Wang](https://github.com/kashwy)). +* * Support limiting of temporary data stored on disk using settings `max_temporary_data_on_disk_size_for_user`/`max_temporary_data_on_disk_size_for_query` . [#40893](https://github.com/ClickHouse/ClickHouse/pull/40893) ([Vladimir C](https://github.com/vdimir)). +* Support Java integers hashing in `javaHash`. [#41131](https://github.com/ClickHouse/ClickHouse/pull/41131) ([JackyWoo](https://github.com/JackyWoo)). +* This PR is to support the OpenSSL in-house build like the BoringSSL submodule. Build flag i.e. ENABLE_CH_BUNDLE_BORINGSSL is used to choose between BoringSSL and OpenSSL. By default, the BoringSSL in-house build will be used. [#41142](https://github.com/ClickHouse/ClickHouse/pull/41142) ([MeenaRenganathan22](https://github.com/MeenaRenganathan22)). +* Composable protocol configuration is added. [#41198](https://github.com/ClickHouse/ClickHouse/pull/41198) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* Add OpenTelemetry support to ON CLUSTER DDL(require `distributed_ddl_entry_format_version` to be set to 4). [#41484](https://github.com/ClickHouse/ClickHouse/pull/41484) ([Frank Chen](https://github.com/FrankChen021)). +* Add setting `format_json_object_each_row_column_for_object_name` to write/parse object name as column value in JSONObjectEachRow format. [#41703](https://github.com/ClickHouse/ClickHouse/pull/41703) ([Kruglov Pavel](https://github.com/Avogar)). +* adds Morton Coding (ZCurve) encode/decode functions. [#41753](https://github.com/ClickHouse/ClickHouse/pull/41753) ([Constantine Peresypkin](https://github.com/pkit)). +* Implement support for different UUID binary formats with support for the two most prevalent ones: the default big-endian and Microsoft's mixed-endian as specified in [RFC 4122](https://datatracker.ietf.org/doc/html/rfc4122#section-4.1.1). [#42108](https://github.com/ClickHouse/ClickHouse/pull/42108) ([ltrk2](https://github.com/ltrk2)). +* Added an aggregate function `analysisOfVariance` (`anova`) to perform a statistical test over several groups of normally distributed observations to find out whether all groups have the same mean or not. Original PR [#37872](https://github.com/ClickHouse/ClickHouse/issues/37872). [#42131](https://github.com/ClickHouse/ClickHouse/pull/42131) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Add support for `SET setting_name = DEFAULT`. [#42187](https://github.com/ClickHouse/ClickHouse/pull/42187) ([Filatenkov Artur](https://github.com/FArthur-cmd)). +* * Add `URL` Functions which conform rfc. Functions include: `cutToFirstSignificantSubdomainCustomRFC`, `cutToFirstSignificantSubdomainCustomWithWWWRFC`, `cutToFirstSignificantSubdomainRFC`, `cutToFirstSignificantSubdomainWithWWWRFC`, `domainRFC`, `domainWithoutWWWRFC`, `firstSignificantSubdomainCustomRFC`, `firstSignificantSubdomainRFC`, `portRFC`, `topLevelDomainRFC`. [#42274](https://github.com/ClickHouse/ClickHouse/pull/42274) ([Quanfa Fu](https://github.com/dentiscalprum)). +* Added functions (`randUniform`, `randNormal`, `randLogNormal`, `randExponential`, `randChiSquared`, `randStudentT`, `randFisherF`, `randBernoulli`, `randBinomial`, `randNegativeBinomial`, `randPoisson` ) to generate random values according to the specified distributions. This closes [#21834](https://github.com/ClickHouse/ClickHouse/issues/21834). [#42411](https://github.com/ClickHouse/ClickHouse/pull/42411) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). + +#### Performance Improvement +* Implement operator precedence element parser to resolve stack overflow issues and make the required stack size smaller. [#34892](https://github.com/ClickHouse/ClickHouse/pull/34892) ([Nikolay Degterinsky](https://github.com/evillique)). +* DISTINCT in order optimization leverage sorting properties of data streams. This improvement will enable reading in order for DISTINCT if applicable (before it was necessary to provide ORDER BY for columns in DISTINCT). [#41014](https://github.com/ClickHouse/ClickHouse/pull/41014) ([Igor Nikonov](https://github.com/devcrafter)). +* ColumnVector: optimize UInt8 index with AVX512VBMI. [#41247](https://github.com/ClickHouse/ClickHouse/pull/41247) ([Guo Wangyang](https://github.com/guowangy)). +* The performance experiments of **SSB** (Star Schema Benchmark) on the ICX device (Intel Xeon Platinum 8380 CPU, 80 cores, 160 threads) shows that this change could bring a **2.95x** improvement of the geomean of all subcases' QPS. [#41675](https://github.com/ClickHouse/ClickHouse/pull/41675) ([Zhiguo Zhou](https://github.com/ZhiguoZh)). +* Fixed slowness in JSONExtract with LowCardinality(String) tuples. [#41726](https://github.com/ClickHouse/ClickHouse/pull/41726) ([AlfVII](https://github.com/AlfVII)). +* Add ldapr capabilities to AArch64 builds. This is supported from Graviton 2+, Azure and GCP instances. Only appeared in clang-15 [not so long ago](https://github.com/llvm/llvm-project/commit/9609b5daffe9fd28d83d83da895abc5113f76c24). [#41778](https://github.com/ClickHouse/ClickHouse/pull/41778) ([Daniel Kutenin](https://github.com/danlark1)). +* Improve performance when comparing strings and one argument is empty constant string. [#41870](https://github.com/ClickHouse/ClickHouse/pull/41870) ([Jiebin Sun](https://github.com/jiebinn)). +* optimize insertFrom of ColumnAggregateFunction to share Aggregate State in some cases. [#41960](https://github.com/ClickHouse/ClickHouse/pull/41960) ([flynn](https://github.com/ucasfl)). +* Relax the "Too many parts" threshold. This closes [#6551](https://github.com/ClickHouse/ClickHouse/issues/6551). Now ClickHouse will allow more parts in a partition if the average part size is large enough (at least 10 GiB). This allows to have up to petabytes of data in a single partition of a single table on a single server, which is possible using disk shelves or object storage. [#42002](https://github.com/ClickHouse/ClickHouse/pull/42002) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Make writing to AzureBlobStorage more efficient (respect `max_single_part_upload_size` instead of writing a block per each buffer size). Inefficiency mentioned in [#41754](https://github.com/ClickHouse/ClickHouse/issues/41754). [#42041](https://github.com/ClickHouse/ClickHouse/pull/42041) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Make thread ids in the process list and query_log unique to avoid waste. [#42180](https://github.com/ClickHouse/ClickHouse/pull/42180) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### Improvement +* Added new infrastructure for query analysis and planning under `allow_experimental_analyzer` setting. [#31796](https://github.com/ClickHouse/ClickHouse/pull/31796) ([Maksim Kita](https://github.com/kitaisreal)). +* * Support expression `(EXPLAIN SELECT ...)` in a subquery. Queries like `SELECT * FROM (EXPLAIN PIPELINE SELECT col FROM TABLE ORDER BY col)` became valid. [#40630](https://github.com/ClickHouse/ClickHouse/pull/40630) ([Vladimir C](https://github.com/vdimir)). +* Currently changing `async_insert_max_data_size` or `async_insert_busy_timeout_ms` in scope of query makes no sense and this leads to bad user experience. E.g. user wants to insert data rarely and he doesn't have an access to server config to tune default settings. [#40668](https://github.com/ClickHouse/ClickHouse/pull/40668) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Embedded Keeper will always start in the background allowing ClickHouse to start without achieving quorum. [#40991](https://github.com/ClickHouse/ClickHouse/pull/40991) ([Antonio Andelic](https://github.com/antonio2368)). +* Improvements for reading from remote filesystems, made threadpool size for reads/writes configurable. Closes [#41070](https://github.com/ClickHouse/ClickHouse/issues/41070). [#41011](https://github.com/ClickHouse/ClickHouse/pull/41011) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Made reestablishing a new connection more reactive in case of expiration of the previous one. Previously there was a task which spawns every minute by default and thus a table could be in readonly state for about this time. [#41092](https://github.com/ClickHouse/ClickHouse/pull/41092) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Support all combinators combination in WindowTransform/arratReduce*/initializeAggregation/aggregate functions versioning. Previously combinators like `ForEach/Resample/Map` didn't work in these places, using them led to exception like`State function ... inserts results into non-state column`. [#41107](https://github.com/ClickHouse/ClickHouse/pull/41107) ([Kruglov Pavel](https://github.com/Avogar)). +* Now projections can be used with zero copy replication. [#41147](https://github.com/ClickHouse/ClickHouse/pull/41147) ([alesapin](https://github.com/alesapin)). +* - Add function tryDecrypt that returns NULL when decrypt fail (e.g. decrypt with incorrect key) instead of throwing exception. [#41206](https://github.com/ClickHouse/ClickHouse/pull/41206) ([Duc Canh Le](https://github.com/canhld94)). +* Add the `unreserved_space` column to the `system.disks` table to check how much space is not taken by reservations per disk. [#41254](https://github.com/ClickHouse/ClickHouse/pull/41254) ([filimonov](https://github.com/filimonov)). +* Support s3 authorisation headers from ast arguments. [#41261](https://github.com/ClickHouse/ClickHouse/pull/41261) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Add setting 'allow_implicit_no_password' that forbids creating a user with no password unless 'IDENTIFIED WITH no_password' is explicitly specified. [#41341](https://github.com/ClickHouse/ClickHouse/pull/41341) ([Nikolay Degterinsky](https://github.com/evillique)). +* keeper-improvement: add support for uploading snapshots to S3. S3 information can be defined inside `keeper_server.s3_snapshot`. [#41342](https://github.com/ClickHouse/ClickHouse/pull/41342) ([Antonio Andelic](https://github.com/antonio2368)). +* Add support for MultiRead in Keeper and internal ZooKeeper client. [#41410](https://github.com/ClickHouse/ClickHouse/pull/41410) ([Antonio Andelic](https://github.com/antonio2368)). +* add a support for decimal type comparing with floating point literal in IN operator. [#41544](https://github.com/ClickHouse/ClickHouse/pull/41544) ([liang.huang](https://github.com/lhuang09287750)). +* Allow readable size values in cache config. [#41688](https://github.com/ClickHouse/ClickHouse/pull/41688) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Check file path for path traversal attacks in errors logger for input formats. [#41694](https://github.com/ClickHouse/ClickHouse/pull/41694) ([Kruglov Pavel](https://github.com/Avogar)). +* ClickHouse could cache stale DNS entries for some period of time (15 seconds by default) until the cache won't be updated asynchronously. During these period ClickHouse can nevertheless try to establish a connection and produce errors. This behaviour is fixed. [#41707](https://github.com/ClickHouse/ClickHouse/pull/41707) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Add interactive history search with fzf-like utility (fzf/sk) for `clickhouse-client`/`clickhouse-local` (note you can use `FZF_DEFAULT_OPTS`/`SKIM_DEFAULT_OPTIONS` to additionally configure the behavior). [#41730](https://github.com/ClickHouse/ClickHouse/pull/41730) ([Azat Khuzhin](https://github.com/azat)). +* For client when connecting to a secure server with invalid certificate only allow to proceed with '--accept-certificate' flag. [#41743](https://github.com/ClickHouse/ClickHouse/pull/41743) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* Add function "tryBase58Decode()", similar to the existing function "tryBase64Decode()". [#41824](https://github.com/ClickHouse/ClickHouse/pull/41824) ([Robert Schulze](https://github.com/rschu1ze)). +* Improve feedback when replacing partition with different primary key. Fixes [#34798](https://github.com/ClickHouse/ClickHouse/issues/34798). [#41838](https://github.com/ClickHouse/ClickHouse/pull/41838) ([Salvatore](https://github.com/tbsal)). +* Replace back `clickhouse su` command with `sudo -u` in start in order to respect limits in `/etc/security/limits.conf`. [#41847](https://github.com/ClickHouse/ClickHouse/pull/41847) ([Eugene Konkov](https://github.com/ekonkov)). +* Fix parallel parsing: segmentator now checks max_block_size. [#41852](https://github.com/ClickHouse/ClickHouse/pull/41852) ([Vitaly Baranov](https://github.com/vitlibar)). +* Don't report TABLE_IS_DROPPED exception in order to skip table in case is was just dropped. [#41908](https://github.com/ClickHouse/ClickHouse/pull/41908) ([AlfVII](https://github.com/AlfVII)). +* Improve option enable_extended_results_for_datetime_functions to return results of type DateTime64 for functions toStartOfDay, toStartOfHour, toStartOfFifteenMinutes, toStartOfTenMinutes, toStartOfFiveMinutes, toStartOfMinute and timeSlot. [#41910](https://github.com/ClickHouse/ClickHouse/pull/41910) ([Roman Vasin](https://github.com/rvasin)). +* Improve DateTime type inference for text formats. Now it respect setting `date_time_input_format` and doesn't try to infer datetimes from numbers as timestamps. Closes [#41389](https://github.com/ClickHouse/ClickHouse/issues/41389) Closes [#42206](https://github.com/ClickHouse/ClickHouse/issues/42206). [#41912](https://github.com/ClickHouse/ClickHouse/pull/41912) ([Kruglov Pavel](https://github.com/Avogar)). +* Remove confusing warning when inserting with `perform_ttl_move_on_insert`=false. [#41980](https://github.com/ClickHouse/ClickHouse/pull/41980) ([Vitaly Baranov](https://github.com/vitlibar)). +* Allow user to write `countState(*)` similar to `count(*)`. This closes [#9338](https://github.com/ClickHouse/ClickHouse/issues/9338). [#41983](https://github.com/ClickHouse/ClickHouse/pull/41983) ([Amos Bird](https://github.com/amosbird)). +* - Fix rankCorr size overflow. [#42020](https://github.com/ClickHouse/ClickHouse/pull/42020) ([Duc Canh Le](https://github.com/canhld94)). +* Added an option to specify an arbitrary string as an environment name in the Sentry's config for more handy reports. [#42037](https://github.com/ClickHouse/ClickHouse/pull/42037) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Added system table `asynchronous_insert_log `. It contains information about asynchronous inserts (including results of queries in fire-and-forget mode (with `wait_for_async_insert=0`)) for better introspection. [#42040](https://github.com/ClickHouse/ClickHouse/pull/42040) ([Anton Popov](https://github.com/CurtizJ)). +* Fix parsing out-of-range Date from CSV:. [#42044](https://github.com/ClickHouse/ClickHouse/pull/42044) ([Andrey Zvonov](https://github.com/zvonand)). +* parseDataTimeBestEffort support comma between date and time. Closes [#42038](https://github.com/ClickHouse/ClickHouse/issues/42038). [#42049](https://github.com/ClickHouse/ClickHouse/pull/42049) ([flynn](https://github.com/ucasfl)). +* Add support for methods lz4, bz2, snappy in 'Accept-Encoding'. [#42071](https://github.com/ClickHouse/ClickHouse/pull/42071) ([Nikolay Degterinsky](https://github.com/evillique)). +* Various minor fixes for BLAKE3 function. [#42073](https://github.com/ClickHouse/ClickHouse/pull/42073) ([BoloniniD](https://github.com/BoloniniD)). +* Improved stale replica recovery process for `ReplicatedMergeTree`. If lost replica have some parts which absent on a healthy replica, but these parts should appear in future according to replication queue of the healthy replica, then lost replica will keep such parts instead of detaching them. [#42134](https://github.com/ClickHouse/ClickHouse/pull/42134) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Support BACKUP to S3 with as-is path/data structure. [#42232](https://github.com/ClickHouse/ClickHouse/pull/42232) ([Azat Khuzhin](https://github.com/azat)). +* Add a possibility to use Date32 arguments for date_diff function. Fix issue in date_diff function when using DateTime64 arguments with start date before Unix epoch and end date after Unix epoch. [#42308](https://github.com/ClickHouse/ClickHouse/pull/42308) ([Roman Vasin](https://github.com/rvasin)). +* When uploading big parts to minio, 'Complete Multipart Upload' can take a long time. Minio sends heartbeats every 10 seconds (see https://github.com/minio/minio/pull/7198). But clickhouse times out earlier, because the default send/receive timeout is [set](https://github.com/ClickHouse/ClickHouse/blob/cc24fcd6d5dfb67f5f66f5483e986bd1010ad9cf/src/IO/S3/PocoHTTPClient.cpp#L123) to 5 seconds. [#42321](https://github.com/ClickHouse/ClickHouse/pull/42321) ([filimonov](https://github.com/filimonov)). +* Add `S3` as a new type of the destination of backups. [#42333](https://github.com/ClickHouse/ClickHouse/pull/42333) ([Vitaly Baranov](https://github.com/vitlibar)). +* Fix rarely invalid cast of aggregate state types with complex types such as Decimal. This fixes [#42408](https://github.com/ClickHouse/ClickHouse/issues/42408). [#42417](https://github.com/ClickHouse/ClickHouse/pull/42417) ([Amos Bird](https://github.com/amosbird)). +* Support skipping cache completely (both download to cache and reading cached data) in case the requested read range exceeds the threshold defined by cache setting `bypass_cache_threashold`, requires to be enabled with `enable_bypass_cache_with_threshold`). [#42418](https://github.com/ClickHouse/ClickHouse/pull/42418) ([Han Shukai](https://github.com/KinderRiven)). +* Merge parts if every part in the range is older than a certain threshold. The threshold can be set by using `min_age_to_force_merge_seconds`. This closes [#35836](https://github.com/ClickHouse/ClickHouse/issues/35836). [#42423](https://github.com/ClickHouse/ClickHouse/pull/42423) ([Antonio Andelic](https://github.com/antonio2368)). +* Enabled CompiledExpressionCache in clickhouse-local. [#42477](https://github.com/ClickHouse/ClickHouse/pull/42477) ([AlfVII](https://github.com/AlfVII)). +* Remove support for the `{database}` macro from the client's prompt. It was displayed incorrectly if the database was unspecified and it was not updated on `USE` statements. This closes [#25891](https://github.com/ClickHouse/ClickHouse/issues/25891). [#42508](https://github.com/ClickHouse/ClickHouse/pull/42508) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* - Improve the time to recover lost keeper connections. [#42541](https://github.com/ClickHouse/ClickHouse/pull/42541) ([Raúl Marín](https://github.com/Algunenano)). +* Allow to use Date32 arguments for dateName function. [#42554](https://github.com/ClickHouse/ClickHouse/pull/42554) ([Roman Vasin](https://github.com/rvasin)). + +#### Bug Fix +* Now filters with NULL literals will be used during index analysis. This closes https://github.com/ClickHouse/ClickHouse/pull/41814 [#34063](https://github.com/ClickHouse/ClickHouse/issues/34063). [#41842](https://github.com/ClickHouse/ClickHouse/pull/41842) ([Amos Bird](https://github.com/amosbird)). +* - Choose correct aggregation method for LowCardinality with BigInt. [#42342](https://github.com/ClickHouse/ClickHouse/pull/42342) ([Duc Canh Le](https://github.com/canhld94)). +* Fix using subqueries in row policy filters. This PR fixes [#32463](https://github.com/ClickHouse/ClickHouse/issues/32463). [#42562](https://github.com/ClickHouse/ClickHouse/pull/42562) ([Vitaly Baranov](https://github.com/vitlibar)). + +#### Build/Testing/Packaging Improvement +* Added support of WHERE clause generation to AST Fuzzer and possibility to add or remove ORDER BY and WHERE clause. [#38519](https://github.com/ClickHouse/ClickHouse/pull/38519) ([Ilya Yatsishin](https://github.com/qoega)). +* Aarch64 binaries now require at least ARMv8.2, released in 2016. Most notably, this enables use of ARM LSE, i.e. native atomic operations. Also, CMake build option "NO_ARMV81_OR_HIGHER" has been added to allow compilation of binaries for older ARMv8.0 hardware, e.g. Raspberry Pi 4. [#41610](https://github.com/ClickHouse/ClickHouse/pull/41610) ([Robert Schulze](https://github.com/rschu1ze)). +* After updating runners to 22.04 cgroups stopped to work in privileged mode, here's the issue https://github.com/moby/moby/issues/42275#issuecomment-1115055846. [#41857](https://github.com/ClickHouse/ClickHouse/pull/41857) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Allow building ClickHouse with Musl (small changes after it was already supported but broken). [#41987](https://github.com/ClickHouse/ClickHouse/pull/41987) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* - Add the `$CLICKHOUSE_CRONFILE` file checking to avoid running the `sed` command to get the file not found error. [#42081](https://github.com/ClickHouse/ClickHouse/pull/42081) ([Chun-Sheng, Li](https://github.com/peter279k)). +* Update cctz to the latest master, update tzdb to 2020e. [#42273](https://github.com/ClickHouse/ClickHouse/pull/42273) ([Dom Del Nano](https://github.com/ddelnano)). +* Update tzdata to 2022e to support the new timezone changes. Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Jordan and Syria switch from +02/+03 with DST to year-round +03. (https://data.iana.org/time-zones/tzdb/NEWS). This closes [#42252](https://github.com/ClickHouse/ClickHouse/issues/42252). [#42327](https://github.com/ClickHouse/ClickHouse/pull/42327) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix power8 support. [#42462](https://github.com/ClickHouse/ClickHouse/pull/42462) ([Boris Kuschel](https://github.com/bkuschel)). + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Several fixes for DiskWeb. [#41652](https://github.com/ClickHouse/ClickHouse/pull/41652) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fixes issue when docker run will fail if "https_port" is not present in config. [#41693](https://github.com/ClickHouse/ClickHouse/pull/41693) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* Mutations were not cancelled properly on server shutdown or `SYSTEM STOP MERGES` query and cancellation might take long time, it's fixed. [#41699](https://github.com/ClickHouse/ClickHouse/pull/41699) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Fix wrong result of queries with `ORDER BY` or `GROUP BY` by columns from prefix of sorting key, wrapped into monotonic functions, with enable "read in order" optimization (settings `optimize_read_in_order` and `optimize_aggregation_in_order`). [#41701](https://github.com/ClickHouse/ClickHouse/pull/41701) ([Anton Popov](https://github.com/CurtizJ)). +* Fix possible crash in `SELECT` from `Merge` table with enabled `optimize_monotonous_functions_in_order_by` setting. Fixes [#41269](https://github.com/ClickHouse/ClickHouse/issues/41269). [#41740](https://github.com/ClickHouse/ClickHouse/pull/41740) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fixed "Part ... intersects part ..." error that might happen in extremely rare cases if replica was restarted just after detaching some part as broken. [#41741](https://github.com/ClickHouse/ClickHouse/pull/41741) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Don't allow to create or alter merge tree tables with virtual column name _row_exists, which is reserved for lightweight delete. Fixed [#41716](https://github.com/ClickHouse/ClickHouse/issues/41716). [#41763](https://github.com/ClickHouse/ClickHouse/pull/41763) ([Jianmei Zhang](https://github.com/zhangjmruc)). +* Fix a bug that CORS headers are missing in some HTTP responses. [#41792](https://github.com/ClickHouse/ClickHouse/pull/41792) ([Frank Chen](https://github.com/FrankChen021)). +* 22.9 might fail to startup `ReplicatedMergeTree` table if that table was created by 20.3 or older version and was never altered, it's fixed. Fixes [#41742](https://github.com/ClickHouse/ClickHouse/issues/41742). [#41796](https://github.com/ClickHouse/ClickHouse/pull/41796) ([Alexander Tokmakov](https://github.com/tavplubix)). +* When the batch sending fails for some reason, it cannot be automatically recovered, and if it is not processed in time, it will lead to accumulation, and the printed error message will become longer and longer, which will cause the http thread to block. [#41813](https://github.com/ClickHouse/ClickHouse/pull/41813) ([zhongyuankai](https://github.com/zhongyuankai)). +* Fix compact parts with compressed marks setting. Fixes [#41783](https://github.com/ClickHouse/ClickHouse/issues/41783) and [#41746](https://github.com/ClickHouse/ClickHouse/issues/41746). [#41823](https://github.com/ClickHouse/ClickHouse/pull/41823) ([alesapin](https://github.com/alesapin)). +* Old versions of Replicated database doesn't have a special marker in [Zoo]Keeper. We need to check only whether the node contains come obscure data instead of special mark. [#41875](https://github.com/ClickHouse/ClickHouse/pull/41875) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Fix possible exception in fs cache. [#41884](https://github.com/ClickHouse/ClickHouse/pull/41884) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix use_environment_credentials for s3 table function. [#41970](https://github.com/ClickHouse/ClickHouse/pull/41970) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fixed "Directory already exists and is not empty" error on detaching broken part that might prevent `ReplicatedMergeTree` table from starting replication. Fixes [#40957](https://github.com/ClickHouse/ClickHouse/issues/40957). [#41981](https://github.com/ClickHouse/ClickHouse/pull/41981) ([Alexander Tokmakov](https://github.com/tavplubix)). +* toDateTime64() now returns the same output with negative integer and float arguments. [#42025](https://github.com/ClickHouse/ClickHouse/pull/42025) ([Robert Schulze](https://github.com/rschu1ze)). +* Fix write into AzureBlobStorage. Partially closes [#41754](https://github.com/ClickHouse/ClickHouse/issues/41754). [#42034](https://github.com/ClickHouse/ClickHouse/pull/42034) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix the bzip2 decoding issue for specific bzip2 files. [#42046](https://github.com/ClickHouse/ClickHouse/pull/42046) ([Nikolay Degterinsky](https://github.com/evillique)). +* - Fix SQL function "toLastDayOfMonth()" with setting "enable_extended_results_for_datetime_functions = 1" at the beginning of the extended range (January 1900). - Fix SQL function "toRelativeWeekNum()" with setting "enable_extended_results_for_datetime_functions = 1" at the end of extended range (December 2299). - Improve the performance of for SQL functions "toISOYear()", "toFirstDayNumOfISOYearIndex()" and "toYearWeekOfNewyearMode()" by avoiding unnecessary index arithmetics. [#42084](https://github.com/ClickHouse/ClickHouse/pull/42084) ([Roman Vasin](https://github.com/rvasin)). +* The maximum size of fetches for each table accidentally was set to 8 while the pool size could be bigger. Now the maximum size of fetches for table is equal to the pool size. [#42090](https://github.com/ClickHouse/ClickHouse/pull/42090) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* A table might be shut down and a dictionary might be detached before checking if can be dropped without breaking dependencies between table, it's fixed. Fixes [#41982](https://github.com/ClickHouse/ClickHouse/issues/41982). [#42106](https://github.com/ClickHouse/ClickHouse/pull/42106) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Fix bad inefficiency of `remote_filesystem_read_method=read` with filesystem cache. Closes [#42125](https://github.com/ClickHouse/ClickHouse/issues/42125). [#42129](https://github.com/ClickHouse/ClickHouse/pull/42129) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix possible timeout exception for distributed queries with use_hedged_requests=0. [#42130](https://github.com/ClickHouse/ClickHouse/pull/42130) ([Azat Khuzhin](https://github.com/azat)). +* Fixed a minor bug inside function `runningDifference` in case of using it with `Date32` type. Previously `Date` was used and it may cause some logical errors like `Bad cast from type DB::ColumnVector to DB::ColumnVector'`. [#42143](https://github.com/ClickHouse/ClickHouse/pull/42143) ([Alfred Xu](https://github.com/sperlingxx)). +* Fix reusing of files > 4GB from base backup. [#42146](https://github.com/ClickHouse/ClickHouse/pull/42146) ([Azat Khuzhin](https://github.com/azat)). +* DISTINCT in order fails with LOGICAL_ERROR if first column in sorting key contains function. [#42186](https://github.com/ClickHouse/ClickHouse/pull/42186) ([Igor Nikonov](https://github.com/devcrafter)). +* Fix a bug with projections and the `aggregate_functions_null_for_empty` setting. This bug is very rare and appears only if you enable the `aggregate_functions_null_for_empty` setting in the server's config. This closes [#41647](https://github.com/ClickHouse/ClickHouse/issues/41647). [#42198](https://github.com/ClickHouse/ClickHouse/pull/42198) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* - Fix read from buffer with read in order desc. [#42236](https://github.com/ClickHouse/ClickHouse/pull/42236) ([Duc Canh Le](https://github.com/canhld94)). +* Fix a bug which prevents ClickHouse to start when background_pool_size setting is set on default profile but background_merges_mutations_concurrency_ratio is not. [#42315](https://github.com/ClickHouse/ClickHouse/pull/42315) ([nvartolomei](https://github.com/nvartolomei)). +* `ALTER UPDATE` of attached part (with columns different from table schema) could create an invalid `columns.txt` metadata on disk. Reading from such part could fail with errors or return invalid data. Fixes [#42161](https://github.com/ClickHouse/ClickHouse/issues/42161). [#42319](https://github.com/ClickHouse/ClickHouse/pull/42319) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Setting `additional_table_filters` were not applied to `Distributed` storage. Fixes [#41692](https://github.com/ClickHouse/ClickHouse/issues/41692). [#42322](https://github.com/ClickHouse/ClickHouse/pull/42322) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fix a data race in query finish/cancel. This closes [#42346](https://github.com/ClickHouse/ClickHouse/issues/42346). [#42362](https://github.com/ClickHouse/ClickHouse/pull/42362) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* This reverts [#40217](https://github.com/ClickHouse/ClickHouse/issues/40217) which introduced a regression in date/time functions. [#42367](https://github.com/ClickHouse/ClickHouse/pull/42367) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix assert cast in join on falsy condition, Close [#42380](https://github.com/ClickHouse/ClickHouse/issues/42380). [#42407](https://github.com/ClickHouse/ClickHouse/pull/42407) ([Vladimir C](https://github.com/vdimir)). +* Fix buffer overflow in the processing of Decimal data types. This closes [#42451](https://github.com/ClickHouse/ClickHouse/issues/42451). [#42465](https://github.com/ClickHouse/ClickHouse/pull/42465) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* `AggregateFunctionQuantile` now correctly works with UInt128 columns. Previously, the quantile state interpreted `UInt128` columns as `Int128` which could have led to incorrect results. [#42473](https://github.com/ClickHouse/ClickHouse/pull/42473) ([Antonio Andelic](https://github.com/antonio2368)). +* Fix bad_assert during INSERT into Annoy indexes over non-Float32 columns. [#42485](https://github.com/ClickHouse/ClickHouse/pull/42485) ([Robert Schulze](https://github.com/rschu1ze)). +* This closes [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix function `arrayElement` with type `Map` with `Nullable` values and `Nullable` index. [#42623](https://github.com/ClickHouse/ClickHouse/pull/42623) ([Anton Popov](https://github.com/CurtizJ)). + +#### Bug Fix (user-visible misbehaviour in official stable or prestable release) + +* Fix unexpected table loading error when partition key contains alias function names during server upgrade. [#36379](https://github.com/ClickHouse/ClickHouse/pull/36379) ([Amos Bird](https://github.com/amosbird)). + +#### Build Improvement + +* Fixed SipHash Endian issue for s390x platform. [#41372](https://github.com/ClickHouse/ClickHouse/pull/41372) ([Harry Lee](https://github.com/HarryLeeIBM)). +* Enable lib base64 for ppc64le platform. [#41974](https://github.com/ClickHouse/ClickHouse/pull/41974) ([Suzy Wang](https://github.com/SuzyWangIBMer)). +* Fixed Endian issue in T64 compression codec on s390x. [#42314](https://github.com/ClickHouse/ClickHouse/pull/42314) ([Harry Lee](https://github.com/HarryLeeIBM)). + +#### NO CL ENTRY + +* NO CL ENTRY: 'Revert "Disable parallel s3 multipart upload for part moves."'. [#41681](https://github.com/ClickHouse/ClickHouse/pull/41681) ([Alexander Tokmakov](https://github.com/tavplubix)). +* NO CL ENTRY: 'Revert "Attempt to fix abort from parallel parsing"'. [#42545](https://github.com/ClickHouse/ClickHouse/pull/42545) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* NO CL ENTRY: 'Revert "Low cardinality cases moved to the function for its corresponding type"'. [#42633](https://github.com/ClickHouse/ClickHouse/pull/42633) ([Anton Popov](https://github.com/CurtizJ)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Test for ignore function in PARTITION KEY [#39875](https://github.com/ClickHouse/ClickHouse/pull/39875) ([UnamedRus](https://github.com/UnamedRus)). +* Add fuzzer for table definitions [#40096](https://github.com/ClickHouse/ClickHouse/pull/40096) ([Anton Popov](https://github.com/CurtizJ)). +* Add missing tests for legacy geobase [#40684](https://github.com/ClickHouse/ClickHouse/pull/40684) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Remove obsolete comment from the config.xml [#41518](https://github.com/ClickHouse/ClickHouse/pull/41518) ([filimonov](https://github.com/filimonov)). +* Resurrect parallel distributed insert select with s3Cluster [#41535](https://github.com/ClickHouse/ClickHouse/pull/41535) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Update runners to a recent version to install on 22.04 [#41556](https://github.com/ClickHouse/ClickHouse/pull/41556) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Refactor wiping sensitive information from logs. [#41562](https://github.com/ClickHouse/ClickHouse/pull/41562) ([Vitaly Baranov](https://github.com/vitlibar)). +* Better S3 logs [#41587](https://github.com/ClickHouse/ClickHouse/pull/41587) ([Nikolay Degterinsky](https://github.com/evillique)). +* Fix typos in JSON formats after [#40910](https://github.com/ClickHouse/ClickHouse/issues/40910) [#41614](https://github.com/ClickHouse/ClickHouse/pull/41614) ([Kruglov Pavel](https://github.com/Avogar)). +* Fix drop for KeeperMap [#41616](https://github.com/ClickHouse/ClickHouse/pull/41616) ([Antonio Andelic](https://github.com/antonio2368)). +* increase default max_suspicious_broken_parts to 100 [#41619](https://github.com/ClickHouse/ClickHouse/pull/41619) ([Denny Crane](https://github.com/den-crane)). +* Release AWS SDK log level + replace one exception [#41649](https://github.com/ClickHouse/ClickHouse/pull/41649) ([alesapin](https://github.com/alesapin)). +* Fix a destruction order for views ThreadStatus [#41650](https://github.com/ClickHouse/ClickHouse/pull/41650) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Add very explicit logging on disk choice for fetch [#41653](https://github.com/ClickHouse/ClickHouse/pull/41653) ([alesapin](https://github.com/alesapin)). +* Fix race between ~BackgroundSchedulePool and ~DNSCacheUpdater [#41654](https://github.com/ClickHouse/ClickHouse/pull/41654) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Add changelog for 22.9 [#41668](https://github.com/ClickHouse/ClickHouse/pull/41668) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Update version after release [#41670](https://github.com/ClickHouse/ClickHouse/pull/41670) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix error message [#41680](https://github.com/ClickHouse/ClickHouse/pull/41680) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Add test for setting output_format_json_validate_utf8 [#41691](https://github.com/ClickHouse/ClickHouse/pull/41691) ([Kruglov Pavel](https://github.com/Avogar)). +* Resolve findings from clang-tidy [#41702](https://github.com/ClickHouse/ClickHouse/pull/41702) ([ltrk2](https://github.com/ltrk2)). +* Ignore Keeper errors from ReplicatedMergeTreeAttachThread in stress tests [#41717](https://github.com/ClickHouse/ClickHouse/pull/41717) ([Antonio Andelic](https://github.com/antonio2368)). +* Collect logs in Stress test using clickhouse-local [#41721](https://github.com/ClickHouse/ClickHouse/pull/41721) ([Antonio Andelic](https://github.com/antonio2368)). +* Disable flaky `test_merge_tree_azure_blob_storage` [#41722](https://github.com/ClickHouse/ClickHouse/pull/41722) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Update version_date.tsv and changelogs after v22.9.2.7-stable [#41724](https://github.com/ClickHouse/ClickHouse/pull/41724) ([robot-clickhouse](https://github.com/robot-clickhouse)). +* Fix part removal retries [#41728](https://github.com/ClickHouse/ClickHouse/pull/41728) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Try fix azure tests [#41731](https://github.com/ClickHouse/ClickHouse/pull/41731) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix test build [#41732](https://github.com/ClickHouse/ClickHouse/pull/41732) ([Robert Schulze](https://github.com/rschu1ze)). +* Change logging levels in cache [#41733](https://github.com/ClickHouse/ClickHouse/pull/41733) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Revert of "Revert the revert of "ColumnVector: optimize filter with AVX512 VBMI2 compress store" [#40033](https://github.com/ClickHouse/ClickHouse/issues/40033)" [#41752](https://github.com/ClickHouse/ClickHouse/pull/41752) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix SET query parameters formatting [#41755](https://github.com/ClickHouse/ClickHouse/pull/41755) ([Nikolay Degterinsky](https://github.com/evillique)). +* Support to run testcases on macOS [#41760](https://github.com/ClickHouse/ClickHouse/pull/41760) ([Frank Chen](https://github.com/FrankChen021)). +* Bump LLVM from 12 to 13 [#41762](https://github.com/ClickHouse/ClickHouse/pull/41762) ([Robert Schulze](https://github.com/rschu1ze)). +* ColumnVector: re-enable AVX512_VBMI/AVX512_VBMI2 optimized filter and index [#41765](https://github.com/ClickHouse/ClickHouse/pull/41765) ([Guo Wangyang](https://github.com/guowangy)). +* Update 02354_annoy.sql [#41767](https://github.com/ClickHouse/ClickHouse/pull/41767) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Fix the typo preventing building latest images [#41769](https://github.com/ClickHouse/ClickHouse/pull/41769) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Make automatic download script choose between ARMv8.0 or ARMv8.2 builds [#41775](https://github.com/ClickHouse/ClickHouse/pull/41775) ([Robert Schulze](https://github.com/rschu1ze)). +* Fix tests for docker-ci [#41777](https://github.com/ClickHouse/ClickHouse/pull/41777) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Possible fix for KeeperMap drop [#41784](https://github.com/ClickHouse/ClickHouse/pull/41784) ([Antonio Andelic](https://github.com/antonio2368)). +* Fix drop of completely dropped table [#41789](https://github.com/ClickHouse/ClickHouse/pull/41789) ([alesapin](https://github.com/alesapin)). +* Log git hash during startup [#41790](https://github.com/ClickHouse/ClickHouse/pull/41790) ([Robert Schulze](https://github.com/rschu1ze)). +* Revert "ColumnVector: optimize UInt8 index with AVX512VBMI ([#41247](https://github.com/ClickHouse/ClickHouse/issues/41247))" [#41797](https://github.com/ClickHouse/ClickHouse/pull/41797) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Small fix in dashboard [#41798](https://github.com/ClickHouse/ClickHouse/pull/41798) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Keep the most important log in stress tests [#41821](https://github.com/ClickHouse/ClickHouse/pull/41821) ([alesapin](https://github.com/alesapin)). +* Use copy for some operations instead of hardlinks [#41832](https://github.com/ClickHouse/ClickHouse/pull/41832) ([alesapin](https://github.com/alesapin)). +* Remove unused variable in registerStorageMergeTree.cpp [#41839](https://github.com/ClickHouse/ClickHouse/pull/41839) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix Jepsen [#41845](https://github.com/ClickHouse/ClickHouse/pull/41845) ([Antonio Andelic](https://github.com/antonio2368)). +* Increase `request_timeout_ms` for s3 tests in CI [#41853](https://github.com/ClickHouse/ClickHouse/pull/41853) ([Kseniia Sumarokova](https://github.com/kssenii)). +* tests: fix debug symbols (and possible crashes) for backward compatiblity check [#41854](https://github.com/ClickHouse/ClickHouse/pull/41854) ([Azat Khuzhin](https://github.com/azat)). +* Remove two redundant lines [#41856](https://github.com/ClickHouse/ClickHouse/pull/41856) ([alesapin](https://github.com/alesapin)). +* Infer Object type only when allow_experimental_object_type is enabled [#41858](https://github.com/ClickHouse/ClickHouse/pull/41858) ([Kruglov Pavel](https://github.com/Avogar)). +* Add default UNION/EXCEPT/INTERSECT to the echo query text [#41862](https://github.com/ClickHouse/ClickHouse/pull/41862) ([Nikolay Degterinsky](https://github.com/evillique)). +* Consolidate CMake-generated config headers [#41873](https://github.com/ClickHouse/ClickHouse/pull/41873) ([Robert Schulze](https://github.com/rschu1ze)). +* Fix 02267_file_globs_schema_inference.sql flakiness [#41877](https://github.com/ClickHouse/ClickHouse/pull/41877) ([Kruglov Pavel](https://github.com/Avogar)). +* Docs: Remove obsolete modelEvaluate() mention [#41878](https://github.com/ClickHouse/ClickHouse/pull/41878) ([Robert Schulze](https://github.com/rschu1ze)). +* Better exception message for duplicate column names in schema inference [#41885](https://github.com/ClickHouse/ClickHouse/pull/41885) ([Kruglov Pavel](https://github.com/Avogar)). +* Docs: Reference external papers as DOIs [#41886](https://github.com/ClickHouse/ClickHouse/pull/41886) ([Robert Schulze](https://github.com/rschu1ze)). +* Make LDAPR a prerequisite for downloading the ARMv8.2 build [#41897](https://github.com/ClickHouse/ClickHouse/pull/41897) ([Robert Schulze](https://github.com/rschu1ze)). +* Another sync replicas in test_recovery_replica [#41898](https://github.com/ClickHouse/ClickHouse/pull/41898) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* remove unused code [#41921](https://github.com/ClickHouse/ClickHouse/pull/41921) ([flynn](https://github.com/ucasfl)). +* Move all queries for MV creation to the end of queue during recovering [#41932](https://github.com/ClickHouse/ClickHouse/pull/41932) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Fix broken test_disks_app_func [#41933](https://github.com/ClickHouse/ClickHouse/pull/41933) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Temporarily disable ThreadFuzzer with TSan [#41943](https://github.com/ClickHouse/ClickHouse/pull/41943) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Enable some disabled S3 tests [#41945](https://github.com/ClickHouse/ClickHouse/pull/41945) ([alesapin](https://github.com/alesapin)). +* QOL log improvements [#41947](https://github.com/ClickHouse/ClickHouse/pull/41947) ([Raúl Marín](https://github.com/Algunenano)). +* Fix non-deterministic test results [#41948](https://github.com/ClickHouse/ClickHouse/pull/41948) ([Robert Schulze](https://github.com/rschu1ze)). +* Earlier throw exception in PullingAsyncPipelineExecutor. [#41949](https://github.com/ClickHouse/ClickHouse/pull/41949) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Fix linker error [#41950](https://github.com/ClickHouse/ClickHouse/pull/41950) ([ltrk2](https://github.com/ltrk2)). +* Bump LLVM from 13 to 14 [#41951](https://github.com/ClickHouse/ClickHouse/pull/41951) ([Robert Schulze](https://github.com/rschu1ze)). +* Update version_date.tsv and changelogs after v22.3.13.80-lts [#41953](https://github.com/ClickHouse/ClickHouse/pull/41953) ([robot-clickhouse](https://github.com/robot-clickhouse)). +* Update version_date.tsv and changelogs after v22.7.6.74-stable [#41954](https://github.com/ClickHouse/ClickHouse/pull/41954) ([robot-clickhouse](https://github.com/robot-clickhouse)). +* Update version_date.tsv and changelogs after v22.8.6.71-lts [#41955](https://github.com/ClickHouse/ClickHouse/pull/41955) ([robot-clickhouse](https://github.com/robot-clickhouse)). +* Update version_date.tsv and changelogs after v22.9.3.18-stable [#41956](https://github.com/ClickHouse/ClickHouse/pull/41956) ([robot-clickhouse](https://github.com/robot-clickhouse)). +* Add a warning message to release.py script, require release type [#41975](https://github.com/ClickHouse/ClickHouse/pull/41975) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Rename max_temp_data_on_disk -> max_temporary_data_on_disk [#41984](https://github.com/ClickHouse/ClickHouse/pull/41984) ([Vladimir C](https://github.com/vdimir)). +* Add more checkStackSize calls [#41991](https://github.com/ClickHouse/ClickHouse/pull/41991) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix test 02403_big_http_chunk_size [#41996](https://github.com/ClickHouse/ClickHouse/pull/41996) ([Vitaly Baranov](https://github.com/vitlibar)). +* More sane behavior of part number thresholds override in query level settings [#42001](https://github.com/ClickHouse/ClickHouse/pull/42001) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Remove useless code [#42004](https://github.com/ClickHouse/ClickHouse/pull/42004) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Refactoring: Uninline some error handling methods [#42010](https://github.com/ClickHouse/ClickHouse/pull/42010) ([Robert Schulze](https://github.com/rschu1ze)). +* Fix warning that ENABLE_REPLXX is unused [#42013](https://github.com/ClickHouse/ClickHouse/pull/42013) ([Robert Schulze](https://github.com/rschu1ze)). +* Drop leftovers of libexecinfo [#42014](https://github.com/ClickHouse/ClickHouse/pull/42014) ([Robert Schulze](https://github.com/rschu1ze)). +* More detailed exception message [#42022](https://github.com/ClickHouse/ClickHouse/pull/42022) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Build against an LLVM version which has clang[-extra-tools], lldb and lld removed [#42023](https://github.com/ClickHouse/ClickHouse/pull/42023) ([Robert Schulze](https://github.com/rschu1ze)). +* Add log message and lower the retry timeout in MergeTreeRestartingThread [#42026](https://github.com/ClickHouse/ClickHouse/pull/42026) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Update amqp-cpp [#42031](https://github.com/ClickHouse/ClickHouse/pull/42031) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix No such key during table drop [#42036](https://github.com/ClickHouse/ClickHouse/pull/42036) ([alesapin](https://github.com/alesapin)). +* Temporarily disable too aggressive tests [#42050](https://github.com/ClickHouse/ClickHouse/pull/42050) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Fix style check [#42055](https://github.com/ClickHouse/ClickHouse/pull/42055) ([Anton Popov](https://github.com/CurtizJ)). +* Function name normalization fix functions header [#42063](https://github.com/ClickHouse/ClickHouse/pull/42063) ([Maksim Kita](https://github.com/kitaisreal)). +* remove unused virtual keyword [#42065](https://github.com/ClickHouse/ClickHouse/pull/42065) ([flynn](https://github.com/ucasfl)). +* Fix crash in `SummingMergeTree` with `LowCardinality` [#42066](https://github.com/ClickHouse/ClickHouse/pull/42066) ([Anton Popov](https://github.com/CurtizJ)). +* Fix drop of completely dropped table [#42067](https://github.com/ClickHouse/ClickHouse/pull/42067) ([alesapin](https://github.com/alesapin)). +* Fix assertion in bloom filter index [#42072](https://github.com/ClickHouse/ClickHouse/pull/42072) ([Anton Popov](https://github.com/CurtizJ)). +* Ignore core.autocrlf for tests references [#42076](https://github.com/ClickHouse/ClickHouse/pull/42076) ([Azat Khuzhin](https://github.com/azat)). +* Fix progress for INSERT SELECT [#42078](https://github.com/ClickHouse/ClickHouse/pull/42078) ([Azat Khuzhin](https://github.com/azat)). +* Avoid adding extra new line after using fuzzy history search [#42080](https://github.com/ClickHouse/ClickHouse/pull/42080) ([Azat Khuzhin](https://github.com/azat)). +* Add `at` to runner AMI, bump gh runner version [#42082](https://github.com/ClickHouse/ClickHouse/pull/42082) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Use send_metadata instead of send_object_metadata [#42085](https://github.com/ClickHouse/ClickHouse/pull/42085) ([Elena Torró](https://github.com/elenatorro)). +* Docs: Preparations to remove misc statements page [#42086](https://github.com/ClickHouse/ClickHouse/pull/42086) ([Robert Schulze](https://github.com/rschu1ze)). +* Followup for TemporaryDataOnDisk [#42103](https://github.com/ClickHouse/ClickHouse/pull/42103) ([Vladimir C](https://github.com/vdimir)). +* Disable 02122_join_group_by_timeout for debug [#42104](https://github.com/ClickHouse/ClickHouse/pull/42104) ([Vladimir C](https://github.com/vdimir)). +* Update version_date.tsv and changelogs after v22.6.9.11-stable [#42114](https://github.com/ClickHouse/ClickHouse/pull/42114) ([robot-clickhouse](https://github.com/robot-clickhouse)). +* JIT compilation migration to LLVM 15 [#42123](https://github.com/ClickHouse/ClickHouse/pull/42123) ([Maksim Kita](https://github.com/kitaisreal)). +* Fix build without TSA [#42128](https://github.com/ClickHouse/ClickHouse/pull/42128) ([Raúl Marín](https://github.com/Algunenano)). +* Update codespell-ignore-words.list [#42132](https://github.com/ClickHouse/ClickHouse/pull/42132) ([Dan Roscigno](https://github.com/DanRoscigno)). +* Add null pointer checks [#42135](https://github.com/ClickHouse/ClickHouse/pull/42135) ([ltrk2](https://github.com/ltrk2)). +* Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787) [#42136](https://github.com/ClickHouse/ClickHouse/pull/42136) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Follow up for [#42129](https://github.com/ClickHouse/ClickHouse/issues/42129) [#42144](https://github.com/ClickHouse/ClickHouse/pull/42144) ([Kseniia Sumarokova](https://github.com/kssenii)). +* Fix checking parent for old-format parts [#42147](https://github.com/ClickHouse/ClickHouse/pull/42147) ([alesapin](https://github.com/alesapin)). +* Revert "Resurrect parallel distributed insert select with s3Cluster [#42150](https://github.com/ClickHouse/ClickHouse/pull/42150) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Docs: Add "TABLE" to CHECK/DESCRIBE statements in sidebar [#42152](https://github.com/ClickHouse/ClickHouse/pull/42152) ([Robert Schulze](https://github.com/rschu1ze)). +* Add logging during merge tree startup [#42163](https://github.com/ClickHouse/ClickHouse/pull/42163) ([alesapin](https://github.com/alesapin)). +* Abort instead of `__builtin_unreachable` in debug builds [#42168](https://github.com/ClickHouse/ClickHouse/pull/42168) ([Alexander Tokmakov](https://github.com/tavplubix)). +* [RFC] Enable -Wshorten-64-to-32 [#42190](https://github.com/ClickHouse/ClickHouse/pull/42190) ([Azat Khuzhin](https://github.com/azat)). +* Fix dialect setting description [#42196](https://github.com/ClickHouse/ClickHouse/pull/42196) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* Add a test for #658 [#42197](https://github.com/ClickHouse/ClickHouse/pull/42197) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* use alias for MergeMutateSelectedEntry share ptr [#42211](https://github.com/ClickHouse/ClickHouse/pull/42211) ([Tian Xinhui](https://github.com/xinhuitian)). +* Fix LLVM build [#42216](https://github.com/ClickHouse/ClickHouse/pull/42216) ([Raúl Marín](https://github.com/Algunenano)). +* Exclude comments from style-check defined extern [#42217](https://github.com/ClickHouse/ClickHouse/pull/42217) ([Vladimir C](https://github.com/vdimir)). +* Update safeExit.cpp [#42220](https://github.com/ClickHouse/ClickHouse/pull/42220) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Disable concurrent parts removal [#42222](https://github.com/ClickHouse/ClickHouse/pull/42222) ([alesapin](https://github.com/alesapin)). +* Fail fast on empty URL in HDFS [#42223](https://github.com/ClickHouse/ClickHouse/pull/42223) ([Ilya Yatsishin](https://github.com/qoega)). +* Add a test for [#2389](https://github.com/ClickHouse/ClickHouse/issues/2389) [#42235](https://github.com/ClickHouse/ClickHouse/pull/42235) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Use MultiRead where possible [#42243](https://github.com/ClickHouse/ClickHouse/pull/42243) ([Antonio Andelic](https://github.com/antonio2368)). +* Minor cleanups of LLVM integration [#42249](https://github.com/ClickHouse/ClickHouse/pull/42249) ([Robert Schulze](https://github.com/rschu1ze)). +* remove useless code [#42253](https://github.com/ClickHouse/ClickHouse/pull/42253) ([flynn](https://github.com/ucasfl)). +* Early return of corner cases in selectPartsToMutate function [#42254](https://github.com/ClickHouse/ClickHouse/pull/42254) ([Tian Xinhui](https://github.com/xinhuitian)). +* Refactor the implementation of user-defined functions [#42263](https://github.com/ClickHouse/ClickHouse/pull/42263) ([Vitaly Baranov](https://github.com/vitlibar)). +* assert unused value in test_replicated_merge_tree_compatibility [#42266](https://github.com/ClickHouse/ClickHouse/pull/42266) ([nvartolomei](https://github.com/nvartolomei)). +* Fix Date Interval add/minus over DataTypeDate32 [#42279](https://github.com/ClickHouse/ClickHouse/pull/42279) ([Alfred Xu](https://github.com/sperlingxx)). +* Fix log-level in `clickhouse-disks` [#42302](https://github.com/ClickHouse/ClickHouse/pull/42302) ([Nikolay Degterinsky](https://github.com/evillique)). +* Remove forgotten debug logging [#42313](https://github.com/ClickHouse/ClickHouse/pull/42313) ([Kruglov Pavel](https://github.com/Avogar)). +* Fix another trash in zero-copy replication [#42317](https://github.com/ClickHouse/ClickHouse/pull/42317) ([alesapin](https://github.com/alesapin)). +* go update for diagnostics tool [#42325](https://github.com/ClickHouse/ClickHouse/pull/42325) ([Dale McDiarmid](https://github.com/gingerwizard)). +* Better logging for asynchronous inserts [#42345](https://github.com/ClickHouse/ClickHouse/pull/42345) ([Anton Popov](https://github.com/CurtizJ)). +* Use nfpm packager for archlinux packages [#42349](https://github.com/ClickHouse/ClickHouse/pull/42349) ([Azat Khuzhin](https://github.com/azat)). +* Bump llvm/clang to 15.0.2 [#42351](https://github.com/ClickHouse/ClickHouse/pull/42351) ([Azat Khuzhin](https://github.com/azat)). +* Make getResource() independent from the order of the sections [#42353](https://github.com/ClickHouse/ClickHouse/pull/42353) ([Azat Khuzhin](https://github.com/azat)). +* Smaller threshold for multipart upload part size increase [#42392](https://github.com/ClickHouse/ClickHouse/pull/42392) ([alesapin](https://github.com/alesapin)). +* Better error message for unsupported delimiters in custom formats [#42406](https://github.com/ClickHouse/ClickHouse/pull/42406) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Fix formatting of `ALTER FREEZE` [#42409](https://github.com/ClickHouse/ClickHouse/pull/42409) ([Anton Popov](https://github.com/CurtizJ)). +* Replace table name in ast fuzzer more often [#42413](https://github.com/ClickHouse/ClickHouse/pull/42413) ([Anton Popov](https://github.com/CurtizJ)). +* Add *-15 tools to cmake.tools for GCC build [#42430](https://github.com/ClickHouse/ClickHouse/pull/42430) ([Ilya Yatsishin](https://github.com/qoega)). +* Deactivate tasks in ReplicatedMergeTree until startup [#42441](https://github.com/ClickHouse/ClickHouse/pull/42441) ([alesapin](https://github.com/alesapin)). +* Revert "Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787)" [#42442](https://github.com/ClickHouse/ClickHouse/pull/42442) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Update woboq_codebrowser location [#42448](https://github.com/ClickHouse/ClickHouse/pull/42448) ([Azat Khuzhin](https://github.com/azat)). +* add mdx and jsx to list of doc files [#42454](https://github.com/ClickHouse/ClickHouse/pull/42454) ([Dan Roscigno](https://github.com/DanRoscigno)). +* Remove code browser docs [#42455](https://github.com/ClickHouse/ClickHouse/pull/42455) ([Dan Roscigno](https://github.com/DanRoscigno)). +* Better workaround for emitting .debug_aranges section [#42457](https://github.com/ClickHouse/ClickHouse/pull/42457) ([Azat Khuzhin](https://github.com/azat)). +* Fix flaky test [#42459](https://github.com/ClickHouse/ClickHouse/pull/42459) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix UBSan report in Julian Day functions [#42464](https://github.com/ClickHouse/ClickHouse/pull/42464) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* rename filesystem_query_cache [#42472](https://github.com/ClickHouse/ClickHouse/pull/42472) ([Han Shukai](https://github.com/KinderRiven)). +* Add convenience typedefs for Date/Date32/DateTime/DateTime64 columns [#42476](https://github.com/ClickHouse/ClickHouse/pull/42476) ([Robert Schulze](https://github.com/rschu1ze)). +* Add error "Destination table is myself" to exception list in BC check [#42479](https://github.com/ClickHouse/ClickHouse/pull/42479) ([Kruglov Pavel](https://github.com/Avogar)). +* Get current clickhouse version without sending query in BC check [#42483](https://github.com/ClickHouse/ClickHouse/pull/42483) ([Kruglov Pavel](https://github.com/Avogar)). +* Fix logical error from welchTTest [#42487](https://github.com/ClickHouse/ClickHouse/pull/42487) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Attempt to fix abort from parallel parsing [#42496](https://github.com/ClickHouse/ClickHouse/pull/42496) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Increase threshold for using physical cores for `max_threads` [#42503](https://github.com/ClickHouse/ClickHouse/pull/42503) ([Nikita Taranov](https://github.com/nickitat)). +* Add a test for [#16827](https://github.com/ClickHouse/ClickHouse/issues/16827) [#42511](https://github.com/ClickHouse/ClickHouse/pull/42511) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Add a test for [#13653](https://github.com/ClickHouse/ClickHouse/issues/13653) [#42512](https://github.com/ClickHouse/ClickHouse/pull/42512) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix aliases [#42514](https://github.com/ClickHouse/ClickHouse/pull/42514) ([Nikolay Degterinsky](https://github.com/evillique)). +* tests: fix 00705_drop_create_merge_tree flakiness [#42522](https://github.com/ClickHouse/ClickHouse/pull/42522) ([Azat Khuzhin](https://github.com/azat)). +* Fix sanitizer reports in integration tests [#42529](https://github.com/ClickHouse/ClickHouse/pull/42529) ([Azat Khuzhin](https://github.com/azat)). +* Fix `KeeperTCPHandler` data race [#42532](https://github.com/ClickHouse/ClickHouse/pull/42532) ([Antonio Andelic](https://github.com/antonio2368)). +* Disable `test_storage_nats`, because it's permanently broken [#42535](https://github.com/ClickHouse/ClickHouse/pull/42535) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Better logs in clickhouse-disks [#42549](https://github.com/ClickHouse/ClickHouse/pull/42549) ([Nikolay Degterinsky](https://github.com/evillique)). +* add lib_fuzzer and lib_fuzzer_no_main to llvm-project build [#42550](https://github.com/ClickHouse/ClickHouse/pull/42550) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* Some polishing: replicated merge tree [#42560](https://github.com/ClickHouse/ClickHouse/pull/42560) ([Igor Nikonov](https://github.com/devcrafter)). +* Temporarily disable flaky `test_replicated_merge_tree_hdfs_zero_copy` [#42563](https://github.com/ClickHouse/ClickHouse/pull/42563) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Adapt internal data structures to 512-bit era [#42564](https://github.com/ClickHouse/ClickHouse/pull/42564) ([Nikita Taranov](https://github.com/nickitat)). +* Fix strange code in date monotonicity [#42574](https://github.com/ClickHouse/ClickHouse/pull/42574) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Clear thread::id when ThreadFromGlobalPool exits. [#42577](https://github.com/ClickHouse/ClickHouse/pull/42577) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* ci/stress: fix memory limits overrides for hung check [#42585](https://github.com/ClickHouse/ClickHouse/pull/42585) ([Azat Khuzhin](https://github.com/azat)). +* tests: avoid model overlap for obfuscator [#42586](https://github.com/ClickHouse/ClickHouse/pull/42586) ([Azat Khuzhin](https://github.com/azat)). +* Fix possible segfault in expression parser [#42598](https://github.com/ClickHouse/ClickHouse/pull/42598) ([Nikolay Degterinsky](https://github.com/evillique)). +* Fix incorrect trace log line on dict reload [#42609](https://github.com/ClickHouse/ClickHouse/pull/42609) ([filimonov](https://github.com/filimonov)). +* Fix flaky 02458_datediff_date32 test [#42611](https://github.com/ClickHouse/ClickHouse/pull/42611) ([Roman Vasin](https://github.com/rvasin)). +* Revert revert 41268 disable s3 parallel write for part moves to disk s3 [#42617](https://github.com/ClickHouse/ClickHouse/pull/42617) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Try to fix data race on zookeeper vs DDLWorker at server shutdown. [#42620](https://github.com/ClickHouse/ClickHouse/pull/42620) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Add a template for installation issues [#42626](https://github.com/ClickHouse/ClickHouse/pull/42626) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Fix typo in cmake code related to fuzzing [#42627](https://github.com/ClickHouse/ClickHouse/pull/42627) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). +* Fix build [#42635](https://github.com/ClickHouse/ClickHouse/pull/42635) ([Anton Popov](https://github.com/CurtizJ)). +* Add .rgignore for test data [#42639](https://github.com/ClickHouse/ClickHouse/pull/42639) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 65ec5ddec01..7bbd8547506 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -1,3 +1,4 @@ +v22.10.1.1875-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 v22.9.2.7-stable 2022-09-23 v22.9.1.2603-stable 2022-09-22 From 7973b4e5e50d31e0529574a3ce34a72b6844c147 Mon Sep 17 00:00:00 2001 From: Ilya Yatsishin <2159081+qoega@users.noreply.github.com> Date: Wed, 26 Oct 2022 11:09:01 +0200 Subject: [PATCH 098/239] Fix error handling --- tests/ci/clickhouse_helper.py | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/tests/ci/clickhouse_helper.py b/tests/ci/clickhouse_helper.py index a81334860d1..c82d9da05e9 100644 --- a/tests/ci/clickhouse_helper.py +++ b/tests/ci/clickhouse_helper.py @@ -37,12 +37,8 @@ class ClickHouseHelper: url, params=params, data=json_str, headers=auth ) except Exception as e: - logging.warning( - "Received exception while sending data to %s on %s attempt: %s", - url, - i, - e, - ) + error = f"Received exception while sending data to {url} on {i} attempt: {e}" + logging.warning(error) continue logging.info("Response content '%s'", response.content) From a8f3b39dcc23ef389cf68cf7c5cfccbe98c3c6ee Mon Sep 17 00:00:00 2001 From: "Mikhail f. Shiryaev" Date: Wed, 26 Oct 2022 12:07:42 +0200 Subject: [PATCH 099/239] Fix execution of version_helper.py to use git tweaks --- tests/ci/version_helper.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/ci/version_helper.py b/tests/ci/version_helper.py index 966858c0747..162bab6a50a 100755 --- a/tests/ci/version_helper.py +++ b/tests/ci/version_helper.py @@ -344,7 +344,7 @@ def main(): update_contributors() return - version = get_version_from_repo(args.version_path) + version = get_version_from_repo(args.version_path, Git(True)) if args.update: version = version.update(args.update) From 1b6293f6db33b474f518d40784a6cfb758c6f306 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 12:43:29 +0200 Subject: [PATCH 100/239] MergeTree indexes use ActionsDAG --- .../QueryPlan/ReadFromMergeTree.cpp | 7 +- src/Storages/MergeTree/KeyCondition.cpp | 581 ++++-------------- src/Storages/MergeTree/KeyCondition.h | 82 +-- .../MergeTreeIndexConditionBloomFilter.cpp | 325 +++++----- .../MergeTreeIndexConditionBloomFilter.h | 24 +- .../MergeTree/MergeTreeIndexFullText.cpp | 151 ++--- .../MergeTree/MergeTreeIndexFullText.h | 8 +- .../MergeTree/MergeTreeIndexUtils.cpp | 47 ++ src/Storages/MergeTree/MergeTreeIndexUtils.h | 13 + src/Storages/MergeTree/RPNBuilder.cpp | 380 ++++++++++++ src/Storages/MergeTree/RPNBuilder.h | 276 +++++++-- 11 files changed, 1063 insertions(+), 831 deletions(-) create mode 100644 src/Storages/MergeTree/MergeTreeIndexUtils.cpp create mode 100644 src/Storages/MergeTree/MergeTreeIndexUtils.h create mode 100644 src/Storages/MergeTree/RPNBuilder.cpp diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.cpp b/src/Processors/QueryPlan/ReadFromMergeTree.cpp index 62b6eddf6ce..71f8378ae41 100644 --- a/src/Processors/QueryPlan/ReadFromMergeTree.cpp +++ b/src/Processors/QueryPlan/ReadFromMergeTree.cpp @@ -906,8 +906,11 @@ MergeTreeDataSelectAnalysisResultPtr ReadFromMergeTree::selectRangesToRead( for (const auto & node : added_filter_nodes.nodes) nodes.nodes.push_back(node); - key_condition.emplace( - std::move(nodes), query_info.syntax_analyzer_result, query_info.prepared_sets, context, primary_key_columns, primary_key.expression); + key_condition.emplace(std::move(nodes), + context, + primary_key_columns, + primary_key.expression, + query_info.syntax_analyzer_result->getArrayJoinSourceNameSet()); } else { diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp index b282fb79514..596345cade7 100644 --- a/src/Storages/MergeTree/KeyCondition.cpp +++ b/src/Storages/MergeTree/KeyCondition.cpp @@ -27,6 +27,7 @@ #include #include #include +#include #include #include @@ -112,289 +113,6 @@ static String firstStringThatIsGreaterThanAllStringsWithPrefix(const String & pr return res; } -static void appendColumnNameWithoutAlias(const ActionsDAG::Node & node, WriteBuffer & out, bool legacy = false) -{ - switch (node.type) - { - case (ActionsDAG::ActionType::INPUT): - writeString(node.result_name, out); - break; - case (ActionsDAG::ActionType::COLUMN): - { - /// If it was created from ASTLiteral, then result_name can be an alias. - /// We need to convert value back to string here. - if (const auto * column_const = typeid_cast(node.column.get())) - writeString(applyVisitor(FieldVisitorToString(), column_const->getField()), out); - /// It may be possible that column is ColumnSet - else - writeString(node.result_name, out); - break; - } - case (ActionsDAG::ActionType::ALIAS): - appendColumnNameWithoutAlias(*node.children.front(), out, legacy); - break; - case (ActionsDAG::ActionType::ARRAY_JOIN): - writeCString("arrayJoin(", out); - appendColumnNameWithoutAlias(*node.children.front(), out, legacy); - writeChar(')', out); - break; - case (ActionsDAG::ActionType::FUNCTION): - { - auto name = node.function_base->getName(); - if (legacy && name == "modulo") - writeCString("moduleLegacy", out); - else - writeString(name, out); - - writeChar('(', out); - bool first = true; - for (const auto * arg : node.children) - { - if (!first) - writeCString(", ", out); - first = false; - - appendColumnNameWithoutAlias(*arg, out, legacy); - } - writeChar(')', out); - } - } -} - -static std::string getColumnNameWithoutAlias(const ActionsDAG::Node & node, bool legacy = false) -{ - WriteBufferFromOwnString out; - appendColumnNameWithoutAlias(node, out, legacy); - return std::move(out.str()); -} - -class KeyCondition::Tree -{ -public: - explicit Tree(const IAST * ast_) : ast(ast_) { assert(ast); } - explicit Tree(const ActionsDAG::Node * dag_) : dag(dag_) { assert(dag); } - - std::string getColumnName() const - { - if (ast) - return ast->getColumnNameWithoutAlias(); - else - return getColumnNameWithoutAlias(*dag); - } - - std::string getColumnNameLegacy() const - { - if (ast) - { - auto adjusted_ast = ast->clone(); - KeyDescription::moduloToModuloLegacyRecursive(adjusted_ast); - return adjusted_ast->getColumnNameWithoutAlias(); - } - else - return getColumnNameWithoutAlias(*dag, true); - } - - bool isFunction() const - { - if (ast) - return typeid_cast(ast); - else - return dag->type == ActionsDAG::ActionType::FUNCTION; - } - - bool isConstant() const - { - if (ast) - return typeid_cast(ast); - else - return dag->column && isColumnConst(*dag->column); - } - - ColumnWithTypeAndName getConstant() const - { - if (!isConstant()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "KeyCondition::Tree node is not a constant"); - - ColumnWithTypeAndName res; - - if (ast) - { - const auto * literal = assert_cast(ast); - res.type = applyVisitor(FieldToDataType(), literal->value); - res.column = res.type->createColumnConst(0, literal->value); - - } - else - { - res.type = dag->result_type; - res.column = dag->column; - } - - return res; - } - - bool tryGetConstant(const Block & block_with_constants, Field & out_value, DataTypePtr & out_type) const - { - if (ast) - { - // Constant expr should use alias names if any - String column_name = ast->getColumnName(); - - if (const auto * lit = ast->as()) - { - /// By default block_with_constants has only one column named "_dummy". - /// If block contains only constants it's may not be preprocessed by - // ExpressionAnalyzer, so try to look up in the default column. - if (!block_with_constants.has(column_name)) - column_name = "_dummy"; - - /// Simple literal - out_value = lit->value; - out_type = block_with_constants.getByName(column_name).type; - - /// If constant is not Null, we can assume it's type is not Nullable as well. - if (!out_value.isNull()) - out_type = removeNullable(out_type); - - return true; - } - else if (block_with_constants.has(column_name) && isColumnConst(*block_with_constants.getByName(column_name).column)) - { - /// An expression which is dependent on constants only - const auto & expr_info = block_with_constants.getByName(column_name); - out_value = (*expr_info.column)[0]; - out_type = expr_info.type; - - if (!out_value.isNull()) - out_type = removeNullable(out_type); - - return true; - } - } - else - { - if (dag->column && isColumnConst(*dag->column)) - { - out_value = (*dag->column)[0]; - out_type = dag->result_type; - - if (!out_value.isNull()) - out_type = removeNullable(out_type); - - return true; - } - } - - return false; - } - - ConstSetPtr tryGetPreparedSet( - const PreparedSetsPtr & sets, - const std::vector & indexes_mapping, - const DataTypes & data_types) const - { - if (sets && ast) - { - if (ast->as() || ast->as()) - return sets->get(PreparedSetKey::forSubquery(*ast)); - - /// We have `PreparedSetKey::forLiteral` but it is useless here as we don't have enough information - /// about types in left argument of the IN operator. Instead, we manually iterate through all the sets - /// and find the one for the right arg based on the AST structure (getTreeHash), after that we check - /// that the types it was prepared with are compatible with the types of the primary key. - auto types_match = [&indexes_mapping, &data_types](const SetPtr & candidate_set) - { - assert(indexes_mapping.size() == data_types.size()); - - for (size_t i = 0; i < indexes_mapping.size(); ++i) - { - if (!candidate_set->areTypesEqual(indexes_mapping[i].tuple_index, data_types[i])) - return false; - } - - return true; - }; - - for (const auto & set : sets->getByTreeHash(ast->getTreeHash())) - { - if (types_match(set)) - return set; - } - } - else if (dag->column) - { - const IColumn * col = dag->column.get(); - if (const auto * col_const = typeid_cast(col)) - col = &col_const->getDataColumn(); - - if (const auto * col_set = typeid_cast(col)) - { - auto set = col_set->getData(); - if (set->isCreated()) - return set; - } - } - - return nullptr; - } - - FunctionTree asFunction() const; - -protected: - const IAST * ast = nullptr; - const ActionsDAG::Node * dag = nullptr; -}; - -class KeyCondition::FunctionTree : public KeyCondition::Tree -{ -public: - std::string getFunctionName() const - { - if (ast) - return assert_cast(ast)->name; - else - return dag->function_base->getName(); - } - - size_t numArguments() const - { - if (ast) - { - const auto * func = assert_cast(ast); - return func->arguments ? func->arguments->children.size() : 0; - } - else - return dag->children.size(); - } - - Tree getArgumentAt(size_t idx) const - { - if (ast) - return Tree(assert_cast(ast)->arguments->children[idx].get()); - else - return Tree(dag->children[idx]); - } - -private: - using Tree::Tree; - - friend class Tree; -}; - - -KeyCondition::FunctionTree KeyCondition::Tree::asFunction() const -{ - if (!isFunction()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "KeyCondition::Tree node is not a function"); - - if (ast) - return KeyCondition::FunctionTree(ast); - else - return KeyCondition::FunctionTree(dag); -} - - -/// A dictionary containing actions to the corresponding functions to turn them into `RPNElement` const KeyCondition::AtomMap KeyCondition::atom_map { { @@ -869,16 +587,17 @@ static NameSet getAllSubexpressionNames(const ExpressionActions & key_expr) KeyCondition::KeyCondition( const ASTPtr & query, const ASTs & additional_filter_asts, - TreeRewriterResultPtr syntax_analyzer_result, - PreparedSetsPtr prepared_sets_, + Block block_with_constants, + PreparedSetsPtr prepared_sets, ContextPtr context, const Names & key_column_names, const ExpressionActionsPtr & key_expr_, + NameSet array_joined_column_names_, bool single_point_, bool strict_) : key_expr(key_expr_) , key_subexpr_names(getAllSubexpressionNames(*key_expr)) - , prepared_sets(prepared_sets_) + , array_joined_column_names(std::move(array_joined_column_names_)) , single_point(single_point_) , strict(strict_) { @@ -889,82 +608,64 @@ KeyCondition::KeyCondition( key_columns[name] = i; } - if (!syntax_analyzer_result) + auto filter_node = buildFilterNode(query, additional_filter_asts); + + if (!filter_node) { rpn.emplace_back(RPNElement::FUNCTION_UNKNOWN); return; } - /** Evaluation of expressions that depend only on constants. - * For the index to be used, if it is written, for example `WHERE Date = toDate(now())`. + /** When non-strictly monotonic functions are employed in functional index (e.g. ORDER BY toStartOfHour(dateTime)), + * the use of NOT operator in predicate will result in the indexing algorithm leave out some data. + * This is caused by rewriting in KeyCondition::tryParseAtomFromAST of relational operators to less strict + * when parsing the AST into internal RPN representation. + * To overcome the problem, before parsing the AST we transform it to its semantically equivalent form where all NOT's + * are pushed down and applied (when possible) to leaf nodes. */ - Block block_with_constants = getBlockWithConstants(query, syntax_analyzer_result, context); + auto inverted_filter_node = cloneASTWithInversionPushDown(filter_node); - if (syntax_analyzer_result) - { - for (const auto & [name, _] : syntax_analyzer_result->array_join_result_to_source) - array_joined_columns.insert(name); - } + RPNBuilder builder( + inverted_filter_node, + std::move(context), + std::move(block_with_constants), + std::move(prepared_sets), + [&](const RPNBuilderTreeNode & node, RPNElement & out) { return extractAtomFromTree(node, out); }); + rpn = std::move(builder).extractRPN(); +} - const ASTSelectQuery & select = query->as(); - - ASTs filters; - if (select.where()) - filters.push_back(select.where()); - - if (select.prewhere()) - filters.push_back(select.prewhere()); - - for (const auto & filter_ast : additional_filter_asts) - filters.push_back(filter_ast); - - if (!filters.empty()) - { - ASTPtr filter_query; - if (filters.size() == 1) - { - filter_query = filters.front(); - } - else - { - auto function = std::make_shared(); - - function->name = "and"; - function->arguments = std::make_shared(); - function->children.push_back(function->arguments); - function->arguments->children = std::move(filters); - - filter_query = function; - } - - /** When non-strictly monotonic functions are employed in functional index (e.g. ORDER BY toStartOfHour(dateTime)), - * the use of NOT operator in predicate will result in the indexing algorithm leave out some data. - * This is caused by rewriting in KeyCondition::tryParseAtomFromAST of relational operators to less strict - * when parsing the AST into internal RPN representation. - * To overcome the problem, before parsing the AST we transform it to its semantically equivalent form where all NOT's - * are pushed down and applied (when possible) to leaf nodes. - */ - auto ast = cloneASTWithInversionPushDown(filter_query); - traverseAST(Tree(ast.get()), context, block_with_constants); - } - else - { - rpn.emplace_back(RPNElement::FUNCTION_UNKNOWN); - } +KeyCondition::KeyCondition( + const SelectQueryInfo & query_info, + ContextPtr context, + const Names & key_column_names, + const ExpressionActionsPtr & key_expr_, + bool single_point_, + bool strict_) + : KeyCondition( + query_info.query, + query_info.filter_asts, + KeyCondition::getBlockWithConstants(query_info.query, query_info.syntax_analyzer_result, context), + query_info.prepared_sets, + context, + key_column_names, + key_expr_, + query_info.syntax_analyzer_result->getArrayJoinSourceNameSet(), + single_point_, + strict_) +{ } KeyCondition::KeyCondition( ActionDAGNodes dag_nodes, - TreeRewriterResultPtr syntax_analyzer_result, - PreparedSetsPtr prepared_sets_, ContextPtr context, const Names & key_column_names, const ExpressionActionsPtr & key_expr_, + NameSet array_joined_column_names_, bool single_point_, bool strict_) : key_expr(key_expr_) , key_subexpr_names(getAllSubexpressionNames(*key_expr)) - , prepared_sets(prepared_sets_) + , array_joined_column_names(std::move(array_joined_column_names_)) , single_point(single_point_) , strict(strict_) { @@ -975,29 +676,23 @@ KeyCondition::KeyCondition( key_columns[name] = i; } - if (!syntax_analyzer_result) + if (dag_nodes.nodes.empty()) { rpn.emplace_back(RPNElement::FUNCTION_UNKNOWN); return; } - for (const auto & [name, _] : syntax_analyzer_result->array_join_result_to_source) - array_joined_columns.insert(name); + auto inverted_dag = cloneASTWithInversionPushDown(std::move(dag_nodes.nodes), context); + assert(inverted_dag->getOutputs().size() == 1); - if (!dag_nodes.nodes.empty()) + const auto * inverted_dag_filter_node = inverted_dag->getOutputs()[0]; + + RPNBuilder builder(inverted_dag_filter_node, context, [&](const RPNBuilderTreeNode & node, RPNElement & out) { - auto inverted_dag = cloneASTWithInversionPushDown(std::move(dag_nodes.nodes), context); + return extractAtomFromTree(node, out); + }); - // std::cerr << "========== inverted dag: " << inverted_dag->dumpDAG() << std::endl; - - Block empty; - for (const auto * node : inverted_dag->getOutputs()) - traverseAST(Tree(node), context, empty); - } - else - { - rpn.emplace_back(RPNElement::FUNCTION_UNKNOWN); - } + rpn = std::move(builder).extractRPN(); } bool KeyCondition::addCondition(const String & column, const Range & range) @@ -1009,12 +704,12 @@ bool KeyCondition::addCondition(const String & column, const Range & range) return true; } -/** Computes value of constant expression and its data type. - * Returns false, if expression isn't constant. - */ bool KeyCondition::getConstant(const ASTPtr & expr, Block & block_with_constants, Field & out_value, DataTypePtr & out_type) { - return Tree(expr.get()).tryGetConstant(block_with_constants, out_value, out_type); + RPNBuilderTreeContext tree_context(nullptr, block_with_constants, nullptr); + RPNBuilderTreeNode node(expr.get(), tree_context); + + return node.tryGetConstant(out_value, out_type); } @@ -1098,39 +793,6 @@ static FieldRef applyFunction(const FunctionBasePtr & func, const DataTypePtr & return {field.columns, field.row_idx, result_idx}; } -void KeyCondition::traverseAST(const Tree & node, ContextPtr context, Block & block_with_constants) -{ - RPNElement element; - - if (node.isFunction()) - { - auto func = node.asFunction(); - if (tryParseLogicalOperatorFromAST(func, element)) - { - size_t num_args = func.numArguments(); - for (size_t i = 0; i < num_args; ++i) - { - traverseAST(func.getArgumentAt(i), context, block_with_constants); - - /** The first part of the condition is for the correct support of `and` and `or` functions of arbitrary arity - * - in this case `n - 1` elements are added (where `n` is the number of arguments). - */ - if (i != 0 || element.function == RPNElement::FUNCTION_NOT) - rpn.emplace_back(element); - } - - return; - } - } - - if (!tryParseAtomFromAST(node, context, block_with_constants, element)) - { - element.function = RPNElement::FUNCTION_UNKNOWN; - } - - rpn.emplace_back(std::move(element)); -} - /** The key functional expression constraint may be inferred from a plain column in the expression. * For example, if the key contains `toStartOfHour(Timestamp)` and query contains `WHERE Timestamp >= now()`, * it can be assumed that if `toStartOfHour()` is monotonic on [now(), inf), the `toStartOfHour(Timestamp) >= toStartOfHour(now())` @@ -1251,7 +913,7 @@ bool KeyCondition::transformConstantWithValidFunctions( } bool KeyCondition::canConstantBeWrappedByMonotonicFunctions( - const Tree & node, + const RPNBuilderTreeNode & node, size_t & out_key_column_num, DataTypePtr & out_key_column_type, Field & out_value, @@ -1259,7 +921,7 @@ bool KeyCondition::canConstantBeWrappedByMonotonicFunctions( { String expr_name = node.getColumnName(); - if (array_joined_columns.contains(expr_name)) + if (array_joined_column_names.contains(expr_name)) return false; if (!key_subexpr_names.contains(expr_name)) @@ -1286,11 +948,15 @@ bool KeyCondition::canConstantBeWrappedByMonotonicFunctions( /// Looking for possible transformation of `column = constant` into `partition_expr = function(constant)` bool KeyCondition::canConstantBeWrappedByFunctions( - const Tree & node, size_t & out_key_column_num, DataTypePtr & out_key_column_type, Field & out_value, DataTypePtr & out_type) + const RPNBuilderTreeNode & node, + size_t & out_key_column_num, + DataTypePtr & out_key_column_type, + Field & out_value, + DataTypePtr & out_type) { String expr_name = node.getColumnName(); - if (array_joined_columns.contains(expr_name)) + if (array_joined_column_names.contains(expr_name)) return false; if (!key_subexpr_names.contains(expr_name)) @@ -1304,7 +970,7 @@ bool KeyCondition::canConstantBeWrappedByFunctions( /// The case `f(modulo(...))` for totally monotonic `f ` is considered to be rare. /// /// Note: for negative values, we can filter more partitions then needed. - expr_name = node.getColumnNameLegacy(); + expr_name = node.getColumnNameWithModuloLegacy(); if (!key_subexpr_names.contains(expr_name)) return false; @@ -1321,8 +987,7 @@ bool KeyCondition::canConstantBeWrappedByFunctions( } bool KeyCondition::tryPrepareSetIndex( - const FunctionTree & func, - ContextPtr context, + const RPNBuilderFunctionTreeNode & func, RPNElement & out, size_t & out_key_column_num) { @@ -1332,13 +997,12 @@ bool KeyCondition::tryPrepareSetIndex( std::vector indexes_mapping; DataTypes data_types; - auto get_key_tuple_position_mapping = [&](const Tree & node, size_t tuple_index) + auto get_key_tuple_position_mapping = [&](const RPNBuilderTreeNode & node, size_t tuple_index) { MergeTreeSetIndex::KeyTuplePositionMapping index_mapping; index_mapping.tuple_index = tuple_index; DataTypePtr data_type; - if (isKeyPossiblyWrappedByMonotonicFunctions( - node, context, index_mapping.key_index, data_type, index_mapping.functions)) + if (isKeyPossiblyWrappedByMonotonicFunctions(node, index_mapping.key_index, data_type, index_mapping.functions)) { indexes_mapping.push_back(index_mapping); data_types.push_back(data_type); @@ -1352,25 +1016,29 @@ bool KeyCondition::tryPrepareSetIndex( { /// Note: in case of ActionsDAG, tuple may be a constant. /// In this case, there is no keys in tuple. So, we don't have to check it. - auto left_arg_tuple = left_arg.asFunction(); + auto left_arg_tuple = left_arg.toFunctionNode(); if (left_arg_tuple.getFunctionName() == "tuple") { - left_args_count = left_arg_tuple.numArguments(); + left_args_count = left_arg_tuple.getArgumentsSize(); for (size_t i = 0; i < left_args_count; ++i) get_key_tuple_position_mapping(left_arg_tuple.getArgumentAt(i), i); } else + { get_key_tuple_position_mapping(left_arg, 0); + } } else + { get_key_tuple_position_mapping(left_arg, 0); + } if (indexes_mapping.empty()) return false; const auto right_arg = func.getArgumentAt(1); - auto prepared_set = right_arg.tryGetPreparedSet(prepared_sets, indexes_mapping, data_types); + auto prepared_set = right_arg.tryGetPreparedSet(indexes_mapping, data_types); if (!prepared_set) return false; @@ -1462,13 +1130,12 @@ private: bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctions( - const Tree & node, - ContextPtr context, + const RPNBuilderTreeNode & node, size_t & out_key_column_num, DataTypePtr & out_key_res_column_type, MonotonicFunctionsChain & out_functions_chain) { - std::vector chain_not_tested_for_monotonicity; + std::vector chain_not_tested_for_monotonicity; DataTypePtr key_column_type; if (!isKeyPossiblyWrappedByMonotonicFunctionsImpl(node, out_key_column_num, key_column_type, chain_not_tested_for_monotonicity)) @@ -1477,17 +1144,17 @@ bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctions( for (auto it = chain_not_tested_for_monotonicity.rbegin(); it != chain_not_tested_for_monotonicity.rend(); ++it) { auto function = *it; - auto func_builder = FunctionFactory::instance().tryGet(function.getFunctionName(), context); + auto func_builder = FunctionFactory::instance().tryGet(function.getFunctionName(), node.getTreeContext().getQueryContext()); if (!func_builder) return false; ColumnsWithTypeAndName arguments; ColumnWithTypeAndName const_arg; FunctionWithOptionalConstArg::Kind kind = FunctionWithOptionalConstArg::Kind::NO_CONST; - if (function.numArguments() == 2) + if (function.getArgumentsSize() == 2) { if (function.getArgumentAt(0).isConstant()) { - const_arg = function.getArgumentAt(0).getConstant(); + const_arg = function.getArgumentAt(0).getConstantColumn(); arguments.push_back(const_arg); arguments.push_back({ nullptr, key_column_type, "" }); kind = FunctionWithOptionalConstArg::Kind::LEFT_CONST; @@ -1495,7 +1162,7 @@ bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctions( else if (function.getArgumentAt(1).isConstant()) { arguments.push_back({ nullptr, key_column_type, "" }); - const_arg = function.getArgumentAt(1).getConstant(); + const_arg = function.getArgumentAt(1).getConstantColumn(); arguments.push_back(const_arg); kind = FunctionWithOptionalConstArg::Kind::RIGHT_CONST; } @@ -1521,10 +1188,10 @@ bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctions( } bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctionsImpl( - const Tree & node, + const RPNBuilderTreeNode & node, size_t & out_key_column_num, DataTypePtr & out_key_column_type, - std::vector & out_functions_chain) + std::vector & out_functions_chain) { /** By itself, the key column can be a functional expression. for example, `intHash32(UserID)`. * Therefore, use the full name of the expression for search. @@ -1534,7 +1201,7 @@ bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctionsImpl( // Key columns should use canonical names for index analysis String name = node.getColumnName(); - if (array_joined_columns.contains(name)) + if (array_joined_column_names.contains(name)) return false; auto it = key_columns.find(name); @@ -1547,37 +1214,39 @@ bool KeyCondition::isKeyPossiblyWrappedByMonotonicFunctionsImpl( if (node.isFunction()) { - auto func = node.asFunction(); + auto function_node = node.toFunctionNode(); - size_t num_args = func.numArguments(); - if (num_args > 2 || num_args == 0) + size_t arguments_size = function_node.getArgumentsSize(); + if (arguments_size > 2 || arguments_size == 0) return false; - out_functions_chain.push_back(func); - bool ret = false; - if (num_args == 2) + out_functions_chain.push_back(function_node); + + bool result = false; + if (arguments_size == 2) { - if (func.getArgumentAt(0).isConstant()) + if (function_node.getArgumentAt(0).isConstant()) { - ret = isKeyPossiblyWrappedByMonotonicFunctionsImpl(func.getArgumentAt(1), out_key_column_num, out_key_column_type, out_functions_chain); + result = isKeyPossiblyWrappedByMonotonicFunctionsImpl(function_node.getArgumentAt(1), out_key_column_num, out_key_column_type, out_functions_chain); } - else if (func.getArgumentAt(1).isConstant()) + else if (function_node.getArgumentAt(1).isConstant()) { - ret = isKeyPossiblyWrappedByMonotonicFunctionsImpl(func.getArgumentAt(0), out_key_column_num, out_key_column_type, out_functions_chain); + result = isKeyPossiblyWrappedByMonotonicFunctionsImpl(function_node.getArgumentAt(0), out_key_column_num, out_key_column_type, out_functions_chain); } } else { - ret = isKeyPossiblyWrappedByMonotonicFunctionsImpl(func.getArgumentAt(0), out_key_column_num, out_key_column_type, out_functions_chain); + result = isKeyPossiblyWrappedByMonotonicFunctionsImpl(function_node.getArgumentAt(0), out_key_column_num, out_key_column_type, out_functions_chain); } - return ret; + + return result; } return false; } -static void castValueToType(const DataTypePtr & desired_type, Field & src_value, const DataTypePtr & src_type, const KeyCondition::Tree & node) +static void castValueToType(const DataTypePtr & desired_type, Field & src_value, const DataTypePtr & src_type, const String & node_column_name) { try { @@ -1587,13 +1256,13 @@ static void castValueToType(const DataTypePtr & desired_type, Field & src_value, { throw Exception("Key expression contains comparison between inconvertible types: " + desired_type->getName() + " and " + src_type->getName() + - " inside " + node.getColumnName(), + " inside " + node_column_name, ErrorCodes::BAD_TYPE_OF_FIELD); } } -bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Block & block_with_constants, RPNElement & out) +bool KeyCondition::extractAtomFromTree(const RPNBuilderTreeNode & node, RPNElement & out) { /** Functions < > = != <= >= in `notIn` isNull isNotNull, where one argument is a constant, and the other is one of columns of key, * or itself, wrapped in a chain of possibly-monotonic functions, @@ -1603,8 +1272,8 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl DataTypePtr const_type; if (node.isFunction()) { - auto func = node.asFunction(); - size_t num_args = func.numArguments(); + auto func = node.toFunctionNode(); + size_t num_args = func.getArgumentsSize(); DataTypePtr key_expr_type; /// Type of expression containing key column size_t key_column_num = -1; /// Number of a key column (inside key_column_names array) @@ -1616,7 +1285,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl if (num_args == 1) { - if (!(isKeyPossiblyWrappedByMonotonicFunctions(func.getArgumentAt(0), context, key_column_num, key_expr_type, chain))) + if (!(isKeyPossiblyWrappedByMonotonicFunctions(func.getArgumentAt(0), key_column_num, key_expr_type, chain))) return false; if (key_column_num == static_cast(-1)) @@ -1647,7 +1316,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl if (functionIsInOrGlobalInOperator(func_name)) { - if (tryPrepareSetIndex(func, context, out, key_column_num)) + if (tryPrepareSetIndex(func, out, key_column_num)) { key_arg_pos = 0; is_set_const = true; @@ -1655,7 +1324,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl else return false; } - else if (func.getArgumentAt(1).tryGetConstant(block_with_constants, const_value, const_type)) + else if (func.getArgumentAt(1).tryGetConstant(const_value, const_type)) { /// If the const operand is null, the atom will be always false if (const_value.isNull()) @@ -1664,7 +1333,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl return true; } - if (isKeyPossiblyWrappedByMonotonicFunctions(func.getArgumentAt(0), context, key_column_num, key_expr_type, chain)) + if (isKeyPossiblyWrappedByMonotonicFunctions(func.getArgumentAt(0), key_column_num, key_expr_type, chain)) { key_arg_pos = 0; } @@ -1685,7 +1354,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl else return false; } - else if (func.getArgumentAt(0).tryGetConstant(block_with_constants, const_value, const_type)) + else if (func.getArgumentAt(0).tryGetConstant(const_value, const_type)) { /// If the const operand is null, the atom will be always false if (const_value.isNull()) @@ -1694,7 +1363,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl return true; } - if (isKeyPossiblyWrappedByMonotonicFunctions(func.getArgumentAt(1), context, key_column_num, key_expr_type, chain)) + if (isKeyPossiblyWrappedByMonotonicFunctions(func.getArgumentAt(1), key_column_num, key_expr_type, chain)) { key_arg_pos = 1; } @@ -1774,7 +1443,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl if (!const_type->equals(*common_type)) { - castValueToType(common_type, const_value, const_type, node); + castValueToType(common_type, const_value, const_type, node.getColumnName()); // Need to set is_constant_transformed unless we're doing exact conversion if (!key_expr_type_not_null->equals(*common_type)) @@ -1819,7 +1488,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl return atom_it->second(out, const_value); } - else if (node.tryGetConstant(block_with_constants, const_value, const_type)) + else if (node.tryGetConstant(const_value, const_type)) { /// For cases where it says, for example, `WHERE 0 AND something` @@ -1842,32 +1511,6 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl return false; } -bool KeyCondition::tryParseLogicalOperatorFromAST(const FunctionTree & func, RPNElement & out) -{ - /// Functions AND, OR, NOT. - /// Also a special function `indexHint` - works as if instead of calling a function there are just parentheses - /// (or, the same thing - calling the function `and` from one argument). - - if (func.getFunctionName() == "not") - { - if (func.numArguments() != 1) - return false; - - out.function = RPNElement::FUNCTION_NOT; - } - else - { - if (func.getFunctionName() == "and" || func.getFunctionName() == "indexHint") - out.function = RPNElement::FUNCTION_AND; - else if (func.getFunctionName() == "or") - out.function = RPNElement::FUNCTION_OR; - else - return false; - } - - return true; -} - String KeyCondition::toString() const { String res; diff --git a/src/Storages/MergeTree/KeyCondition.h b/src/Storages/MergeTree/KeyCondition.h index d00a25a1077..fe1bffa9305 100644 --- a/src/Storages/MergeTree/KeyCondition.h +++ b/src/Storages/MergeTree/KeyCondition.h @@ -2,11 +2,16 @@ #include -#include #include -#include -#include +#include + +#include +#include +#include + +#include +#include namespace DB { @@ -205,45 +210,37 @@ public: class KeyCondition { public: - /// Does not take into account the SAMPLE section. all_columns - the set of all columns of the table. + /// Construct key condition from AST SELECT query WHERE, PREWHERE and additional filters KeyCondition( const ASTPtr & query, const ASTs & additional_filter_asts, - TreeRewriterResultPtr syntax_analyzer_result, + Block block_with_constants, PreparedSetsPtr prepared_sets_, ContextPtr context, const Names & key_column_names, const ExpressionActionsPtr & key_expr, + NameSet array_joined_column_names, bool single_point_ = false, bool strict_ = false); + /** Construct key condition from AST SELECT query WHERE, PREWHERE and additional filters. + * Select query, additional filters, prepared sets are initialized using query info. + */ KeyCondition( const SelectQueryInfo & query_info, ContextPtr context, const Names & key_column_names, const ExpressionActionsPtr & key_expr_, bool single_point_ = false, - bool strict_ = false) - : KeyCondition( - query_info.query, - query_info.filter_asts, - query_info.syntax_analyzer_result, - query_info.prepared_sets, - context, - key_column_names, - key_expr_, - single_point_, - strict_) - { - } + bool strict_ = false); + /// Construct key condition from ActionsDAG nodes KeyCondition( ActionDAGNodes dag_nodes, - TreeRewriterResultPtr syntax_analyzer_result, - PreparedSetsPtr prepared_sets_, ContextPtr context, const Names & key_column_names, const ExpressionActionsPtr & key_expr, + NameSet array_joined_column_names, bool single_point_ = false, bool strict_ = false); @@ -275,6 +272,7 @@ public: /// Checks that the index can not be used /// FUNCTION_UNKNOWN will be AND'ed (if any). bool alwaysUnknownOrTrue() const; + /// Checks that the index can not be used /// Does not allow any FUNCTION_UNKNOWN (will instantly return true). bool anyUnknownOrAlwaysTrue() const; @@ -313,10 +311,18 @@ public: * Returns false, if expression isn't constant. */ static bool getConstant( - const ASTPtr & expr, Block & block_with_constants, Field & out_value, DataTypePtr & out_type); + const ASTPtr & expr, + Block & block_with_constants, + Field & out_value, + DataTypePtr & out_type); + /** Calculate expressions, that depend only on constants. + * For index to work when something like "WHERE Date = toDate(now())" is written. + */ static Block getBlockWithConstants( - const ASTPtr & query, const TreeRewriterResultPtr & syntax_analyzer_result, ContextPtr context); + const ASTPtr & query, + const TreeRewriterResultPtr & syntax_analyzer_result, + ContextPtr context); static std::optional applyMonotonicFunctionsChainToRange( Range key_range, @@ -373,14 +379,11 @@ private: using RPN = std::vector; using ColumnIndices = std::map; - using AtomMap = std::unordered_map; public: + using AtomMap = std::unordered_map; static const AtomMap atom_map; - class Tree; - class FunctionTree; - private: BoolMask checkInRange( size_t used_key_size, @@ -390,9 +393,7 @@ private: bool right_bounded, BoolMask initial_mask) const; - void traverseAST(const Tree & node, ContextPtr context, Block & block_with_constants); - bool tryParseAtomFromAST(const Tree & node, ContextPtr context, Block & block_with_constants, RPNElement & out); - static bool tryParseLogicalOperatorFromAST(const FunctionTree & func, RPNElement & out); + bool extractAtomFromTree(const RPNBuilderTreeNode & node, RPNElement & out); /** Is node the key column * or expression in which column of key is wrapped by chain of functions, @@ -401,17 +402,16 @@ private: * and fills chain of possibly-monotonic functions. */ bool isKeyPossiblyWrappedByMonotonicFunctions( - const Tree & node, - ContextPtr context, + const RPNBuilderTreeNode & node, size_t & out_key_column_num, DataTypePtr & out_key_res_column_type, MonotonicFunctionsChain & out_functions_chain); bool isKeyPossiblyWrappedByMonotonicFunctionsImpl( - const Tree & node, + const RPNBuilderTreeNode & node, size_t & out_key_column_num, DataTypePtr & out_key_column_type, - std::vector & out_functions_chain); + std::vector & out_functions_chain); bool transformConstantWithValidFunctions( const String & expr_name, @@ -422,21 +422,24 @@ private: std::function always_monotonic) const; bool canConstantBeWrappedByMonotonicFunctions( - const Tree & node, + const RPNBuilderTreeNode & node, size_t & out_key_column_num, DataTypePtr & out_key_column_type, Field & out_value, DataTypePtr & out_type); bool canConstantBeWrappedByFunctions( - const Tree & node, size_t & out_key_column_num, DataTypePtr & out_key_column_type, Field & out_value, DataTypePtr & out_type); + const RPNBuilderTreeNode & node, + size_t & out_key_column_num, + DataTypePtr & out_key_column_type, + Field & out_value, + DataTypePtr & out_type); /// If it's possible to make an RPNElement /// that will filter values (possibly tuples) by the content of 'prepared_set', /// do it and return true. bool tryPrepareSetIndex( - const FunctionTree & func, - ContextPtr context, + const RPNBuilderFunctionTreeNode & func, RPNElement & out, size_t & out_key_column_num); @@ -472,11 +475,12 @@ private: /// All intermediate columns are used to calculate key_expr. const NameSet key_subexpr_names; - NameSet array_joined_columns; - PreparedSetsPtr prepared_sets; + /// Array joined column names + NameSet array_joined_column_names; // If true, always allow key_expr to be wrapped by function bool single_point; + // If true, do not use always_monotonic information to transform constants bool strict; }; diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp index 3dd0568107e..db1b9762668 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp @@ -6,11 +6,13 @@ #include #include #include +#include #include #include #include #include #include +#include #include #include #include @@ -28,19 +30,7 @@ namespace ErrorCodes namespace { -PreparedSetKey getPreparedSetKey(const ASTPtr & node, const DataTypePtr & data_type) -{ - /// If the data type is tuple, let's try unbox once - if (node->as() || node->as()) - return PreparedSetKey::forSubquery(*node); - - if (const auto * date_type_tuple = typeid_cast(&*data_type)) - return PreparedSetKey::forLiteral(*node, date_type_tuple->getElements()); - - return PreparedSetKey::forLiteral(*node, DataTypes(1, data_type)); -} - -ColumnWithTypeAndName getPreparedSetInfo(const SetPtr & prepared_set) +ColumnWithTypeAndName getPreparedSetInfo(const ConstSetPtr & prepared_set) { if (prepared_set->getDataTypes().size() == 1) return {prepared_set->getSetElements()[0], prepared_set->getElementsTypes()[0], "dummy"}; @@ -110,8 +100,22 @@ MergeTreeIndexConditionBloomFilter::MergeTreeIndexConditionBloomFilter( const SelectQueryInfo & info_, ContextPtr context_, const Block & header_, size_t hash_functions_) : WithContext(context_), header(header_), query_info(info_), hash_functions(hash_functions_) { - auto atom_from_ast = [this](auto & node, auto, auto & constants, auto & out) { return traverseAtomAST(node, constants, out); }; - rpn = std::move(RPNBuilder(info_, getContext(), atom_from_ast).extractRPN()); + ASTPtr filter_node = buildFilterNode(query_info.query); + + if (!filter_node) + { + rpn.push_back(RPNElement::FUNCTION_UNKNOWN); + return; + } + + auto block_with_constants = KeyCondition::getBlockWithConstants(query_info.query, query_info.syntax_analyzer_result, context_); + RPNBuilder builder( + filter_node, + context_, + std::move(block_with_constants), + query_info.prepared_sets, + [&](const RPNBuilderTreeNode & node, RPNElement & out) { return extractAtomFromTree(node, out); }); + rpn = std::move(builder).extractRPN(); } bool MergeTreeIndexConditionBloomFilter::alwaysUnknownOrTrue() const @@ -235,12 +239,13 @@ bool MergeTreeIndexConditionBloomFilter::mayBeTrueOnGranule(const MergeTreeIndex return rpn_stack[0].can_be_true; } -bool MergeTreeIndexConditionBloomFilter::traverseAtomAST(const ASTPtr & node, Block & block_with_constants, RPNElement & out) +bool MergeTreeIndexConditionBloomFilter::extractAtomFromTree(const RPNBuilderTreeNode & node, RPNElement & out) { { Field const_value; DataTypePtr const_type; - if (KeyCondition::getConstant(node, block_with_constants, const_value, const_type)) + + if (node.tryGetConstant(const_value, const_type)) { if (const_value.getType() == Field::Types::UInt64) { @@ -262,56 +267,57 @@ bool MergeTreeIndexConditionBloomFilter::traverseAtomAST(const ASTPtr & node, Bl } } - return traverseFunction(node, block_with_constants, out, nullptr); + return traverseFunction(node, out, nullptr /*parent*/); } -bool MergeTreeIndexConditionBloomFilter::traverseFunction(const ASTPtr & node, Block & block_with_constants, RPNElement & out, const ASTPtr & parent) +bool MergeTreeIndexConditionBloomFilter::traverseFunction(const RPNBuilderTreeNode & node, RPNElement & out, const RPNBuilderTreeNode * parent) { bool maybe_useful = false; - if (const auto * function = node->as()) + if (node.isFunction()) { - if (!function->arguments) + const auto function = node.toFunctionNode(); + auto arguments_size = function.getArgumentsSize(); + auto function_name = function.getFunctionName(); + + if (arguments_size != 2) return false; - const ASTs & arguments = function->arguments->children; - for (const auto & arg : arguments) + auto lhs_argument = function.getArgumentAt(0); + auto rhs_argument = function.getArgumentAt(1); + + auto lhs_argument_column_name = lhs_argument.getColumnName(); + + if (functionIsInOrGlobalInOperator(function_name)) { - if (traverseFunction(arg, block_with_constants, out, node)) - maybe_useful = true; - } + ConstSetPtr prepared_set = rhs_argument.tryGetPreparedSet(); - if (arguments.size() != 2) - return false; - - if (functionIsInOrGlobalInOperator(function->name)) - { - auto prepared_set = getPreparedSet(arguments[1]); - - if (prepared_set) + if (prepared_set && prepared_set->hasExplicitSetElements()) { - if (traverseASTIn(function->name, arguments[0], prepared_set, out)) + const auto prepared_info = getPreparedSetInfo(prepared_set); + if (traverseTreeIn(function_name, lhs_argument, prepared_set, prepared_info.type, prepared_info.column, out)) maybe_useful = true; } } - else if (function->name == "equals" || - function->name == "notEquals" || - function->name == "has" || - function->name == "mapContains" || - function->name == "indexOf" || - function->name == "hasAny" || - function->name == "hasAll") + else if (function_name == "equals" || + function_name == "notEquals" || + function_name == "has" || + function_name == "mapContains" || + function_name == "indexOf" || + function_name == "hasAny" || + function_name == "hasAll") { Field const_value; DataTypePtr const_type; - if (KeyCondition::getConstant(arguments[1], block_with_constants, const_value, const_type)) + + if (rhs_argument.tryGetConstant(const_value, const_type)) { - if (traverseASTEquals(function->name, arguments[0], const_type, const_value, out, parent)) + if (traverseTreeEquals(function_name, lhs_argument, const_type, const_value, out, parent)) maybe_useful = true; } - else if (KeyCondition::getConstant(arguments[0], block_with_constants, const_value, const_type)) + else if (lhs_argument.tryGetConstant(const_value, const_type)) { - if (traverseASTEquals(function->name, arguments[1], const_type, const_value, out, parent)) + if (traverseTreeEquals(function_name, rhs_argument, const_type, const_value, out, parent)) maybe_useful = true; } } @@ -320,28 +326,20 @@ bool MergeTreeIndexConditionBloomFilter::traverseFunction(const ASTPtr & node, B return maybe_useful; } -bool MergeTreeIndexConditionBloomFilter::traverseASTIn( +bool MergeTreeIndexConditionBloomFilter::traverseTreeIn( const String & function_name, - const ASTPtr & key_ast, - const SetPtr & prepared_set, - RPNElement & out) -{ - const auto prepared_info = getPreparedSetInfo(prepared_set); - return traverseASTIn(function_name, key_ast, prepared_set, prepared_info.type, prepared_info.column, out); -} - -bool MergeTreeIndexConditionBloomFilter::traverseASTIn( - const String & function_name, - const ASTPtr & key_ast, - const SetPtr & prepared_set, + const RPNBuilderTreeNode & key_node, + const ConstSetPtr & prepared_set, const DataTypePtr & type, const ColumnPtr & column, RPNElement & out) { - if (header.has(key_ast->getColumnName())) + auto key_node_column_name = key_node.getColumnName(); + + if (header.has(key_node_column_name)) { size_t row_size = column->size(); - size_t position = header.getPositionByName(key_ast->getColumnName()); + size_t position = header.getPositionByName(key_node_column_name); const DataTypePtr & index_type = header.getByPosition(position).type; const auto & converted_column = castColumn(ColumnWithTypeAndName{column, type, ""}, index_type); out.predicate.emplace_back(std::make_pair(position, BloomFilterHash::hashWithColumn(index_type, converted_column, 0, row_size))); @@ -355,30 +353,33 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTIn( return true; } - if (const auto * function = key_ast->as()) + if (key_node.isFunction()) { + auto key_node_function = key_node.toFunctionNode(); + auto key_node_function_name = key_node_function.getFunctionName(); + size_t key_node_function_arguments_size = key_node_function.getArgumentsSize(); + WhichDataType which(type); - if (which.isTuple() && function->name == "tuple") + if (which.isTuple() && function_name == "tuple") { const auto & tuple_column = typeid_cast(column.get()); const auto & tuple_data_type = typeid_cast(type.get()); - const ASTs & arguments = typeid_cast(*function->arguments).children; - if (tuple_data_type->getElements().size() != arguments.size() || tuple_column->getColumns().size() != arguments.size()) + if (tuple_data_type->getElements().size() != key_node_function_arguments_size || tuple_column->getColumns().size() != key_node_function_arguments_size) throw Exception("Illegal types of arguments of function " + function_name, ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); bool match_with_subtype = false; const auto & sub_columns = tuple_column->getColumns(); const auto & sub_data_types = tuple_data_type->getElements(); - for (size_t index = 0; index < arguments.size(); ++index) - match_with_subtype |= traverseASTIn(function_name, arguments[index], nullptr, sub_data_types[index], sub_columns[index], out); + for (size_t index = 0; index < key_node_function_arguments_size; ++index) + match_with_subtype |= traverseTreeIn(function_name, key_node_function.getArgumentAt(index), nullptr, sub_data_types[index], sub_columns[index], out); return match_with_subtype; } - if (function->name == "arrayElement") + if (key_node_function_name == "arrayElement") { /** Try to parse arrayElement for mapKeys index. * It is important to ignore keys like column_map['Key'] IN ('') because if key does not exists in map @@ -388,9 +389,6 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTIn( * that way we skip necessary granules where map key does not exists. */ - if (!prepared_set) - return false; - auto default_column_to_check = type->createColumnConstWithDefaultValue(1)->convertToFullColumnIfConst(); ColumnWithTypeAndName default_column_with_type_to_check { default_column_to_check, type, "" }; ColumnsWithTypeAndName default_columns_with_type_to_check = {default_column_with_type_to_check}; @@ -400,11 +398,7 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTIn( if (set_contain_default_value) return false; - const auto * column_ast_identifier = function->arguments.get()->children[0].get()->as(); - if (!column_ast_identifier) - return false; - - const auto & col_name = column_ast_identifier->name(); + const auto & col_name = key_node_function.getArgumentAt(0).getColumnName(); auto map_keys_index_column_name = fmt::format("mapKeys({})", col_name); auto map_values_index_column_name = fmt::format("mapValues({})", col_name); @@ -412,16 +406,17 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTIn( { /// For mapKeys we serialize key argument with bloom filter - auto & argument = function->arguments.get()->children[1]; + auto first_argument = key_node_function.getArgumentAt(1); - if (const auto * literal = argument->as()) + Field constant_value; + DataTypePtr constant_type; + + if (first_argument.tryGetConstant(constant_value, constant_type)) { size_t position = header.getPositionByName(map_keys_index_column_name); const DataTypePtr & index_type = header.getByPosition(position).type; - - auto element_key = literal->value; const DataTypePtr actual_type = BloomFilter::getPrimitiveType(index_type); - out.predicate.emplace_back(std::make_pair(position, BloomFilterHash::hashWithField(actual_type.get(), element_key))); + out.predicate.emplace_back(std::make_pair(position, BloomFilterHash::hashWithField(actual_type.get(), constant_value))); } else { @@ -459,74 +454,97 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTIn( } -static bool indexOfCanUseBloomFilter(const ASTPtr & parent) +static bool indexOfCanUseBloomFilter(const RPNBuilderTreeNode * parent) { if (!parent) return true; + if (!parent->isFunction()) + return false; + + auto function = parent->toFunctionNode(); + auto function_name = function.getFunctionName(); + /// `parent` is a function where `indexOf` is located. /// Example: `indexOf(arr, x) = 1`, parent is a function named `equals`. - if (const auto * function = parent->as()) + if (function_name == "and") { - if (function->name == "and") + return true; + } + else if (function_name == "equals" /// notEquals is not applicable + || function_name == "greater" || function_name == "greaterOrEquals" + || function_name == "less" || function_name == "lessOrEquals") + { + size_t function_arguments_size = function.getArgumentsSize(); + if (function_arguments_size != 2) + return false; + + /// We don't allow constant expressions like `indexOf(arr, x) = 1 + 0` but it's negligible. + + /// We should return true when the corresponding expression implies that the array contains the element. + /// Example: when `indexOf(arr, x)` > 10 is written, it means that arr definitely should contain the element + /// (at least at 11th position but it does not matter). + + bool reversed = false; + Field constant_value; + DataTypePtr constant_type; + + if (function.getArgumentAt(0).tryGetConstant(constant_value, constant_type)) { + reversed = true; + } + else if (function.getArgumentAt(1).tryGetConstant(constant_value, constant_type)) + { + } + else + { + return false; + } + + Field zero(0); + bool constant_equal_zero = applyVisitor(FieldVisitorAccurateEquals(), constant_value, zero); + + if (function_name == "equals" && !constant_equal_zero) + { + /// indexOf(...) = c, c != 0 return true; } - else if (function->name == "equals" /// notEquals is not applicable - || function->name == "greater" || function->name == "greaterOrEquals" - || function->name == "less" || function->name == "lessOrEquals") + else if (function_name == "notEquals" && constant_equal_zero) { - if (function->arguments->children.size() != 2) - return false; - - /// We don't allow constant expressions like `indexOf(arr, x) = 1 + 0` but it's negligible. - - /// We should return true when the corresponding expression implies that the array contains the element. - /// Example: when `indexOf(arr, x)` > 10 is written, it means that arr definitely should contain the element - /// (at least at 11th position but it does not matter). - - bool reversed = false; - const ASTLiteral * constant = nullptr; - - if (const ASTLiteral * left = function->arguments->children[0]->as()) - { - constant = left; - reversed = true; - } - else if (const ASTLiteral * right = function->arguments->children[1]->as()) - { - constant = right; - } - else - return false; - - Field zero(0); - return (function->name == "equals" /// indexOf(...) = c, c != 0 - && !applyVisitor(FieldVisitorAccurateEquals(), constant->value, zero)) - || (function->name == "notEquals" /// indexOf(...) != c, c = 0 - && applyVisitor(FieldVisitorAccurateEquals(), constant->value, zero)) - || (function->name == (reversed ? "less" : "greater") /// indexOf(...) > c, c >= 0 - && !applyVisitor(FieldVisitorAccurateLess(), constant->value, zero)) - || (function->name == (reversed ? "lessOrEquals" : "greaterOrEquals") /// indexOf(...) >= c, c > 0 - && applyVisitor(FieldVisitorAccurateLess(), zero, constant->value)); + /// indexOf(...) != c, c = 0 + return true; } + else if (function_name == (reversed ? "less" : "greater") && !applyVisitor(FieldVisitorAccurateLess(), constant_value, zero)) + { + /// indexOf(...) > c, c >= 0 + return true; + } + else if (function_name == (reversed ? "lessOrEquals" : "greaterOrEquals") && applyVisitor(FieldVisitorAccurateLess(), zero, constant_value)) + { + /// indexOf(...) >= c, c > 0 + return true; + } + + return false; } return false; } -bool MergeTreeIndexConditionBloomFilter::traverseASTEquals( +bool MergeTreeIndexConditionBloomFilter::traverseTreeEquals( const String & function_name, - const ASTPtr & key_ast, + const RPNBuilderTreeNode & key_node, const DataTypePtr & value_type, const Field & value_field, RPNElement & out, - const ASTPtr & parent) + const RPNBuilderTreeNode * parent) { - if (header.has(key_ast->getColumnName())) + auto key_column_name = key_node.getColumnName(); + + if (header.has(key_column_name)) { - size_t position = header.getPositionByName(key_ast->getColumnName()); + size_t position = header.getPositionByName(key_column_name); const DataTypePtr & index_type = header.getByPosition(position).type; const auto * array_type = typeid_cast(index_type.get()); @@ -602,13 +620,7 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTEquals( if (function_name == "mapContains" || function_name == "has") { - const auto * key_ast_identifier = key_ast.get()->as(); - if (!key_ast_identifier) - return false; - - const auto & col_name = key_ast_identifier->name(); - auto map_keys_index_column_name = fmt::format("mapKeys({})", col_name); - + auto map_keys_index_column_name = fmt::format("mapKeys({})", key_column_name); if (!header.has(map_keys_index_column_name)) return false; @@ -629,29 +641,32 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTEquals( return true; } - if (const auto * function = key_ast->as()) + if (key_node.isFunction()) { WhichDataType which(value_type); - if (which.isTuple() && function->name == "tuple") + auto key_node_function = key_node.toFunctionNode(); + auto key_node_function_name = key_node_function.getFunctionName(); + size_t key_node_function_arguments_size = key_node_function.getArgumentsSize(); + + if (which.isTuple() && key_node_function_name == "tuple") { const Tuple & tuple = value_field.get(); const auto * value_tuple_data_type = typeid_cast(value_type.get()); - const ASTs & arguments = typeid_cast(*function->arguments).children; - if (tuple.size() != arguments.size()) + if (tuple.size() != key_node_function_arguments_size) throw Exception("Illegal types of arguments of function " + function_name, ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); bool match_with_subtype = false; const DataTypes & subtypes = value_tuple_data_type->getElements(); for (size_t index = 0; index < tuple.size(); ++index) - match_with_subtype |= traverseASTEquals(function_name, arguments[index], subtypes[index], tuple[index], out, key_ast); + match_with_subtype |= traverseTreeEquals(function_name, key_node_function.getArgumentAt(index), subtypes[index], tuple[index], out, &key_node); return match_with_subtype; } - if (function->name == "arrayElement" && (function_name == "equals" || function_name == "notEquals")) + if (key_node_function_name == "arrayElement" && (function_name == "equals" || function_name == "notEquals")) { /** Try to parse arrayElement for mapKeys index. * It is important to ignore keys like column_map['Key'] = '' because if key does not exists in map @@ -663,27 +678,22 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTEquals( if (value_field == value_type->getDefault()) return false; - const auto * column_ast_identifier = function->arguments.get()->children[0].get()->as(); - if (!column_ast_identifier) - return false; - - const auto & col_name = column_ast_identifier->name(); + auto first_argument = key_node_function.getArgumentAt(0); + const auto col_name = first_argument.getColumnName(); auto map_keys_index_column_name = fmt::format("mapKeys({})", col_name); auto map_values_index_column_name = fmt::format("mapValues({})", col_name); size_t position = 0; Field const_value = value_field; + DataTypePtr const_type; if (header.has(map_keys_index_column_name)) { position = header.getPositionByName(map_keys_index_column_name); + auto second_argument = key_node_function.getArgumentAt(1); - auto & argument = function->arguments.get()->children[1]; - - if (const auto * literal = argument->as()) - const_value = literal->value; - else + if (!second_argument.tryGetConstant(const_value, const_type)) return false; } else if (header.has(map_values_index_column_name)) @@ -708,23 +718,4 @@ bool MergeTreeIndexConditionBloomFilter::traverseASTEquals( return false; } -SetPtr MergeTreeIndexConditionBloomFilter::getPreparedSet(const ASTPtr & node) -{ - if (header.has(node->getColumnName())) - { - const auto & column_and_type = header.getByName(node->getColumnName()); - auto set_key = getPreparedSetKey(node, column_and_type.type); - if (auto prepared_set = query_info.prepared_sets->get(set_key)) - return prepared_set; - } - else - { - for (const auto & set : query_info.prepared_sets->getByTreeHash(node->getTreeHash())) - if (set->hasExplicitSetElements()) - return set; - } - - return DB::SetPtr(); -} - } diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h index 27fd701c67b..5d7ea371a83 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.h @@ -62,35 +62,27 @@ private: const size_t hash_functions; std::vector rpn; - SetPtr getPreparedSet(const ASTPtr & node); - bool mayBeTrueOnGranule(const MergeTreeIndexGranuleBloomFilter * granule) const; - bool traverseAtomAST(const ASTPtr & node, Block & block_with_constants, RPNElement & out); + bool extractAtomFromTree(const RPNBuilderTreeNode & node, RPNElement & out); - bool traverseFunction(const ASTPtr & node, Block & block_with_constants, RPNElement & out, const ASTPtr & parent); + bool traverseFunction(const RPNBuilderTreeNode & node, RPNElement & out, const RPNBuilderTreeNode * parent); - bool traverseASTIn( + bool traverseTreeIn( const String & function_name, - const ASTPtr & key_ast, - const SetPtr & prepared_set, - RPNElement & out); - - bool traverseASTIn( - const String & function_name, - const ASTPtr & key_ast, - const SetPtr & prepared_set, + const RPNBuilderTreeNode & key_node, + const ConstSetPtr & prepared_set, const DataTypePtr & type, const ColumnPtr & column, RPNElement & out); - bool traverseASTEquals( + bool traverseTreeEquals( const String & function_name, - const ASTPtr & key_ast, + const RPNBuilderTreeNode & key_node, const DataTypePtr & value_type, const Field & value_field, RPNElement & out, - const ASTPtr & parent); + const RPNBuilderTreeNode * parent); }; } diff --git a/src/Storages/MergeTree/MergeTreeIndexFullText.cpp b/src/Storages/MergeTree/MergeTreeIndexFullText.cpp index ff924290783..9074485a020 100644 --- a/src/Storages/MergeTree/MergeTreeIndexFullText.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexFullText.cpp @@ -11,9 +11,11 @@ #include #include #include +#include #include #include #include +#include #include #include @@ -148,13 +150,22 @@ MergeTreeConditionFullText::MergeTreeConditionFullText( , token_extractor(token_extactor_) , prepared_sets(query_info.prepared_sets) { - rpn = std::move( - RPNBuilder( - query_info, context, - [this] (const ASTPtr & node, ContextPtr /* context */, Block & block_with_constants, RPNElement & out) -> bool - { - return this->traverseAtomAST(node, block_with_constants, out); - }).extractRPN()); + ASTPtr filter_node = buildFilterNode(query_info.query); + + if (!filter_node) + { + rpn.push_back(RPNElement::FUNCTION_UNKNOWN); + return; + } + + auto block_with_constants = KeyCondition::getBlockWithConstants(query_info.query, query_info.syntax_analyzer_result, context); + RPNBuilder builder( + filter_node, + context, + std::move(block_with_constants), + query_info.prepared_sets, + [&](const RPNBuilderTreeNode & node, RPNElement & out) { return extractAtomFromTree(node, out); }); + rpn = std::move(builder).extractRPN(); } bool MergeTreeConditionFullText::alwaysUnknownOrTrue() const @@ -306,13 +317,13 @@ bool MergeTreeConditionFullText::getKey(const std::string & key_column_name, siz return true; } -bool MergeTreeConditionFullText::traverseAtomAST(const ASTPtr & node, Block & block_with_constants, RPNElement & out) +bool MergeTreeConditionFullText::extractAtomFromTree(const RPNBuilderTreeNode & node, RPNElement & out) { { Field const_value; DataTypePtr const_type; - if (KeyCondition::getConstant(node, block_with_constants, const_value, const_type)) + if (node.tryGetConstant(const_value, const_type)) { /// Check constant like in KeyCondition if (const_value.getType() == Field::Types::UInt64 @@ -329,53 +340,56 @@ bool MergeTreeConditionFullText::traverseAtomAST(const ASTPtr & node, Block & bl } } - if (const auto * function = node->as()) + if (node.isFunction()) { - if (!function->arguments) + auto function_node = node.toFunctionNode(); + auto function_name = function_node.getFunctionName(); + + size_t arguments_size = function_node.getArgumentsSize(); + if (arguments_size != 2) return false; - const ASTs & arguments = function->arguments->children; + auto left_argument = function_node.getArgumentAt(0); + auto right_argument = function_node.getArgumentAt(1); - if (arguments.size() != 2) - return false; - - if (functionIsInOrGlobalInOperator(function->name)) + if (functionIsInOrGlobalInOperator(function_name)) { - if (tryPrepareSetBloomFilter(arguments, out)) + if (tryPrepareSetBloomFilter(left_argument, right_argument, out)) { - if (function->name == "notIn") + if (function_name == "notIn") { out.function = RPNElement::FUNCTION_NOT_IN; return true; } - else if (function->name == "in") + else if (function_name == "in") { out.function = RPNElement::FUNCTION_IN; return true; } } } - else if (function->name == "equals" || - function->name == "notEquals" || - function->name == "has" || - function->name == "mapContains" || - function->name == "like" || - function->name == "notLike" || - function->name == "hasToken" || - function->name == "startsWith" || - function->name == "endsWith" || - function->name == "multiSearchAny") + else if (function_name == "equals" || + function_name == "notEquals" || + function_name == "has" || + function_name == "mapContains" || + function_name == "like" || + function_name == "notLike" || + function_name == "hasToken" || + function_name == "startsWith" || + function_name == "endsWith" || + function_name == "multiSearchAny") { Field const_value; DataTypePtr const_type; - if (KeyCondition::getConstant(arguments[1], block_with_constants, const_value, const_type)) + + if (right_argument.tryGetConstant(const_value, const_type)) { - if (traverseASTEquals(function->name, arguments[0], const_type, const_value, out)) + if (traverseTreeEquals(function_name, left_argument, const_type, const_value, out)) return true; } - else if (KeyCondition::getConstant(arguments[0], block_with_constants, const_value, const_type) && (function->name == "equals" || function->name == "notEquals")) + else if (left_argument.tryGetConstant(const_value, const_type) && (function_name == "equals" || function_name == "notEquals")) { - if (traverseASTEquals(function->name, arguments[1], const_type, const_value, out)) + if (traverseTreeEquals(function_name, right_argument, const_type, const_value, out)) return true; } } @@ -384,9 +398,9 @@ bool MergeTreeConditionFullText::traverseAtomAST(const ASTPtr & node, Block & bl return false; } -bool MergeTreeConditionFullText::traverseASTEquals( +bool MergeTreeConditionFullText::traverseTreeEquals( const String & function_name, - const ASTPtr & key_ast, + const RPNBuilderTreeNode & key_node, const DataTypePtr & value_type, const Field & value_field, RPNElement & out) @@ -397,13 +411,17 @@ bool MergeTreeConditionFullText::traverseASTEquals( Field const_value = value_field; + auto column_name = key_node.getColumnName(); size_t key_column_num = 0; - bool key_exists = getKey(key_ast->getColumnName(), key_column_num); - bool map_key_exists = getKey(fmt::format("mapKeys({})", key_ast->getColumnName()), key_column_num); + bool key_exists = getKey(column_name, key_column_num); + bool map_key_exists = getKey(fmt::format("mapKeys({})", column_name), key_column_num); - if (const auto * function = key_ast->as()) + if (key_node.isFunction()) { - if (function->name == "arrayElement") + auto key_function_node = key_node.toFunctionNode(); + auto key_function_node_function_name = key_function_node.getFunctionName(); + + if (key_function_node_function_name == "arrayElement") { /** Try to parse arrayElement for mapKeys index. * It is important to ignore keys like column_map['Key'] = '' because if key does not exists in map @@ -415,11 +433,8 @@ bool MergeTreeConditionFullText::traverseASTEquals( if (value_field == value_type->getDefault()) return false; - const auto * column_ast_identifier = function->arguments.get()->children[0].get()->as(); - if (!column_ast_identifier) - return false; - - const auto & map_column_name = column_ast_identifier->name(); + auto first_argument = key_function_node.getArgumentAt(0); + const auto & map_column_name = first_argument.getColumnName(); size_t map_keys_key_column_num = 0; auto map_keys_index_column_name = fmt::format("mapKeys({})", map_column_name); @@ -431,12 +446,11 @@ bool MergeTreeConditionFullText::traverseASTEquals( if (map_keys_exists) { - auto & argument = function->arguments.get()->children[1]; + auto second_argument = key_function_node.getArgumentAt(1); + DataTypePtr const_type; - if (const auto * literal = argument->as()) + if (second_argument.tryGetConstant(const_value, const_type)) { - auto element_key = literal->value; - const_value = element_key; key_column_num = map_keys_key_column_num; key_exists = true; } @@ -567,33 +581,37 @@ bool MergeTreeConditionFullText::traverseASTEquals( } bool MergeTreeConditionFullText::tryPrepareSetBloomFilter( - const ASTs & args, + const RPNBuilderTreeNode & left_argument, + const RPNBuilderTreeNode & right_argument, RPNElement & out) { - const ASTPtr & left_arg = args[0]; - const ASTPtr & right_arg = args[1]; - std::vector key_tuple_mapping; DataTypes data_types; - const auto * left_arg_tuple = typeid_cast(left_arg.get()); - if (left_arg_tuple && left_arg_tuple->name == "tuple") + if (left_argument.isFunction()) { - const auto & tuple_elements = left_arg_tuple->arguments->children; - for (size_t i = 0; i < tuple_elements.size(); ++i) + auto left_argument_function_node = left_argument.toFunctionNode(); + auto left_argument_function_node_name = left_argument_function_node.getFunctionName(); + + if (left_argument_function_node_name == "tuple") { - size_t key = 0; - if (getKey(tuple_elements[i]->getColumnName(), key)) + size_t left_argument_function_node_arguments_size = left_argument_function_node.getArgumentsSize(); + + for (size_t i = 0; i < left_argument_function_node_arguments_size; ++i) { - key_tuple_mapping.emplace_back(i, key); - data_types.push_back(index_data_types[key]); + size_t key = 0; + if (getKey(left_argument_function_node.getArgumentAt(i).getColumnName(), key)) + { + key_tuple_mapping.emplace_back(i, key); + data_types.push_back(index_data_types[key]); + } } } } else { size_t key = 0; - if (getKey(left_arg->getColumnName(), key)) + if (getKey(left_argument.getColumnName(), key)) { key_tuple_mapping.emplace_back(0, key); data_types.push_back(index_data_types[key]); @@ -603,19 +621,10 @@ bool MergeTreeConditionFullText::tryPrepareSetBloomFilter( if (key_tuple_mapping.empty()) return false; - PreparedSetKey set_key; - if (typeid_cast(right_arg.get()) || typeid_cast(right_arg.get())) - set_key = PreparedSetKey::forSubquery(*right_arg); - else - set_key = PreparedSetKey::forLiteral(*right_arg, data_types); - - auto prepared_set = prepared_sets->get(set_key); + auto prepared_set = right_argument.tryGetPreparedSet(data_types); if (!prepared_set) return false; - if (!prepared_set->hasExplicitSetElements()) - return false; - for (const auto & data_type : prepared_set->getDataTypes()) if (data_type->getTypeId() != TypeIndex::String && data_type->getTypeId() != TypeIndex::FixedString) return false; diff --git a/src/Storages/MergeTree/MergeTreeIndexFullText.h b/src/Storages/MergeTree/MergeTreeIndexFullText.h index bb4f52a463e..ad487816aef 100644 --- a/src/Storages/MergeTree/MergeTreeIndexFullText.h +++ b/src/Storages/MergeTree/MergeTreeIndexFullText.h @@ -122,17 +122,17 @@ private: using RPN = std::vector; - bool traverseAtomAST(const ASTPtr & node, Block & block_with_constants, RPNElement & out); + bool extractAtomFromTree(const RPNBuilderTreeNode & node, RPNElement & out); - bool traverseASTEquals( + bool traverseTreeEquals( const String & function_name, - const ASTPtr & key_ast, + const RPNBuilderTreeNode & key_node, const DataTypePtr & value_type, const Field & value_field, RPNElement & out); bool getKey(const std::string & key_column_name, size_t & key_column_num); - bool tryPrepareSetBloomFilter(const ASTs & args, RPNElement & out); + bool tryPrepareSetBloomFilter(const RPNBuilderTreeNode & left_argument, const RPNBuilderTreeNode & right_argument, RPNElement & out); static bool createFunctionEqualsCondition( RPNElement & out, const Field & value, const BloomFilterParameters & params, TokenExtractorPtr token_extractor); diff --git a/src/Storages/MergeTree/MergeTreeIndexUtils.cpp b/src/Storages/MergeTree/MergeTreeIndexUtils.cpp new file mode 100644 index 00000000000..652f0c853d4 --- /dev/null +++ b/src/Storages/MergeTree/MergeTreeIndexUtils.cpp @@ -0,0 +1,47 @@ +#include + +#include +#include +#include + +namespace DB +{ + +ASTPtr buildFilterNode(const ASTPtr & select_query, ASTs additional_filters) +{ + auto & select_query_typed = select_query->as(); + + ASTs filters; + if (select_query_typed.where()) + filters.push_back(select_query_typed.where()); + + if (select_query_typed.prewhere()) + filters.push_back(select_query_typed.prewhere()); + + filters.insert(filters.end(), additional_filters.begin(), additional_filters.end()); + + if (filters.empty()) + return nullptr; + + ASTPtr filter_node; + + if (filters.size() == 1) + { + filter_node = filters.front(); + } + else + { + auto function = std::make_shared(); + + function->name = "and"; + function->arguments = std::make_shared(); + function->children.push_back(function->arguments); + function->arguments->children = std::move(filters); + + filter_node = std::move(function); + } + + return filter_node; +} + +} diff --git a/src/Storages/MergeTree/MergeTreeIndexUtils.h b/src/Storages/MergeTree/MergeTreeIndexUtils.h new file mode 100644 index 00000000000..23cd92f3d99 --- /dev/null +++ b/src/Storages/MergeTree/MergeTreeIndexUtils.h @@ -0,0 +1,13 @@ +#pragma once + +#include + +namespace DB +{ + +/** Build AST filter node for index analysis from WHERE and PREWHERE sections of select query and additional filters. + * If select query does not have WHERE or PREWHERE and additional filters are empty null is returned. + */ +ASTPtr buildFilterNode(const ASTPtr & select_query, ASTs additional_filters = {}); + +} diff --git a/src/Storages/MergeTree/RPNBuilder.cpp b/src/Storages/MergeTree/RPNBuilder.cpp new file mode 100644 index 00000000000..362e02e8498 --- /dev/null +++ b/src/Storages/MergeTree/RPNBuilder.cpp @@ -0,0 +1,380 @@ +#include + +#include + +#include +#include +#include +#include + +#include +#include + +#include +#include + +#include + +#include + + +namespace DB +{ + +namespace +{ + +void appendColumnNameWithoutAlias(const ActionsDAG::Node & node, WriteBuffer & out, bool legacy = false) +{ + switch (node.type) + { + case (ActionsDAG::ActionType::INPUT): + writeString(node.result_name, out); + break; + case (ActionsDAG::ActionType::COLUMN): + { + /// If it was created from ASTLiteral, then result_name can be an alias. + /// We need to convert value back to string here. + if (const auto * column_const = typeid_cast(node.column.get())) + writeString(applyVisitor(FieldVisitorToString(), column_const->getField()), out); + /// It may be possible that column is ColumnSet + else + writeString(node.result_name, out); + break; + } + case (ActionsDAG::ActionType::ALIAS): + appendColumnNameWithoutAlias(*node.children.front(), out, legacy); + break; + case (ActionsDAG::ActionType::ARRAY_JOIN): + writeCString("arrayJoin(", out); + appendColumnNameWithoutAlias(*node.children.front(), out, legacy); + writeChar(')', out); + break; + case (ActionsDAG::ActionType::FUNCTION): + { + auto name = node.function_base->getName(); + if (legacy && name == "modulo") + writeCString("moduleLegacy", out); + else + writeString(name, out); + + writeChar('(', out); + bool first = true; + for (const auto * arg : node.children) + { + if (!first) + writeCString(", ", out); + first = false; + + appendColumnNameWithoutAlias(*arg, out, legacy); + } + writeChar(')', out); + } + } +} + +String getColumnNameWithoutAlias(const ActionsDAG::Node & node, bool legacy = false) +{ + WriteBufferFromOwnString out; + appendColumnNameWithoutAlias(node, out, legacy); + return std::move(out.str()); +} + +} + +RPNBuilderTreeContext::RPNBuilderTreeContext(ContextPtr query_context_) + : query_context(std::move(query_context_)) +{} + +RPNBuilderTreeContext::RPNBuilderTreeContext(ContextPtr query_context_, Block block_with_constants_, PreparedSetsPtr prepared_sets_) + : query_context(std::move(query_context_)) + , block_with_constants(std::move(block_with_constants_)) + , prepared_sets(std::move(prepared_sets_)) +{} + +RPNBuilderTreeNode::RPNBuilderTreeNode(const ActionsDAG::Node * dag_node_, RPNBuilderTreeContext & tree_context_) + : dag_node(dag_node_) + , tree_context(tree_context_) +{ + assert(dag_node); +} + +RPNBuilderTreeNode::RPNBuilderTreeNode(const IAST * ast_node_, RPNBuilderTreeContext & tree_context_) + : ast_node(ast_node_) + , tree_context(tree_context_) +{ + assert(ast_node); +} + +std::string RPNBuilderTreeNode::getColumnName() const +{ + if (ast_node) + return ast_node->getColumnNameWithoutAlias(); + else + return getColumnNameWithoutAlias(*dag_node); +} + +std::string RPNBuilderTreeNode::getColumnNameWithModuloLegacy() const +{ + if (ast_node) + { + auto adjusted_ast = ast_node->clone(); + KeyDescription::moduloToModuloLegacyRecursive(adjusted_ast); + return adjusted_ast->getColumnNameWithoutAlias(); + } + else + { + return getColumnNameWithoutAlias(*dag_node, true); + } +} + +bool RPNBuilderTreeNode::isFunction() const +{ + if (ast_node) + return typeid_cast(ast_node); + else + return dag_node->type == ActionsDAG::ActionType::FUNCTION; +} + +bool RPNBuilderTreeNode::isConstant() const +{ + if (ast_node) + return typeid_cast(ast_node); + else + return dag_node->column && isColumnConst(*dag_node->column); +} + +ColumnWithTypeAndName RPNBuilderTreeNode::getConstantColumn() const +{ + if (!isConstant()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "RPNBuilderTree node is not a constant"); + + ColumnWithTypeAndName res; + + if (ast_node) + { + const auto * literal = assert_cast(ast_node); + res.type = applyVisitor(FieldToDataType(), literal->value); + res.column = res.type->createColumnConst(0, literal->value); + } + else + { + res.type = dag_node->result_type; + res.column = dag_node->column; + } + + return res; +} + +bool RPNBuilderTreeNode::tryGetConstant(Field & output_value, DataTypePtr & output_type) const +{ + if (!isConstant()) + return false; + + if (ast_node) + { + // Constant expr should use alias names if any + String column_name = ast_node->getColumnName(); + const auto & block_with_constants = tree_context.getBlockWithConstants(); + + if (const auto * lit = ast_node->as()) + { + /// By default block_with_constants has only one column named "_dummy". + /// If block contains only constants it's may not be preprocessed by + // ExpressionAnalyzer, so try to look up in the default column. + if (!block_with_constants.has(column_name)) + column_name = "_dummy"; + + /// Simple literal + output_value = lit->value; + output_type = block_with_constants.getByName(column_name).type; + + /// If constant is not Null, we can assume it's type is not Nullable as well. + if (!output_value.isNull()) + output_type = removeNullable(output_type); + + return true; + } + else if (block_with_constants.has(column_name) && + isColumnConst(*block_with_constants.getByName(column_name).column)) + { + /// An expression which is dependent on constants only + const auto & expr_info = block_with_constants.getByName(column_name); + output_value = (*expr_info.column)[0]; + output_type = expr_info.type; + + if (!output_value.isNull()) + output_type = removeNullable(output_type); + + return true; + } + } + else + { + if (dag_node->column && isColumnConst(*dag_node->column)) + { + output_value = (*dag_node->column)[0]; + output_type = dag_node->result_type; + + if (!output_value.isNull()) + output_type = removeNullable(output_type); + + return true; + } + } + + return false; +} + +namespace +{ + +ConstSetPtr tryGetSetFromDAGNode(const ActionsDAG::Node * dag_node) +{ + if (!dag_node->column) + return {}; + + const IColumn * col = dag_node->column.get(); + if (const auto * col_const = typeid_cast(col)) + col = &col_const->getDataColumn(); + + if (const auto * col_set = typeid_cast(col)) + { + auto set = col_set->getData(); + + if (set->isCreated()) + return set; + } + + return {}; +} + +} + +ConstSetPtr RPNBuilderTreeNode::tryGetPreparedSet() const +{ + const auto & prepared_sets = getTreeContext().getPreparedSets(); + + if (ast_node && prepared_sets) + { + auto prepared_sets_with_same_hash = prepared_sets->getByTreeHash(ast_node->getTreeHash()); + for (auto & set : prepared_sets_with_same_hash) + if (set->isCreated()) + return set; + } + else if (dag_node) + { + return tryGetSetFromDAGNode(dag_node); + } + + return {}; +} + +ConstSetPtr RPNBuilderTreeNode::tryGetPreparedSet(const DataTypes & data_types) const +{ + const auto & prepared_sets = getTreeContext().getPreparedSets(); + + if (prepared_sets && ast_node) + { + if (ast_node->as() || ast_node->as()) + return prepared_sets->get(PreparedSetKey::forSubquery(*ast_node)); + + return prepared_sets->get(PreparedSetKey::forLiteral(*ast_node, data_types)); + } + else if (dag_node) + { + return tryGetSetFromDAGNode(dag_node); + } + + return nullptr; +} + +ConstSetPtr RPNBuilderTreeNode::tryGetPreparedSet( + const std::vector & indexes_mapping, + const DataTypes & data_types) const +{ + const auto & prepared_sets = getTreeContext().getPreparedSets(); + + if (prepared_sets && ast_node) + { + if (ast_node->as() || ast_node->as()) + return prepared_sets->get(PreparedSetKey::forSubquery(*ast_node)); + + /// We have `PreparedSetKey::forLiteral` but it is useless here as we don't have enough information + /// about types in left argument of the IN operator. Instead, we manually iterate through all the sets + /// and find the one for the right arg based on the AST structure (getTreeHash), after that we check + /// that the types it was prepared with are compatible with the types of the primary key. + auto types_match = [&indexes_mapping, &data_types](const SetPtr & candidate_set) + { + assert(indexes_mapping.size() == data_types.size()); + + for (size_t i = 0; i < indexes_mapping.size(); ++i) + { + if (!candidate_set->areTypesEqual(indexes_mapping[i].tuple_index, data_types[i])) + return false; + } + + return true; + }; + + auto tree_hash = ast_node->getTreeHash(); + for (const auto & set : prepared_sets->getByTreeHash(tree_hash)) + { + if (types_match(set)) + return set; + } + } + else if (dag_node->column) + { + return tryGetSetFromDAGNode(dag_node); + } + + return nullptr; +} + +RPNBuilderFunctionTreeNode RPNBuilderTreeNode::toFunctionNode() const +{ + if (!isFunction()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "RPNBuilderTree node is not a constant"); + + if (this->ast_node) + return RPNBuilderFunctionTreeNode(this->ast_node, tree_context); + else + return RPNBuilderFunctionTreeNode(this->dag_node, tree_context); +} + +std::string RPNBuilderFunctionTreeNode::getFunctionName() const +{ + if (ast_node) + return assert_cast(ast_node)->name; + else + return dag_node->function_base->getName(); +} + +size_t RPNBuilderFunctionTreeNode::getArgumentsSize() const +{ + if (ast_node) + { + const auto * ast_function = assert_cast(ast_node); + return ast_function->arguments ? ast_function->arguments->children.size() : 0; + } + else + { + return dag_node->children.size(); + } +} + +RPNBuilderTreeNode RPNBuilderFunctionTreeNode::getArgumentAt(size_t index) const +{ + if (ast_node) + { + const auto * ast_function = assert_cast(ast_node); + return RPNBuilderTreeNode(ast_function->arguments->children[index].get(), tree_context); + } + else + { + return RPNBuilderTreeNode(dag_node->children[index], tree_context); + } +} + +} diff --git a/src/Storages/MergeTree/RPNBuilder.h b/src/Storages/MergeTree/RPNBuilder.h index 27b616dc301..d4fc09dcdab 100644 --- a/src/Storages/MergeTree/RPNBuilder.h +++ b/src/Storages/MergeTree/RPNBuilder.h @@ -1,111 +1,262 @@ #pragma once #include -#include -#include -#include -#include -#include -#include -#include +#include +#include +#include +#include namespace DB { -/// Builds reverse polish notation -template -class RPNBuilder : WithContext +/** Context of RPNBuilderTree. + * + * For AST tree context, precalculated block with constansts and prepared sets are required for index analysis. + * For DAG tree precalculated block with constants and prepared sets are not required, because constants and sets already + * calculated inside COLUMN actions dag node. + */ +class RPNBuilderTreeContext { public: - using RPN = std::vector; - using AtomFromASTFunc = std::function< - bool(const ASTPtr & node, ContextPtr context, Block & block_with_constants, RPNElement & out)>; + /// Construct RPNBuilderTreeContext for ActionsDAG tree + explicit RPNBuilderTreeContext(ContextPtr query_context_); - RPNBuilder(const SelectQueryInfo & query_info, ContextPtr context_, const AtomFromASTFunc & atom_from_ast_) - : WithContext(context_), atom_from_ast(atom_from_ast_) + /// Construct RPNBuilderTreeContext for AST tree + explicit RPNBuilderTreeContext(ContextPtr query_context_, Block block_with_constants_, PreparedSetsPtr prepared_sets_); + + /// Get query context + const ContextPtr & getQueryContext() const { - /** Evaluation of expressions that depend only on constants. - * For the index to be used, if it is written, for example `WHERE Date = toDate(now())`. - */ - block_with_constants = KeyCondition::getBlockWithConstants(query_info.query, query_info.syntax_analyzer_result, getContext()); - - /// Transform WHERE section to Reverse Polish notation - const ASTSelectQuery & select = typeid_cast(*query_info.query); - if (select.where()) - { - traverseAST(select.where()); - - if (select.prewhere()) - { - traverseAST(select.prewhere()); - rpn.emplace_back(RPNElement::FUNCTION_AND); - } - } - else if (select.prewhere()) - { - traverseAST(select.prewhere()); - } - else - { - rpn.emplace_back(RPNElement::FUNCTION_UNKNOWN); - } + return query_context; } - RPN && extractRPN() { return std::move(rpn); } + /** Get block with constants. + * Valid only for AST tree. + */ + const Block & getBlockWithConstants() const + { + return block_with_constants; + } + + /** Get prepared sets. + * Valid only for AST tree. + */ + const PreparedSetsPtr & getPreparedSets() const + { + return prepared_sets; + } private: - void traverseAST(const ASTPtr & node) + /// Valid for both AST and ActionDAG tree + ContextPtr query_context; + + /// Valid only for AST tree + Block block_with_constants; + + /// Valid only for AST tree + PreparedSetsPtr prepared_sets; +}; + +class RPNBuilderFunctionTreeNode; + +/** RPNBuilderTreeNode is wrapper around DAG or AST node. + * It defines unified interface for index analysis. + */ +class RPNBuilderTreeNode +{ +public: + /// Construct RPNBuilderTreeNode with non null dag node and tree context + explicit RPNBuilderTreeNode(const ActionsDAG::Node * dag_node_, RPNBuilderTreeContext & tree_context_); + + /// Construct RPNBuilderTreeNode with non null ast node and tree context + explicit RPNBuilderTreeNode(const IAST * ast_node_, RPNBuilderTreeContext & tree_context_); + + /// Get column name + std::string getColumnName() const; + + /** Get column name. + * Function `modulo` is replaced with `moduloLegacy`. + */ + std::string getColumnNameWithModuloLegacy() const; + + /// Is node function + bool isFunction() const; + + /// Is node constant + bool isConstant() const; + + /** Get constant as constant column. + * Node must be constant before calling these method, otherwise logical exception is thrown. + */ + ColumnWithTypeAndName getConstantColumn() const; + + /** Try get constant from node. If node is constant returns true, and constant value and constant type output parameters are set. + * Otherwise false is returned. + */ + bool tryGetConstant(Field & output_value, DataTypePtr & output_type) const; + + /// Try get prepared set from node + ConstSetPtr tryGetPreparedSet() const; + + /// Try get prepared set from node that match data types + ConstSetPtr tryGetPreparedSet(const DataTypes & data_types) const; + + /// Try get prepared set from node that match indexes mapping and data types + ConstSetPtr tryGetPreparedSet( + const std::vector & indexes_mapping, + const DataTypes & data_types) const; + + /** Convert node to function node. + * Node must be function before calling these method, otherwise exception is thrown. + */ + RPNBuilderFunctionTreeNode toFunctionNode() const; + + /// Get tree context + const RPNBuilderTreeContext & getTreeContext() const + { + return tree_context; + } + + /// Get tree context + RPNBuilderTreeContext & getTreeContext() + { + return tree_context; + } + +protected: + const IAST * ast_node = nullptr; + const ActionsDAG::Node * dag_node = nullptr; + RPNBuilderTreeContext & tree_context; +}; + +/** RPNBuilderFunctionTreeNode is wrapper around RPNBuilderTreeNode with function type. + * It provide additional functionality that is specific for function. + */ +class RPNBuilderFunctionTreeNode : public RPNBuilderTreeNode +{ +public: + /// Get function name + std::string getFunctionName() const; + + /// Get function arguments size + size_t getArgumentsSize() const; + + /// Get argument at index + RPNBuilderTreeNode getArgumentAt(size_t index) const; + + using RPNBuilderTreeNode::RPNBuilderTreeNode; +}; + +/** RPN Builder build stack of reverse polish notation elements (RPNElements) required for index analysis. + * + * RPNBuilder client must provide RPNElement type that has following interface: + * + * struct RPNElementInterface + * { + * enum Function + * { + * FUNCTION_UNKNOWN, /// Can take any value. + * /// Operators of the logical expression. + * FUNCTION_NOT, + * FUNCTION_AND, + * FUNCTION_OR, + * ... + * }; + * + * RPNElementInterface(); + * + * Function function = FUNCTION_UNKNOWN; + * + * } + * + * RPNBuilder take care of building stack of RPNElements with `NOT`, `AND`, `OR` types. + * In addition client must provide ExtractAtomFromTreeFunction that returns true and RPNElement as output parameter, + * if it can convert RPNBuilderTree node to RPNElement, false otherwise. + */ +template +class RPNBuilder +{ +public: + using RPNElements = std::vector; + using ExtractAtomFromTreeFunction = std::function; + + explicit RPNBuilder(const ActionsDAG::Node * & filter_actions_dag_node, + ContextPtr query_context_, + const ExtractAtomFromTreeFunction & extract_atom_from_tree_function_) + : tree_context(std::move(query_context_)) + , extract_atom_from_tree_function(extract_atom_from_tree_function_) + { + traverseTree(RPNBuilderTreeNode(filter_actions_dag_node, tree_context)); + } + + RPNBuilder(const ASTPtr & filter_node, + ContextPtr query_context_, + Block block_with_constants_, + PreparedSetsPtr prepared_sets_, + const ExtractAtomFromTreeFunction & extract_atom_from_tree_function_) + : tree_context(std::move(query_context_), std::move(block_with_constants_), std::move(prepared_sets_)) + , extract_atom_from_tree_function(extract_atom_from_tree_function_) + { + traverseTree(RPNBuilderTreeNode(filter_node.get(), tree_context)); + } + + RPNElements && extractRPN() && { return std::move(rpn_elements); } + +private: + void traverseTree(const RPNBuilderTreeNode & node) { RPNElement element; - if (ASTFunction * func = typeid_cast(&*node)) + if (node.isFunction()) { - if (operatorFromAST(func, element)) + auto function_node = node.toFunctionNode(); + + if (extractLogicalOperatorFromTree(function_node, element)) { - auto & args = typeid_cast(*func->arguments).children; - for (size_t i = 0, size = args.size(); i < size; ++i) + size_t arguments_size = function_node.getArgumentsSize(); + + for (size_t argument_index = 0; argument_index < arguments_size; ++argument_index) { - traverseAST(args[i]); + auto function_node_argument = function_node.getArgumentAt(argument_index); + traverseTree(function_node_argument); /** The first part of the condition is for the correct support of `and` and `or` functions of arbitrary arity * - in this case `n - 1` elements are added (where `n` is the number of arguments). */ - if (i != 0 || element.function == RPNElement::FUNCTION_NOT) - rpn.emplace_back(std::move(element)); + if (argument_index != 0 || element.function == RPNElement::FUNCTION_NOT) + rpn_elements.emplace_back(std::move(element)); } return; } } - if (!atom_from_ast(node, getContext(), block_with_constants, element)) - { + if (!extract_atom_from_tree_function(node, element)) element.function = RPNElement::FUNCTION_UNKNOWN; - } - rpn.emplace_back(std::move(element)); + rpn_elements.emplace_back(std::move(element)); } - bool operatorFromAST(const ASTFunction * func, RPNElement & out) + bool extractLogicalOperatorFromTree(const RPNBuilderFunctionTreeNode & function_node, RPNElement & out) { /// Functions AND, OR, NOT. /// Also a special function `indexHint` - works as if instead of calling a function there are just parentheses /// (or, the same thing - calling the function `and` from one argument). - const ASTs & args = typeid_cast(*func->arguments).children; - if (func->name == "not") + auto function_name = function_node.getFunctionName(); + if (function_name == "not") { - if (args.size() != 1) + if (function_node.getArgumentsSize() != 1) return false; out.function = RPNElement::FUNCTION_NOT; } else { - if (func->name == "and" || func->name == "indexHint") + if (function_name == "and" || function_name == "indexHint") out.function = RPNElement::FUNCTION_AND; - else if (func->name == "or") + else if (function_name == "or") out.function = RPNElement::FUNCTION_OR; else return false; @@ -114,10 +265,9 @@ private: return true; } - const AtomFromASTFunc & atom_from_ast; - Block block_with_constants; - RPN rpn; + RPNBuilderTreeContext tree_context; + const ExtractAtomFromTreeFunction & extract_atom_from_tree_function; + RPNElements rpn_elements; }; - } From 14d0f6457b864db108b3834302f4d72a74f3d71f Mon Sep 17 00:00:00 2001 From: vdimir Date: Tue, 25 Oct 2022 17:18:32 +0000 Subject: [PATCH 101/239] Add tests and doc for some url-related functions --- .../URL/cutToFirstSignificantSubdomain.cpp | 32 +++++- .../cutToFirstSignificantSubdomainCustom.cpp | 39 +++++++- src/Functions/URL/domain.cpp | 20 +++- src/Functions/URL/domainWithoutWWW.cpp | 19 +++- .../URL/firstSignificantSubdomain.cpp | 24 ++++- src/Functions/URL/port.cpp | 14 ++- src/Functions/URL/topLevelDomain.cpp | 19 +++- tests/performance/url_hits.xml | 4 + .../0_stateless/00398_url_functions.reference | 17 ++++ ...nctions.sql => 00398_url_functions.sql.j2} | 98 +++++++++---------- .../queries/0_stateless/01284_port.reference | 24 +++++ tests/queries/0_stateless/01284_port.sql | 34 ------- tests/queries/0_stateless/01284_port.sql.j2 | 39 ++++++++ .../0_stateless/01601_custom_tld.reference | 89 +++++++++++++++++ .../queries/0_stateless/01601_custom_tld.sql | 57 ----------- .../0_stateless/01601_custom_tld.sql.j2 | 61 ++++++++++++ ...new_functions_must_be_documented.reference | 18 ---- 17 files changed, 428 insertions(+), 180 deletions(-) rename tests/queries/0_stateless/{00398_url_functions.sql => 00398_url_functions.sql.j2} (69%) delete mode 100644 tests/queries/0_stateless/01284_port.sql create mode 100644 tests/queries/0_stateless/01284_port.sql.j2 delete mode 100644 tests/queries/0_stateless/01601_custom_tld.sql create mode 100644 tests/queries/0_stateless/01601_custom_tld.sql.j2 diff --git a/src/Functions/URL/cutToFirstSignificantSubdomain.cpp b/src/Functions/URL/cutToFirstSignificantSubdomain.cpp index 10c41b6a4c3..7bf09d1eb00 100644 --- a/src/Functions/URL/cutToFirstSignificantSubdomain.cpp +++ b/src/Functions/URL/cutToFirstSignificantSubdomain.cpp @@ -43,10 +43,34 @@ using FunctionCutToFirstSignificantSubdomainWithWWWRFC = FunctionStringToString< REGISTER_FUNCTION(CutToFirstSignificantSubdomain) { - factory.registerFunction(); - factory.registerFunction(); - factory.registerFunction(); - factory.registerFunction(); + factory.registerFunction( + { + R"(Returns the part of the domain that includes top-level subdomains up to the "first significant subdomain" (see documentation of the `firstSignificantSubdomain`).)", + Documentation::Examples{ + {"cutToFirstSignificantSubdomain1", "SELECT cutToFirstSignificantSubdomain('https://news.clickhouse.com.tr/')"}, + {"cutToFirstSignificantSubdomain2", "SELECT cutToFirstSignificantSubdomain('www.tr')"}, + {"cutToFirstSignificantSubdomain3", "SELECT cutToFirstSignificantSubdomain('tr')"}, + }, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Returns the part of the domain that includes top-level subdomains up to the "first significant subdomain", without stripping "www".)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Similar to `cutToFirstSignificantSubdomain` but follows stricter rules to be compatible with RFC 3986 and less performant.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Similar to `cutToFirstSignificantSubdomainWithWWW` but follows stricter rules to be compatible with RFC 3986 and less performant.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp b/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp index 521216c84a7..e81921d69ff 100644 --- a/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp +++ b/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp @@ -42,10 +42,41 @@ using FunctionCutToFirstSignificantSubdomainCustomWithWWWRFC = FunctionCutToFirs REGISTER_FUNCTION(CutToFirstSignificantSubdomainCustom) { - factory.registerFunction(); - factory.registerFunction(); - factory.registerFunction(); - factory.registerFunction(); + factory.registerFunction( + { + R"( +Returns the part of the domain that includes top-level subdomains up to the first significant subdomain. Accepts custom TLD list name. + +Can be useful if you need fresh TLD list or you have custom. + )", + Documentation::Examples{ + {"cutToFirstSignificantSubdomainCustom", "SELECT cutToFirstSignificantSubdomainCustom('bar.foo.there-is-no-such-domain', 'public_suffix_list');"}, + }, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"( +Returns the part of the domain that includes top-level subdomains up to the first significant subdomain without stripping `www`. +Accepts custom TLD list name from config. + +Can be useful if you need fresh TLD list or you have custom. + )", + Documentation::Examples{{"cutToFirstSignificantSubdomainCustomWithWWW", "SELECT cutToFirstSignificantSubdomainCustomWithWWW('www.foo', 'public_suffix_list')"}}, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Similar to `cutToFirstSignificantSubdomainCustom` but follows stricter rules according to RFC 3986.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Similar to `cutToFirstSignificantSubdomainCustomWithWWW` but follows stricter rules according to RFC 3986.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/src/Functions/URL/domain.cpp b/src/Functions/URL/domain.cpp index e7fead24dc9..fce7cea4693 100644 --- a/src/Functions/URL/domain.cpp +++ b/src/Functions/URL/domain.cpp @@ -14,8 +14,24 @@ using FunctionDomainRFC = FunctionStringToString(); - factory.registerFunction(); + factory.registerFunction( + { + R"( +Extracts the hostname from a URL. + +The URL can be specified with or without a scheme. +If the argument can't be parsed as URL, the function returns an empty string. + )", + Documentation::Examples{{"domain", "SELECT domain('svn+ssh://some.svn-hosting.com:80/repo/trunk')"}}, + Documentation::Categories{"URL"} + }); + + factory.registerFunction( + { + R"(Similar to `domain` but follows stricter rules to be compatible with RFC 3986 and less performant.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/src/Functions/URL/domainWithoutWWW.cpp b/src/Functions/URL/domainWithoutWWW.cpp index 2fa9159d7af..48401e5e6e5 100644 --- a/src/Functions/URL/domainWithoutWWW.cpp +++ b/src/Functions/URL/domainWithoutWWW.cpp @@ -14,8 +14,23 @@ using FunctionDomainWithoutWWWRFC = FunctionStringToString(); - factory.registerFunction(); + factory.registerFunction( + { + R"( +Extracts the hostname from a URL, removing the leading "www." if present. + +The URL can be specified with or without a scheme. +If the argument can't be parsed as URL, the function returns an empty string. + )", + Documentation::Examples{{"domainWithoutWWW", "SELECT domainWithoutWWW('https://www.clickhouse.com')"}}, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Similar to `domainWithoutWWW` but follows stricter rules to be compatible with RFC 3986 and less performant.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/src/Functions/URL/firstSignificantSubdomain.cpp b/src/Functions/URL/firstSignificantSubdomain.cpp index 902a4f43fba..62307ef816c 100644 --- a/src/Functions/URL/firstSignificantSubdomain.cpp +++ b/src/Functions/URL/firstSignificantSubdomain.cpp @@ -14,8 +14,28 @@ using FunctionFirstSignificantSubdomainRFC = FunctionStringToString(); - factory.registerFunction(); + factory.registerFunction( + { + R"( +Returns the "first significant subdomain". + +The first significant subdomain is a second-level domain if it is 'com', 'net', 'org', or 'co'. +Otherwise, it is a third-level domain. + +For example, firstSignificantSubdomain('https://news.clickhouse.com/') = 'clickhouse', firstSignificantSubdomain ('https://news.clickhouse.com.tr/') = 'clickhouse'. + +The list of "insignificant" second-level domains and other implementation details may change in the future. + )", + Documentation::Examples{{"firstSignificantSubdomain", "SELECT firstSignificantSubdomain('https://news.clickhouse.com/')"}}, + Documentation::Categories{"URL"} + }); + + factory.registerFunction( + { + R"(Returns the "first significant subdomain" according to RFC 1034.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/src/Functions/URL/port.cpp b/src/Functions/URL/port.cpp index f716f3e454b..52fa4077c18 100644 --- a/src/Functions/URL/port.cpp +++ b/src/Functions/URL/port.cpp @@ -139,8 +139,18 @@ struct FunctionPortRFC : public FunctionPortImpl REGISTER_FUNCTION(Port) { - factory.registerFunction(); - factory.registerFunction(); + factory.registerFunction( + { + R"(Returns the port or `default_port` if there is no port in the URL (or in case of validation error).)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); + factory.registerFunction( + { + R"(Similar to `port`, but conforms to RFC 3986.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/src/Functions/URL/topLevelDomain.cpp b/src/Functions/URL/topLevelDomain.cpp index f5610ed93b7..ed9b40d4b73 100644 --- a/src/Functions/URL/topLevelDomain.cpp +++ b/src/Functions/URL/topLevelDomain.cpp @@ -53,8 +53,23 @@ using FunctionTopLevelDomainRFC = FunctionStringToString(); - factory.registerFunction(); + factory.registerFunction( + { + R"( +Extracts the the top-level domain from a URL. + +Returns an empty string if the argument cannot be parsed as a URL or does not contain a top-level domain. + )", + Documentation::Examples{{"topLevelDomain", "SELECT topLevelDomain('svn+ssh://www.some.svn-hosting.com:80/repo/trunk')"}}, + Documentation::Categories{"URL"} + }); + + factory.registerFunction( + { + R"(Similar to topLevelDomain, but conforms to RFC 3986.)", + Documentation::Examples{}, + Documentation::Categories{"URL"} + }); } } diff --git a/tests/performance/url_hits.xml b/tests/performance/url_hits.xml index 4a07c38b83f..46b39f3a6e9 100644 --- a/tests/performance/url_hits.xml +++ b/tests/performance/url_hits.xml @@ -13,10 +13,14 @@ protocol domain + domainRFC domainWithoutWWW + domainWithoutWWWRFC topLevelDomain firstSignificantSubdomain + firstSignificantSubdomainRFC cutToFirstSignificantSubdomain + cutToFirstSignificantSubdomainRFC path pathFull queryString diff --git a/tests/queries/0_stateless/00398_url_functions.reference b/tests/queries/0_stateless/00398_url_functions.reference index 2e5a97b380e..39d740e55cd 100644 --- a/tests/queries/0_stateless/00398_url_functions.reference +++ b/tests/queries/0_stateless/00398_url_functions.reference @@ -124,8 +124,25 @@ example.com example.com com +example.com +example.com +example.com +example.com +example.com +example.com +example.com +example.com +example.com +com + ====CUT TO FIRST SIGNIFICANT SUBDOMAIN WITH WWW==== +www.com +example.com +example.com +example.com +example.com + www.com example.com example.com diff --git a/tests/queries/0_stateless/00398_url_functions.sql b/tests/queries/0_stateless/00398_url_functions.sql.j2 similarity index 69% rename from tests/queries/0_stateless/00398_url_functions.sql rename to tests/queries/0_stateless/00398_url_functions.sql.j2 index cbefde7515a..dd7da2ce6ad 100644 --- a/tests/queries/0_stateless/00398_url_functions.sql +++ b/tests/queries/0_stateless/00398_url_functions.sql.j2 @@ -7,42 +7,28 @@ SELECT protocol('http://127.0.0.1:443/') AS Scheme; SELECT protocol('//127.0.0.1:443/') AS Scheme; SELECT '====HOST===='; -SELECT domain('http://paul@www.example.com:80/') AS Host; -SELECT domain('user:password@example.com:8080') AS Host; -SELECT domain('http://user:password@example.com:8080') AS Host; -SELECT domain('http://user:password@example.com:8080/path?query=value#fragment') AS Host; -SELECT domain('newuser:@example.com') AS Host; -SELECT domain('http://:pass@example.com') AS Host; -SELECT domain(':newpass@example.com') AS Host; -SELECT domain('http://user:pass@example@.com') AS Host; -SELECT domain('http://user:pass:example.com') AS Host; -SELECT domain('http:/paul/example/com') AS Host; -SELECT domain('http://www.example.com?q=4') AS Host; -SELECT domain('http://127.0.0.1:443/') AS Host; -SELECT domain('//www.example.com') AS Host; -SELECT domain('//paul@www.example.com') AS Host; -SELECT domain('www.example.com') as Host; -SELECT domain('example.com') as Host; -SELECT domainWithoutWWW('//paul@www.example.com') AS Host; -SELECT domainWithoutWWW('http://paul@www.example.com:80/') AS Host; -SELECT domainRFC('http://paul@www.example.com:80/') AS Host; -SELECT domainRFC('user:password@example.com:8080') AS Host; -SELECT domainRFC('http://user:password@example.com:8080') AS Host; -SELECT domainRFC('http://user:password@example.com:8080/path?query=value#fragment') AS Host; -SELECT domainRFC('newuser:@example.com') AS Host; -SELECT domainRFC('http://:pass@example.com') AS Host; -SELECT domainRFC(':newpass@example.com') AS Host; -SELECT domainRFC('http://user:pass@example@.com') AS Host; -SELECT domainRFC('http://user:pass:example.com') AS Host; -SELECT domainRFC('http:/paul/example/com') AS Host; -SELECT domainRFC('http://www.example.com?q=4') AS Host; -SELECT domainRFC('http://127.0.0.1:443/') AS Host; -SELECT domainRFC('//www.example.com') AS Host; -SELECT domainRFC('//paul@www.example.com') AS Host; -SELECT domainRFC('www.example.com') as Host; -SELECT domainRFC('example.com') as Host; -SELECT domainWithoutWWWRFC('//paul@www.example.com') AS Host; -SELECT domainWithoutWWWRFC('http://paul@www.example.com:80/') AS Host; +{% for suffix in ['', 'RFC'] -%} + +SELECT domain{{ suffix }}('http://paul@www.example.com:80/') AS Host; +SELECT domain{{ suffix }}('user:password@example.com:8080') AS Host; +SELECT domain{{ suffix }}('http://user:password@example.com:8080') AS Host; +SELECT domain{{ suffix }}('http://user:password@example.com:8080/path?query=value#fragment') AS Host; +SELECT domain{{ suffix }}('newuser:@example.com') AS Host; +SELECT domain{{ suffix }}('http://:pass@example.com') AS Host; +SELECT domain{{ suffix }}(':newpass@example.com') AS Host; +SELECT domain{{ suffix }}('http://user:pass@example@.com') AS Host; +SELECT domain{{ suffix }}('http://user:pass:example.com') AS Host; +SELECT domain{{ suffix }}('http:/paul/example/com') AS Host; +SELECT domain{{ suffix }}('http://www.example.com?q=4') AS Host; +SELECT domain{{ suffix }}('http://127.0.0.1:443/') AS Host; +SELECT domain{{ suffix }}('//www.example.com') AS Host; +SELECT domain{{ suffix }}('//paul@www.example.com') AS Host; +SELECT domain{{ suffix }}('www.example.com') as Host; +SELECT domain{{ suffix }}('example.com') as Host; +SELECT domainWithoutWWW{{ suffix }}('//paul@www.example.com') AS Host; +SELECT domainWithoutWWW{{ suffix }}('http://paul@www.example.com:80/') AS Host; + +{% endfor %} SELECT '====NETLOC===='; SELECT netloc('http://paul@www.example.com:80/') AS Netloc; @@ -121,25 +107,31 @@ SELECT decodeURLComponent(encodeURLComponent('http://paul@127.0.0.1/?query=hello SELECT decodeURLFormComponent(encodeURLFormComponent('http://paul@127.0.0.1/?query=hello world foo+bar#a=b')); SELECT '====CUT TO FIRST SIGNIFICANT SUBDOMAIN===='; -SELECT cutToFirstSignificantSubdomain('http://www.example.com'); -SELECT cutToFirstSignificantSubdomain('http://www.example.com:1234'); -SELECT cutToFirstSignificantSubdomain('http://www.example.com/a/b/c'); -SELECT cutToFirstSignificantSubdomain('http://www.example.com/a/b/c?a=b'); -SELECT cutToFirstSignificantSubdomain('http://www.example.com/a/b/c?a=b#d=f'); -SELECT cutToFirstSignificantSubdomain('http://paul@www.example.com/a/b/c?a=b#d=f'); -SELECT cutToFirstSignificantSubdomain('//paul@www.example.com/a/b/c?a=b#d=f'); -SELECT cutToFirstSignificantSubdomain('www.example.com'); -SELECT cutToFirstSignificantSubdomain('example.com'); -SELECT cutToFirstSignificantSubdomain('www.com'); -SELECT cutToFirstSignificantSubdomain('com'); + +{% for suffix in ['', 'RFC'] -%} +SELECT cutToFirstSignificantSubdomain{{ suffix }}('http://www.example.com'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('http://www.example.com:1234'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('http://www.example.com/a/b/c'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('http://www.example.com/a/b/c?a=b'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('http://www.example.com/a/b/c?a=b#d=f'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('http://paul@www.example.com/a/b/c?a=b#d=f'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('//paul@www.example.com/a/b/c?a=b#d=f'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('www.example.com'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('example.com'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('www.com'); +SELECT cutToFirstSignificantSubdomain{{ suffix }}('com'); +{% endfor %} SELECT '====CUT TO FIRST SIGNIFICANT SUBDOMAIN WITH WWW===='; -SELECT cutToFirstSignificantSubdomainWithWWW('http://com'); -SELECT cutToFirstSignificantSubdomainWithWWW('http://www.com'); -SELECT cutToFirstSignificantSubdomainWithWWW('http://www.example.com'); -SELECT cutToFirstSignificantSubdomainWithWWW('http://www.foo.example.com'); -SELECT cutToFirstSignificantSubdomainWithWWW('http://www.example.com:1'); -SELECT cutToFirstSignificantSubdomainWithWWW('http://www.example.com/'); + +{% for suffix in ['', 'RFC'] -%} +SELECT cutToFirstSignificantSubdomainWithWWW{{ suffix }}('http://com'); +SELECT cutToFirstSignificantSubdomainWithWWW{{ suffix }}('http://www.com'); +SELECT cutToFirstSignificantSubdomainWithWWW{{ suffix }}('http://www.example.com'); +SELECT cutToFirstSignificantSubdomainWithWWW{{ suffix }}('http://www.foo.example.com'); +SELECT cutToFirstSignificantSubdomainWithWWW{{ suffix }}('http://www.example.com:1'); +SELECT cutToFirstSignificantSubdomainWithWWW{{ suffix }}('http://www.example.com/'); +{% endfor %} SELECT '====CUT WWW===='; SELECT cutWWW('http://www.example.com'); diff --git a/tests/queries/0_stateless/01284_port.reference b/tests/queries/0_stateless/01284_port.reference index 7e776595065..5b7b58bc7e4 100644 --- a/tests/queries/0_stateless/01284_port.reference +++ b/tests/queries/0_stateless/01284_port.reference @@ -22,3 +22,27 @@ ipv6 0 host-no-dot 0 +ipv4 +0 +80 +80 +80 +80 +hostname +0 +80 +80 +80 +80 +default-port +80 +80 +ipv6 +0 +0 +0 +0 +0 +0 +host-no-dot +0 diff --git a/tests/queries/0_stateless/01284_port.sql b/tests/queries/0_stateless/01284_port.sql deleted file mode 100644 index 9c31a5d42ad..00000000000 --- a/tests/queries/0_stateless/01284_port.sql +++ /dev/null @@ -1,34 +0,0 @@ -select 'ipv4'; -select port('http://127.0.0.1/'); -select port('http://127.0.0.1:80'); -select port('http://127.0.0.1:80/'); -select port('//127.0.0.1:80/'); -select port('127.0.0.1:80'); -select 'hostname'; -select port('http://foobar.com/'); -select port('http://foobar.com:80'); -select port('http://foobar.com:80/'); -select port('//foobar.com:80/'); -select port('foobar.com:80'); - -select 'default-port'; -select port('http://127.0.0.1/', toUInt16(80)); -select port('http://foobar.com/', toUInt16(80)); - --- unsupported -/* ILLEGAL_TYPE_OF_ARGUMENT */ select port(toFixedString('', 1)); -- { serverError 43; } -/* ILLEGAL_TYPE_OF_ARGUMENT */ select port('', 1); -- { serverError 43; } -/* NUMBER_OF_ARGUMENTS_DOESNT_MATCH */ select port('', 1, 1); -- { serverError 42; } - --- --- Known limitations of domain() (getURLHost()) --- -select 'ipv6'; -select port('http://[2001:db8::8a2e:370:7334]/'); -select port('http://[2001:db8::8a2e:370:7334]:80'); -select port('http://[2001:db8::8a2e:370:7334]:80/'); -select port('//[2001:db8::8a2e:370:7334]:80/'); -select port('[2001:db8::8a2e:370:7334]:80'); -select port('2001:db8::8a2e:370:7334:80'); -select 'host-no-dot'; -select port('//foobar:80/'); diff --git a/tests/queries/0_stateless/01284_port.sql.j2 b/tests/queries/0_stateless/01284_port.sql.j2 new file mode 100644 index 00000000000..6f78b3b8e3b --- /dev/null +++ b/tests/queries/0_stateless/01284_port.sql.j2 @@ -0,0 +1,39 @@ +{% for suffix in ['', 'RFC'] -%} + +select 'ipv4'; +select port{{ suffix }}('http://127.0.0.1/'); +select port{{ suffix }}('http://127.0.0.1:80'); +select port{{ suffix }}('http://127.0.0.1:80/'); +select port{{ suffix }}('//127.0.0.1:80/'); +select port{{ suffix }}('127.0.0.1:80'); + +select 'hostname'; +select port{{ suffix }}('http://foobar.com/'); +select port{{ suffix }}('http://foobar.com:80'); +select port{{ suffix }}('http://foobar.com:80/'); +select port{{ suffix }}('//foobar.com:80/'); +select port{{ suffix }}('foobar.com:80'); + +select 'default-port'; +select port{{ suffix }}('http://127.0.0.1/', toUInt16(80)); +select port{{ suffix }}('http://foobar.com/', toUInt16(80)); + +-- unsupported +/* ILLEGAL_TYPE_OF_ARGUMENT */ select port(toFixedString('', 1)); -- { serverError 43; } +/* ILLEGAL_TYPE_OF_ARGUMENT */ select port{{ suffix }}('', 1); -- { serverError 43; } +/* NUMBER_OF_ARGUMENTS_DOESNT_MATCH */ select port{{ suffix }}('', 1, 1); -- { serverError 42; } + +-- +-- Known limitations of domain() (getURLHost()) +-- +select 'ipv6'; +select port{{ suffix }}('http://[2001:db8::8a2e:370:7334]/'); +select port{{ suffix }}('http://[2001:db8::8a2e:370:7334]:80'); +select port{{ suffix }}('http://[2001:db8::8a2e:370:7334]:80/'); +select port{{ suffix }}('//[2001:db8::8a2e:370:7334]:80/'); +select port{{ suffix }}('[2001:db8::8a2e:370:7334]:80'); +select port{{ suffix }}('2001:db8::8a2e:370:7334:80'); +select 'host-no-dot'; +select port{{ suffix }}('//foobar:80/'); + +{%- endfor %} diff --git a/tests/queries/0_stateless/01601_custom_tld.reference b/tests/queries/0_stateless/01601_custom_tld.reference index 981067606a2..7ef6eb7d3a2 100644 --- a/tests/queries/0_stateless/01601_custom_tld.reference +++ b/tests/queries/0_stateless/01601_custom_tld.reference @@ -89,3 +89,92 @@ select cutToFirstSignificantSubdomainCustom('city.kawasaki.jp', 'public_suffix_l city.kawasaki.jp select cutToFirstSignificantSubdomainCustom('some.city.kawasaki.jp', 'public_suffix_list'); city.kawasaki.jp +select '-- no-tld'; +-- no-tld +-- even if there is no TLD, 2-nd level by default anyway +-- FIXME: make this behavior optional (so that TLD for host never changed, either empty or something real) +select cutToFirstSignificantSubdomainRFC('there-is-no-such-domain'); + +select cutToFirstSignificantSubdomainRFC('foo.there-is-no-such-domain'); +foo.there-is-no-such-domain +select cutToFirstSignificantSubdomainRFC('bar.foo.there-is-no-such-domain'); +foo.there-is-no-such-domain +select cutToFirstSignificantSubdomainCustomRFC('there-is-no-such-domain', 'public_suffix_list'); + +select cutToFirstSignificantSubdomainCustomRFC('foo.there-is-no-such-domain', 'public_suffix_list'); +foo.there-is-no-such-domain +select cutToFirstSignificantSubdomainCustomRFC('bar.foo.there-is-no-such-domain', 'public_suffix_list'); +foo.there-is-no-such-domain +select firstSignificantSubdomainCustomRFC('bar.foo.there-is-no-such-domain', 'public_suffix_list'); +foo +select '-- generic'; +-- generic +select firstSignificantSubdomainCustomRFC('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel +kernel +select cutToFirstSignificantSubdomainCustomRFC('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel.biz.ss +kernel.biz.ss +select '-- difference'; +-- difference +-- biz.ss is not in the default TLD list, hence: +select cutToFirstSignificantSubdomainRFC('foo.kernel.biz.ss'); -- biz.ss +biz.ss +select cutToFirstSignificantSubdomainCustomRFC('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel.biz.ss +kernel.biz.ss +select '-- 3+level'; +-- 3+level +select cutToFirstSignificantSubdomainCustomRFC('xx.blogspot.co.at', 'public_suffix_list'); -- xx.blogspot.co.at +xx.blogspot.co.at +select firstSignificantSubdomainCustomRFC('xx.blogspot.co.at', 'public_suffix_list'); -- blogspot +blogspot +select cutToFirstSignificantSubdomainCustomRFC('foo.bar.xx.blogspot.co.at', 'public_suffix_list'); -- xx.blogspot.co.at +xx.blogspot.co.at +select firstSignificantSubdomainCustomRFC('foo.bar.xx.blogspot.co.at', 'public_suffix_list'); -- blogspot +blogspot +select '-- url'; +-- url +select cutToFirstSignificantSubdomainCustomRFC('http://foobar.com', 'public_suffix_list'); +foobar.com +select cutToFirstSignificantSubdomainCustomRFC('http://foobar.com/foo', 'public_suffix_list'); +foobar.com +select cutToFirstSignificantSubdomainCustomRFC('http://bar.foobar.com/foo', 'public_suffix_list'); +foobar.com +select cutToFirstSignificantSubdomainCustomRFC('http://xx.blogspot.co.at', 'public_suffix_list'); +xx.blogspot.co.at +select '-- www'; +-- www +select cutToFirstSignificantSubdomainCustomWithWWWRFC('http://www.foo', 'public_suffix_list'); +www.foo +select cutToFirstSignificantSubdomainCustomRFC('http://www.foo', 'public_suffix_list'); +foo +select '-- vector'; +-- vector +select cutToFirstSignificantSubdomainCustomRFC('http://xx.blogspot.co.at/' || toString(number), 'public_suffix_list') from numbers(1); +xx.blogspot.co.at +select cutToFirstSignificantSubdomainCustomRFC('there-is-no-such-domain' || toString(number), 'public_suffix_list') from numbers(1); + +select '-- no new line'; +-- no new line +select cutToFirstSignificantSubdomainCustomRFC('foo.bar', 'no_new_line_list'); +foo.bar +select cutToFirstSignificantSubdomainCustomRFC('a.foo.bar', 'no_new_line_list'); +a.foo.bar +select cutToFirstSignificantSubdomainCustomRFC('a.foo.baz', 'no_new_line_list'); +foo.baz +select '-- asterisk'; +-- asterisk +select cutToFirstSignificantSubdomainCustomRFC('foo.something.sheffield.sch.uk', 'public_suffix_list'); +something.sheffield.sch.uk +select cutToFirstSignificantSubdomainCustomRFC('something.sheffield.sch.uk', 'public_suffix_list'); +something.sheffield.sch.uk +select cutToFirstSignificantSubdomainCustomRFC('sheffield.sch.uk', 'public_suffix_list'); +sheffield.sch.uk +select '-- exclamation mark'; +-- exclamation mark +select cutToFirstSignificantSubdomainCustomRFC('foo.kawasaki.jp', 'public_suffix_list'); +foo.kawasaki.jp +select cutToFirstSignificantSubdomainCustomRFC('foo.foo.kawasaki.jp', 'public_suffix_list'); +foo.foo.kawasaki.jp +select cutToFirstSignificantSubdomainCustomRFC('city.kawasaki.jp', 'public_suffix_list'); +city.kawasaki.jp +select cutToFirstSignificantSubdomainCustomRFC('some.city.kawasaki.jp', 'public_suffix_list'); +city.kawasaki.jp diff --git a/tests/queries/0_stateless/01601_custom_tld.sql b/tests/queries/0_stateless/01601_custom_tld.sql deleted file mode 100644 index 69ae209af2c..00000000000 --- a/tests/queries/0_stateless/01601_custom_tld.sql +++ /dev/null @@ -1,57 +0,0 @@ --- { echo } - -select '-- no-tld'; --- even if there is no TLD, 2-nd level by default anyway --- FIXME: make this behavior optional (so that TLD for host never changed, either empty or something real) -select cutToFirstSignificantSubdomain('there-is-no-such-domain'); -select cutToFirstSignificantSubdomain('foo.there-is-no-such-domain'); -select cutToFirstSignificantSubdomain('bar.foo.there-is-no-such-domain'); -select cutToFirstSignificantSubdomainCustom('there-is-no-such-domain', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('foo.there-is-no-such-domain', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('bar.foo.there-is-no-such-domain', 'public_suffix_list'); -select firstSignificantSubdomainCustom('bar.foo.there-is-no-such-domain', 'public_suffix_list'); - -select '-- generic'; -select firstSignificantSubdomainCustom('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel -select cutToFirstSignificantSubdomainCustom('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel.biz.ss - -select '-- difference'; --- biz.ss is not in the default TLD list, hence: -select cutToFirstSignificantSubdomain('foo.kernel.biz.ss'); -- biz.ss -select cutToFirstSignificantSubdomainCustom('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel.biz.ss - -select '-- 3+level'; -select cutToFirstSignificantSubdomainCustom('xx.blogspot.co.at', 'public_suffix_list'); -- xx.blogspot.co.at -select firstSignificantSubdomainCustom('xx.blogspot.co.at', 'public_suffix_list'); -- blogspot -select cutToFirstSignificantSubdomainCustom('foo.bar.xx.blogspot.co.at', 'public_suffix_list'); -- xx.blogspot.co.at -select firstSignificantSubdomainCustom('foo.bar.xx.blogspot.co.at', 'public_suffix_list'); -- blogspot - -select '-- url'; -select cutToFirstSignificantSubdomainCustom('http://foobar.com', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('http://foobar.com/foo', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('http://bar.foobar.com/foo', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('http://xx.blogspot.co.at', 'public_suffix_list'); - -select '-- www'; -select cutToFirstSignificantSubdomainCustomWithWWW('http://www.foo', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('http://www.foo', 'public_suffix_list'); - -select '-- vector'; -select cutToFirstSignificantSubdomainCustom('http://xx.blogspot.co.at/' || toString(number), 'public_suffix_list') from numbers(1); -select cutToFirstSignificantSubdomainCustom('there-is-no-such-domain' || toString(number), 'public_suffix_list') from numbers(1); - -select '-- no new line'; -select cutToFirstSignificantSubdomainCustom('foo.bar', 'no_new_line_list'); -select cutToFirstSignificantSubdomainCustom('a.foo.bar', 'no_new_line_list'); -select cutToFirstSignificantSubdomainCustom('a.foo.baz', 'no_new_line_list'); - -select '-- asterisk'; -select cutToFirstSignificantSubdomainCustom('foo.something.sheffield.sch.uk', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('something.sheffield.sch.uk', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('sheffield.sch.uk', 'public_suffix_list'); - -select '-- exclamation mark'; -select cutToFirstSignificantSubdomainCustom('foo.kawasaki.jp', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('foo.foo.kawasaki.jp', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('city.kawasaki.jp', 'public_suffix_list'); -select cutToFirstSignificantSubdomainCustom('some.city.kawasaki.jp', 'public_suffix_list'); diff --git a/tests/queries/0_stateless/01601_custom_tld.sql.j2 b/tests/queries/0_stateless/01601_custom_tld.sql.j2 new file mode 100644 index 00000000000..1e0982ea1b7 --- /dev/null +++ b/tests/queries/0_stateless/01601_custom_tld.sql.j2 @@ -0,0 +1,61 @@ +-- { echo } + +{% for suffix in ['', 'RFC'] -%} + +select '-- no-tld'; +-- even if there is no TLD, 2-nd level by default anyway +-- FIXME: make this behavior optional (so that TLD for host never changed, either empty or something real) +select cutToFirstSignificantSubdomain{{ suffix }}('there-is-no-such-domain'); +select cutToFirstSignificantSubdomain{{ suffix }}('foo.there-is-no-such-domain'); +select cutToFirstSignificantSubdomain{{ suffix }}('bar.foo.there-is-no-such-domain'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('there-is-no-such-domain', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.there-is-no-such-domain', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('bar.foo.there-is-no-such-domain', 'public_suffix_list'); +select firstSignificantSubdomainCustom{{ suffix }}('bar.foo.there-is-no-such-domain', 'public_suffix_list'); + +select '-- generic'; +select firstSignificantSubdomainCustom{{ suffix }}('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel.biz.ss + +select '-- difference'; +-- biz.ss is not in the default TLD list, hence: +select cutToFirstSignificantSubdomain{{ suffix }}('foo.kernel.biz.ss'); -- biz.ss +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.kernel.biz.ss', 'public_suffix_list'); -- kernel.biz.ss + +select '-- 3+level'; +select cutToFirstSignificantSubdomainCustom{{ suffix }}('xx.blogspot.co.at', 'public_suffix_list'); -- xx.blogspot.co.at +select firstSignificantSubdomainCustom{{ suffix }}('xx.blogspot.co.at', 'public_suffix_list'); -- blogspot +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.bar.xx.blogspot.co.at', 'public_suffix_list'); -- xx.blogspot.co.at +select firstSignificantSubdomainCustom{{ suffix }}('foo.bar.xx.blogspot.co.at', 'public_suffix_list'); -- blogspot + +select '-- url'; +select cutToFirstSignificantSubdomainCustom{{ suffix }}('http://foobar.com', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('http://foobar.com/foo', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('http://bar.foobar.com/foo', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('http://xx.blogspot.co.at', 'public_suffix_list'); + +select '-- www'; +select cutToFirstSignificantSubdomainCustomWithWWW{{ suffix }}('http://www.foo', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('http://www.foo', 'public_suffix_list'); + +select '-- vector'; +select cutToFirstSignificantSubdomainCustom{{ suffix }}('http://xx.blogspot.co.at/' || toString(number), 'public_suffix_list') from numbers(1); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('there-is-no-such-domain' || toString(number), 'public_suffix_list') from numbers(1); + +select '-- no new line'; +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.bar', 'no_new_line_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('a.foo.bar', 'no_new_line_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('a.foo.baz', 'no_new_line_list'); + +select '-- asterisk'; +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.something.sheffield.sch.uk', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('something.sheffield.sch.uk', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('sheffield.sch.uk', 'public_suffix_list'); + +select '-- exclamation mark'; +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.kawasaki.jp', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('foo.foo.kawasaki.jp', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('city.kawasaki.jp', 'public_suffix_list'); +select cutToFirstSignificantSubdomainCustom{{ suffix }}('some.city.kawasaki.jp', 'public_suffix_list'); + +{% endfor %} diff --git a/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference b/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference index 3fd12051f4a..040a8c8d317 100644 --- a/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference +++ b/tests/queries/0_stateless/02415_all_new_functions_must_be_documented.reference @@ -219,14 +219,6 @@ cutFragment cutIPv6 cutQueryString cutQueryStringAndFragment -cutToFirstSignificantSubdomain -cutToFirstSignificantSubdomainCustom -cutToFirstSignificantSubdomainCustomRFC -cutToFirstSignificantSubdomainCustomWithWWW -cutToFirstSignificantSubdomainCustomWithWWWRFC -cutToFirstSignificantSubdomainRFC -cutToFirstSignificantSubdomainWithWWW -cutToFirstSignificantSubdomainWithWWWRFC cutURLParameter cutWWW dateDiff @@ -284,10 +276,6 @@ dictGetUUIDOrDefault dictHas dictIsIn divide -domain -domainRFC -domainWithoutWWW -domainWithoutWWWRFC dotProduct dumpColumnStructure e @@ -336,10 +324,8 @@ filesystemAvailable filesystemCapacity filesystemFree finalizeAggregation -firstSignificantSubdomain firstSignificantSubdomainCustom firstSignificantSubdomainCustomRFC -firstSignificantSubdomainRFC flattenTuple floor format @@ -600,8 +586,6 @@ polygonsUnionCartesian polygonsUnionSpherical polygonsWithinCartesian polygonsWithinSpherical -port -portRFC position positionCaseInsensitive positionCaseInsensitiveUTF8 @@ -906,8 +890,6 @@ toYear toYearWeek today tokens -topLevelDomain -topLevelDomainRFC transactionID transactionLatestSnapshot transactionOldestSnapshot From e0852ba028b59db7e538cf0265c768b66fa04513 Mon Sep 17 00:00:00 2001 From: "Mikhail f. Shiryaev" Date: Wed, 26 Oct 2022 13:15:32 +0200 Subject: [PATCH 102/239] Always run BuilderReport in all CI types --- .github/workflows/backport_branches.yml | 2 ++ .github/workflows/master.yml | 2 ++ .github/workflows/release_branches.yml | 2 ++ 3 files changed, 6 insertions(+) diff --git a/.github/workflows/backport_branches.yml b/.github/workflows/backport_branches.yml index 4c8d023f2ec..30a77a9b27f 100644 --- a/.github/workflows/backport_branches.yml +++ b/.github/workflows/backport_branches.yml @@ -466,6 +466,7 @@ jobs: - BuilderDebTsan - BuilderDebDebug runs-on: [self-hosted, style-checker] + if: ${{ success() || failure() }} steps: - name: Set envs run: | @@ -504,6 +505,7 @@ jobs: - BuilderBinDarwin - BuilderBinDarwinAarch64 runs-on: [self-hosted, style-checker] + if: ${{ success() || failure() }} steps: - name: Set envs run: | diff --git a/.github/workflows/master.yml b/.github/workflows/master.yml index 3d22cb984dd..fba8a975ca6 100644 --- a/.github/workflows/master.yml +++ b/.github/workflows/master.yml @@ -974,6 +974,7 @@ jobs: - BuilderDebTsan - BuilderDebUBsan runs-on: [self-hosted, style-checker] + if: ${{ success() || failure() }} steps: - name: Set envs run: | @@ -1021,6 +1022,7 @@ jobs: - BuilderBinClangTidy - BuilderDebShared runs-on: [self-hosted, style-checker] + if: ${{ success() || failure() }} steps: - name: Set envs run: | diff --git a/.github/workflows/release_branches.yml b/.github/workflows/release_branches.yml index 8f42ca92646..abe85d3e72d 100644 --- a/.github/workflows/release_branches.yml +++ b/.github/workflows/release_branches.yml @@ -541,6 +541,7 @@ jobs: - BuilderDebMsan - BuilderDebDebug runs-on: [self-hosted, style-checker] + if: ${{ success() || failure() }} steps: - name: Set envs run: | @@ -580,6 +581,7 @@ jobs: - BuilderBinDarwin - BuilderBinDarwinAarch64 runs-on: [self-hosted, style-checker] + if: ${{ success() || failure() }} steps: - name: Set envs run: | From 4e2f56ac34e6a92da3db446be4fe12d3ada63296 Mon Sep 17 00:00:00 2001 From: "Mikhail f. Shiryaev" Date: Wed, 26 Oct 2022 13:24:51 +0200 Subject: [PATCH 103/239] Try to force StyleCheck running --- .github/workflows/pull_request.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/pull_request.yml b/.github/workflows/pull_request.yml index 2795dc62d6d..3951f99b16b 100644 --- a/.github/workflows/pull_request.yml +++ b/.github/workflows/pull_request.yml @@ -112,7 +112,7 @@ jobs: StyleCheck: needs: DockerHubPush runs-on: [self-hosted, style-checker] - if: ${{ success() || failure() }} + if: ${{ success() || failure() || always() }} steps: - name: Set envs run: | From c47ca522f40178e3e5b4cd1a675499e2d42327b8 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 13:27:14 +0200 Subject: [PATCH 104/239] Fixed style check --- src/Storages/MergeTree/RPNBuilder.cpp | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/Storages/MergeTree/RPNBuilder.cpp b/src/Storages/MergeTree/RPNBuilder.cpp index 362e02e8498..05cc6492624 100644 --- a/src/Storages/MergeTree/RPNBuilder.cpp +++ b/src/Storages/MergeTree/RPNBuilder.cpp @@ -21,6 +21,11 @@ namespace DB { +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; +} + namespace { From 0f93c75dd92f0dd4dc8ded906f39828af570b441 Mon Sep 17 00:00:00 2001 From: DanRoscigno Date: Wed, 26 Oct 2022 09:12:42 -0400 Subject: [PATCH 105/239] add recent merges to CHANGELOG --- CHANGELOG.md | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/CHANGELOG.md b/CHANGELOG.md index 22f6afc4901..27b3358b26e 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -31,6 +31,8 @@ * Add OpenTelemetry support to ON CLUSTER DDL (require `distributed_ddl_entry_format_version` to be set to 4). [#41484](https://github.com/ClickHouse/ClickHouse/pull/41484) ([Frank Chen](https://github.com/FrankChen021)). * Added system table `asynchronous_insert_log`. It contains information about asynchronous inserts (including results of queries in fire-and-forget mode (with `wait_for_async_insert=0`)) for better introspection. [#42040](https://github.com/ClickHouse/ClickHouse/pull/42040) ([Anton Popov](https://github.com/CurtizJ)). * Add support for methods `lz4`, `bz2`, `snappy` in HTTP's `Accept-Encoding` which is a non-standard extension to HTTP protocol. [#42071](https://github.com/ClickHouse/ClickHouse/pull/42071) ([Nikolay Degterinsky](https://github.com/evillique)). +* Adds Morton Coding (ZCurve) encode/decode functions. [#41753](https://github.com/ClickHouse/ClickHouse/pull/41753) ([Constantine Peresypkin](https://github.com/pkit)). +* Add support for `SET setting_name = DEFAULT`. [#42187](https://github.com/ClickHouse/ClickHouse/pull/42187) ([Filatenkov Artur](https://github.com/FArthur-cmd)). #### Experimental Feature * Added new infrastructure for query analysis and planning under the `allow_experimental_analyzer` setting. [#31796](https://github.com/ClickHouse/ClickHouse/pull/31796) ([Maksim Kita](https://github.com/kitaisreal)). @@ -86,6 +88,9 @@ Only allow clients connecting to a secure server with an invalid certificate onl * Fix rarely invalid cast of aggregate state types with complex types such as Decimal. This fixes [#42408](https://github.com/ClickHouse/ClickHouse/issues/42408). [#42417](https://github.com/ClickHouse/ClickHouse/pull/42417) ([Amos Bird](https://github.com/amosbird)). * Allow to use `Date32` arguments for `dateName` function. [#42554](https://github.com/ClickHouse/ClickHouse/pull/42554) ([Roman Vasin](https://github.com/rvasin)). * Now filters with NULL literals will be used during index analysis. [#34063](https://github.com/ClickHouse/ClickHouse/issues/34063). [#41842](https://github.com/ClickHouse/ClickHouse/pull/41842) ([Amos Bird](https://github.com/amosbird)). +* Merge parts if every part in the range is older than a certain threshold. The threshold can be set by using `min_age_to_force_merge_seconds`. This closes [#35836](https://github.com/ClickHouse/ClickHouse/issues/35836). [#42423](https://github.com/ClickHouse/ClickHouse/pull/42423) ([Antonio Andelic](https://github.com/antonio2368)). This is continuation of [#39550i](https://github.com/ClickHouse/ClickHouse/pull/39550) by [@fastio](https://github.com/fastio) who implemented most of the logic. +* Added new infrastructure for query analysis and planning under `allow_experimental_analyzer` setting. [#31796](https://github.com/ClickHouse/ClickHouse/pull/31796) ([Maksim Kita](https://github.com/kitaisreal)). +* Improve the time to recover lost keeper connections. [#42541](https://github.com/ClickHouse/ClickHouse/pull/42541) ([Raúl Marín](https://github.com/Algunenano)). #### Build/Testing/Packaging Improvement * Add fuzzer for table definitions [#40096](https://github.com/ClickHouse/ClickHouse/pull/40096) ([Anton Popov](https://github.com/CurtizJ)). This represents the biggest advancement for ClickHouse testing in this year so far. @@ -139,6 +144,7 @@ Only allow clients connecting to a secure server with an invalid certificate onl * Fix bad_cast assert during INSERT into `Annoy` indexes over non-Float32 columns. `Annoy` indices is an experimental feature. [#42485](https://github.com/ClickHouse/ClickHouse/pull/42485) ([Robert Schulze](https://github.com/rschu1ze)). * Arithmetic operator with Date or DateTime and 128 or 256-bit integer was referencing uninitialized memory. [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). * Fix unexpected table loading error when partition key contains alias function names during server upgrade. [#36379](https://github.com/ClickHouse/ClickHouse/pull/36379) ([Amos Bird](https://github.com/amosbird)). +* Fixes a crash in `JSONExtract` with `LowCardinality`. [#42633](https://github.com/ClickHouse/ClickHouse/pull/42633) ([Anton Popov](https://github.com/CurtizJ)). ### ClickHouse release 22.9, 2022-09-22 From 5bc8b267b5f00e160963f7cdc0e96439010198db Mon Sep 17 00:00:00 2001 From: DanRoscigno Date: Wed, 26 Oct 2022 09:16:03 -0400 Subject: [PATCH 106/239] remove newline --- CHANGELOG.md | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index 27b3358b26e..3948a24d52a 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -68,8 +68,7 @@ * Allow readable size values (like `1TB`) in cache config. [#41688](https://github.com/ClickHouse/ClickHouse/pull/41688) ([Kseniia Sumarokova](https://github.com/kssenii)). * ClickHouse could cache stale DNS entries for some period of time (15 seconds by default) until the cache won't be updated asynchronously. During these periods ClickHouse can nevertheless try to establish a connection and produce errors. This behavior is fixed. [#41707](https://github.com/ClickHouse/ClickHouse/pull/41707) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). * Add interactive history search with fzf-like utility (fzf/sk) for `clickhouse-client`/`clickhouse-local` (note you can use `FZF_DEFAULT_OPTS`/`SKIM_DEFAULT_OPTIONS` to additionally configure the behavior). [#41730](https://github.com/ClickHouse/ClickHouse/pull/41730) ([Azat Khuzhin](https://github.com/azat)). -* -Only allow clients connecting to a secure server with an invalid certificate only to proceed with the '--accept-certificate' flag. [#41743](https://github.com/ClickHouse/ClickHouse/pull/41743) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). +* Only allow clients connecting to a secure server with an invalid certificate only to proceed with the '--accept-certificate' flag. [#41743](https://github.com/ClickHouse/ClickHouse/pull/41743) ([Yakov Olkhovskiy](https://github.com/yakov-olkhovskiy)). * Add function `tryBase58Decode`, similar to the existing function `tryBase64Decode`. [#41824](https://github.com/ClickHouse/ClickHouse/pull/41824) ([Robert Schulze](https://github.com/rschu1ze)). * Improve feedback when replacing partition with different primary key. Fixes [#34798](https://github.com/ClickHouse/ClickHouse/issues/34798). [#41838](https://github.com/ClickHouse/ClickHouse/pull/41838) ([Salvatore](https://github.com/tbsal)). * Fix parallel parsing: segmentator now checks `max_block_size`. This fixed memory overallocation in case of parallel parsing and small LIMIT. [#41852](https://github.com/ClickHouse/ClickHouse/pull/41852) ([Vitaly Baranov](https://github.com/vitlibar)). From d3461abd1f0c43055daaff0def45c32de2b203f1 Mon Sep 17 00:00:00 2001 From: "Mikhail f. Shiryaev" Date: Wed, 26 Oct 2022 15:27:09 +0200 Subject: [PATCH 107/239] Update version to 22.11.1.1 --- cmake/autogenerated_versions.txt | 10 ++++---- .../StorageSystemContributors.generated.cpp | 24 +++++++++++++++++++ 2 files changed, 29 insertions(+), 5 deletions(-) diff --git a/cmake/autogenerated_versions.txt b/cmake/autogenerated_versions.txt index 9728451f38a..11b37f5a7c8 100644 --- a/cmake/autogenerated_versions.txt +++ b/cmake/autogenerated_versions.txt @@ -2,11 +2,11 @@ # NOTE: has nothing common with DBMS_TCP_PROTOCOL_VERSION, # only DBMS_TCP_PROTOCOL_VERSION should be incremented on protocol changes. -SET(VERSION_REVISION 54467) +SET(VERSION_REVISION 54468) SET(VERSION_MAJOR 22) -SET(VERSION_MINOR 10) +SET(VERSION_MINOR 11) SET(VERSION_PATCH 1) -SET(VERSION_GITHASH 3030d4c7ff09ec44ab07d0a8069ea923227288a1) -SET(VERSION_DESCRIBE v22.10.1.1-testing) -SET(VERSION_STRING 22.10.1.1) +SET(VERSION_GITHASH 98ab5a3c189232ea2a3dddb9d2be7196ae8b3434) +SET(VERSION_DESCRIBE v22.11.1.1-testing) +SET(VERSION_STRING 22.11.1.1) # end of autochange diff --git a/src/Storages/System/StorageSystemContributors.generated.cpp b/src/Storages/System/StorageSystemContributors.generated.cpp index 25eae3b83b6..e1f4f7b82bf 100644 --- a/src/Storages/System/StorageSystemContributors.generated.cpp +++ b/src/Storages/System/StorageSystemContributors.generated.cpp @@ -12,6 +12,7 @@ const char * auto_contributors[] { "821008736@qq.com", "ANDREI STAROVEROV", "Aaron Katz", + "Adam Rutkowski", "Adri Fernandez", "Ahmed Dardery", "Aimiyoo", @@ -76,11 +77,15 @@ const char * auto_contributors[] { "Alexey Elymanov", "Alexey Gusev", "Alexey Ilyukhov", + "Alexey Ivanov", "Alexey Milovidov", "Alexey Tronov", "Alexey Vasiliev", "Alexey Zatelepin", "Alexsey Shestakov", + "AlfVII", + "Alfonso Martinez", + "Alfred Xu", "Ali Demirci", "Aliaksandr Pliutau", "Aliaksandr Shylau", @@ -196,6 +201,7 @@ const char * auto_contributors[] { "Brian Hunter", "Bulat Gaifullin", "Carbyn", + "Carlos Rodríguez Hernández", "Caspian", "Chao Ma", "Chao Wang", @@ -222,6 +228,7 @@ const char * auto_contributors[] { "DIAOZHAFENG", "Dale McDiarmid", "Dale Mcdiarmid", + "Dalitso Banda", "Dan Roscigno", "DanRoscigno", "Daniel Bershatsky", @@ -267,6 +274,7 @@ const char * auto_contributors[] { "Dmitry S..ky / skype: dvska-at-skype", "Dmitry Ukolov", "Doge", + "Dom Del Nano", "Dongdong Yang", "DoomzD", "Dr. Strange Looker", @@ -276,6 +284,7 @@ const char * auto_contributors[] { "Egor Savin", "Ekaterina", "Eldar Zaitov", + "Elena", "Elena Baskakova", "Elghazal Ahmed", "Elizaveta Mironyuk", @@ -342,6 +351,7 @@ const char * auto_contributors[] { "Grigory Pervakov", "GruffGemini", "Guillaume Tassery", + "Guo Wangyang", "Guo Wei (William)", "Haavard Kvaalen", "Habibullah Oladepo", @@ -349,6 +359,7 @@ const char * auto_contributors[] { "Hakob Saghatelyan", "Hamoon", "Han Fei", + "Han Shukai", "Harry Lee", "Harry-Lee", "HarryLeeIBM", @@ -404,6 +415,7 @@ const char * auto_contributors[] { "Jack Song", "JackyWoo", "Jacob Hayes", + "Jacob Herrington", "Jake Liu", "Jakub Kuklis", "James Maidment", @@ -419,6 +431,7 @@ const char * auto_contributors[] { "Jiading Guo", "Jiang Tao", "Jianmei Zhang", + "Jiebin Sun", "Jochen Schalanda", "John", "John Hummel", @@ -432,6 +445,7 @@ const char * auto_contributors[] { "Julian Gilyadov", "Julian Zhou", "Julio Jimenez", + "Jus", "Justin Hilliard", "Kang Liu", "Karl Pietrzak", @@ -652,6 +666,7 @@ const char * auto_contributors[] { "OuO", "PHO", "Pablo Alegre", + "Pablo Marcos", "Paramtamtam", "Patrick Zippenfenig", "Paul Loyd", @@ -681,6 +696,7 @@ const char * auto_contributors[] { "Prashant Shahi", "Pxl", "Pysaoke", + "Quanfa Fu", "Quid37", "Rafael Acevedo", "Rafael David Tinoco", @@ -693,6 +709,7 @@ const char * auto_contributors[] { "RedClusive", "RegulusZ", "Reilee", + "Reinaldy Rafli", "Reto Kromer", "Ri", "Rich Raposa", @@ -726,6 +743,7 @@ const char * auto_contributors[] { "Sachin", "Safronov Michail", "SaltTan", + "Salvatore Mesoraca", "Sami Kerola", "Samuel Chou", "San", @@ -927,6 +945,7 @@ const char * auto_contributors[] { "ZhiYong Wang", "Zhichang Yu", "Zhichun Wu", + "Zhiguo Zhou", "Zhipeng", "Zijie Lu", "Zoran Pandovski", @@ -950,6 +969,7 @@ const char * auto_contributors[] { "alexander goryanets", "alexander kozhikhov", "alexey-milovidov", + "alexeyerm", "alexeypavlenko", "alfredlu", "amesaru", @@ -1131,6 +1151,7 @@ const char * auto_contributors[] { "jennyma", "jetgm", "jewisliu", + "jferroal", "jiahui-97", "jianmei zhang", "jinjunzh", @@ -1236,6 +1257,7 @@ const char * auto_contributors[] { "mo-avatar", "morty", "moscas", + "mosinnik", "mreddy017", "msaf1980", "msirm", @@ -1321,6 +1343,7 @@ const char * auto_contributors[] { "simon-says", "snyk-bot", "songenjie", + "sperlingxx", "spff", "spongedc", "spume", @@ -1422,6 +1445,7 @@ const char * auto_contributors[] { "zhongyuankai", "zhoubintao", "zhukai", + "zimv", "zkun", "zlx19950903", "zombee0", From 42f5a3b2f886711459771526e012710d068c2fd0 Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Wed, 26 Oct 2022 13:34:29 +0000 Subject: [PATCH 108/239] Update version_date.tsv and changelogs after v22.10.1.1877-stable --- docker/server/Dockerfile.alpine | 2 +- docker/server/Dockerfile.ubuntu | 2 +- .../{v22.10.1.1875-stable.md => v22.10.1.1877-stable.md} | 3 ++- utils/list-versions/version_date.tsv | 2 +- 4 files changed, 5 insertions(+), 4 deletions(-) rename docs/changelogs/{v22.10.1.1875-stable.md => v22.10.1.1877-stable.md} (99%) diff --git a/docker/server/Dockerfile.alpine b/docker/server/Dockerfile.alpine index 9b633b66188..cf4eb3fe645 100644 --- a/docker/server/Dockerfile.alpine +++ b/docker/server/Dockerfile.alpine @@ -33,7 +33,7 @@ RUN arch=${TARGETARCH:-amd64} \ # lts / testing / prestable / etc ARG REPO_CHANNEL="stable" ARG REPOSITORY="https://packages.clickhouse.com/tgz/${REPO_CHANNEL}" -ARG VERSION="22.10.1.1875" +ARG VERSION="22.10.1.1877" ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static" # user/group precreated explicitly with fixed uid/gid on purpose. diff --git a/docker/server/Dockerfile.ubuntu b/docker/server/Dockerfile.ubuntu index 35c78763a31..d26657a7979 100644 --- a/docker/server/Dockerfile.ubuntu +++ b/docker/server/Dockerfile.ubuntu @@ -21,7 +21,7 @@ RUN sed -i "s|http://archive.ubuntu.com|${apt_archive}|g" /etc/apt/sources.list ARG REPO_CHANNEL="stable" ARG REPOSITORY="deb https://packages.clickhouse.com/deb ${REPO_CHANNEL} main" -ARG VERSION="22.10.1.1875" +ARG VERSION="22.10.1.1877" ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static" # set non-empty deb_location_url url to create a docker image diff --git a/docs/changelogs/v22.10.1.1875-stable.md b/docs/changelogs/v22.10.1.1877-stable.md similarity index 99% rename from docs/changelogs/v22.10.1.1875-stable.md rename to docs/changelogs/v22.10.1.1877-stable.md index 49f93168a00..77e540ce928 100644 --- a/docs/changelogs/v22.10.1.1875-stable.md +++ b/docs/changelogs/v22.10.1.1877-stable.md @@ -5,7 +5,7 @@ sidebar_label: 2022 # 2022 Changelog -### ClickHouse release v22.10.1.1875-stable (011ae8675a2) FIXME as compared to v22.9.1.2603-stable (3030d4c7ff0) +### ClickHouse release v22.10.1.1877-stable (98ab5a3c189) FIXME as compared to v22.9.1.2603-stable (3030d4c7ff0) #### Backward Incompatible Change * Rename cache commands: `show caches` -> `show filesystem caches`, `describe cache` -> `describe filesystem cache`. [#41508](https://github.com/ClickHouse/ClickHouse/pull/41508) ([Kseniia Sumarokova](https://github.com/kssenii)). @@ -348,4 +348,5 @@ sidebar_label: 2022 * Fix typo in cmake code related to fuzzing [#42627](https://github.com/ClickHouse/ClickHouse/pull/42627) ([Nikita Mikhaylov](https://github.com/nikitamikhaylov)). * Fix build [#42635](https://github.com/ClickHouse/ClickHouse/pull/42635) ([Anton Popov](https://github.com/CurtizJ)). * Add .rgignore for test data [#42639](https://github.com/ClickHouse/ClickHouse/pull/42639) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Fix flaky 02457_datediff_via_unix_epoch test [#42655](https://github.com/ClickHouse/ClickHouse/pull/42655) ([Roman Vasin](https://github.com/rvasin)). diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 7bbd8547506..e72fce63fda 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -1,4 +1,4 @@ -v22.10.1.1875-stable 2022-10-26 +v22.10.1.1877-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 v22.9.2.7-stable 2022-09-23 v22.9.1.2603-stable 2022-09-22 From 9b4da8c89ae79f0f6e4da4c8c20d862ef52a48e9 Mon Sep 17 00:00:00 2001 From: "Mikhail f. Shiryaev" Date: Wed, 26 Oct 2022 15:41:12 +0200 Subject: [PATCH 109/239] Update SECURITY.md --- SECURITY.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/SECURITY.md b/SECURITY.md index fb6caa92cb8..0fb333c8ea3 100644 --- a/SECURITY.md +++ b/SECURITY.md @@ -10,9 +10,11 @@ The following versions of ClickHouse server are currently being supported with s | Version | Supported | |:-|:-| +| 22.10 | ✔️ | +| 22.9 | ✔️ | | 22.8 | ✔️ | -| 22.7 | ✔️ | -| 22.6 | ✔️ | +| 22.7 | ❌ | +| 22.6 | ❌ | | 22.5 | ❌ | | 22.4 | ❌ | | 22.3 | ✔️ | From 1c17e9d45446f7d6e3dd4d6dfbc4aa64f203a074 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 15:54:37 +0200 Subject: [PATCH 110/239] Fixed tests --- .../optimizePrimaryKeyCondition.cpp | 16 +++--- .../QueryPlan/ReadFromMergeTree.cpp | 6 ++- src/Storages/MergeTree/RPNBuilder.cpp | 51 +++++++++++++------ src/Storages/MergeTree/RPNBuilder.h | 2 +- 4 files changed, 50 insertions(+), 25 deletions(-) diff --git a/src/Processors/QueryPlan/Optimizations/optimizePrimaryKeyCondition.cpp b/src/Processors/QueryPlan/Optimizations/optimizePrimaryKeyCondition.cpp index 7d682c408e5..984c76701ba 100644 --- a/src/Processors/QueryPlan/Optimizations/optimizePrimaryKeyCondition.cpp +++ b/src/Processors/QueryPlan/Optimizations/optimizePrimaryKeyCondition.cpp @@ -17,7 +17,7 @@ void optimizePrimaryKeyCondition(QueryPlan::Node & root) size_t next_child = 0; }; - std::deque stack; + std::vector stack; stack.push_back({.node = &root}); while (!stack.empty()) @@ -27,29 +27,29 @@ void optimizePrimaryKeyCondition(QueryPlan::Node & root) /// Traverse all children first. if (frame.next_child < frame.node->children.size()) { - stack.push_back({.node = frame.node->children[frame.next_child]}); - + auto next_frame = Frame{.node = frame.node->children[frame.next_child]}; ++frame.next_child; + stack.push_back(next_frame); continue; } - auto add_filter = [&](auto & storage) + auto add_read_from_storage_filter = [&](auto & storage) { - for (auto iter=stack.rbegin() + 1; iter!=stack.rend(); ++iter) + for (auto iter = stack.rbegin() + 1; iter != stack.rend(); ++iter) { if (auto * filter_step = typeid_cast(iter->node->step.get())) storage.addFilter(filter_step->getExpression(), filter_step->getFilterColumnName()); else if (typeid_cast(iter->node->step.get())) - ; + continue; else break; } }; if (auto * read_from_merge_tree = typeid_cast(frame.node->step.get())) - add_filter(*read_from_merge_tree); + add_read_from_storage_filter(*read_from_merge_tree); else if (auto * read_from_merge = typeid_cast(frame.node->step.get())) - add_filter(*read_from_merge); + add_read_from_storage_filter(*read_from_merge); stack.pop_back(); } diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.cpp b/src/Processors/QueryPlan/ReadFromMergeTree.cpp index 71f8378ae41..53568b6a79e 100644 --- a/src/Processors/QueryPlan/ReadFromMergeTree.cpp +++ b/src/Processors/QueryPlan/ReadFromMergeTree.cpp @@ -906,11 +906,15 @@ MergeTreeDataSelectAnalysisResultPtr ReadFromMergeTree::selectRangesToRead( for (const auto & node : added_filter_nodes.nodes) nodes.nodes.push_back(node); + NameSet array_join_name_set; + if (query_info.syntax_analyzer_result) + array_join_name_set = query_info.syntax_analyzer_result->getArrayJoinSourceNameSet(); + key_condition.emplace(std::move(nodes), context, primary_key_columns, primary_key.expression, - query_info.syntax_analyzer_result->getArrayJoinSourceNameSet()); + array_join_name_set); } else { diff --git a/src/Storages/MergeTree/RPNBuilder.cpp b/src/Storages/MergeTree/RPNBuilder.cpp index 05cc6492624..4cc311b8b9e 100644 --- a/src/Storages/MergeTree/RPNBuilder.cpp +++ b/src/Storages/MergeTree/RPNBuilder.cpp @@ -33,10 +33,10 @@ void appendColumnNameWithoutAlias(const ActionsDAG::Node & node, WriteBuffer & o { switch (node.type) { - case (ActionsDAG::ActionType::INPUT): + case ActionsDAG::ActionType::INPUT: writeString(node.result_name, out); break; - case (ActionsDAG::ActionType::COLUMN): + case ActionsDAG::ActionType::COLUMN: { /// If it was created from ASTLiteral, then result_name can be an alias. /// We need to convert value back to string here. @@ -47,15 +47,15 @@ void appendColumnNameWithoutAlias(const ActionsDAG::Node & node, WriteBuffer & o writeString(node.result_name, out); break; } - case (ActionsDAG::ActionType::ALIAS): + case ActionsDAG::ActionType::ALIAS: appendColumnNameWithoutAlias(*node.children.front(), out, legacy); break; - case (ActionsDAG::ActionType::ARRAY_JOIN): + case ActionsDAG::ActionType::ARRAY_JOIN: writeCString("arrayJoin(", out); appendColumnNameWithoutAlias(*node.children.front(), out, legacy); writeChar(')', out); break; - case (ActionsDAG::ActionType::FUNCTION): + case ActionsDAG::ActionType::FUNCTION: { auto name = node.function_base->getName(); if (legacy && name == "modulo") @@ -144,9 +144,23 @@ bool RPNBuilderTreeNode::isFunction() const bool RPNBuilderTreeNode::isConstant() const { if (ast_node) - return typeid_cast(ast_node); + { + bool is_literal = typeid_cast(ast_node); + if (is_literal) + return true; + + String column_name = ast_node->getColumnName(); + const auto & block_with_constants = tree_context.getBlockWithConstants(); + + if (block_with_constants.has(column_name) && isColumnConst(*block_with_constants.getByName(column_name).column)) + return true; + + return false; + } else + { return dag_node->column && isColumnConst(*dag_node->column); + } } ColumnWithTypeAndName RPNBuilderTreeNode::getConstantColumn() const @@ -154,28 +168,35 @@ ColumnWithTypeAndName RPNBuilderTreeNode::getConstantColumn() const if (!isConstant()) throw Exception(ErrorCodes::LOGICAL_ERROR, "RPNBuilderTree node is not a constant"); - ColumnWithTypeAndName res; + ColumnWithTypeAndName result; if (ast_node) { const auto * literal = assert_cast(ast_node); - res.type = applyVisitor(FieldToDataType(), literal->value); - res.column = res.type->createColumnConst(0, literal->value); + if (literal) + { + result.type = applyVisitor(FieldToDataType(), literal->value); + result.column = result.type->createColumnConst(0, literal->value); + + return result; + } + + String column_name = ast_node->getColumnName(); + const auto & block_with_constants = tree_context.getBlockWithConstants(); + + return block_with_constants.getByName(column_name); } else { - res.type = dag_node->result_type; - res.column = dag_node->column; + result.type = dag_node->result_type; + result.column = dag_node->column; } - return res; + return result; } bool RPNBuilderTreeNode::tryGetConstant(Field & output_value, DataTypePtr & output_type) const { - if (!isConstant()) - return false; - if (ast_node) { // Constant expr should use alias names if any diff --git a/src/Storages/MergeTree/RPNBuilder.h b/src/Storages/MergeTree/RPNBuilder.h index d4fc09dcdab..f6fc4a70102 100644 --- a/src/Storages/MergeTree/RPNBuilder.h +++ b/src/Storages/MergeTree/RPNBuilder.h @@ -181,7 +181,7 @@ public: using RPNElements = std::vector; using ExtractAtomFromTreeFunction = std::function; - explicit RPNBuilder(const ActionsDAG::Node * & filter_actions_dag_node, + explicit RPNBuilder(const ActionsDAG::Node * filter_actions_dag_node, ContextPtr query_context_, const ExtractAtomFromTreeFunction & extract_atom_from_tree_function_) : tree_context(std::move(query_context_)) From 2b44a00fbc39db06f0c2e11db35b5608b9341aa1 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 16:47:40 +0200 Subject: [PATCH 111/239] UNION node remove unnecessary union modes --- src/Analyzer/QueryTreeBuilder.cpp | 23 ++++++++--------- src/Analyzer/UnionNode.cpp | 42 ++++++++----------------------- src/Analyzer/UnionNode.h | 25 +++--------------- 3 files changed, 23 insertions(+), 67 deletions(-) diff --git a/src/Analyzer/QueryTreeBuilder.cpp b/src/Analyzer/QueryTreeBuilder.cpp index 890aa2b01a2..b68d7d3d387 100644 --- a/src/Analyzer/QueryTreeBuilder.cpp +++ b/src/Analyzer/QueryTreeBuilder.cpp @@ -145,12 +145,10 @@ QueryTreeNodePtr QueryTreeBuilder::buildSelectWithUnionExpression(const ASTPtr & if (select_lists.children.size() == 1) return buildSelectOrUnionExpression(select_lists.children[0], is_subquery, cte_name); - auto union_node = std::make_shared(); + auto union_node = std::make_shared(select_with_union_query_typed.union_mode); union_node->setIsSubquery(is_subquery); union_node->setIsCTE(!cte_name.empty()); union_node->setCTEName(cte_name); - union_node->setUnionMode(select_with_union_query_typed.union_mode); - union_node->setUnionModes(select_with_union_query_typed.list_of_modes); union_node->setOriginalAST(select_with_union_query); size_t select_lists_children_size = select_lists.children.size(); @@ -173,23 +171,22 @@ QueryTreeNodePtr QueryTreeBuilder::buildSelectIntersectExceptQuery(const ASTPtr if (select_lists.size() == 1) return buildSelectExpression(select_lists[0], is_subquery, cte_name); - auto union_node = std::make_shared(); - union_node->setIsSubquery(is_subquery); - union_node->setIsCTE(!cte_name.empty()); - union_node->setCTEName(cte_name); - + SelectUnionMode union_mode; if (select_intersect_except_query_typed.final_operator == ASTSelectIntersectExceptQuery::Operator::INTERSECT_ALL) - union_node->setUnionMode(SelectUnionMode::INTERSECT_ALL); + union_mode = SelectUnionMode::INTERSECT_ALL; else if (select_intersect_except_query_typed.final_operator == ASTSelectIntersectExceptQuery::Operator::INTERSECT_DISTINCT) - union_node->setUnionMode(SelectUnionMode::INTERSECT_DISTINCT); + union_mode = SelectUnionMode::INTERSECT_DISTINCT; else if (select_intersect_except_query_typed.final_operator == ASTSelectIntersectExceptQuery::Operator::EXCEPT_ALL) - union_node->setUnionMode(SelectUnionMode::EXCEPT_ALL); + union_mode = SelectUnionMode::EXCEPT_ALL; else if (select_intersect_except_query_typed.final_operator == ASTSelectIntersectExceptQuery::Operator::EXCEPT_DISTINCT) - union_node->setUnionMode(SelectUnionMode::EXCEPT_DISTINCT); + union_mode = SelectUnionMode::EXCEPT_DISTINCT; else throw Exception(ErrorCodes::LOGICAL_ERROR, "UNION type is not initialized"); - union_node->setUnionModes(SelectUnionModes(select_lists.size() - 1, union_node->getUnionMode())); + auto union_node = std::make_shared(union_mode); + union_node->setIsSubquery(is_subquery); + union_node->setIsCTE(!cte_name.empty()); + union_node->setCTEName(cte_name); union_node->setOriginalAST(select_intersect_except_query); size_t select_lists_size = select_lists.size(); diff --git a/src/Analyzer/UnionNode.cpp b/src/Analyzer/UnionNode.cpp index b8ed46c645e..74992652a1c 100644 --- a/src/Analyzer/UnionNode.cpp +++ b/src/Analyzer/UnionNode.cpp @@ -32,9 +32,15 @@ namespace ErrorCodes extern const int TYPE_MISMATCH; } -UnionNode::UnionNode() +UnionNode::UnionNode(SelectUnionMode union_mode_) : IQueryTreeNode(children_size) + , union_mode(union_mode_) { + if (union_mode == SelectUnionMode::UNION_DEFAULT || + union_mode == SelectUnionMode::EXCEPT_DEFAULT || + union_mode == SelectUnionMode::INTERSECT_DEFAULT) + throw Exception(ErrorCodes::BAD_ARGUMENTS, "UNION mode must be normalized"); + children[queries_child_index] = std::make_shared(); } @@ -109,20 +115,6 @@ void UnionNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, s buffer << ", union_mode: " << toString(union_mode); - size_t union_modes_size = union_modes.size(); - buffer << '\n' << std::string(indent + 2, ' ') << "UNION MODES " << union_modes_size << '\n'; - - for (size_t i = 0; i < union_modes_size; ++i) - { - buffer << std::string(indent + 4, ' '); - - auto query_union_mode = union_modes[i]; - buffer << toString(query_union_mode); - - if (i + 1 != union_modes_size) - buffer << '\n'; - } - buffer << '\n' << std::string(indent + 2, ' ') << "QUERIES\n"; getQueriesNode()->dumpTreeImpl(buffer, format_state, indent + 4); } @@ -145,7 +137,7 @@ bool UnionNode::isEqualImpl(const IQueryTreeNode & rhs) const return false; return is_subquery == rhs_typed.is_subquery && is_cte == rhs_typed.is_cte && cte_name == rhs_typed.cte_name && - union_mode == rhs_typed.union_mode && union_modes == rhs_typed.union_modes; + union_mode == rhs_typed.union_mode; } void UnionNode::updateTreeHashImpl(HashState & state) const @@ -158,10 +150,6 @@ void UnionNode::updateTreeHashImpl(HashState & state) const state.update(static_cast(union_mode)); - state.update(union_modes.size()); - for (const auto & query_union_mode : union_modes) - state.update(static_cast(query_union_mode)); - if (constant_value) { auto constant_dump = applyVisitor(FieldVisitorToString(), constant_value->getValue()); @@ -179,14 +167,11 @@ void UnionNode::updateTreeHashImpl(HashState & state) const QueryTreeNodePtr UnionNode::cloneImpl() const { - auto result_union_node = std::make_shared(); + auto result_union_node = std::make_shared(union_mode); result_union_node->is_subquery = is_subquery; result_union_node->is_cte = is_cte; result_union_node->cte_name = cte_name; - result_union_node->union_mode = union_mode; - result_union_node->union_modes = union_modes; - result_union_node->union_modes_set = union_modes_set; result_union_node->constant_value = constant_value; result_union_node->table_expression_modifiers = table_expression_modifiers; @@ -197,14 +182,7 @@ ASTPtr UnionNode::toASTImpl() const { auto select_with_union_query = std::make_shared(); select_with_union_query->union_mode = union_mode; - - if (union_mode != SelectUnionMode::UNION_DEFAULT && - union_mode != SelectUnionMode::EXCEPT_DEFAULT && - union_mode != SelectUnionMode::INTERSECT_DEFAULT) - select_with_union_query->is_normalized = true; - - select_with_union_query->list_of_modes = union_modes; - select_with_union_query->set_of_modes = union_modes_set; + select_with_union_query->is_normalized = true; select_with_union_query->children.push_back(getQueriesNode()->toAST()); select_with_union_query->list_of_selects = select_with_union_query->children.back(); diff --git a/src/Analyzer/UnionNode.h b/src/Analyzer/UnionNode.h index 05e70b87a27..f192741faa0 100644 --- a/src/Analyzer/UnionNode.h +++ b/src/Analyzer/UnionNode.h @@ -19,6 +19,7 @@ namespace ErrorCodes } /** Union node represents union of queries in query tree. + * Union node must be initialized with normalized union mode. * * Example: (SELECT id FROM test_table) UNION ALL (SELECT id FROM test_table_2); * Example: (SELECT id FROM test_table) UNION DISTINCT (SELECT id FROM test_table_2); @@ -41,7 +42,8 @@ using UnionNodePtr = std::shared_ptr; class UnionNode final : public IQueryTreeNode { public: - explicit UnionNode(); + /// Construct union node with normalized union mode + explicit UnionNode(SelectUnionMode union_mode_); /// Returns true if union node is subquery, false otherwise bool isSubquery() const @@ -85,25 +87,6 @@ public: return union_mode; } - /// Set union mode value - void setUnionMode(SelectUnionMode union_mode_value) - { - union_mode = union_mode_value; - } - - /// Get union modes - const SelectUnionModes & getUnionModes() const - { - return union_modes; - } - - /// Set union modes value - void setUnionModes(const SelectUnionModes & union_modes_value) - { - union_modes = union_modes_value; - union_modes_set = SelectUnionModesSet(union_modes.begin(), union_modes.end()); - } - /// Get union node queries const ListNode & getQueries() const { @@ -189,8 +172,6 @@ private: bool is_cte = false; std::string cte_name; SelectUnionMode union_mode; - SelectUnionModes union_modes; - SelectUnionModesSet union_modes_set; ConstantValuePtr constant_value; std::optional table_expression_modifiers; From 6a0e3fc31da1c07e50798b77a1bd27b569b3ede7 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 17:19:29 +0200 Subject: [PATCH 112/239] Subqueries remove unnecessary table expression modifiers --- src/Analyzer/Passes/QueryAnalysisPass.cpp | 67 +++++++++---------- src/Analyzer/QueryNode.cpp | 17 ----- src/Analyzer/QueryNode.h | 19 ------ src/Analyzer/QueryTreeBuilder.cpp | 26 ++++--- src/Analyzer/UnionNode.cpp | 17 ----- src/Analyzer/UnionNode.h | 19 ------ src/Analyzer/Utils.cpp | 5 -- ...eries_table_expression_modifiers.reference | 0 ..._subqueries_table_expression_modifiers.sql | 17 +++++ 9 files changed, 65 insertions(+), 122 deletions(-) create mode 100644 tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.reference create mode 100644 tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.sql diff --git a/src/Analyzer/Passes/QueryAnalysisPass.cpp b/src/Analyzer/Passes/QueryAnalysisPass.cpp index 6b91d0f8053..019b002c527 100644 --- a/src/Analyzer/Passes/QueryAnalysisPass.cpp +++ b/src/Analyzer/Passes/QueryAnalysisPass.cpp @@ -1621,34 +1621,7 @@ void QueryAnalyzer::validateTableExpressionModifiers(const QueryTreeNodePtr & ta table_expression_node->formatASTForErrorMessage(), scope.scope_node->formatASTForErrorMessage()); - if (query_node || union_node) - { - auto table_expression_modifiers = query_node ? query_node->getTableExpressionModifiers() : union_node->getTableExpressionModifiers(); - - if (table_expression_modifiers.has_value()) - { - String table_expression_modifiers_error_message; - - if (table_expression_modifiers->hasFinal()) - { - table_expression_modifiers_error_message += "FINAL"; - - if (table_expression_modifiers->hasSampleSizeRatio()) - table_expression_modifiers_error_message += ", SAMPLE"; - } - else if (table_expression_modifiers->hasSampleSizeRatio()) - { - table_expression_modifiers_error_message += "SAMPLE"; - } - - throw Exception(ErrorCodes::UNSUPPORTED_METHOD, - "Table expression modifiers {} are not supported for subquery {}. In scope {}", - table_expression_modifiers_error_message, - table_expression_node->formatASTForErrorMessage(), - scope.scope_node->formatASTForErrorMessage()); - } - } - else if (table_node || table_function_node) + if (table_node || table_function_node) { auto table_expression_modifiers = table_node ? table_node->getTableExpressionModifiers() : table_function_node->getTableExpressionModifiers(); @@ -4661,17 +4634,37 @@ void QueryAnalyzer::initializeQueryJoinTreeNode(QueryTreeNodePtr & join_tree_nod auto table_expression_modifiers = from_table_identifier.getTableExpressionModifiers(); - if (auto * resolved_identifier_query_node = resolved_identifier->as()) + auto * resolved_identifier_query_node = resolved_identifier->as(); + auto * resolved_identifier_union_node = resolved_identifier->as(); + + if (resolved_identifier_query_node || resolved_identifier_union_node) { - resolved_identifier_query_node->setIsCTE(false); + if (resolved_identifier_query_node) + resolved_identifier_query_node->setIsCTE(false); + else + resolved_identifier_union_node->setIsCTE(false); + if (table_expression_modifiers.has_value()) - resolved_identifier_query_node->setTableExpressionModifiers(*table_expression_modifiers); - } - else if (auto * resolved_identifier_union_node = resolved_identifier->as()) - { - resolved_identifier_union_node->setIsCTE(false); - if (table_expression_modifiers.has_value()) - resolved_identifier_union_node->setTableExpressionModifiers(*table_expression_modifiers); + { + String table_expression_modifiers_error_message; + + if (table_expression_modifiers->hasFinal()) + { + table_expression_modifiers_error_message += "FINAL"; + + if (table_expression_modifiers->hasSampleSizeRatio()) + table_expression_modifiers_error_message += ", SAMPLE"; + } + else if (table_expression_modifiers->hasSampleSizeRatio()) + { + table_expression_modifiers_error_message += "SAMPLE"; + } + + throw Exception(ErrorCodes::UNSUPPORTED_METHOD, + "Table expression modifiers {} are not supported for subquery {}", + table_expression_modifiers_error_message, + resolved_identifier->formatASTForErrorMessage()); + } } else if (auto * resolved_identifier_table_node = resolved_identifier->as()) { diff --git a/src/Analyzer/QueryNode.cpp b/src/Analyzer/QueryNode.cpp index 557baad2654..c5bbc193544 100644 --- a/src/Analyzer/QueryNode.cpp +++ b/src/Analyzer/QueryNode.cpp @@ -74,12 +74,6 @@ void QueryNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, s buffer << ", constant_value_type: " << constant_value->getType()->getName(); } - if (table_expression_modifiers) - { - buffer << ", "; - table_expression_modifiers->dump(buffer); - } - if (hasWith()) { buffer << '\n' << std::string(indent + 2, ' ') << "WITH\n"; @@ -195,13 +189,6 @@ bool QueryNode::isEqualImpl(const IQueryTreeNode & rhs) const else if (!constant_value && rhs_typed.constant_value) return false; - if (table_expression_modifiers && rhs_typed.table_expression_modifiers && table_expression_modifiers != rhs_typed.table_expression_modifiers) - return false; - else if (table_expression_modifiers && !rhs_typed.table_expression_modifiers) - return false; - else if (!table_expression_modifiers && rhs_typed.table_expression_modifiers) - return false; - return is_subquery == rhs_typed.is_subquery && is_cte == rhs_typed.is_cte && cte_name == rhs_typed.cte_name && @@ -250,9 +237,6 @@ void QueryNode::updateTreeHashImpl(HashState & state) const state.update(constant_value_type_name.size()); state.update(constant_value_type_name); } - - if (table_expression_modifiers) - table_expression_modifiers->updateTreeHash(state); } QueryTreeNodePtr QueryNode::cloneImpl() const @@ -270,7 +254,6 @@ QueryTreeNodePtr QueryNode::cloneImpl() const result_query_node->cte_name = cte_name; result_query_node->projection_columns = projection_columns; result_query_node->constant_value = constant_value; - result_query_node->table_expression_modifiers = table_expression_modifiers; return result_query_node; } diff --git a/src/Analyzer/QueryNode.h b/src/Analyzer/QueryNode.h index 6bb6613fc2b..1bb381c95c9 100644 --- a/src/Analyzer/QueryNode.h +++ b/src/Analyzer/QueryNode.h @@ -176,24 +176,6 @@ public: is_group_by_with_grouping_sets = is_group_by_with_grouping_sets_value; } - /// Return true if query node has table expression modifiers, false otherwise - bool hasTableExpressionModifiers() const - { - return table_expression_modifiers.has_value(); - } - - /// Get table expression modifiers - const std::optional & getTableExpressionModifiers() const - { - return table_expression_modifiers; - } - - /// Set table expression modifiers - void setTableExpressionModifiers(TableExpressionModifiers table_expression_modifiers_value) - { - table_expression_modifiers = std::move(table_expression_modifiers_value); - } - /// Returns true if query node WITH section is not empty, false otherwise bool hasWith() const { @@ -602,7 +584,6 @@ private: std::string cte_name; NamesAndTypes projection_columns; ConstantValuePtr constant_value; - std::optional table_expression_modifiers; SettingsChanges settings_changes; static constexpr size_t with_child_index = 0; diff --git a/src/Analyzer/QueryTreeBuilder.cpp b/src/Analyzer/QueryTreeBuilder.cpp index b68d7d3d387..c5bb3ae7074 100644 --- a/src/Analyzer/QueryTreeBuilder.cpp +++ b/src/Analyzer/QueryTreeBuilder.cpp @@ -673,14 +673,24 @@ QueryTreeNodePtr QueryTreeBuilder::buildJoinTree(const ASTPtr & tables_in_select if (table_expression_modifiers) { - if (auto * query_node = node->as()) - query_node->setTableExpressionModifiers(*table_expression_modifiers); - else if (auto * union_node = node->as()) - union_node->setTableExpressionModifiers(*table_expression_modifiers); - else - throw Exception(ErrorCodes::LOGICAL_ERROR, - "Unexpected table expression subquery node. Expected union or query. Actual {}", - node->formatASTForErrorMessage()); + String table_expression_modifiers_error_message; + + if (table_expression_modifiers->hasFinal()) + { + table_expression_modifiers_error_message += "FINAL"; + + if (table_expression_modifiers->hasSampleSizeRatio()) + table_expression_modifiers_error_message += ", SAMPLE"; + } + else if (table_expression_modifiers->hasSampleSizeRatio()) + { + table_expression_modifiers_error_message += "SAMPLE"; + } + + throw Exception(ErrorCodes::UNSUPPORTED_METHOD, + "Table expression modifiers {} are not supported for subquery {}", + table_expression_modifiers_error_message, + node->formatASTForErrorMessage()); } table_expressions.push_back(std::move(node)); diff --git a/src/Analyzer/UnionNode.cpp b/src/Analyzer/UnionNode.cpp index 74992652a1c..039077bd08a 100644 --- a/src/Analyzer/UnionNode.cpp +++ b/src/Analyzer/UnionNode.cpp @@ -107,12 +107,6 @@ void UnionNode::dumpTreeImpl(WriteBuffer & buffer, FormatState & format_state, s buffer << ", constant_value_type: " << constant_value->getType()->getName(); } - if (table_expression_modifiers) - { - buffer << ", "; - table_expression_modifiers->dump(buffer); - } - buffer << ", union_mode: " << toString(union_mode); buffer << '\n' << std::string(indent + 2, ' ') << "QUERIES\n"; @@ -129,13 +123,6 @@ bool UnionNode::isEqualImpl(const IQueryTreeNode & rhs) const else if (!constant_value && rhs_typed.constant_value) return false; - if (table_expression_modifiers && rhs_typed.table_expression_modifiers && table_expression_modifiers != rhs_typed.table_expression_modifiers) - return false; - else if (table_expression_modifiers && !rhs_typed.table_expression_modifiers) - return false; - else if (!table_expression_modifiers && rhs_typed.table_expression_modifiers) - return false; - return is_subquery == rhs_typed.is_subquery && is_cte == rhs_typed.is_cte && cte_name == rhs_typed.cte_name && union_mode == rhs_typed.union_mode; } @@ -160,9 +147,6 @@ void UnionNode::updateTreeHashImpl(HashState & state) const state.update(constant_value_type_name.size()); state.update(constant_value_type_name); } - - if (table_expression_modifiers) - table_expression_modifiers->updateTreeHash(state); } QueryTreeNodePtr UnionNode::cloneImpl() const @@ -173,7 +157,6 @@ QueryTreeNodePtr UnionNode::cloneImpl() const result_union_node->is_cte = is_cte; result_union_node->cte_name = cte_name; result_union_node->constant_value = constant_value; - result_union_node->table_expression_modifiers = table_expression_modifiers; return result_union_node; } diff --git a/src/Analyzer/UnionNode.h b/src/Analyzer/UnionNode.h index f192741faa0..9ef76591597 100644 --- a/src/Analyzer/UnionNode.h +++ b/src/Analyzer/UnionNode.h @@ -111,24 +111,6 @@ public: return children[queries_child_index]; } - /// Return true if union node has table expression modifiers, false otherwise - bool hasTableExpressionModifiers() const - { - return table_expression_modifiers.has_value(); - } - - /// Get table expression modifiers - const std::optional & getTableExpressionModifiers() const - { - return table_expression_modifiers; - } - - /// Set table expression modifiers - void setTableExpressionModifiers(TableExpressionModifiers table_expression_modifiers_value) - { - table_expression_modifiers = std::move(table_expression_modifiers_value); - } - /// Compute union node projection columns NamesAndTypes computeProjectionColumns() const; @@ -173,7 +155,6 @@ private: std::string cte_name; SelectUnionMode union_mode; ConstantValuePtr constant_value; - std::optional table_expression_modifiers; static constexpr size_t queries_child_index = 0; static constexpr size_t children_size = queries_child_index + 1; diff --git a/src/Analyzer/Utils.cpp b/src/Analyzer/Utils.cpp index 5f0d682865f..b504a5b5787 100644 --- a/src/Analyzer/Utils.cpp +++ b/src/Analyzer/Utils.cpp @@ -98,11 +98,6 @@ static ASTPtr convertIntoTableExpressionAST(const QueryTreeNodePtr & table_expre if (node_type == QueryTreeNodeType::QUERY || node_type == QueryTreeNodeType::UNION) { - if (auto * query_node = table_expression_node->as()) - table_expression_modifiers = query_node->getTableExpressionModifiers(); - else if (auto * union_node = table_expression_node->as()) - table_expression_modifiers = union_node->getTableExpressionModifiers(); - result_table_expression->subquery = result_table_expression->children.back(); } else if (node_type == QueryTreeNodeType::TABLE || node_type == QueryTreeNodeType::IDENTIFIER) diff --git a/tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.reference b/tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.sql b/tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.sql new file mode 100644 index 00000000000..456783cad26 --- /dev/null +++ b/tests/queries/0_stateless/02474_analyzer_subqueries_table_expression_modifiers.sql @@ -0,0 +1,17 @@ +SET allow_experimental_analyzer = 1; + +SELECT * FROM (SELECT 1) FINAL; -- { serverError 1 } +SELECT * FROM (SELECT 1) SAMPLE 1/2; -- { serverError 1 } +SELECT * FROM (SELECT 1) FINAL SAMPLE 1/2; -- { serverError 1 } + +WITH cte_subquery AS (SELECT 1) SELECT * FROM cte_subquery FINAL; -- { serverError 1 } +WITH cte_subquery AS (SELECT 1) SELECT * FROM cte_subquery SAMPLE 1/2; -- { serverError 1 } +WITH cte_subquery AS (SELECT 1) SELECT * FROM cte_subquery FINAL SAMPLE 1/2; -- { serverError 1 } + +SELECT * FROM (SELECT 1 UNION ALL SELECT 1) FINAL; -- { serverError 1 } +SELECT * FROM (SELECT 1 UNION ALL SELECT 1) SAMPLE 1/2; -- { serverError 1 } +SELECT * FROM (SELECT 1 UNION ALL SELECT 1) FINAL SAMPLE 1/2; -- { serverError 1 } + +WITH cte_subquery AS (SELECT 1 UNION ALL SELECT 1) SELECT * FROM cte_subquery FINAL; -- { serverError 1 } +WITH cte_subquery AS (SELECT 1 UNION ALL SELECT 1) SELECT * FROM cte_subquery SAMPLE 1/2; -- { serverError 1 } +WITH cte_subquery AS (SELECT 1 UNION ALL SELECT 1) SELECT * FROM cte_subquery FINAL SAMPLE 1/2; -- { serverError 1 } From f5f6f1b5933b4d19702f9fabf63fbb4dc02f21a6 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Sat, 22 Oct 2022 13:39:41 +0200 Subject: [PATCH 113/239] Add profile events for jemalloc purge Signed-off-by: Azat Khuzhin --- src/Common/MemoryTracker.cpp | 6 ++++++ src/Common/ProfileEvents.cpp | 2 ++ 2 files changed, 8 insertions(+) diff --git a/src/Common/MemoryTracker.cpp b/src/Common/MemoryTracker.cpp index 8bd31681706..41634e8f561 100644 --- a/src/Common/MemoryTracker.cpp +++ b/src/Common/MemoryTracker.cpp @@ -10,6 +10,7 @@ #include #include #include +#include #include #include "config.h" @@ -86,6 +87,8 @@ inline std::string_view toDescription(OvercommitResult result) namespace ProfileEvents { extern const Event QueryMemoryLimitExceeded; + extern const Event MemoryAllocatorPurge; + extern const Event MemoryAllocatorPurgeTimeMicroseconds; } using namespace std::chrono_literals; @@ -229,7 +232,10 @@ void MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceeded, MemoryT { if (free_memory_in_allocator_arenas.exchange(-current_free_memory_in_allocator_arenas) > 0) { + Stopwatch watch; mallctl("arena." STRINGIFY(MALLCTL_ARENAS_ALL) ".purge", nullptr, nullptr, nullptr, 0); + ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurge); + ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurgeTimeMicroseconds, watch.elapsedMicroseconds()); } } diff --git a/src/Common/ProfileEvents.cpp b/src/Common/ProfileEvents.cpp index 46bec669626..2f801e496fa 100644 --- a/src/Common/ProfileEvents.cpp +++ b/src/Common/ProfileEvents.cpp @@ -229,6 +229,8 @@ The server successfully detected this situation and will download merged part fr M(UserTimeMicroseconds, "Total time spent in processing (queries and other tasks) threads executing CPU instructions in user space. This include time CPU pipeline was stalled due to cache misses, branch mispredictions, hyper-threading, etc.") \ M(SystemTimeMicroseconds, "Total time spent in processing (queries and other tasks) threads executing CPU instructions in OS kernel space. This include time CPU pipeline was stalled due to cache misses, branch mispredictions, hyper-threading, etc.") \ M(MemoryOvercommitWaitTimeMicroseconds, "Total time spent in waiting for memory to be freed in OvercommitTracker.") \ + M(MemoryAllocatorPurge, "Total number of times memory allocator purge was requested") \ + M(MemoryAllocatorPurgeTimeMicroseconds, "Total number of times memory allocator purge was requested") \ M(SoftPageFaults, "The number of soft page faults in query execution threads. Soft page fault usually means a miss in the memory allocator cache which required a new memory mapping from the OS and subsequent allocation of a page of physical memory.") \ M(HardPageFaults, "The number of hard page faults in query execution threads. High values indicate either that you forgot to turn off swap on your server, or eviction of memory pages of the ClickHouse binary during very high memory pressure, or successful usage of the 'mmap' read method for the tables data.") \ \ From 2faefc0c0973b666397b6ac4a96c6732706e250b Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Sat, 22 Oct 2022 13:40:37 +0200 Subject: [PATCH 114/239] Add amount of memory used by allocator into logs Signed-off-by: Azat Khuzhin --- src/Interpreters/AsynchronousMetrics.cpp | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/Interpreters/AsynchronousMetrics.cpp b/src/Interpreters/AsynchronousMetrics.cpp index 338ae1bbbfd..2e43a594960 100644 --- a/src/Interpreters/AsynchronousMetrics.cpp +++ b/src/Interpreters/AsynchronousMetrics.cpp @@ -713,9 +713,10 @@ void AsynchronousMetrics::update(TimePoint update_time) /// Log only if difference is high. This is for convenience. The threshold is arbitrary. if (difference >= 1048576 || difference <= -1048576) LOG_TRACE(log, - "MemoryTracking: was {}, peak {}, will set to {} (RSS), difference: {}", + "MemoryTracking: was {}, peak {}, free memory in arenas {}, will set to {} (RSS), difference: {}", ReadableSize(amount), ReadableSize(peak), + ReadableSize(free_memory_in_allocator_arenas), ReadableSize(rss), ReadableSize(difference)); From 7b69a70e828a6fab644086aa589d4352c5d86883 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Sat, 22 Oct 2022 18:39:36 +0200 Subject: [PATCH 115/239] Fix frequent memory drift message and clarify things in comments Somethine like: 2022.09.28 06:33:34.001433 [ 3133669 ] {} AsynchronousMetrics: MemoryTracking: was 562.20 MiB, peak 562.21 MiB, will set to 562.20 MiB (RSS), difference: -70.46 MiB 2022.09.28 06:33:35.001639 [ 3133669 ] {} AsynchronousMetrics: MemoryTracking: was 562.20 MiB, peak 562.21 MiB, will set to 562.20 MiB (RSS), difference: -70.45 MiB Signed-off-by: Azat Khuzhin --- src/Common/MemoryTracker.cpp | 2 +- src/Interpreters/AsynchronousMetrics.cpp | 12 +++++++++--- 2 files changed, 10 insertions(+), 4 deletions(-) diff --git a/src/Common/MemoryTracker.cpp b/src/Common/MemoryTracker.cpp index 41634e8f561..b530410ec63 100644 --- a/src/Common/MemoryTracker.cpp +++ b/src/Common/MemoryTracker.cpp @@ -438,7 +438,7 @@ void MemoryTracker::reset() void MemoryTracker::setRSS(Int64 rss_, Int64 free_memory_in_allocator_arenas_) { - Int64 new_amount = rss_; // - free_memory_in_allocator_arenas_; + Int64 new_amount = rss_; total_memory_tracker.amount.store(new_amount, std::memory_order_relaxed); free_memory_in_allocator_arenas.store(free_memory_in_allocator_arenas_, std::memory_order_relaxed); diff --git a/src/Interpreters/AsynchronousMetrics.cpp b/src/Interpreters/AsynchronousMetrics.cpp index 2e43a594960..488ac77e956 100644 --- a/src/Interpreters/AsynchronousMetrics.cpp +++ b/src/Interpreters/AsynchronousMetrics.cpp @@ -703,12 +703,18 @@ void AsynchronousMetrics::update(TimePoint update_time) Int64 free_memory_in_allocator_arenas = 0; #if USE_JEMALLOC - /// This is a memory which is kept by allocator. - /// Will subsract it from RSS to decrease memory drift. + /// According to jemalloc man, pdirty is: + /// + /// Number of pages within unused extents that are potentially + /// dirty, and for which madvise() or similar has not been called. + /// + /// So they will be subtracted from RSS to make accounting more + /// accurate, since those pages are not really RSS but a memory + /// that can be used at anytime via jemalloc. free_memory_in_allocator_arenas = je_malloc_pdirty * getPageSize(); #endif - Int64 difference = rss - free_memory_in_allocator_arenas - amount; + Int64 difference = rss - amount; /// Log only if difference is high. This is for convenience. The threshold is arbitrary. if (difference >= 1048576 || difference <= -1048576) From 1520bcd53c27c2c24a3e85c313d593cc1871a136 Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Wed, 26 Oct 2022 15:50:10 +0000 Subject: [PATCH 116/239] Fix error code. --- src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp | 4 ++-- src/Storages/StorageMerge.cpp | 6 ------ 2 files changed, 2 insertions(+), 8 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp b/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp index 475407a402b..d8daec3b88e 100644 --- a/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp +++ b/src/Storages/MergeTree/MergeTreeBaseSelectProcessor.cpp @@ -607,7 +607,7 @@ Block MergeTreeBaseSelectProcessor::transformHeader( if (!row_level_column.type->canBeUsedInBooleanContext()) { throw Exception("Invalid type for filter in PREWHERE: " + row_level_column.type->getName(), - ErrorCodes::LOGICAL_ERROR); + ErrorCodes::ILLEGAL_TYPE_OF_COLUMN_FOR_FILTER); } block.erase(prewhere_info->row_level_column_name); @@ -620,7 +620,7 @@ Block MergeTreeBaseSelectProcessor::transformHeader( if (!prewhere_column.type->canBeUsedInBooleanContext()) { throw Exception("Invalid type for filter in PREWHERE: " + prewhere_column.type->getName(), - ErrorCodes::LOGICAL_ERROR); + ErrorCodes::ILLEGAL_TYPE_OF_COLUMN_FOR_FILTER); } if (prewhere_info->remove_prewhere_column) diff --git a/src/Storages/StorageMerge.cpp b/src/Storages/StorageMerge.cpp index c68e9103704..6d81b424f51 100644 --- a/src/Storages/StorageMerge.cpp +++ b/src/Storages/StorageMerge.cpp @@ -547,12 +547,6 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( } else { - { - /// Analyze query to check that types are valid (e.g. in PREWHERE). - InterpreterSelectQuery interpreter - (modified_query_info.query, modified_context, SelectQueryOptions(processed_stage).ignoreProjections()); - } - storage->read( plan, real_column_names, From 09fe9c3ed130cbf2f70bcf5018063d0593a17a6b Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Wed, 26 Oct 2022 16:07:56 +0000 Subject: [PATCH 117/239] Use {} in exceptions --- src/Functions/formatDateTime.cpp | 99 ++++++++++++++------------------ 1 file changed, 42 insertions(+), 57 deletions(-) diff --git a/src/Functions/formatDateTime.cpp b/src/Functions/formatDateTime.cpp index 4f28dae7a66..c5240abf7a1 100644 --- a/src/Functions/formatDateTime.cpp +++ b/src/Functions/formatDateTime.cpp @@ -317,44 +317,39 @@ public: if constexpr (support_integer) { if (arguments.size() != 1 && arguments.size() != 2 && arguments.size() != 3) - throw Exception( - "Number of arguments for function " + getName() + " doesn't match: passed " + toString(arguments.size()) - + ", should be 1, 2 or 3", - ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH); + throw Exception(ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH, + "Number of arguments for function {} doesn't match: passed {}, should be 1, 2 or 3", + getName(), arguments.size()); if (arguments.size() == 1 && !isInteger(arguments[0].type)) - throw Exception( - "Illegal type " + arguments[0].type->getName() + " of 1 argument of function " + getName() - + " when arguments size is 1. Should be integer", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Illegal type {} of 1 argument of function {} when arguments size is 1. Should be integer", + arguments[0].type->getName(), getName()); if (arguments.size() > 1 && !(isInteger(arguments[0].type) || isDate(arguments[0].type) || isDateTime(arguments[0].type) || isDate32(arguments[0].type) || isDateTime64(arguments[0].type))) - throw Exception( - "Illegal type " + arguments[0].type->getName() + " of 1 argument of function " + getName() - + " when arguments size is 2 or 3. Should be a integer or a date with time", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Illegal type {} of 1 argument of function {} when arguments size is 2 or 3. Should be a integer or a date with time", + arguments[0].type->getName(), getName()); } else { if (arguments.size() != 2 && arguments.size() != 3) - throw Exception( - "Number of arguments for function " + getName() + " doesn't match: passed " + toString(arguments.size()) - + ", should be 2 or 3", - ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH); + throw Exception(ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH, + "Number of arguments for function {} doesn't match: passed {}, should be 2 or 3", + getName(), arguments.size()); if (!isDate(arguments[0].type) && !isDateTime(arguments[0].type) && !isDate32(arguments[0].type) && !isDateTime64(arguments[0].type)) - throw Exception( - "Illegal type " + arguments[0].type->getName() + " of 1 argument of function " + getName() - + ". Should be a date or a date with time", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Illegal type {} of 1 argument of function {}. Should be a date or a date with time", + arguments[0].type->getName(), getName()); } if (arguments.size() == 2 && !WhichDataType(arguments[1].type).isString()) - throw Exception( - "Illegal type " + arguments[1].type->getName() + " of 2 argument of function " + getName() + ". Must be String.", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Illegal type {} of 2 argument of function {}. Must be String.", + arguments[1].type->getName(), getName()); if (arguments.size() == 3 && !WhichDataType(arguments[2].type).isString()) - throw Exception( - "Illegal type " + arguments[2].type->getName() + " of 3 argument of function " + getName() + ". Must be String.", - ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); + throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, + "Illegal type {} of 3 argument of function {}. Must be String.", + arguments[2].type->getName(), getName()); if (arguments.size() == 1) return std::make_shared(); @@ -375,10 +370,9 @@ public: return true; })) { - throw Exception( - "Illegal column " + arguments[0].column->getName() + " of function " + getName() - + ", must be Integer or DateTime when arguments size is 1.", - ErrorCodes::ILLEGAL_COLUMN); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Illegal column {} of function {}, must be Integer or DateTime when arguments size is 1.", + arguments[0].column->getName(), getName()); } } else @@ -387,10 +381,9 @@ public: { using FromDataType = std::decay_t; if (!(res = executeType(arguments, result_type))) - throw Exception( - "Illegal column " + arguments[0].column->getName() + " of function " + getName() - + ", must be Integer or DateTime.", - ErrorCodes::ILLEGAL_COLUMN); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Illegal column {} of function {}, must be Integer or DateTime.", + arguments[0].column->getName(), getName()); return true; })) { @@ -398,10 +391,9 @@ public: || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)))) - throw Exception( - "Illegal column " + arguments[0].column->getName() + " of function " + getName() - + ", must be Integer or DateTime.", - ErrorCodes::ILLEGAL_COLUMN); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Illegal column {} of function {}, must be Integer or DateTime.", + arguments[0].column->getName(), getName()); } } } @@ -411,10 +403,9 @@ public: || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)) || (res = executeType(arguments, result_type)))) - throw Exception( - "Illegal column " + arguments[0].column->getName() + " of function " + getName() - + ", must be Date or DateTime.", - ErrorCodes::ILLEGAL_COLUMN); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Illegal column {} of function {}, must be Date or DateTime.", + arguments[0].column->getName(), getName()); } return res; @@ -429,10 +420,9 @@ public: const ColumnConst * pattern_column = checkAndGetColumnConst(arguments[1].column.get()); if (!pattern_column) - throw Exception("Illegal column " + arguments[1].column->getName() - + " of second ('format') argument of function " + getName() - + ". Must be constant string.", - ErrorCodes::ILLEGAL_COLUMN); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Illegal column {} of second ('format') argument of function {}. Must be constant string.", + arguments[1].column->getName(), getName()); String pattern = pattern_column->getValue(); @@ -500,13 +490,6 @@ public: instruction.perform(pos, static_cast(c.whole), time_zone); } } - else if constexpr (std::is_same_v) - { - for (auto & instruction : instructions) - { - instruction.perform(pos, static_cast(vec[i]), time_zone); - } - } else { for (auto & instruction : instructions) @@ -723,12 +706,14 @@ public: // Unimplemented case 'U': [[fallthrough]]; case 'W': - throw Exception("Wrong pattern '" + pattern + "', symbol '" + *pos + " is not implemented ' for function " + getName(), - ErrorCodes::NOT_IMPLEMENTED); + throw Exception(ErrorCodes::NOT_IMPLEMENTED, + "Wrong pattern '{}', symbol '{}' is not implemented for function {}", + pattern, *pos, getName()); default: - throw Exception( - "Wrong pattern '" + pattern + "', unexpected symbol '" + *pos + "' for function " + getName(), ErrorCodes::ILLEGAL_COLUMN); + throw Exception(ErrorCodes::ILLEGAL_COLUMN, + "Wrong pattern '{}', unexpected symbol '{}' for function {}", + pattern, *pos, getName()); } ++pos; From 22011aeaec07b9b14f4a6bc6855860ab0be10eda Mon Sep 17 00:00:00 2001 From: DanRoscigno Date: Wed, 26 Oct 2022 12:39:17 -0400 Subject: [PATCH 118/239] remove frontmatter --- docs/en/operations/_backup.md | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/docs/en/operations/_backup.md b/docs/en/operations/_backup.md index c543c49a083..d694c51cee6 100644 --- a/docs/en/operations/_backup.md +++ b/docs/en/operations/_backup.md @@ -1,9 +1,5 @@ ---- -slug: /en/operations/backup -sidebar_position: 49 -sidebar_label: Data backup and restore -title: Data backup and restore ---- + +[//]: # (This file is included in Manage > Backups) - [Backup to a local disk](#backup-to-a-local-disk) - [Configuring backup/restore to use an S3 endpoint](#configuring-backuprestore-to-use-an-s3-endpoint) From f04f6638189bd4a0f9c9defdb43c66b3e8c5ef37 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 18:55:14 +0200 Subject: [PATCH 119/239] Fixed tests --- src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp index db1b9762668..1413f084536 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp @@ -361,7 +361,7 @@ bool MergeTreeIndexConditionBloomFilter::traverseTreeIn( WhichDataType which(type); - if (which.isTuple() && function_name == "tuple") + if (which.isTuple() && key_node_function_name == "tuple") { const auto & tuple_column = typeid_cast(column.get()); const auto & tuple_data_type = typeid_cast(type.get()); @@ -388,6 +388,8 @@ bool MergeTreeIndexConditionBloomFilter::traverseTreeIn( * We cannot skip keys that does not exist in map if comparison is with default type value because * that way we skip necessary granules where map key does not exists. */ + if (!prepared_set) + return false; auto default_column_to_check = type->createColumnConstWithDefaultValue(1)->convertToFullColumnIfConst(); ColumnWithTypeAndName default_column_with_type_to_check { default_column_to_check, type, "" }; From 31a0044981e7da291487a48a8871dc8fed9e12e9 Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Wed, 26 Oct 2022 17:14:38 +0000 Subject: [PATCH 120/239] Update version_date.tsv and changelogs after v22.9.4.32-stable --- docs/changelogs/v22.9.4.32-stable.md | 33 ++++++++++++++++++++++++++++ utils/list-versions/version_date.tsv | 1 + 2 files changed, 34 insertions(+) create mode 100644 docs/changelogs/v22.9.4.32-stable.md diff --git a/docs/changelogs/v22.9.4.32-stable.md b/docs/changelogs/v22.9.4.32-stable.md new file mode 100644 index 00000000000..d6c3f4ba498 --- /dev/null +++ b/docs/changelogs/v22.9.4.32-stable.md @@ -0,0 +1,33 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.9.4.32-stable (3db8bcf1a70) FIXME as compared to v22.9.3.18-stable (0cb4b15d2fa) + +#### Bug Fix +* Backported in [#42435](https://github.com/ClickHouse/ClickHouse/issues/42435): - Choose correct aggregation method for LowCardinality with BigInt. [#42342](https://github.com/ClickHouse/ClickHouse/pull/42342) ([Duc Canh Le](https://github.com/canhld94)). + +#### Build/Testing/Packaging Improvement +* Backported in [#42297](https://github.com/ClickHouse/ClickHouse/issues/42297): Update cctz to the latest master, update tzdb to 2020e. [#42273](https://github.com/ClickHouse/ClickHouse/pull/42273) ([Dom Del Nano](https://github.com/ddelnano)). +* Backported in [#42361](https://github.com/ClickHouse/ClickHouse/issues/42361): Update tzdata to 2022e to support the new timezone changes. Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Jordan and Syria switch from +02/+03 with DST to year-round +03. (https://data.iana.org/time-zones/tzdb/NEWS). This closes [#42252](https://github.com/ClickHouse/ClickHouse/issues/42252). [#42327](https://github.com/ClickHouse/ClickHouse/pull/42327) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42122](https://github.com/ClickHouse/ClickHouse/issues/42122): Fixed "Part ... intersects part ..." error that might happen in extremely rare cases if replica was restarted just after detaching some part as broken. [#41741](https://github.com/ClickHouse/ClickHouse/pull/41741) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Backported in [#41938](https://github.com/ClickHouse/ClickHouse/issues/41938): Don't allow to create or alter merge tree tables with virtual column name _row_exists, which is reserved for lightweight delete. Fixed [#41716](https://github.com/ClickHouse/ClickHouse/issues/41716). [#41763](https://github.com/ClickHouse/ClickHouse/pull/41763) ([Jianmei Zhang](https://github.com/zhangjmruc)). +* Backported in [#42179](https://github.com/ClickHouse/ClickHouse/issues/42179): Fix reusing of files > 4GB from base backup. [#42146](https://github.com/ClickHouse/ClickHouse/pull/42146) ([Azat Khuzhin](https://github.com/azat)). +* Backported in [#42301](https://github.com/ClickHouse/ClickHouse/issues/42301): Fix a bug with projections and the `aggregate_functions_null_for_empty` setting. This bug is very rare and appears only if you enable the `aggregate_functions_null_for_empty` setting in the server's config. This closes [#41647](https://github.com/ClickHouse/ClickHouse/issues/41647). [#42198](https://github.com/ClickHouse/ClickHouse/pull/42198) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42388](https://github.com/ClickHouse/ClickHouse/issues/42388): `ALTER UPDATE` of attached part (with columns different from table schema) could create an invalid `columns.txt` metadata on disk. Reading from such part could fail with errors or return invalid data. Fixes [#42161](https://github.com/ClickHouse/ClickHouse/issues/42161). [#42319](https://github.com/ClickHouse/ClickHouse/pull/42319) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Backported in [#42500](https://github.com/ClickHouse/ClickHouse/issues/42500): Setting `additional_table_filters` were not applied to `Distributed` storage. Fixes [#41692](https://github.com/ClickHouse/ClickHouse/issues/41692). [#42322](https://github.com/ClickHouse/ClickHouse/pull/42322) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Backported in [#42581](https://github.com/ClickHouse/ClickHouse/issues/42581): This reverts [#40217](https://github.com/ClickHouse/ClickHouse/issues/40217) which introduced a regression in date/time functions. [#42367](https://github.com/ClickHouse/ClickHouse/pull/42367) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42572](https://github.com/ClickHouse/ClickHouse/issues/42572): Fix buffer overflow in the processing of Decimal data types. This closes [#42451](https://github.com/ClickHouse/ClickHouse/issues/42451). [#42465](https://github.com/ClickHouse/ClickHouse/pull/42465) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42595](https://github.com/ClickHouse/ClickHouse/issues/42595): This closes [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Add a warning message to release.py script, require release type [#41975](https://github.com/ClickHouse/ClickHouse/pull/41975) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787) [#42136](https://github.com/ClickHouse/ClickHouse/pull/42136) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index e72fce63fda..7807fa32cbc 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -1,4 +1,5 @@ v22.10.1.1877-stable 2022-10-26 +v22.9.4.32-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 v22.9.2.7-stable 2022-09-23 v22.9.1.2603-stable 2022-09-22 From 18c0e7b2285061187aa3daf6fdc19b89539df57e Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Wed, 26 Oct 2022 17:19:01 +0000 Subject: [PATCH 121/239] Update version_date.tsv and changelogs after v22.8.7.34-lts --- docs/changelogs/v22.8.7.34-lts.md | 37 ++++++++++++++++++++++++++++ utils/list-versions/version_date.tsv | 3 +++ 2 files changed, 40 insertions(+) create mode 100644 docs/changelogs/v22.8.7.34-lts.md diff --git a/docs/changelogs/v22.8.7.34-lts.md b/docs/changelogs/v22.8.7.34-lts.md new file mode 100644 index 00000000000..0dc899f4717 --- /dev/null +++ b/docs/changelogs/v22.8.7.34-lts.md @@ -0,0 +1,37 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.8.7.34-lts (3c38e5e8ab9) FIXME as compared to v22.8.6.71-lts (7bf38a43e30) + +#### Improvement +* Backported in [#42096](https://github.com/ClickHouse/ClickHouse/issues/42096): Replace back `clickhouse su` command with `sudo -u` in start in order to respect limits in `/etc/security/limits.conf`. [#41847](https://github.com/ClickHouse/ClickHouse/pull/41847) ([Eugene Konkov](https://github.com/ekonkov)). + +#### Bug Fix +* Backported in [#42434](https://github.com/ClickHouse/ClickHouse/issues/42434): - Choose correct aggregation method for LowCardinality with BigInt. [#42342](https://github.com/ClickHouse/ClickHouse/pull/42342) ([Duc Canh Le](https://github.com/canhld94)). + +#### Build/Testing/Packaging Improvement +* Backported in [#42296](https://github.com/ClickHouse/ClickHouse/issues/42296): Update cctz to the latest master, update tzdb to 2020e. [#42273](https://github.com/ClickHouse/ClickHouse/pull/42273) ([Dom Del Nano](https://github.com/ddelnano)). +* Backported in [#42360](https://github.com/ClickHouse/ClickHouse/issues/42360): Update tzdata to 2022e to support the new timezone changes. Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Jordan and Syria switch from +02/+03 with DST to year-round +03. (https://data.iana.org/time-zones/tzdb/NEWS). This closes [#42252](https://github.com/ClickHouse/ClickHouse/issues/42252). [#42327](https://github.com/ClickHouse/ClickHouse/pull/42327) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42489](https://github.com/ClickHouse/ClickHouse/issues/42489): Removed skipping of mutations in unaffected partitions of `MergeTree` tables, because this feature never worked correctly and might cause resurrection of finished mutations. [#40589](https://github.com/ClickHouse/ClickHouse/pull/40589) ([Alexander Tokmakov](https://github.com/tavplubix)). +* Backported in [#42121](https://github.com/ClickHouse/ClickHouse/issues/42121): Fixed "Part ... intersects part ..." error that might happen in extremely rare cases if replica was restarted just after detaching some part as broken. [#41741](https://github.com/ClickHouse/ClickHouse/pull/41741) ([Alexander Tokmakov](https://github.com/tavplubix)). +* - Prevent crash when passing wrong aggregation states to groupBitmap*. [#41972](https://github.com/ClickHouse/ClickHouse/pull/41972) ([Raúl Marín](https://github.com/Algunenano)). +* - Fix read bytes/rows in X-ClickHouse-Summary with materialized views. [#41973](https://github.com/ClickHouse/ClickHouse/pull/41973) ([Raúl Marín](https://github.com/Algunenano)). +* Backported in [#42269](https://github.com/ClickHouse/ClickHouse/issues/42269): Fix reusing of files > 4GB from base backup. [#42146](https://github.com/ClickHouse/ClickHouse/pull/42146) ([Azat Khuzhin](https://github.com/azat)). +* Backported in [#42300](https://github.com/ClickHouse/ClickHouse/issues/42300): Fix a bug with projections and the `aggregate_functions_null_for_empty` setting. This bug is very rare and appears only if you enable the `aggregate_functions_null_for_empty` setting in the server's config. This closes [#41647](https://github.com/ClickHouse/ClickHouse/issues/41647). [#42198](https://github.com/ClickHouse/ClickHouse/pull/42198) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42387](https://github.com/ClickHouse/ClickHouse/issues/42387): `ALTER UPDATE` of attached part (with columns different from table schema) could create an invalid `columns.txt` metadata on disk. Reading from such part could fail with errors or return invalid data. Fixes [#42161](https://github.com/ClickHouse/ClickHouse/issues/42161). [#42319](https://github.com/ClickHouse/ClickHouse/pull/42319) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Backported in [#42499](https://github.com/ClickHouse/ClickHouse/issues/42499): Setting `additional_table_filters` were not applied to `Distributed` storage. Fixes [#41692](https://github.com/ClickHouse/ClickHouse/issues/41692). [#42322](https://github.com/ClickHouse/ClickHouse/pull/42322) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Backported in [#42571](https://github.com/ClickHouse/ClickHouse/issues/42571): Fix buffer overflow in the processing of Decimal data types. This closes [#42451](https://github.com/ClickHouse/ClickHouse/issues/42451). [#42465](https://github.com/ClickHouse/ClickHouse/pull/42465) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42594](https://github.com/ClickHouse/ClickHouse/issues/42594): This closes [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Add a warning message to release.py script, require release type [#41975](https://github.com/ClickHouse/ClickHouse/pull/41975) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787) [#42136](https://github.com/ClickHouse/ClickHouse/pull/42136) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index e72fce63fda..0470152ecff 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -1,13 +1,16 @@ v22.10.1.1877-stable 2022-10-26 +v22.9.4.32-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 v22.9.2.7-stable 2022-09-23 v22.9.1.2603-stable 2022-09-22 +v22.8.7.34-lts 2022-10-26 v22.8.6.71-lts 2022-09-30 v22.8.5.29-lts 2022-09-13 v22.8.4.7-lts 2022-08-31 v22.8.3.13-lts 2022-08-29 v22.8.2.11-lts 2022-08-23 v22.8.1.2097-lts 2022-08-18 +v22.7.7.24-stable 2022-10-26 v22.7.6.74-stable 2022-09-30 v22.7.5.13-stable 2022-08-29 v22.7.4.16-stable 2022-08-23 From 4c8f0e7591f4248248984df84ecd545e90b20455 Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Wed, 26 Oct 2022 17:28:41 +0000 Subject: [PATCH 122/239] Update version_date.tsv and changelogs after v22.7.7.24-stable --- docs/changelogs/v22.7.7.24-stable.md | 29 ++++++++++++++++++++++++++++ utils/list-versions/version_date.tsv | 3 +++ 2 files changed, 32 insertions(+) create mode 100644 docs/changelogs/v22.7.7.24-stable.md diff --git a/docs/changelogs/v22.7.7.24-stable.md b/docs/changelogs/v22.7.7.24-stable.md new file mode 100644 index 00000000000..d7b83775502 --- /dev/null +++ b/docs/changelogs/v22.7.7.24-stable.md @@ -0,0 +1,29 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.7.7.24-stable (02ad1f979a8) FIXME as compared to v22.7.6.74-stable (c00ffb3c11a) + +#### Bug Fix +* Backported in [#42433](https://github.com/ClickHouse/ClickHouse/issues/42433): - Choose correct aggregation method for LowCardinality with BigInt. [#42342](https://github.com/ClickHouse/ClickHouse/pull/42342) ([Duc Canh Le](https://github.com/canhld94)). + +#### Build/Testing/Packaging Improvement +* Backported in [#42329](https://github.com/ClickHouse/ClickHouse/issues/42329): Update cctz to the latest master, update tzdb to 2020e. [#42273](https://github.com/ClickHouse/ClickHouse/pull/42273) ([Dom Del Nano](https://github.com/ddelnano)). +* Backported in [#42359](https://github.com/ClickHouse/ClickHouse/issues/42359): Update tzdata to 2022e to support the new timezone changes. Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Jordan and Syria switch from +02/+03 with DST to year-round +03. (https://data.iana.org/time-zones/tzdb/NEWS). This closes [#42252](https://github.com/ClickHouse/ClickHouse/issues/42252). [#42327](https://github.com/ClickHouse/ClickHouse/pull/42327) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42268](https://github.com/ClickHouse/ClickHouse/issues/42268): Fix reusing of files > 4GB from base backup. [#42146](https://github.com/ClickHouse/ClickHouse/pull/42146) ([Azat Khuzhin](https://github.com/azat)). +* Backported in [#42299](https://github.com/ClickHouse/ClickHouse/issues/42299): Fix a bug with projections and the `aggregate_functions_null_for_empty` setting. This bug is very rare and appears only if you enable the `aggregate_functions_null_for_empty` setting in the server's config. This closes [#41647](https://github.com/ClickHouse/ClickHouse/issues/41647). [#42198](https://github.com/ClickHouse/ClickHouse/pull/42198) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42386](https://github.com/ClickHouse/ClickHouse/issues/42386): `ALTER UPDATE` of attached part (with columns different from table schema) could create an invalid `columns.txt` metadata on disk. Reading from such part could fail with errors or return invalid data. Fixes [#42161](https://github.com/ClickHouse/ClickHouse/issues/42161). [#42319](https://github.com/ClickHouse/ClickHouse/pull/42319) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Backported in [#42498](https://github.com/ClickHouse/ClickHouse/issues/42498): Setting `additional_table_filters` were not applied to `Distributed` storage. Fixes [#41692](https://github.com/ClickHouse/ClickHouse/issues/41692). [#42322](https://github.com/ClickHouse/ClickHouse/pull/42322) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). +* Backported in [#42593](https://github.com/ClickHouse/ClickHouse/issues/42593): This closes [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Add a warning message to release.py script, require release type [#41975](https://github.com/ClickHouse/ClickHouse/pull/41975) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787) [#42136](https://github.com/ClickHouse/ClickHouse/pull/42136) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index e72fce63fda..0470152ecff 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -1,13 +1,16 @@ v22.10.1.1877-stable 2022-10-26 +v22.9.4.32-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 v22.9.2.7-stable 2022-09-23 v22.9.1.2603-stable 2022-09-22 +v22.8.7.34-lts 2022-10-26 v22.8.6.71-lts 2022-09-30 v22.8.5.29-lts 2022-09-13 v22.8.4.7-lts 2022-08-31 v22.8.3.13-lts 2022-08-29 v22.8.2.11-lts 2022-08-23 v22.8.1.2097-lts 2022-08-18 +v22.7.7.24-stable 2022-10-26 v22.7.6.74-stable 2022-09-30 v22.7.5.13-stable 2022-08-29 v22.7.4.16-stable 2022-08-23 From acc03cf52e7079110cc99f2c17854672697e4f80 Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Wed, 26 Oct 2022 13:44:46 -0400 Subject: [PATCH 123/239] Sonar Cloud Workflow (#42534) * Sonar Cloud Workflow Signed-off-by: Julio Jimenez * build command Signed-off-by: Julio Jimenez * try builder Signed-off-by: Julio Jimenez * build Signed-off-by: Julio Jimenez * ok Signed-off-by: Julio Jimenez * ok Signed-off-by: Julio Jimenez * not sure why python was there Signed-off-by: Julio Jimenez * ccache Signed-off-by: Julio Jimenez * install python3 Signed-off-by: Julio Jimenez * env vars and simpler cmake Signed-off-by: Julio Jimenez * ninjaless Signed-off-by: Julio Jimenez * mandatory properties Signed-off-by: Julio Jimenez * typo Signed-off-by: Julio Jimenez Signed-off-by: Julio Jimenez --- .github/workflows/sonar.yml | 64 +++++++++++++++++++++++++++++++++++++ 1 file changed, 64 insertions(+) create mode 100644 .github/workflows/sonar.yml diff --git a/.github/workflows/sonar.yml b/.github/workflows/sonar.yml new file mode 100644 index 00000000000..316f1e90d40 --- /dev/null +++ b/.github/workflows/sonar.yml @@ -0,0 +1,64 @@ +name: Sonar Cloud +on: + push: + branches: + - master + pull_request: + types: [opened, synchronize, reopened] +env: + CC: clang-15 + CXX: clang++-15 +jobs: + sonar_cloud: + name: Sonar Cloud + runs-on: [self-hosted, builder] + env: + SONAR_SCANNER_VERSION: 4.7.0.2747 + SONAR_SERVER_URL: "https://sonarcloud.io" + BUILD_WRAPPER_OUT_DIR: build_wrapper_output_directory # Directory where build-wrapper output will be placed + steps: + - uses: actions/checkout@v2 + with: + fetch-depth: 0 # Shallow clones should be disabled for a better relevancy of analysis + submodules: true + - name: Set up JDK 11 + uses: actions/setup-java@v1 + with: + java-version: 11 + - name: Download and set up sonar-scanner + env: + SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip + run: | + mkdir -p $HOME/.sonar + curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} + unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ + echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> $GITHUB_PATH + - name: Download and set up build-wrapper + env: + BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip + run: | + curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} + unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ + echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH + - name: Set Up Build Tools + run: | + sudo apt-get update + sudo apt-get install -yq git cmake ccache python3 ninja-build + sudo bash -c "$(wget -O - https://apt.llvm.org/llvm.sh)" + - name: Run build-wrapper + run: | + mkdir build + cd build + cmake .. + cd .. + build-wrapper-linux-x86-64 --out-dir ${{ env.BUILD_WRAPPER_OUT_DIR }} cmake --build build/ + - name: Run sonar-scanner + env: + GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} + SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }} + run: | + sonar-scanner \ + --define sonar.host.url="${{ env.SONAR_SERVER_URL }}" \ + --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ + --define sonar.projectKey="clickhouse-java" \ + --define sonar.organization="ClickHouse" From 313cbb1058a13abb03cc0c839c7ace6b6e8ba68a Mon Sep 17 00:00:00 2001 From: Nikolai Kochetov Date: Wed, 26 Oct 2022 17:49:36 +0000 Subject: [PATCH 124/239] Fix test. --- src/Storages/StorageMerge.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Storages/StorageMerge.cpp b/src/Storages/StorageMerge.cpp index 6d81b424f51..9891340a0d0 100644 --- a/src/Storages/StorageMerge.cpp +++ b/src/Storages/StorageMerge.cpp @@ -520,8 +520,6 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( modified_select.setFinal(); } - modified_select.replaceDatabaseAndTable(database_name, table_name); - auto storage_stage = storage->getQueryProcessingStage(modified_context, QueryProcessingStage::Complete, storage_snapshot, modified_query_info); if (processed_stage <= storage_stage) @@ -571,6 +569,8 @@ QueryPipelineBuilderPtr ReadFromMerge::createSources( } else if (processed_stage > storage_stage) { + modified_select.replaceDatabaseAndTable(database_name, table_name); + /// Maximum permissible parallelism is streams_num modified_context->setSetting("max_threads", streams_num); modified_context->setSetting("max_streams_to_max_threads_ratio", 1); From c8444f751f968ebd3b048d6edcc9d90c1c6ac02d Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Wed, 26 Oct 2022 14:07:10 -0400 Subject: [PATCH 125/239] Move SonarCloud Job to nightly Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 53 +++++++++++++++++++++++++++++ .github/workflows/sonar.yml | 64 ----------------------------------- 2 files changed, 53 insertions(+), 64 deletions(-) delete mode 100644 .github/workflows/sonar.yml diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 801f7eda94a..834f74822a1 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -122,3 +122,56 @@ jobs: docker ps --quiet | xargs --no-run-if-empty docker kill ||: docker ps --all --quiet | xargs --no-run-if-empty docker rm -f ||: sudo rm -fr "$TEMP_PATH" "$CACHES_PATH" + SonarCloud: + name: Sonar Cloud + runs-on: [self-hosted, builder] + env: + SONAR_SCANNER_VERSION: 4.7.0.2747 + SONAR_SERVER_URL: "https://sonarcloud.io" + BUILD_WRAPPER_OUT_DIR: build_wrapper_output_directory # Directory where build-wrapper output will be placed + steps: + - uses: actions/checkout@v2 + with: + fetch-depth: 0 # Shallow clones should be disabled for a better relevancy of analysis + submodules: true + - name: Set up JDK 11 + uses: actions/setup-java@v1 + with: + java-version: 11 + - name: Download and set up sonar-scanner + env: + SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip + run: | + mkdir -p $HOME/.sonar + curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} + unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ + echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> $GITHUB_PATH + - name: Download and set up build-wrapper + env: + BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip + run: | + curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} + unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ + echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH + - name: Set Up Build Tools + run: | + sudo apt-get update + sudo apt-get install -yq git cmake ccache python3 ninja-build + sudo bash -c "$(wget -O - https://apt.llvm.org/llvm.sh)" + - name: Run build-wrapper + run: | + mkdir build + cd build + cmake .. + cd .. + build-wrapper-linux-x86-64 --out-dir ${{ env.BUILD_WRAPPER_OUT_DIR }} cmake --build build/ + - name: Run sonar-scanner + env: + GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} + SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }} + run: | + sonar-scanner \ + --define sonar.host.url="${{ env.SONAR_SERVER_URL }}" \ + --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ + --define sonar.projectKey="ClickHouse_ClickHouse" \ + --define sonar.organization="clickhouse-java" \ No newline at end of file diff --git a/.github/workflows/sonar.yml b/.github/workflows/sonar.yml deleted file mode 100644 index 316f1e90d40..00000000000 --- a/.github/workflows/sonar.yml +++ /dev/null @@ -1,64 +0,0 @@ -name: Sonar Cloud -on: - push: - branches: - - master - pull_request: - types: [opened, synchronize, reopened] -env: - CC: clang-15 - CXX: clang++-15 -jobs: - sonar_cloud: - name: Sonar Cloud - runs-on: [self-hosted, builder] - env: - SONAR_SCANNER_VERSION: 4.7.0.2747 - SONAR_SERVER_URL: "https://sonarcloud.io" - BUILD_WRAPPER_OUT_DIR: build_wrapper_output_directory # Directory where build-wrapper output will be placed - steps: - - uses: actions/checkout@v2 - with: - fetch-depth: 0 # Shallow clones should be disabled for a better relevancy of analysis - submodules: true - - name: Set up JDK 11 - uses: actions/setup-java@v1 - with: - java-version: 11 - - name: Download and set up sonar-scanner - env: - SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip - run: | - mkdir -p $HOME/.sonar - curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} - unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ - echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> $GITHUB_PATH - - name: Download and set up build-wrapper - env: - BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip - run: | - curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} - unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ - echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH - - name: Set Up Build Tools - run: | - sudo apt-get update - sudo apt-get install -yq git cmake ccache python3 ninja-build - sudo bash -c "$(wget -O - https://apt.llvm.org/llvm.sh)" - - name: Run build-wrapper - run: | - mkdir build - cd build - cmake .. - cd .. - build-wrapper-linux-x86-64 --out-dir ${{ env.BUILD_WRAPPER_OUT_DIR }} cmake --build build/ - - name: Run sonar-scanner - env: - GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} - SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }} - run: | - sonar-scanner \ - --define sonar.host.url="${{ env.SONAR_SERVER_URL }}" \ - --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ - --define sonar.projectKey="clickhouse-java" \ - --define sonar.organization="ClickHouse" From 488c2200466571ebe97dec9ac98925d6eb4a8f48 Mon Sep 17 00:00:00 2001 From: Nikolay Degterinsky Date: Thu, 27 Oct 2022 01:45:38 +0000 Subject: [PATCH 126/239] Fix bug in ParserFunction --- src/Parsers/ExpressionListParsers.cpp | 10 ++++++++++ .../02474_fix_function_parser_bug.reference | 0 .../0_stateless/02474_fix_function_parser_bug.sql | 1 + 3 files changed, 11 insertions(+) create mode 100644 tests/queries/0_stateless/02474_fix_function_parser_bug.reference create mode 100644 tests/queries/0_stateless/02474_fix_function_parser_bug.sql diff --git a/src/Parsers/ExpressionListParsers.cpp b/src/Parsers/ExpressionListParsers.cpp index 4af4dabb12e..2f39162e104 100644 --- a/src/Parsers/ExpressionListParsers.cpp +++ b/src/Parsers/ExpressionListParsers.cpp @@ -830,6 +830,16 @@ public: explicit FunctionLayer(String function_name_, bool allow_function_parameters_ = true) : function_name(function_name_), allow_function_parameters(allow_function_parameters_){} + bool getResult(ASTPtr & node) override + { + // FunctionLayer can be the only layer in our Layers stack, + // so we need to check that we exited the main cycle properly + if (!finished) + return false; + + return Layer::getResult(node); + } + bool parse(IParser::Pos & pos, Expected & expected, Action & action) override { /// | 0 | 1 | 2 | diff --git a/tests/queries/0_stateless/02474_fix_function_parser_bug.reference b/tests/queries/0_stateless/02474_fix_function_parser_bug.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02474_fix_function_parser_bug.sql b/tests/queries/0_stateless/02474_fix_function_parser_bug.sql new file mode 100644 index 00000000000..12e9e03f151 --- /dev/null +++ b/tests/queries/0_stateless/02474_fix_function_parser_bug.sql @@ -0,0 +1 @@ +CREATE DATABASE conv_mian ENGINE QALL(COLUMNS('|T.D'),¸mp} -- { clientError 62 } From f31061c68f0c27e4ce1bbec3ed3dcb0822423417 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Thu, 27 Oct 2022 09:04:46 +0300 Subject: [PATCH 127/239] Revert "Sonar Cloud Workflow (#42534)" This reverts commit acc03cf52e7079110cc99f2c17854672697e4f80. --- .github/workflows/sonar.yml | 64 ------------------------------------- 1 file changed, 64 deletions(-) delete mode 100644 .github/workflows/sonar.yml diff --git a/.github/workflows/sonar.yml b/.github/workflows/sonar.yml deleted file mode 100644 index 316f1e90d40..00000000000 --- a/.github/workflows/sonar.yml +++ /dev/null @@ -1,64 +0,0 @@ -name: Sonar Cloud -on: - push: - branches: - - master - pull_request: - types: [opened, synchronize, reopened] -env: - CC: clang-15 - CXX: clang++-15 -jobs: - sonar_cloud: - name: Sonar Cloud - runs-on: [self-hosted, builder] - env: - SONAR_SCANNER_VERSION: 4.7.0.2747 - SONAR_SERVER_URL: "https://sonarcloud.io" - BUILD_WRAPPER_OUT_DIR: build_wrapper_output_directory # Directory where build-wrapper output will be placed - steps: - - uses: actions/checkout@v2 - with: - fetch-depth: 0 # Shallow clones should be disabled for a better relevancy of analysis - submodules: true - - name: Set up JDK 11 - uses: actions/setup-java@v1 - with: - java-version: 11 - - name: Download and set up sonar-scanner - env: - SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip - run: | - mkdir -p $HOME/.sonar - curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} - unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ - echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> $GITHUB_PATH - - name: Download and set up build-wrapper - env: - BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip - run: | - curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} - unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ - echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH - - name: Set Up Build Tools - run: | - sudo apt-get update - sudo apt-get install -yq git cmake ccache python3 ninja-build - sudo bash -c "$(wget -O - https://apt.llvm.org/llvm.sh)" - - name: Run build-wrapper - run: | - mkdir build - cd build - cmake .. - cd .. - build-wrapper-linux-x86-64 --out-dir ${{ env.BUILD_WRAPPER_OUT_DIR }} cmake --build build/ - - name: Run sonar-scanner - env: - GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} - SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }} - run: | - sonar-scanner \ - --define sonar.host.url="${{ env.SONAR_SERVER_URL }}" \ - --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ - --define sonar.projectKey="clickhouse-java" \ - --define sonar.organization="ClickHouse" From b74cccfa0adf3b4aaf79270fcf5bc72ec17bd3f6 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Thu, 27 Oct 2022 10:19:50 +0300 Subject: [PATCH 128/239] Update CHANGELOG.md --- CHANGELOG.md | 2 -- 1 file changed, 2 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index 3948a24d52a..68767612892 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -88,7 +88,6 @@ * Allow to use `Date32` arguments for `dateName` function. [#42554](https://github.com/ClickHouse/ClickHouse/pull/42554) ([Roman Vasin](https://github.com/rvasin)). * Now filters with NULL literals will be used during index analysis. [#34063](https://github.com/ClickHouse/ClickHouse/issues/34063). [#41842](https://github.com/ClickHouse/ClickHouse/pull/41842) ([Amos Bird](https://github.com/amosbird)). * Merge parts if every part in the range is older than a certain threshold. The threshold can be set by using `min_age_to_force_merge_seconds`. This closes [#35836](https://github.com/ClickHouse/ClickHouse/issues/35836). [#42423](https://github.com/ClickHouse/ClickHouse/pull/42423) ([Antonio Andelic](https://github.com/antonio2368)). This is continuation of [#39550i](https://github.com/ClickHouse/ClickHouse/pull/39550) by [@fastio](https://github.com/fastio) who implemented most of the logic. -* Added new infrastructure for query analysis and planning under `allow_experimental_analyzer` setting. [#31796](https://github.com/ClickHouse/ClickHouse/pull/31796) ([Maksim Kita](https://github.com/kitaisreal)). * Improve the time to recover lost keeper connections. [#42541](https://github.com/ClickHouse/ClickHouse/pull/42541) ([Raúl Marín](https://github.com/Algunenano)). #### Build/Testing/Packaging Improvement @@ -143,7 +142,6 @@ * Fix bad_cast assert during INSERT into `Annoy` indexes over non-Float32 columns. `Annoy` indices is an experimental feature. [#42485](https://github.com/ClickHouse/ClickHouse/pull/42485) ([Robert Schulze](https://github.com/rschu1ze)). * Arithmetic operator with Date or DateTime and 128 or 256-bit integer was referencing uninitialized memory. [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). * Fix unexpected table loading error when partition key contains alias function names during server upgrade. [#36379](https://github.com/ClickHouse/ClickHouse/pull/36379) ([Amos Bird](https://github.com/amosbird)). -* Fixes a crash in `JSONExtract` with `LowCardinality`. [#42633](https://github.com/ClickHouse/ClickHouse/pull/42633) ([Anton Popov](https://github.com/CurtizJ)). ### ClickHouse release 22.9, 2022-09-22 From 0fc0e81c0c3428c865fa719c6a421dc3bcfeed37 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rcio=20Martins?= Date: Thu, 8 Sep 2022 20:36:08 +0200 Subject: [PATCH 129/239] Add Element::type() to JSONParser --- src/Common/JSONParsers/DummyJSONParser.h | 2 ++ src/Common/JSONParsers/ElementTypes.h | 16 ++++++++++++++++ src/Common/JSONParsers/RapidJSONParser.h | 16 +++++++++++++++- src/Common/JSONParsers/SimdJSONParser.h | 17 ++++++++++++++++- 4 files changed, 49 insertions(+), 2 deletions(-) create mode 100644 src/Common/JSONParsers/ElementTypes.h diff --git a/src/Common/JSONParsers/DummyJSONParser.h b/src/Common/JSONParsers/DummyJSONParser.h index 3cedd59decd..50c112affe2 100644 --- a/src/Common/JSONParsers/DummyJSONParser.h +++ b/src/Common/JSONParsers/DummyJSONParser.h @@ -3,6 +3,7 @@ #include #include #include +#include "ElementTypes.h" namespace DB @@ -25,6 +26,7 @@ struct DummyJSONParser { public: Element() = default; + static ElementType type() { return ElementType::NULL_VALUE; } static bool isInt64() { return false; } static bool isUInt64() { return false; } static bool isDouble() { return false; } diff --git a/src/Common/JSONParsers/ElementTypes.h b/src/Common/JSONParsers/ElementTypes.h new file mode 100644 index 00000000000..0dcfa7115aa --- /dev/null +++ b/src/Common/JSONParsers/ElementTypes.h @@ -0,0 +1,16 @@ +#pragma once + +namespace DB +{ +// Enum values match simdjson's for fast conversion +enum class ElementType { + ARRAY = '[', + OBJECT = '{', + INT64 = 'l', + UINT64 = 'u', + DOUBLE = 'd', + STRING = '"', + BOOL = 't', + NULL_VALUE = 'n' +}; +} diff --git a/src/Common/JSONParsers/RapidJSONParser.h b/src/Common/JSONParsers/RapidJSONParser.h index 01730bc0692..16a87a0e0bf 100644 --- a/src/Common/JSONParsers/RapidJSONParser.h +++ b/src/Common/JSONParsers/RapidJSONParser.h @@ -6,7 +6,7 @@ # include # include # include - +# include "ElementTypes.h" namespace DB { @@ -26,6 +26,20 @@ struct RapidJSONParser ALWAYS_INLINE Element() = default; ALWAYS_INLINE Element(const rapidjson::Value & value_) : ptr(&value_) {} /// NOLINT + ALWAYS_INLINE ElementType type() const { + switch (ptr->GetType()) { + case rapidjson::kNumberType: return ptr->IsDouble() ? ElementType::DOUBLE : (ptr->IsUint64() ? ElementType::UINT64 : ElementType::INT64); + case rapidjson::kStringType: return ElementType::STRING; + case rapidjson::kArrayType: return ElementType::ARRAY; + case rapidjson::kObjectType: return ElementType::OBJECT; + case rapidjson::kTrueType: return ElementType::BOOL; + case rapidjson::kFalseType: return ElementType::BOOL; + case rapidjson::kNullType: + default: + return ElementType::NULL_VALUE; + } + } + ALWAYS_INLINE bool isInt64() const { return ptr->IsInt64(); } ALWAYS_INLINE bool isUInt64() const { return ptr->IsUint64(); } ALWAYS_INLINE bool isDouble() const { return ptr->IsDouble(); } diff --git a/src/Common/JSONParsers/SimdJSONParser.h b/src/Common/JSONParsers/SimdJSONParser.h index 14eb3cd6d78..dd2077ba768 100644 --- a/src/Common/JSONParsers/SimdJSONParser.h +++ b/src/Common/JSONParsers/SimdJSONParser.h @@ -7,7 +7,7 @@ # include # include # include - +# include "ElementTypes.h" namespace DB { @@ -31,6 +31,21 @@ struct SimdJSONParser ALWAYS_INLINE Element() {} /// NOLINT ALWAYS_INLINE Element(const simdjson::dom::element & element_) : element(element_) {} /// NOLINT + ALWAYS_INLINE ElementType type() const { + switch (element.type()) { + case simdjson::dom::element_type::INT64: return ElementType::INT64; + case simdjson::dom::element_type::UINT64: return ElementType::UINT64; + case simdjson::dom::element_type::DOUBLE: return ElementType::DOUBLE; + case simdjson::dom::element_type::STRING: return ElementType::STRING; + case simdjson::dom::element_type::ARRAY: return ElementType::ARRAY; + case simdjson::dom::element_type::OBJECT: return ElementType::OBJECT; + case simdjson::dom::element_type::BOOL: return ElementType::BOOL; + case simdjson::dom::element_type::NULL_VALUE: + default: + return ElementType::NULL_VALUE; + } + } + ALWAYS_INLINE bool isInt64() const { return element.type() == simdjson::dom::element_type::INT64; } ALWAYS_INLINE bool isUInt64() const { return element.type() == simdjson::dom::element_type::UINT64; } ALWAYS_INLINE bool isDouble() const { return element.type() == simdjson::dom::element_type::DOUBLE; } From c3bcb1899df9dee8d6a98bfddec9414c3feed247 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rcio=20Martins?= Date: Thu, 8 Sep 2022 20:36:21 +0200 Subject: [PATCH 130/239] Add convenience ctor to ReadBufferFromMemory --- src/IO/ReadBufferFromMemory.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/IO/ReadBufferFromMemory.h b/src/IO/ReadBufferFromMemory.h index dc5c464604b..ad96e4bfa28 100644 --- a/src/IO/ReadBufferFromMemory.h +++ b/src/IO/ReadBufferFromMemory.h @@ -16,6 +16,8 @@ public: requires (sizeof(CharT) == 1) ReadBufferFromMemory(const CharT * buf, size_t size) : SeekableReadBuffer(const_cast(reinterpret_cast(buf)), size, 0) {} + explicit ReadBufferFromMemory(const std::string_view&& str) + : SeekableReadBuffer(const_cast(str.data()), str.size(), 0) {} off_t seek(off_t off, int whence) override; From 40b572c23b317a0c9a40db0733b85bacc6b0a128 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rcio=20Martins?= Date: Fri, 9 Sep 2022 16:52:39 +0200 Subject: [PATCH 131/239] Remove redundant default case in switch statement --- src/Common/JSONParsers/RapidJSONParser.h | 4 +--- src/Common/JSONParsers/SimdJSONParser.h | 4 +--- 2 files changed, 2 insertions(+), 6 deletions(-) diff --git a/src/Common/JSONParsers/RapidJSONParser.h b/src/Common/JSONParsers/RapidJSONParser.h index 16a87a0e0bf..4a0035c8360 100644 --- a/src/Common/JSONParsers/RapidJSONParser.h +++ b/src/Common/JSONParsers/RapidJSONParser.h @@ -34,9 +34,7 @@ struct RapidJSONParser case rapidjson::kObjectType: return ElementType::OBJECT; case rapidjson::kTrueType: return ElementType::BOOL; case rapidjson::kFalseType: return ElementType::BOOL; - case rapidjson::kNullType: - default: - return ElementType::NULL_VALUE; + case rapidjson::kNullType: return ElementType::NULL_VALUE; } } diff --git a/src/Common/JSONParsers/SimdJSONParser.h b/src/Common/JSONParsers/SimdJSONParser.h index dd2077ba768..2af0e1d6c01 100644 --- a/src/Common/JSONParsers/SimdJSONParser.h +++ b/src/Common/JSONParsers/SimdJSONParser.h @@ -40,9 +40,7 @@ struct SimdJSONParser case simdjson::dom::element_type::ARRAY: return ElementType::ARRAY; case simdjson::dom::element_type::OBJECT: return ElementType::OBJECT; case simdjson::dom::element_type::BOOL: return ElementType::BOOL; - case simdjson::dom::element_type::NULL_VALUE: - default: - return ElementType::NULL_VALUE; + case simdjson::dom::element_type::NULL_VALUE: return ElementType::NULL_VALUE; } } From 1ca230ff6d8a578aa60a8298c92fd129cfa57f4c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rcio=20Martins?= Date: Mon, 19 Sep 2022 14:00:48 +0200 Subject: [PATCH 132/239] Allow coercing in JSONExtract* functions --- src/Common/JSONParsers/ElementTypes.h | 3 +- src/Common/JSONParsers/RapidJSONParser.h | 6 +- src/Common/JSONParsers/SimdJSONParser.h | 6 +- src/Functions/FunctionsJSON.cpp | 179 +++++++++++++----- src/IO/readDecimalText.h | 41 ++-- .../00700_to_decimal_or_something.reference | 4 +- .../00918_json_functions.reference | 36 ++++ .../0_stateless/00918_json_functions.sql | 36 ++++ .../01186_conversion_to_nullable.reference | 2 +- ...685_json_extract_double_as_float.reference | 4 +- .../02013_json_function_null_column.reference | 3 +- .../02013_json_function_null_column.sql | 1 + 12 files changed, 239 insertions(+), 82 deletions(-) diff --git a/src/Common/JSONParsers/ElementTypes.h b/src/Common/JSONParsers/ElementTypes.h index 0dcfa7115aa..44e4c850a2f 100644 --- a/src/Common/JSONParsers/ElementTypes.h +++ b/src/Common/JSONParsers/ElementTypes.h @@ -3,7 +3,8 @@ namespace DB { // Enum values match simdjson's for fast conversion -enum class ElementType { +enum class ElementType +{ ARRAY = '[', OBJECT = '{', INT64 = 'l', diff --git a/src/Common/JSONParsers/RapidJSONParser.h b/src/Common/JSONParsers/RapidJSONParser.h index 4a0035c8360..6c5ea938bfe 100644 --- a/src/Common/JSONParsers/RapidJSONParser.h +++ b/src/Common/JSONParsers/RapidJSONParser.h @@ -26,8 +26,10 @@ struct RapidJSONParser ALWAYS_INLINE Element() = default; ALWAYS_INLINE Element(const rapidjson::Value & value_) : ptr(&value_) {} /// NOLINT - ALWAYS_INLINE ElementType type() const { - switch (ptr->GetType()) { + ALWAYS_INLINE ElementType type() const + { + switch (ptr->GetType()) + { case rapidjson::kNumberType: return ptr->IsDouble() ? ElementType::DOUBLE : (ptr->IsUint64() ? ElementType::UINT64 : ElementType::INT64); case rapidjson::kStringType: return ElementType::STRING; case rapidjson::kArrayType: return ElementType::ARRAY; diff --git a/src/Common/JSONParsers/SimdJSONParser.h b/src/Common/JSONParsers/SimdJSONParser.h index 2af0e1d6c01..f0f8f91109f 100644 --- a/src/Common/JSONParsers/SimdJSONParser.h +++ b/src/Common/JSONParsers/SimdJSONParser.h @@ -31,8 +31,10 @@ struct SimdJSONParser ALWAYS_INLINE Element() {} /// NOLINT ALWAYS_INLINE Element(const simdjson::dom::element & element_) : element(element_) {} /// NOLINT - ALWAYS_INLINE ElementType type() const { - switch (element.type()) { + ALWAYS_INLINE ElementType type() const + { + switch (element.type()) + { case simdjson::dom::element_type::INT64: return ElementType::INT64; case simdjson::dom::element_type::UINT64: return ElementType::UINT64; case simdjson::dom::element_type::DOUBLE: return ElementType::DOUBLE; diff --git a/src/Functions/FunctionsJSON.cpp b/src/Functions/FunctionsJSON.cpp index 493fc36ca3c..7c221209071 100644 --- a/src/Functions/FunctionsJSON.cpp +++ b/src/Functions/FunctionsJSON.cpp @@ -25,7 +25,6 @@ #include #include #include -#include #include #include #include @@ -40,6 +39,7 @@ #include #include +#include #include @@ -623,24 +623,32 @@ public: static bool insertResultToColumn(IColumn & dest, const Element & element, std::string_view) { UInt8 type; - if (element.isInt64()) - type = 'i'; - else if (element.isUInt64()) - type = 'u'; - else if (element.isDouble()) - type = 'd'; - else if (element.isBool()) - type = 'b'; - else if (element.isString()) - type = '"'; - else if (element.isArray()) - type = '['; - else if (element.isObject()) - type = '{'; - else if (element.isNull()) - type = 0; - else - return false; + switch (element.type()) + { + case ElementType::INT64: + type = 'i'; + break; + case ElementType::UINT64: + type = 'u'; + break; + case ElementType::DOUBLE: + type = 'd'; + break; + case ElementType::STRING: + type = '"'; + break; + case ElementType::ARRAY: + type = '['; + break; + case ElementType::OBJECT: + type = '{'; + break; + case ElementType::NULL_VALUE: + type = 0; + break; + default: + return false; + } ColumnVector & col_vec = assert_cast &>(dest); col_vec.insertValue(type); @@ -666,34 +674,51 @@ public: { NumberType value; - if (element.isInt64()) + switch (element.type()) { - if (!accurate::convertNumeric(element.getInt64(), value)) + case ElementType::DOUBLE: + if constexpr (std::is_floating_point_v) + { + /// We permit inaccurate conversion of double to float. + /// Example: double 0.1 from JSON is not representable in float. + /// But it will be more convenient for user to perform conversion. + value = static_cast(element.getDouble()); + } + else if (!accurate::convertNumeric(element.getDouble(), value)) + return false; + break; + case ElementType::UINT64: + if (!accurate::convertNumeric(element.getUInt64(), value)) + return false; + break; + case ElementType::INT64: + if (!accurate::convertNumeric(element.getInt64(), value)) + return false; + break; + case ElementType::BOOL: + if constexpr (is_integer && convert_bool_to_integer) + { + value = static_cast(element.getBool()); + break; + } return false; - } - else if (element.isUInt64()) - { - if (!accurate::convertNumeric(element.getUInt64(), value)) - return false; - } - else if (element.isDouble()) - { - if constexpr (std::is_floating_point_v) - { - /// We permit inaccurate conversion of double to float. - /// Example: double 0.1 from JSON is not representable in float. - /// But it will be more convenient for user to perform conversion. - value = static_cast(element.getDouble()); + case ElementType::STRING: { + auto rb = ReadBufferFromMemory{element.getString()}; + if constexpr (std::is_floating_point_v) + { + if (!tryReadFloatText(value, rb) || !rb.eof()) + return false; + } + else + { + if (!tryReadIntText(value, rb) || !rb.eof()) + return false; + } + break; } - else if (!accurate::convertNumeric(element.getDouble(), value)) + default: return false; } - else if (element.isBool() && is_integer && convert_bool_to_integer) - { - value = static_cast(element.getBool()); - } - else - return false; auto & col_vec = assert_cast &>(dest); col_vec.insertValue(value); @@ -719,9 +744,25 @@ using JSONExtractInt64Impl = JSONExtractNumericImpl; template using JSONExtractUInt64Impl = JSONExtractNumericImpl; template +using JSONExtractInt128Impl = JSONExtractNumericImpl; +template +using JSONExtractUInt128Impl = JSONExtractNumericImpl; +template +using JSONExtractInt256Impl = JSONExtractNumericImpl; +template +using JSONExtractUInt256Impl = JSONExtractNumericImpl; +template using JSONExtractFloat32Impl = JSONExtractNumericImpl; template using JSONExtractFloat64Impl = JSONExtractNumericImpl; +template +using JSONExtractDecimal32Impl = JSONExtractNumericImpl; +template +using JSONExtractDecimal64Impl = JSONExtractNumericImpl; +template +using JSONExtractDecimal128Impl = JSONExtractNumericImpl; +template +using JSONExtractDecimal256Impl = JSONExtractNumericImpl; template @@ -739,11 +780,22 @@ public: static bool insertResultToColumn(IColumn & dest, const Element & element, std::string_view) { - if (!element.isBool()) - return false; + bool value; + switch (element.type()) + { + case ElementType::BOOL: + value = element.getBool(); + break; + case ElementType::INT64: + case ElementType::UINT64: + value = element.getUInt64() != 0; + break; + default: + return false; + } auto & col_vec = assert_cast &>(dest); - col_vec.insertValue(static_cast(element.getBool())); + col_vec.insertValue(static_cast(value)); return true; } }; @@ -845,12 +897,35 @@ struct JSONExtractTree explicit DecimalNode(DataTypePtr data_type_) : data_type(data_type_) {} bool insertResultToColumn(IColumn & dest, const Element & element) override { - if (!element.isDouble()) - return false; - const auto * type = assert_cast *>(data_type.get()); - auto result = convertToDecimal, DataTypeDecimal>(element.getDouble(), type->getScale()); - assert_cast &>(dest).insert(result); + + DecimalType value{}; + + switch (element.type()) + { + case ElementType::DOUBLE: + value = convertToDecimal, DataTypeDecimal>( + element.getDouble(), type->getScale()); + break; + case ElementType::UINT64: + value = convertToDecimal, DataTypeDecimal>( + element.getUInt64(), type->getScale()); + break; + case ElementType::INT64: + value = convertToDecimal, DataTypeDecimal>( + element.getInt64(), type->getScale()); + break; + case ElementType::STRING: { + auto rb = ReadBufferFromMemory{element.getString()}; + if (!SerializationDecimal::tryReadText(value, rb, DecimalUtils::max_precision, type->getScale())) + return false; + break; + } + default: + return false; + } + + assert_cast &>(dest).insert(value); return true; } private: @@ -1088,10 +1163,14 @@ struct JSONExtractTree case TypeIndex::UInt16: return std::make_unique>(); case TypeIndex::UInt32: return std::make_unique>(); case TypeIndex::UInt64: return std::make_unique>(); + case TypeIndex::UInt128: return std::make_unique>(); + case TypeIndex::UInt256: return std::make_unique>(); case TypeIndex::Int8: return std::make_unique>(); case TypeIndex::Int16: return std::make_unique>(); case TypeIndex::Int32: return std::make_unique>(); case TypeIndex::Int64: return std::make_unique>(); + case TypeIndex::Int128: return std::make_unique>(); + case TypeIndex::Int256: return std::make_unique>(); case TypeIndex::Float32: return std::make_unique>(); case TypeIndex::Float64: return std::make_unique>(); case TypeIndex::String: return std::make_unique(); diff --git a/src/IO/readDecimalText.h b/src/IO/readDecimalText.h index 64374a20574..054990a8c7e 100644 --- a/src/IO/readDecimalText.h +++ b/src/IO/readDecimalText.h @@ -147,23 +147,30 @@ inline bool readDigits(ReadBuffer & buf, T & x, uint32_t & digits, int32_t & exp return true; } -template -inline void readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_t & scale, bool digits_only = false) +template +inline bool readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_t & scale, bool digits_only = false) { uint32_t digits = precision; int32_t exponent; - readDigits(buf, x, digits, exponent, digits_only); + auto ok = readDigits<_throw_on_error>(buf, x, digits, exponent, digits_only); + + if (!_throw_on_error && !ok) + return false; if (static_cast(digits) + exponent > static_cast(precision - scale)) { - static constexpr const char * pattern = - "Decimal value is too big: {} digits were read: {}e{}." - " Expected to read decimal with scale {} and precision {}"; + if constexpr (_throw_on_error) + { + static constexpr const char * pattern = "Decimal value is too big: {} digits were read: {}e{}." + " Expected to read decimal with scale {} and precision {}"; - if constexpr (is_big_int_v) - throw Exception(fmt::format(pattern, digits, x.value, exponent, scale, precision), ErrorCodes::ARGUMENT_OUT_OF_BOUND); + if constexpr (is_big_int_v) + throw Exception(fmt::format(pattern, digits, x.value, exponent, scale, precision), ErrorCodes::ARGUMENT_OUT_OF_BOUND); + else + throw Exception(fmt::format(pattern, digits, x, exponent, scale, precision), ErrorCodes::ARGUMENT_OUT_OF_BOUND); + } else - throw Exception(fmt::format(pattern, digits, x, exponent, scale, precision), ErrorCodes::ARGUMENT_OUT_OF_BOUND); + return false; } if (static_cast(scale) + exponent < 0) @@ -175,7 +182,7 @@ inline void readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_ /// Too big negative exponent x.value = 0; scale = 0; - return; + return true; } else { @@ -184,26 +191,18 @@ inline void readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_ assert(divisor > 0); /// This is for Clang Static Analyzer. It is not smart enough to infer it automatically. x.value /= divisor; scale = 0; - return; + return true; } } scale += exponent; + return true; } template inline bool tryReadDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_t & scale) { - uint32_t digits = precision; - int32_t exponent; - - if (!readDigits(buf, x, digits, exponent, true) || - static_cast(digits) + exponent > static_cast(precision - scale) || - static_cast(scale) + exponent < 0) - return false; - - scale += exponent; - return true; + return readDecimalText(buf, x, precision, scale, true); } template diff --git a/tests/queries/0_stateless/00700_to_decimal_or_something.reference b/tests/queries/0_stateless/00700_to_decimal_or_something.reference index 89ded7bd6d4..dec36ed5df5 100644 --- a/tests/queries/0_stateless/00700_to_decimal_or_something.reference +++ b/tests/queries/0_stateless/00700_to_decimal_or_something.reference @@ -1,5 +1,5 @@ 1.1 1.1 1.1 -0 +1 0 0.42 0 0.42 0 0.42 @@ -13,7 +13,7 @@ 0 ---- 1.1 1.1 1.1 -\N +1 \N -0.42 \N -0.42 \N -0.42 diff --git a/tests/queries/0_stateless/00918_json_functions.reference b/tests/queries/0_stateless/00918_json_functions.reference index 8e6fc3914e0..7d925c1236d 100644 --- a/tests/queries/0_stateless/00918_json_functions.reference +++ b/tests/queries/0_stateless/00918_json_functions.reference @@ -61,11 +61,47 @@ Friday (1,'417ddc5d-e556-4d27-95dd-a34d84e46a50') hello (3333.6,'test') +(3333.6,'test') +(3333.6333333333,'test') (3333.6333333333,'test') 123456.1234 Decimal(20, 4) +123456.1234 Decimal(20, 4) +123456789012345.12 Decimal(30, 4) +(1234567890.1234567890123456789,'test') Tuple(a Decimal(35, 20), b LowCardinality(String)) +(1234567890.12345678901234567890123456789,'test') Tuple(a Decimal(45, 30), b LowCardinality(String)) 123456789012345.1136 123456789012345.1136 1234567890.12345677879616925706 (1234567890.12345677879616925706,'test') 1234567890.123456695758468374595199311875 (1234567890.123456695758468374595199311875,'test') +-1234567890 Int32 +1234567890 UInt32 +-1234567890123456789 Int64 +1234567890123456789 UInt64 +-1234567890123456789 Int128 +1234567890123456789 UInt128 +-1234567890123456789 Int256 +1234567890123456789 UInt256 +-123456789 Int32 +123456789 UInt32 +-123456789012 Int64 +123456789012 UInt64 +-123456789012 Int128 +123456789012 UInt128 +-123456789012 Int256 +123456789012 UInt256 +-123456789 Int32 +123456789 UInt32 +-1234567890123456789 Int64 +1234567890123456789 UInt64 +-12345678901234567890123456789012345678 Int128 +12345678901234567890123456789012345678 UInt128 +-11345678901234567890123456789012345678901234567890123456789012345678901234567 Int256 +11345678901234567890123456789012345678901234567890123456789012345678901234567 UInt256 +0 Int32 +0 UInt32 +0 Int64 +0 UInt64 +false Bool +true Bool --JSONExtractKeysAndValues-- [('a','hello'),('b','[-100,200,300]')] [('b',[-100,200,300])] diff --git a/tests/queries/0_stateless/00918_json_functions.sql b/tests/queries/0_stateless/00918_json_functions.sql index 87682587c8e..d614d507dda 100644 --- a/tests/queries/0_stateless/00918_json_functions.sql +++ b/tests/queries/0_stateless/00918_json_functions.sql @@ -72,11 +72,47 @@ SELECT JSONExtract('{"a":123456, "b":3.55}', 'Tuple(a LowCardinality(Int32), b D SELECT JSONExtract('{"a":1, "b":"417ddc5d-e556-4d27-95dd-a34d84e46a50"}', 'Tuple(a Int8, b UUID)'); SELECT JSONExtract('{"a": "hello", "b": [-100, 200.0, 300]}', 'a', 'LowCardinality(String)'); SELECT JSONExtract('{"a":3333.6333333333333333333333, "b":"test"}', 'Tuple(a Decimal(10,1), b LowCardinality(String))'); +SELECT JSONExtract('{"a":"3333.6333333333333333333333", "b":"test"}', 'Tuple(a Decimal(10,1), b LowCardinality(String))'); SELECT JSONExtract('{"a":3333.6333333333333333333333, "b":"test"}', 'Tuple(a Decimal(20,10), b LowCardinality(String))'); +SELECT JSONExtract('{"a":"3333.6333333333333333333333", "b":"test"}', 'Tuple(a Decimal(20,10), b LowCardinality(String))'); SELECT JSONExtract('{"a":123456.123456}', 'a', 'Decimal(20, 4)') as a, toTypeName(a); +SELECT JSONExtract('{"a":"123456.123456"}', 'a', 'Decimal(20, 4)') as a, toTypeName(a); +SELECT JSONExtract('{"a":"123456789012345.12"}', 'a', 'Decimal(30, 4)') as a, toTypeName(a); +SELECT JSONExtract('{"a":"1234567890.12345678901234567890", "b":"test"}', 'Tuple(a Decimal(35,20), b LowCardinality(String))') as a, toTypeName(a); +SELECT JSONExtract('{"a":"1234567890.123456789012345678901234567890", "b":"test"}', 'Tuple(a Decimal(45,30), b LowCardinality(String))') as a, toTypeName(a); SELECT toDecimal64(123456789012345.12, 4), JSONExtract('{"a":123456789012345.12}', 'a', 'Decimal(30, 4)'); SELECT toDecimal128(1234567890.12345678901234567890, 20), JSONExtract('{"a":1234567890.12345678901234567890, "b":"test"}', 'Tuple(a Decimal(35,20), b LowCardinality(String))'); SELECT toDecimal256(1234567890.123456789012345678901234567890, 30), JSONExtract('{"a":1234567890.12345678901234567890, "b":"test"}', 'Tuple(a Decimal(45,30), b LowCardinality(String))'); +SELECT JSONExtract('{"a":-1234567890}', 'a', 'Int32') as a, toTypeName(a); +SELECT JSONExtract('{"a":1234567890}', 'a', 'UInt32') as a, toTypeName(a); +SELECT JSONExtract('{"a":-1234567890123456789}', 'a', 'Int64') as a, toTypeName(a); +SELECT JSONExtract('{"a":1234567890123456789}', 'a', 'UInt64') as a, toTypeName(a); +SELECT JSONExtract('{"a":-1234567890123456789}', 'a', 'Int128') as a, toTypeName(a); +SELECT JSONExtract('{"a":1234567890123456789}', 'a', 'UInt128') as a, toTypeName(a); +SELECT JSONExtract('{"a":-1234567890123456789}', 'a', 'Int256') as a, toTypeName(a); +SELECT JSONExtract('{"a":1234567890123456789}', 'a', 'UInt256') as a, toTypeName(a); +SELECT JSONExtract('{"a":-123456789.345}', 'a', 'Int32') as a, toTypeName(a); +SELECT JSONExtract('{"a":123456789.345}', 'a', 'UInt32') as a, toTypeName(a); +SELECT JSONExtract('{"a":-123456789012.345}', 'a', 'Int64') as a, toTypeName(a); +SELECT JSONExtract('{"a":123456789012.345}', 'a', 'UInt64') as a, toTypeName(a); +SELECT JSONExtract('{"a":-123456789012.345}', 'a', 'Int128') as a, toTypeName(a); +SELECT JSONExtract('{"a":123456789012.345}', 'a', 'UInt128') as a, toTypeName(a); +SELECT JSONExtract('{"a":-123456789012.345}', 'a', 'Int256') as a, toTypeName(a); +SELECT JSONExtract('{"a":123456789012.345}', 'a', 'UInt256') as a, toTypeName(a); +SELECT JSONExtract('{"a":"-123456789"}', 'a', 'Int32') as a, toTypeName(a); +SELECT JSONExtract('{"a":"123456789"}', 'a', 'UInt32') as a, toTypeName(a); +SELECT JSONExtract('{"a":"-1234567890123456789"}', 'a', 'Int64') as a, toTypeName(a); +SELECT JSONExtract('{"a":"1234567890123456789"}', 'a', 'UInt64') as a, toTypeName(a); +SELECT JSONExtract('{"a":"-12345678901234567890123456789012345678"}', 'a', 'Int128') as a, toTypeName(a); +SELECT JSONExtract('{"a":"12345678901234567890123456789012345678"}', 'a', 'UInt128') as a, toTypeName(a); +SELECT JSONExtract('{"a":"-11345678901234567890123456789012345678901234567890123456789012345678901234567"}', 'a', 'Int256') as a, toTypeName(a); +SELECT JSONExtract('{"a":"11345678901234567890123456789012345678901234567890123456789012345678901234567"}', 'a', 'UInt256') as a, toTypeName(a); +SELECT JSONExtract('{"a":"-1234567899999"}', 'a', 'Int32') as a, toTypeName(a); +SELECT JSONExtract('{"a":"1234567899999"}', 'a', 'UInt32') as a, toTypeName(a); +SELECT JSONExtract('{"a":"-1234567890123456789999"}', 'a', 'Int64') as a, toTypeName(a); +SELECT JSONExtract('{"a":"1234567890123456789999"}', 'a', 'UInt64') as a, toTypeName(a); +SELECT JSONExtract('{"a":0}', 'a', 'Bool') as a, toTypeName(a); +SELECT JSONExtract('{"a":1}', 'a', 'Bool') as a, toTypeName(a); SELECT '--JSONExtractKeysAndValues--'; SELECT JSONExtractKeysAndValues('{"a": "hello", "b": [-100, 200.0, 300]}', 'String'); diff --git a/tests/queries/0_stateless/01186_conversion_to_nullable.reference b/tests/queries/0_stateless/01186_conversion_to_nullable.reference index 86fa0afff20..e4c1fd7c40b 100644 --- a/tests/queries/0_stateless/01186_conversion_to_nullable.reference +++ b/tests/queries/0_stateless/01186_conversion_to_nullable.reference @@ -26,7 +26,7 @@ \N 42 \N -\N +3.14 42 \N 3.14159 diff --git a/tests/queries/0_stateless/01685_json_extract_double_as_float.reference b/tests/queries/0_stateless/01685_json_extract_double_as_float.reference index f3f4206b425..a24f6569f44 100644 --- a/tests/queries/0_stateless/01685_json_extract_double_as_float.reference +++ b/tests/queries/0_stateless/01685_json_extract_double_as_float.reference @@ -1,7 +1,7 @@ 1.1 1.1 1.1 1.1 0.01 0.01 0.01 0.01 -0 -\N +1 +1 -1e300 -inf 0 diff --git a/tests/queries/0_stateless/02013_json_function_null_column.reference b/tests/queries/0_stateless/02013_json_function_null_column.reference index b580986f548..ab702ab52cb 100644 --- a/tests/queries/0_stateless/02013_json_function_null_column.reference +++ b/tests/queries/0_stateless/02013_json_function_null_column.reference @@ -11,7 +11,7 @@ b c 1 -0 +1 1 a 1 @@ -20,3 +20,4 @@ true 1 a \N +\N diff --git a/tests/queries/0_stateless/02013_json_function_null_column.sql b/tests/queries/0_stateless/02013_json_function_null_column.sql index ef5ea3e1320..bf680dfe0b5 100644 --- a/tests/queries/0_stateless/02013_json_function_null_column.sql +++ b/tests/queries/0_stateless/02013_json_function_null_column.sql @@ -23,3 +23,4 @@ SELECT JSONExtract('[1]', toNullable(1), 'Nullable(Bool)'); SELECT JSONExtract('[1]', toNullable(1), 'Nullable(Float)'); SELECT JSONExtract('["a"]', toNullable(1), 'Nullable(String)'); SELECT JSONExtract('["a"]', toNullable(1), 'Nullable(Int)'); +SELECT JSONExtract('["-a"]', toNullable(1), 'Nullable(Int)'); From aaf71d157ddae80941487299ef807259e97da30f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rcio=20Martins?= Date: Wed, 26 Oct 2022 21:44:17 +0200 Subject: [PATCH 133/239] Use return type template parameter to select throwing behavior --- src/IO/readDecimalText.h | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-) diff --git a/src/IO/readDecimalText.h b/src/IO/readDecimalText.h index 054990a8c7e..9d7f8137136 100644 --- a/src/IO/readDecimalText.h +++ b/src/IO/readDecimalText.h @@ -147,19 +147,21 @@ inline bool readDigits(ReadBuffer & buf, T & x, uint32_t & digits, int32_t & exp return true; } -template -inline bool readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_t & scale, bool digits_only = false) +template +inline ReturnType readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_t & scale, bool digits_only = false) { + static constexpr bool throw_exception = std::is_same_v; + uint32_t digits = precision; int32_t exponent; - auto ok = readDigits<_throw_on_error>(buf, x, digits, exponent, digits_only); + auto ok = readDigits(buf, x, digits, exponent, digits_only); - if (!_throw_on_error && !ok) - return false; + if (!throw_exception && !ok) + return ReturnType(false); if (static_cast(digits) + exponent > static_cast(precision - scale)) { - if constexpr (_throw_on_error) + if constexpr (throw_exception) { static constexpr const char * pattern = "Decimal value is too big: {} digits were read: {}e{}." " Expected to read decimal with scale {} and precision {}"; @@ -170,7 +172,7 @@ inline bool readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_ throw Exception(fmt::format(pattern, digits, x, exponent, scale, precision), ErrorCodes::ARGUMENT_OUT_OF_BOUND); } else - return false; + return ReturnType(false); } if (static_cast(scale) + exponent < 0) @@ -182,7 +184,7 @@ inline bool readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_ /// Too big negative exponent x.value = 0; scale = 0; - return true; + return ReturnType(true); } else { @@ -191,18 +193,18 @@ inline bool readDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_ assert(divisor > 0); /// This is for Clang Static Analyzer. It is not smart enough to infer it automatically. x.value /= divisor; scale = 0; - return true; + return ReturnType(true); } } scale += exponent; - return true; + return ReturnType(true); } template inline bool tryReadDecimalText(ReadBuffer & buf, T & x, uint32_t precision, uint32_t & scale) { - return readDecimalText(buf, x, precision, scale, true); + return readDecimalText(buf, x, precision, scale, true); } template From ec1389cbe7514b0ae4c00b8de228fee682cb573a Mon Sep 17 00:00:00 2001 From: Sergei Trifonov Date: Thu, 27 Oct 2022 10:30:00 +0200 Subject: [PATCH 134/239] Update src/Common/Throttler.cpp Co-authored-by: Yakov Olkhovskiy <99031427+yakov-olkhovskiy@users.noreply.github.com> --- src/Common/Throttler.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/Throttler.cpp b/src/Common/Throttler.cpp index a3ae966c49b..e0f2b63e8ae 100644 --- a/src/Common/Throttler.cpp +++ b/src/Common/Throttler.cpp @@ -65,7 +65,7 @@ void Throttler::add(size_t amount) /// Wait unless there is positive amount of tokens - throttling if (max_speed && tokens_value < 0) { - int64_t sleep_time = static_cast(-tokens_value * max_speed * NS); + int64_t sleep_time = static_cast(-tokens_value / max_speed * NS); accumulated_sleep += sleep_time; sleepForNanoseconds(sleep_time); accumulated_sleep -= sleep_time; From accf78f1fffae8b04ec14f0586e7038f94e59d51 Mon Sep 17 00:00:00 2001 From: serxa Date: Thu, 27 Oct 2022 08:39:08 +0000 Subject: [PATCH 135/239] fix the first add() call --- src/Common/Throttler.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Common/Throttler.cpp b/src/Common/Throttler.cpp index e0f2b63e8ae..169262d30c7 100644 --- a/src/Common/Throttler.cpp +++ b/src/Common/Throttler.cpp @@ -48,9 +48,9 @@ void Throttler::add(size_t amount) { std::lock_guard lock(mutex); auto now = clock_gettime_ns_adjusted(prev_ns); - if (max_speed && prev_ns != 0) + if (max_speed) { - double delta_seconds = static_cast(now - prev_ns) / NS; + double delta_seconds = prev_ns ? static_cast(now - prev_ns) / NS : 0; tokens = std::min(tokens + max_speed * delta_seconds - amount, max_burst); } count += amount; From 2b08fe69634c080f01af7109dcab4ae57afd4c13 Mon Sep 17 00:00:00 2001 From: flynn Date: Thu, 27 Oct 2022 08:56:18 +0000 Subject: [PATCH 136/239] Fix truncate table does not hold lock correctly --- src/Interpreters/InterpreterDropQuery.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Interpreters/InterpreterDropQuery.cpp b/src/Interpreters/InterpreterDropQuery.cpp index 8be3dce7bf1..28f8e43ee9b 100644 --- a/src/Interpreters/InterpreterDropQuery.cpp +++ b/src/Interpreters/InterpreterDropQuery.cpp @@ -213,7 +213,7 @@ BlockIO InterpreterDropQuery::executeToTableImpl(ContextPtr context_, ASTDropQue { /// And for simple MergeTree we can stop merges before acquiring the lock auto merges_blocker = table->getActionLock(ActionLocks::PartsMerge); - auto table_lock = table->lockExclusively(context_->getCurrentQueryId(), context_->getSettingsRef().lock_acquire_timeout); + table_lock = table->lockExclusively(context_->getCurrentQueryId(), context_->getSettingsRef().lock_acquire_timeout); } auto metadata_snapshot = table->getInMemoryMetadataPtr(); From 5fcbc26800d44fb76051265efc007a1640c42f41 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Wed, 26 Oct 2022 16:21:25 +0200 Subject: [PATCH 137/239] OrderByLimitByDuplicateEliminationPass improve performance --- ...OrderByLimitByDuplicateEliminationPass.cpp | 27 +++++++++++++------ 1 file changed, 19 insertions(+), 8 deletions(-) diff --git a/src/Analyzer/Passes/OrderByLimitByDuplicateEliminationPass.cpp b/src/Analyzer/Passes/OrderByLimitByDuplicateEliminationPass.cpp index 0232d8958ff..e4d6633b6e6 100644 --- a/src/Analyzer/Passes/OrderByLimitByDuplicateEliminationPass.cpp +++ b/src/Analyzer/Passes/OrderByLimitByDuplicateEliminationPass.cpp @@ -10,23 +10,34 @@ namespace DB namespace { -struct QueryTreeNodeHash +struct QueryTreeNodeWithHash { - size_t operator()(const IQueryTreeNode * node) const + explicit QueryTreeNodeWithHash(const IQueryTreeNode * node_) + : node(node_) + , hash(node->getTreeHash().first) + {} + + const IQueryTreeNode * node = nullptr; + size_t hash = 0; +}; + +struct QueryTreeNodeWithHashHash +{ + size_t operator()(const QueryTreeNodeWithHash & node_with_hash) const { - return node->getTreeHash().first; + return node_with_hash.hash; } }; -struct QueryTreeNodeEqualTo +struct QueryTreeNodeWithHashEqualTo { - size_t operator()(const IQueryTreeNode * lhs_node, const IQueryTreeNode * rhs_node) const + bool operator()(const QueryTreeNodeWithHash & lhs_node, const QueryTreeNodeWithHash & rhs_node) const { - return lhs_node->isEqual(*rhs_node); + return lhs_node.hash == rhs_node.hash && lhs_node.node->isEqual(*rhs_node.node); } }; -using QueryTreeNodeSet = std::unordered_set; +using QueryTreeNodeWithHashSet = std::unordered_set; class OrderByLimitByDuplicateEliminationVisitor : public InDepthQueryTreeVisitor { @@ -82,7 +93,7 @@ public: } private: - QueryTreeNodeSet unique_expressions_nodes_set; + QueryTreeNodeWithHashSet unique_expressions_nodes_set; }; } From 275d076cb720b132572210fdf3bfe1525b793cc9 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Thu, 27 Oct 2022 11:26:53 +0200 Subject: [PATCH 138/239] Fixed style check --- src/Analyzer/UnionNode.cpp | 1 + 1 file changed, 1 insertion(+) diff --git a/src/Analyzer/UnionNode.cpp b/src/Analyzer/UnionNode.cpp index 039077bd08a..52ad0102d7c 100644 --- a/src/Analyzer/UnionNode.cpp +++ b/src/Analyzer/UnionNode.cpp @@ -30,6 +30,7 @@ namespace DB namespace ErrorCodes { extern const int TYPE_MISMATCH; + extern const int BAD_ARGUMENTS; } UnionNode::UnionNode(SelectUnionMode union_mode_) From b2ab692d87ec8a8fac20b7d8cc92653504bb6aaf Mon Sep 17 00:00:00 2001 From: Nikolay Degterinsky Date: Thu, 27 Oct 2022 10:03:03 +0000 Subject: [PATCH 139/239] Safer getResult() --- src/Parsers/ExpressionListParsers.cpp | 252 +++++++++++++------------- 1 file changed, 128 insertions(+), 124 deletions(-) diff --git a/src/Parsers/ExpressionListParsers.cpp b/src/Parsers/ExpressionListParsers.cpp index 2f39162e104..c362340d013 100644 --- a/src/Parsers/ExpressionListParsers.cpp +++ b/src/Parsers/ExpressionListParsers.cpp @@ -561,13 +561,10 @@ public: virtual bool getResult(ASTPtr & node) { - if (elements.size() == 1) - { - node = std::move(elements[0]); - return true; - } + if (!finished) + return false; - return false; + return getResultImpl(node); } virtual bool parse(IParser::Pos & /*pos*/, Expected & /*expected*/, Action & /*action*/) = 0; @@ -746,6 +743,17 @@ public: Checkpoint current_checkpoint = Checkpoint::None; protected: + virtual bool getResultImpl(ASTPtr & node) + { + if (elements.size() == 1) + { + node = std::move(elements[0]); + return true; + } + + return false; + } + std::vector operators; ASTs operands; ASTs elements; @@ -766,17 +774,12 @@ public: bool getResult(ASTPtr & node) override { /// We can exit the main cycle outside the parse() function, - /// so we need to merge the element here + /// so we need to merge the element here. + /// Because of this 'finished' flag can also not be set. if (!mergeElement()) return false; - if (elements.size() == 1) - { - node = std::move(elements[0]); - return true; - } - - return false; + return Layer::getResultImpl(node); } bool parse(IParser::Pos & pos, Expected & /*expected*/, Action & /*action*/) override @@ -830,16 +833,6 @@ public: explicit FunctionLayer(String function_name_, bool allow_function_parameters_ = true) : function_name(function_name_), allow_function_parameters(allow_function_parameters_){} - bool getResult(ASTPtr & node) override - { - // FunctionLayer can be the only layer in our Layers stack, - // so we need to check that we exited the main cycle properly - if (!finished) - return false; - - return Layer::getResult(node); - } - bool parse(IParser::Pos & pos, Expected & expected, Action & action) override { /// | 0 | 1 | 2 | @@ -1039,17 +1032,6 @@ private: class RoundBracketsLayer : public Layer { public: - bool getResult(ASTPtr & node) override - { - // Round brackets can mean priority operator as well as function tuple() - if (!is_tuple && elements.size() == 1) - node = std::move(elements[0]); - else - node = makeASTFunction("tuple", std::move(elements)); - - return true; - } - bool parse(IParser::Pos & pos, Expected & expected, Action & action) override { if (ParserToken(TokenType::Comma).ignore(pos, expected)) @@ -1079,6 +1061,19 @@ public: return true; } + +protected: + bool getResultImpl(ASTPtr & node) override + { + // Round brackets can mean priority operator as well as function tuple() + if (!is_tuple && elements.size() == 1) + node = std::move(elements[0]); + else + node = makeASTFunction("tuple", std::move(elements)); + + return true; + } + private: bool is_tuple = false; }; @@ -1087,16 +1082,17 @@ private: class ArrayLayer : public LayerWithSeparator { public: - bool getResult(ASTPtr & node) override - { - node = makeASTFunction("array", std::move(elements)); - return true; - } - bool parse(IParser::Pos & pos, Expected & expected, Action & action) override { return LayerWithSeparator::parse(pos, expected, action); } + +protected: + bool getResultImpl(ASTPtr & node) override + { + node = makeASTFunction("array", std::move(elements)); + return true; + } }; /// Layer for arrayElement square brackets operator @@ -1216,23 +1212,6 @@ class ExtractLayer : public LayerWithSeparator(interval_kind.toDateDiffUnit()), elements[0], elements[1]); - else if (elements.size() == 3) - node = makeASTFunction("dateDiff", std::make_shared(interval_kind.toDateDiffUnit()), elements[0], elements[1], elements[2]); - else - return false; - } - else - { - node = makeASTFunction("dateDiff", std::move(elements)); - } - return true; - } - bool parse(IParser::Pos & pos, Expected & expected, Action & action) override { /// 0. Try to parse interval_kind (-> 1) @@ -1709,6 +1693,25 @@ public: return true; } +protected: + bool getResultImpl(ASTPtr & node) override + { + if (parsed_interval_kind) + { + if (elements.size() == 2) + node = makeASTFunction("dateDiff", std::make_shared(interval_kind.toDateDiffUnit()), elements[0], elements[1]); + else if (elements.size() == 3) + node = makeASTFunction("dateDiff", std::make_shared(interval_kind.toDateDiffUnit()), elements[0], elements[1], elements[2]); + else + return false; + } + else + { + node = makeASTFunction("dateDiff", std::move(elements)); + } + return true; + } + private: IntervalKind interval_kind; bool parsed_interval_kind = false; @@ -1892,16 +1895,6 @@ class ViewLayer : public Layer public: explicit ViewLayer(bool if_permitted_) : if_permitted(if_permitted_) {} - bool getResult(ASTPtr & node) override - { - if (if_permitted) - node = makeASTFunction("viewIfPermitted", std::move(elements)); - else - node = makeASTFunction("view", std::move(elements)); - - return true; - } - bool parse(IParser::Pos & pos, Expected & expected, Action & /*action*/) override { /// view(SELECT ...) @@ -1958,6 +1951,17 @@ public: return true; } +protected: + bool getResultImpl(ASTPtr & node) override + { + if (if_permitted) + node = makeASTFunction("viewIfPermitted", std::move(elements)); + else + node = makeASTFunction("view", std::move(elements)); + + return true; + } + private: bool if_permitted; }; From a36119247faaf4a394fb56625184ceebe3623c8a Mon Sep 17 00:00:00 2001 From: Radistka-75 <76100894+Radistka-75@users.noreply.github.com> Date: Thu, 27 Oct 2022 13:06:28 +0300 Subject: [PATCH 140/239] Update in.md Change md-formatting for note --- docs/ru/sql-reference/operators/in.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/ru/sql-reference/operators/in.md b/docs/ru/sql-reference/operators/in.md index 2b3d87a877f..fa679b890a7 100644 --- a/docs/ru/sql-reference/operators/in.md +++ b/docs/ru/sql-reference/operators/in.md @@ -122,9 +122,9 @@ FROM t_null Существует два варианта IN-ов с подзапросами (аналогично для JOIN-ов): обычный `IN` / `JOIN` и `GLOBAL IN` / `GLOBAL JOIN`. Они отличаются способом выполнения при распределённой обработке запроса. - :::note "Attention" - Помните, что алгоритмы, описанные ниже, могут работать иначе в зависимости от [настройки](../../operations/settings/settings.md) `distributed_product_mode`. - ::: +:::note "Attention" +Помните, что алгоритмы, описанные ниже, могут работать иначе в зависимости от [настройки](../../operations/settings/settings.md) `distributed_product_mode`. +::: При использовании обычного IN-а, запрос отправляется на удалённые серверы, и на каждом из них выполняются подзапросы в секциях `IN` / `JOIN`. При использовании `GLOBAL IN` / `GLOBAL JOIN-а`, сначала выполняются все подзапросы для `GLOBAL IN` / `GLOBAL JOIN-ов`, и результаты складываются во временные таблицы. Затем эти временные таблицы передаются на каждый удалённый сервер, и на них выполняются запросы, с использованием этих переданных временных данных. From d489c46f02b709e01bda01ffe7b3daa26cc63305 Mon Sep 17 00:00:00 2001 From: Nikolay Degterinsky Date: Thu, 27 Oct 2022 10:23:41 +0000 Subject: [PATCH 141/239] Better test --- tests/queries/0_stateless/02474_fix_function_parser_bug.sql | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/queries/0_stateless/02474_fix_function_parser_bug.sql b/tests/queries/0_stateless/02474_fix_function_parser_bug.sql index 12e9e03f151..67d97aa1c25 100644 --- a/tests/queries/0_stateless/02474_fix_function_parser_bug.sql +++ b/tests/queries/0_stateless/02474_fix_function_parser_bug.sql @@ -1 +1 @@ -CREATE DATABASE conv_mian ENGINE QALL(COLUMNS('|T.D'),¸mp} -- { clientError 62 } +CREATE DATABASE conv_mian ENGINE QALL(COLUMNS('|T.D'),¸mp} -- { clientError SYNTAX_ERROR } From 0245df6a82f4fff9f45f6a62dfe8a5c9307b0819 Mon Sep 17 00:00:00 2001 From: Nikolay Degterinsky Date: Thu, 27 Oct 2022 10:41:45 +0000 Subject: [PATCH 142/239] Fix docs --- docs/ru/engines/table-engines/integrations/kafka.md | 7 ++++--- .../mergetree-family/aggregatingmergetree.md | 7 ++++--- .../mergetree-family/collapsingmergetree.md | 9 +++++---- .../table-engines/mergetree-family/graphitemergetree.md | 7 ++++--- .../engines/table-engines/mergetree-family/mergetree.md | 7 ++++--- .../table-engines/mergetree-family/summingmergetree.md | 7 ++++--- docs/ru/sql-reference/functions/date-time-functions.md | 4 ++-- 7 files changed, 27 insertions(+), 21 deletions(-) diff --git a/docs/ru/engines/table-engines/integrations/kafka.md b/docs/ru/engines/table-engines/integrations/kafka.md index 37fc902e777..a5f091e1b23 100644 --- a/docs/ru/engines/table-engines/integrations/kafka.md +++ b/docs/ru/engines/table-engines/integrations/kafka.md @@ -87,14 +87,15 @@ SETTINGS Устаревший способ создания таблицы - :::note "Attention" - Не используйте этот метод в новых проектах. По возможности переключите старые проекты на метод, описанный выше. +:::note "Attention" +Не используйте этот метод в новых проектах. По возможности переключите старые проекты на метод, описанный выше. +::: ``` sql Kafka(kafka_broker_list, kafka_topic_list, kafka_group_name, kafka_format [, kafka_row_delimiter, kafka_schema, kafka_num_consumers, kafka_skip_broken_messages]) ``` - ::: + ## Описание {#opisanie} diff --git a/docs/ru/engines/table-engines/mergetree-family/aggregatingmergetree.md b/docs/ru/engines/table-engines/mergetree-family/aggregatingmergetree.md index aa16113192e..86a275767a0 100644 --- a/docs/ru/engines/table-engines/mergetree-family/aggregatingmergetree.md +++ b/docs/ru/engines/table-engines/mergetree-family/aggregatingmergetree.md @@ -39,9 +39,10 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] Устаревший способ создания таблицы - :::note "Attention" - Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. - ::: +:::note "Attention" +Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. +::: + ``` sql CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] ( diff --git a/docs/ru/engines/table-engines/mergetree-family/collapsingmergetree.md b/docs/ru/engines/table-engines/mergetree-family/collapsingmergetree.md index ecaaa6b8417..72b4725c6ed 100644 --- a/docs/ru/engines/table-engines/mergetree-family/collapsingmergetree.md +++ b/docs/ru/engines/table-engines/mergetree-family/collapsingmergetree.md @@ -43,9 +43,10 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] Устаревший способ создания таблицы - :::note "Attention" - Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. - ::: +:::note "Attention" +Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. +::: + ``` sql CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] ( @@ -59,7 +60,7 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] - `sign` — Имя столбца с типом строки: `1` — строка состояния, `-1` — строка отмены состояния. - Тип данных столбца — `Int8`. + Тип данных столбца — `Int8`. diff --git a/docs/ru/engines/table-engines/mergetree-family/graphitemergetree.md b/docs/ru/engines/table-engines/mergetree-family/graphitemergetree.md index 818f85f7e37..324a3fd1633 100644 --- a/docs/ru/engines/table-engines/mergetree-family/graphitemergetree.md +++ b/docs/ru/engines/table-engines/mergetree-family/graphitemergetree.md @@ -55,9 +55,10 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] Устаревший способ создания таблицы - :::note "Attention" - Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. - ::: +:::note "Attention" +Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. +::: + ``` sql CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] ( diff --git a/docs/ru/engines/table-engines/mergetree-family/mergetree.md b/docs/ru/engines/table-engines/mergetree-family/mergetree.md index e01e0006b87..f024d5f1985 100644 --- a/docs/ru/engines/table-engines/mergetree-family/mergetree.md +++ b/docs/ru/engines/table-engines/mergetree-family/mergetree.md @@ -115,9 +115,10 @@ ENGINE MergeTree() PARTITION BY toYYYYMM(EventDate) ORDER BY (CounterID, EventDa Устаревший способ создания таблицы - :::note "Attention" - Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ, описанный выше. - ::: +:::note "Attention" +Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ, описанный выше. +::: + ``` sql CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] ( diff --git a/docs/ru/engines/table-engines/mergetree-family/summingmergetree.md b/docs/ru/engines/table-engines/mergetree-family/summingmergetree.md index 0d9d268fa46..7b69927e161 100644 --- a/docs/ru/engines/table-engines/mergetree-family/summingmergetree.md +++ b/docs/ru/engines/table-engines/mergetree-family/summingmergetree.md @@ -42,9 +42,10 @@ CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] Устаревший способ создания таблицы - :::note "Attention" - Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. - ::: +:::note "Attention" +Не используйте этот способ в новых проектах и по возможности переведите старые проекты на способ описанный выше. +::: + ``` sql CREATE TABLE [IF NOT EXISTS] [db.]table_name [ON CLUSTER cluster] ( diff --git a/docs/ru/sql-reference/functions/date-time-functions.md b/docs/ru/sql-reference/functions/date-time-functions.md index 80e2561a8d7..d586bbd42e4 100644 --- a/docs/ru/sql-reference/functions/date-time-functions.md +++ b/docs/ru/sql-reference/functions/date-time-functions.md @@ -316,9 +316,9 @@ SELECT toStartOfISOYear(toDate('2017-01-01')) AS ISOYear20170101; Возвращается дата. :::note "Attention" - Возвращаемое значение для некорректных дат зависит от реализации. ClickHouse может вернуть нулевую дату, выбросить исключение, или выполнить «естественное» перетекание дат между месяцами. +Возвращаемое значение для некорректных дат зависит от реализации. ClickHouse может вернуть нулевую дату, выбросить исключение, или выполнить «естественное» перетекание дат между месяцами. ::: - + ## toMonday {#tomonday} Округляет дату или дату-с-временем вниз до ближайшего понедельника. From 8bd94979ebfa9fa2e90544910e78bee6928f4568 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Thu, 27 Oct 2022 10:46:41 +0000 Subject: [PATCH 143/239] Add test queries for Date32 --- tests/queries/0_stateless/00718_format_datetime.reference | 2 ++ tests/queries/0_stateless/00718_format_datetime.sql | 4 ++++ 2 files changed, 6 insertions(+) diff --git a/tests/queries/0_stateless/00718_format_datetime.reference b/tests/queries/0_stateless/00718_format_datetime.reference index 4f12a46d7c0..6a0fd451e73 100644 --- a/tests/queries/0_stateless/00718_format_datetime.reference +++ b/tests/queries/0_stateless/00718_format_datetime.reference @@ -28,6 +28,8 @@ PM % no formatting pattern 2018-01-01 00:00:00 +2018-01-01 00:00:00 +1927-01-01 00:00:00 2018-01-01 01:00:00 2018-01-01 04:00:00 +0000 -1100 diff --git a/tests/queries/0_stateless/00718_format_datetime.sql b/tests/queries/0_stateless/00718_format_datetime.sql index 7ed1f0abea4..532a0c02e6f 100644 --- a/tests/queries/0_stateless/00718_format_datetime.sql +++ b/tests/queries/0_stateless/00718_format_datetime.sql @@ -40,6 +40,10 @@ SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%%'); SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), 'no formatting pattern'); SELECT formatDateTime(toDate('2018-01-01'), '%F %T'); + +SELECT formatDateTime(toDate32('2018-01-01'), '%F %T'); +SELECT formatDateTime(toDate32('1927-01-01'), '%F %T'); + SELECT formatDateTime(toDateTime('2018-01-01 01:00:00', 'UTC'), '%F %T', 'UTC'), formatDateTime(toDateTime('2018-01-01 01:00:00', 'UTC'), '%F %T', 'Asia/Istanbul'); From c46ee42a99e785515fe4e4fe1839cdce657b7a58 Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Thu, 27 Oct 2022 11:12:23 +0000 Subject: [PATCH 144/239] Update version_date.tsv and changelogs after v22.8.8.3-lts --- docs/changelogs/v22.8.8.3-lts.md | 13 +++++++++++++ utils/list-versions/version_date.tsv | 1 + 2 files changed, 14 insertions(+) create mode 100644 docs/changelogs/v22.8.8.3-lts.md diff --git a/docs/changelogs/v22.8.8.3-lts.md b/docs/changelogs/v22.8.8.3-lts.md new file mode 100644 index 00000000000..deaab51fce9 --- /dev/null +++ b/docs/changelogs/v22.8.8.3-lts.md @@ -0,0 +1,13 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.8.8.3-lts (ac5a6cababc) FIXME as compared to v22.8.7.34-lts (3c38e5e8ab9) + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42677](https://github.com/ClickHouse/ClickHouse/issues/42677): keeper-fix: fix race in accessing logs while snapshot is being installed. [#40627](https://github.com/ClickHouse/ClickHouse/pull/40627) ([Antonio Andelic](https://github.com/antonio2368)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 0470152ecff..8977c98eb7e 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -3,6 +3,7 @@ v22.9.4.32-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 v22.9.2.7-stable 2022-09-23 v22.9.1.2603-stable 2022-09-22 +v22.8.8.3-lts 2022-10-27 v22.8.7.34-lts 2022-10-26 v22.8.6.71-lts 2022-09-30 v22.8.5.29-lts 2022-09-13 From 1fa24161e79833ebd333c67a56b0f41ff4ff5d68 Mon Sep 17 00:00:00 2001 From: Jordi Villar Date: Thu, 27 Oct 2022 18:34:37 +0200 Subject: [PATCH 145/239] Minor fix implicit cast CaresPTRResolver --- src/Common/CaresPTRResolver.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index c6fe70a09fa..ab4883c104a 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -128,7 +128,7 @@ namespace DB int number_of_fds_ready = 0; if (!readable_sockets.empty()) { - number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), timeout); + number_of_fds_ready = poll(readable_sockets.data(), static_cast(readable_sockets.size()), static_cast(timeout)); } if (number_of_fds_ready > 0) From c4f25228f73a86a626e5491ba7c7b6057ee4d3ed Mon Sep 17 00:00:00 2001 From: serxa Date: Thu, 27 Oct 2022 17:18:50 +0000 Subject: [PATCH 146/239] smaller burst to be closer to old behaviour --- src/Common/Throttler.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/Throttler.cpp b/src/Common/Throttler.cpp index 169262d30c7..b38777efc03 100644 --- a/src/Common/Throttler.cpp +++ b/src/Common/Throttler.cpp @@ -20,7 +20,7 @@ namespace ErrorCodes /// Just 10^9. static constexpr auto NS = 1000000000UL; -static const size_t default_burst_seconds = 10; +static const size_t default_burst_seconds = 1; Throttler::Throttler(size_t max_speed_, const std::shared_ptr & parent_) : max_speed(max_speed_) From 2038db21f485d8b3de230c58e30429012c4d2b1a Mon Sep 17 00:00:00 2001 From: serxa Date: Thu, 27 Oct 2022 17:26:12 +0000 Subject: [PATCH 147/239] lower limit to avoid hitting another bottleneck --- .../0_stateless/01288_shard_max_network_bandwidth.sql | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tests/queries/0_stateless/01288_shard_max_network_bandwidth.sql b/tests/queries/0_stateless/01288_shard_max_network_bandwidth.sql index 969bb0a126c..d2daf48a1cb 100644 --- a/tests/queries/0_stateless/01288_shard_max_network_bandwidth.sql +++ b/tests/queries/0_stateless/01288_shard_max_network_bandwidth.sql @@ -1,7 +1,7 @@ -- Tags: shard --- Limit to 10 MB/sec -SET max_network_bandwidth = 10000000; +-- Limit to 100 KB/sec +SET max_network_bandwidth = 100000; -- Lower max_block_size, so we can start throttling sooner. Otherwise query will be executed too quickly. SET max_block_size = 100; @@ -11,7 +11,7 @@ CREATE TEMPORARY TABLE times (t DateTime); -- rand64 is uncompressable data. Each number will take 8 bytes of bandwidth. -- This query should execute in no less than 1.6 seconds if throttled. INSERT INTO times SELECT now(); -SELECT sum(ignore(*)) FROM (SELECT rand64() FROM remote('127.0.0.{2,3}', numbers(2000000))); +SELECT sum(ignore(*)) FROM (SELECT rand64() FROM remote('127.0.0.{2,3}', numbers(20000))); INSERT INTO times SELECT now(); SELECT max(t) - min(t) >= 1 FROM times; From 073d81c31f612ccaba3ba23678966e202ef17dca Mon Sep 17 00:00:00 2001 From: Jordi Villar Date: Thu, 27 Oct 2022 20:25:49 +0200 Subject: [PATCH 148/239] Use nfds_t instead --- src/Common/CaresPTRResolver.cpp | 2 +- utils/iotest/iotest_nonblock.cpp | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index ab4883c104a..99b4c34dfbd 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -128,7 +128,7 @@ namespace DB int number_of_fds_ready = 0; if (!readable_sockets.empty()) { - number_of_fds_ready = poll(readable_sockets.data(), static_cast(readable_sockets.size()), static_cast(timeout)); + number_of_fds_ready = poll(readable_sockets.data(), static_cast(readable_sockets.size()), static_cast(timeout)); } if (number_of_fds_ready > 0) diff --git a/utils/iotest/iotest_nonblock.cpp b/utils/iotest/iotest_nonblock.cpp index 6db00045e03..32c86282743 100644 --- a/utils/iotest/iotest_nonblock.cpp +++ b/utils/iotest/iotest_nonblock.cpp @@ -101,7 +101,7 @@ int mainImpl(int argc, char ** argv) size_t ops = 0; while (ops < count) { - if (poll(polls.data(), static_cast(descriptors), -1) <= 0) + if (poll(polls.data(), static_cast(descriptors), -1) <= 0) throwFromErrno("poll failed", ErrorCodes::SYSTEM_ERROR); for (size_t i = 0; i < descriptors; ++i) { From a4ce0e344a33a3eb862a8178761bdeddf1022a08 Mon Sep 17 00:00:00 2001 From: Igor Nikonov Date: Thu, 27 Oct 2022 21:11:41 +0000 Subject: [PATCH 149/239] Fix build src/Common/CaresPTRResolver.cpp:126:27 error: implicit conversion loses integer precision: 'int64_t' (aka 'long') to 'int' [-Werror,-Wshorten-64-to-32] --- src/Common/CaresPTRResolver.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Common/CaresPTRResolver.cpp b/src/Common/CaresPTRResolver.cpp index c6fe70a09fa..f5533a416ad 100644 --- a/src/Common/CaresPTRResolver.cpp +++ b/src/Common/CaresPTRResolver.cpp @@ -128,7 +128,7 @@ namespace DB int number_of_fds_ready = 0; if (!readable_sockets.empty()) { - number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), timeout); + number_of_fds_ready = poll(readable_sockets.data(), readable_sockets.size(), static_cast(timeout)); } if (number_of_fds_ready > 0) From 51e59405bc33ce3b2edddb6a08de5b7795e143d5 Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Fri, 28 Oct 2022 09:41:29 +0000 Subject: [PATCH 150/239] Update version_date.tsv and changelogs after v22.3.14.18-lts --- docs/changelogs/v22.3.14.18-lts.md | 26 ++++++++++++++++++++++++++ utils/list-versions/version_date.tsv | 1 + 2 files changed, 27 insertions(+) create mode 100644 docs/changelogs/v22.3.14.18-lts.md diff --git a/docs/changelogs/v22.3.14.18-lts.md b/docs/changelogs/v22.3.14.18-lts.md new file mode 100644 index 00000000000..d0c67a2b241 --- /dev/null +++ b/docs/changelogs/v22.3.14.18-lts.md @@ -0,0 +1,26 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.3.14.18-lts (642946f61b2) FIXME as compared to v22.3.13.80-lts (e2708b01fba) + +#### Bug Fix +* Backported in [#42432](https://github.com/ClickHouse/ClickHouse/issues/42432): - Choose correct aggregation method for LowCardinality with BigInt. [#42342](https://github.com/ClickHouse/ClickHouse/pull/42342) ([Duc Canh Le](https://github.com/canhld94)). + +#### Build/Testing/Packaging Improvement +* Backported in [#42328](https://github.com/ClickHouse/ClickHouse/issues/42328): Update cctz to the latest master, update tzdb to 2020e. [#42273](https://github.com/ClickHouse/ClickHouse/pull/42273) ([Dom Del Nano](https://github.com/ddelnano)). +* Backported in [#42358](https://github.com/ClickHouse/ClickHouse/issues/42358): Update tzdata to 2022e to support the new timezone changes. Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Jordan and Syria switch from +02/+03 with DST to year-round +03. (https://data.iana.org/time-zones/tzdb/NEWS). This closes [#42252](https://github.com/ClickHouse/ClickHouse/issues/42252). [#42327](https://github.com/ClickHouse/ClickHouse/pull/42327) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42298](https://github.com/ClickHouse/ClickHouse/issues/42298): Fix a bug with projections and the `aggregate_functions_null_for_empty` setting. This bug is very rare and appears only if you enable the `aggregate_functions_null_for_empty` setting in the server's config. This closes [#41647](https://github.com/ClickHouse/ClickHouse/issues/41647). [#42198](https://github.com/ClickHouse/ClickHouse/pull/42198) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42592](https://github.com/ClickHouse/ClickHouse/issues/42592): This closes [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Add a warning message to release.py script, require release type [#41975](https://github.com/ClickHouse/ClickHouse/pull/41975) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787) [#42136](https://github.com/ClickHouse/ClickHouse/pull/42136) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 8977c98eb7e..5e50b998ec4 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -36,6 +36,7 @@ v22.4.5.9-stable 2022-05-06 v22.4.4.7-stable 2022-04-29 v22.4.3.3-stable 2022-04-26 v22.4.2.1-stable 2022-04-22 +v22.3.14.18-lts 2022-10-28 v22.3.13.80-lts 2022-09-30 v22.3.12.19-lts 2022-08-29 v22.3.11.12-lts 2022-08-10 From acec4475261df88a9139f2743f4034b6d33b9554 Mon Sep 17 00:00:00 2001 From: Sergei Trifonov Date: Fri, 28 Oct 2022 12:35:18 +0200 Subject: [PATCH 151/239] Fix anchor links --- docs/en/sql-reference/statements/alter/partition.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/docs/en/sql-reference/statements/alter/partition.md b/docs/en/sql-reference/statements/alter/partition.md index da99c52538f..2d89c1d5d18 100644 --- a/docs/en/sql-reference/statements/alter/partition.md +++ b/docs/en/sql-reference/statements/alter/partition.md @@ -39,7 +39,7 @@ ALTER TABLE mt DETACH PARTITION '2020-11-21'; ALTER TABLE mt DETACH PART 'all_2_2_0'; ``` -Read about setting the partition expression in a section [How to specify the partition expression](#alter-how-to-specify-part-expr). +Read about setting the partition expression in a section [How to set the partition expression](#how-to-set-partition-expression). After the query is executed, you can do whatever you want with the data in the `detached` directory — delete it from the file system, or just leave it. @@ -53,7 +53,7 @@ ALTER TABLE table_name [ON CLUSTER cluster] DROP PARTITION|PART partition_expr Deletes the specified partition from the table. This query tags the partition as inactive and deletes data completely, approximately in 10 minutes. -Read about setting the partition expression in a section [How to specify the partition expression](#alter-how-to-specify-part-expr). +Read about setting the partition expression in a section [How to set the partition expression](#how-to-set-partition-expression). The query is replicated – it deletes data on all replicas. @@ -71,7 +71,7 @@ ALTER TABLE table_name [ON CLUSTER cluster] DROP DETACHED PARTITION|PART partiti ``` Removes the specified part or all parts of the specified partition from `detached`. -Read more about setting the partition expression in a section [How to specify the partition expression](#alter-how-to-specify-part-expr). +Read more about setting the partition expression in a section [How to set the partition expression](#how-to-set-partition-expression). ## ATTACH PARTITION\|PART @@ -86,7 +86,7 @@ ALTER TABLE visits ATTACH PARTITION 201901; ALTER TABLE visits ATTACH PART 201901_2_2_0; ``` -Read more about setting the partition expression in a section [How to specify the partition expression](#alter-how-to-specify-part-expr). +Read more about setting the partition expression in a section [How to set the partition expression](#how-to-set-partition-expression). This query is replicated. The replica-initiator checks whether there is data in the `detached` directory. If data exists, the query checks its integrity. If everything is correct, the query adds the data to the table. @@ -166,7 +166,7 @@ This query creates a local backup of a specified partition. If the `PARTITION` c The entire backup process is performed without stopping the server. ::: -Note that for old-styled tables you can specify the prefix of the partition name (for example, `2019`) - then the query creates the backup for all the corresponding partitions. Read about setting the partition expression in a section [How to specify the partition expression](#alter-how-to-specify-part-expr). +Note that for old-styled tables you can specify the prefix of the partition name (for example, `2019`) - then the query creates the backup for all the corresponding partitions. Read about setting the partition expression in a section [How to set the partition expression](#how-to-set-partition-expression). At the time of execution, for a data snapshot, the query creates hardlinks to a table data. Hardlinks are placed in the directory `/var/lib/clickhouse/shadow/N/...`, where: From 4f62bd6ba05f44ab992a4eed53dc55c8ad9727d9 Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Fri, 28 Oct 2022 11:43:06 +0000 Subject: [PATCH 152/239] Update version_date.tsv and changelogs after v22.3.14.23-lts --- docs/changelogs/v22.3.14.23-lts.md | 29 ++++++++++++++++++++++++++++ utils/list-versions/version_date.tsv | 1 + 2 files changed, 30 insertions(+) create mode 100644 docs/changelogs/v22.3.14.23-lts.md diff --git a/docs/changelogs/v22.3.14.23-lts.md b/docs/changelogs/v22.3.14.23-lts.md new file mode 100644 index 00000000000..663d8b43f6f --- /dev/null +++ b/docs/changelogs/v22.3.14.23-lts.md @@ -0,0 +1,29 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.3.14.23-lts (74956bfee4d) FIXME as compared to v22.3.13.80-lts (e2708b01fba) + +#### Improvement +* Backported in [#42527](https://github.com/ClickHouse/ClickHouse/issues/42527): Fix issue with passing MySQL timeouts for MySQL database engine and MySQL table function. Closes [#34168](https://github.com/ClickHouse/ClickHouse/issues/34168)?notification_referrer_id=NT_kwDOAzsV57MzMDMxNjAzNTY5OjU0MjAzODc5. [#40751](https://github.com/ClickHouse/ClickHouse/pull/40751) ([Kseniia Sumarokova](https://github.com/kssenii)). + +#### Bug Fix +* Backported in [#42432](https://github.com/ClickHouse/ClickHouse/issues/42432): - Choose correct aggregation method for LowCardinality with BigInt. [#42342](https://github.com/ClickHouse/ClickHouse/pull/42342) ([Duc Canh Le](https://github.com/canhld94)). + +#### Build/Testing/Packaging Improvement +* Backported in [#42328](https://github.com/ClickHouse/ClickHouse/issues/42328): Update cctz to the latest master, update tzdb to 2020e. [#42273](https://github.com/ClickHouse/ClickHouse/pull/42273) ([Dom Del Nano](https://github.com/ddelnano)). +* Backported in [#42358](https://github.com/ClickHouse/ClickHouse/issues/42358): Update tzdata to 2022e to support the new timezone changes. Palestine transitions are now Saturdays at 02:00. Simplify three Ukraine zones into one. Jordan and Syria switch from +02/+03 with DST to year-round +03. (https://data.iana.org/time-zones/tzdb/NEWS). This closes [#42252](https://github.com/ClickHouse/ClickHouse/issues/42252). [#42327](https://github.com/ClickHouse/ClickHouse/pull/42327) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42298](https://github.com/ClickHouse/ClickHouse/issues/42298): Fix a bug with projections and the `aggregate_functions_null_for_empty` setting. This bug is very rare and appears only if you enable the `aggregate_functions_null_for_empty` setting in the server's config. This closes [#41647](https://github.com/ClickHouse/ClickHouse/issues/41647). [#42198](https://github.com/ClickHouse/ClickHouse/pull/42198) ([Alexey Milovidov](https://github.com/alexey-milovidov)). +* Backported in [#42592](https://github.com/ClickHouse/ClickHouse/issues/42592): This closes [#42453](https://github.com/ClickHouse/ClickHouse/issues/42453). [#42573](https://github.com/ClickHouse/ClickHouse/pull/42573) ([Alexey Milovidov](https://github.com/alexey-milovidov)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Add a warning message to release.py script, require release type [#41975](https://github.com/ClickHouse/ClickHouse/pull/41975) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). +* Revert [#27787](https://github.com/ClickHouse/ClickHouse/issues/27787) [#42136](https://github.com/ClickHouse/ClickHouse/pull/42136) ([Nikolai Kochetov](https://github.com/KochetovNicolai)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 8977c98eb7e..21197713694 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -36,6 +36,7 @@ v22.4.5.9-stable 2022-05-06 v22.4.4.7-stable 2022-04-29 v22.4.3.3-stable 2022-04-26 v22.4.2.1-stable 2022-04-22 +v22.3.14.23-lts 2022-10-28 v22.3.13.80-lts 2022-09-30 v22.3.12.19-lts 2022-08-29 v22.3.11.12-lts 2022-08-10 From 15f3f56812d82243d0ed89a45916eb2bf5e1be70 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Fri, 28 Oct 2022 12:44:47 +0000 Subject: [PATCH 153/239] Change 1 argument of function to first argument of function etc --- src/Functions/formatDateTime.cpp | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/src/Functions/formatDateTime.cpp b/src/Functions/formatDateTime.cpp index c5240abf7a1..9634768e68a 100644 --- a/src/Functions/formatDateTime.cpp +++ b/src/Functions/formatDateTime.cpp @@ -322,11 +322,11 @@ public: getName(), arguments.size()); if (arguments.size() == 1 && !isInteger(arguments[0].type)) throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, - "Illegal type {} of 1 argument of function {} when arguments size is 1. Should be integer", + "Illegal type {} of first argument of function {} when arguments size is 1. Should be integer", arguments[0].type->getName(), getName()); if (arguments.size() > 1 && !(isInteger(arguments[0].type) || isDate(arguments[0].type) || isDateTime(arguments[0].type) || isDate32(arguments[0].type) || isDateTime64(arguments[0].type))) throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, - "Illegal type {} of 1 argument of function {} when arguments size is 2 or 3. Should be a integer or a date with time", + "Illegal type {} of first argument of function {} when arguments size is 2 or 3. Should be a integer or a date with time", arguments[0].type->getName(), getName()); } else @@ -337,18 +337,18 @@ public: getName(), arguments.size()); if (!isDate(arguments[0].type) && !isDateTime(arguments[0].type) && !isDate32(arguments[0].type) && !isDateTime64(arguments[0].type)) throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, - "Illegal type {} of 1 argument of function {}. Should be a date or a date with time", + "Illegal type {} of first argument of function {}. Should be a date or a date with time", arguments[0].type->getName(), getName()); } if (arguments.size() == 2 && !WhichDataType(arguments[1].type).isString()) throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, - "Illegal type {} of 2 argument of function {}. Must be String.", + "Illegal type {} of second argument of function {}. Must be String.", arguments[1].type->getName(), getName()); if (arguments.size() == 3 && !WhichDataType(arguments[2].type).isString()) throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, - "Illegal type {} of 3 argument of function {}. Must be String.", + "Illegal type {} of third argument of function {}. Must be String.", arguments[2].type->getName(), getName()); if (arguments.size() == 1) From fc16752844eb03adfbcb5c4241a6109243ce4e98 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Fri, 28 Oct 2022 12:56:52 +0000 Subject: [PATCH 154/239] Update docs for FROM_UNIXTIME --- docs/en/sql-reference/functions/date-time-functions.md | 4 ++-- docs/ru/sql-reference/functions/date-time-functions.md | 5 ++--- 2 files changed, 4 insertions(+), 5 deletions(-) diff --git a/docs/en/sql-reference/functions/date-time-functions.md b/docs/en/sql-reference/functions/date-time-functions.md index 15fc9ef0c89..f7ea2690b21 100644 --- a/docs/en/sql-reference/functions/date-time-functions.md +++ b/docs/en/sql-reference/functions/date-time-functions.md @@ -1068,7 +1068,7 @@ Example: SELECT timeSlots(toDateTime('2012-01-01 12:20:00'), toUInt32(600)); SELECT timeSlots(toDateTime('1980-12-12 21:01:02', 'UTC'), toUInt32(600), 299); SELECT timeSlots(toDateTime64('1980-12-12 21:01:02.1234', 4, 'UTC'), toDecimal64(600.1, 1), toDecimal64(299, 0)); -``` +``` ``` text ┌─timeSlots(toDateTime('2012-01-01 12:20:00'), toUInt32(600))─┐ │ ['2012-01-01 12:00:00','2012-01-01 12:30:00'] │ @@ -1244,7 +1244,7 @@ Result: └──────────────────────────┘ ``` -When there are two arguments: first is an [Integer](../../sql-reference/data-types/int-uint.md) or [DateTime](../../sql-reference/data-types/datetime.md), second is a constant format string — it acts in the same way as [formatDateTime](#formatdatetime) and return [String](../../sql-reference/data-types/string.md#string) type. +When there are two or three arguments, the first an [Integer](../../sql-reference/data-types/int-uint.md), [Date](../../sql-reference/data-types/date.md), [Date32](../../sql-reference/data-types/date32.md), [DateTime](../../sql-reference/data-types/datetime.md) or [DateTime64](../../sql-reference/data-types/datetime64.md), the second a constant format string and the third an optional constant time zone string — it acts in the same way as [formatDateTime](#formatdatetime) and return [String](../../sql-reference/data-types/string.md#string) type. For example: diff --git a/docs/ru/sql-reference/functions/date-time-functions.md b/docs/ru/sql-reference/functions/date-time-functions.md index 80e2561a8d7..4acd706b3db 100644 --- a/docs/ru/sql-reference/functions/date-time-functions.md +++ b/docs/ru/sql-reference/functions/date-time-functions.md @@ -318,7 +318,7 @@ SELECT toStartOfISOYear(toDate('2017-01-01')) AS ISOYear20170101; :::note "Attention" Возвращаемое значение для некорректных дат зависит от реализации. ClickHouse может вернуть нулевую дату, выбросить исключение, или выполнить «естественное» перетекание дат между месяцами. ::: - + ## toMonday {#tomonday} Округляет дату или дату-с-временем вниз до ближайшего понедельника. @@ -1126,8 +1126,7 @@ SELECT FROM_UNIXTIME(423543535); └──────────────────────────┘ ``` -В случае, когда есть два аргумента: первый типа [Integer](../../sql-reference/data-types/int-uint.md) или [DateTime](../../sql-reference/data-types/datetime.md), а второй является строкой постоянного формата — функция работает также, как [formatDateTime](#formatdatetime), и возвращает значение типа [String](../../sql-reference/data-types/string.md#string). - +В случае, когда есть два или три аргумента: первый типа [Integer](../../sql-reference/data-types/int-uint.md), [Date](../../sql-reference/data-types/date.md), [Date32](../../sql-reference/data-types/date32.md), [DateTime](../../sql-reference/data-types/datetime.md) или [DateTime64](../../sql-reference/data-types/datetime64.md), а второй является строкой постоянного формата и третий является строкой постоянной временной зоны — функция работает также, как [formatDateTime](#formatdatetime), и возвращает значение типа [String](../../sql-reference/data-types/string.md#string). Запрос: From 5755728b33543d472a00c71dc1789d0532d00785 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Fri, 28 Oct 2022 13:11:09 +0000 Subject: [PATCH 155/239] Add Date32 and DateTime64 into exception messages --- src/Functions/formatDateTime.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Functions/formatDateTime.cpp b/src/Functions/formatDateTime.cpp index 9634768e68a..4db04d61d84 100644 --- a/src/Functions/formatDateTime.cpp +++ b/src/Functions/formatDateTime.cpp @@ -371,7 +371,7 @@ public: })) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Illegal column {} of function {}, must be Integer or DateTime when arguments size is 1.", + "Illegal column {} of function {}, must be Integer, Date, Date32, DateTime or DateTime64 when arguments size is 1.", arguments[0].column->getName(), getName()); } } @@ -382,7 +382,7 @@ public: using FromDataType = std::decay_t; if (!(res = executeType(arguments, result_type))) throw Exception(ErrorCodes::ILLEGAL_COLUMN, - "Illegal column {} of function {}, must be Integer or DateTime.", + "Illegal column {} of function {}, must be Integer, Date, Date32, DateTime or DateTime64.", arguments[0].column->getName(), getName()); return true; })) From 7a5432feaa39596be26974ef1d84d6e954deafc1 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Fri, 28 Oct 2022 15:10:53 +0200 Subject: [PATCH 156/239] Fix possible SIGSEGV for web disks when file does not exists It can be triggered in multiple ways, either when file does not exists and you are trying to create MergeTree table from web (that has special code for UUID handling) or by simply OPTIMIZE TABLE FINAL for MergeTree table that is located on web disk, in both cases you will get the following:
stacktrace 2022.10.28 14:08:40.631226 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} ContextAccess (default): Access granted: OPTIMIZE ON default.data_from_web 2022.10.28 14:08:40.632017 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} default.data_from_web (a3e65e1f-5fd4-47ed-9dbd-307f2586b52d) (MergerMutator): Selected 1 parts from all_1_1_0 to all_1_1_0 2022.10.28 14:08:40.632496 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} default.data_from_web (a3e65e1f-5fd4-47ed-9dbd-307f2586b52d): Trying to reserve 1.00 MiB using storage policy from min volume index 0 2022.10.28 14:08:40.632752 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} DiskObjectStorage(DiskWebServer): Reserved 1.00 MiB on remote disk `web_disk`, having unreserved 16.00 EiB. 2022.10.28 14:08:40.634155 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} MergeTask::PrepareStage: Merging 1 parts: from all_1_1_0 to all_1_1_0 into Compact 2022.10.28 14:08:40.634498 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} WebObjectStorage: Loading metadata for directory: http://127.0.0.1:8080/store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1 2022.10.28 14:08:40.635025 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} DiskWeb: Adding directory: store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/ 2022.10.28 14:08:40.635355 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} ReadWriteBufferFromHTTP: Sending request to http://127.0.0.1:8080/store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/.index 2022.10.28 14:08:40.639618 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} DiskWeb: Cannot load disk metadata. Error: Code: 86. DB::Exception: Received error from remote server /store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/.index. HTTP status code: 404 Not Found, body: 404 Not Found

Not Found

The requested resource /store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/.index was not found on this server.

: while loading disk metadata. (RECEIVED_ERROR_FROM_REMOTE_IO_SERVER) (version 22.11.1.1) 2022.10.28 14:08:40.640527 [ 5488 ] {} BaseDaemon: Received signal 11 2022.10.28 14:08:40.641529 [ 9027 ] {} BaseDaemon: ######################################## 2022.10.28 14:08:40.642759 [ 9027 ] {} BaseDaemon: (version 22.11.1.1, build id: 12145DA78CE5E9EBB10A034177FAE5967EF81A4A) (from thread 6043) (query_id: a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1) (query: optimize table data_from_web final) Received signal Segmentation fault (11) 2022.10.28 14:08:40.643260 [ 9027 ] {} BaseDaemon: Address: NULL pointer. Access: read. Unknown si_code. 2022.10.28 14:08:40.643769 [ 9027 ] {} BaseDaemon: Stack trace: 0x7ffff416c0f2 0x7ffff7cd1ca8 0x7ffff679ae5e 0x7fffd52e7906 0x7fffd50c65aa 0x7fffca7a0d42 0x7fffcaee79ec 0x7fffcaf242f8 0x7fffcaf242b5 0x7fffcaf2427d 0x7fffcaf24255 0x7fffcaf2421d 0x7ffff65c3686 0x7ffff65c2295 0x7fffcaeee2a9 0x7fffcaef2c43 0x7fffcaee3c0e 0x7fffcc4a7851 0x7fffcc4a768f 0x7fffcc4abb2d 0x7fffcfdce828 0x7fffd03e3eaa 0x7fffd03dfe3b 0x7fffc8ec42d4 0x7fffc8ed51d2 0x7ffff4bdd839 0x7ffff4bde0a8 0x7ffff48ab261 0x7ffff48a769a 0x7ffff48a6335 0x7ffff409f8fd 0x7ffff4121a60 2022.10.28 14:08:40.644411 [ 9027 ] {} BaseDaemon: 4. ? @ 0x7ffff416c0f2 in ? 2022.10.28 14:08:40.676390 [ 9027 ] {} BaseDaemon: 5. /src/ch/clickhouse/src/Common/StringUtils/StringUtils.cpp:9: detail::startsWith(std::__1::basic_string, std::__1::allocator> const&, char const*, unsigned long) @ 0x1ca8 in /src/ch/clickhouse/.cmake/src/Common/StringUtils/libstring_utilsd.so 2022.10.28 14:08:40.730727 [ 9027 ] {} BaseDaemon: 6. /src/ch/clickhouse/src/Common/StringUtils/StringUtils.h:19: startsWith(std::__1::basic_string, std::__1::allocator> const&, std::__1::basic_string, std::__1::allocator> const&) @ 0x59ae5e in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so 2022.10.28 14:08:40.923955 [ 9027 ] {} BaseDaemon: 7. /src/ch/clickhouse/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp:58: DB::MetadataStorageFromStaticFilesWebServer::exists(std::__1::basic_string, std::__1::allocator> const&) const @ 0x6e7906 in /src/ch/clickhouse/.cmake/src/libdbmsd.so 2022.10.28 14:08:41.291996 [ 9027 ] {} BaseDaemon: 8. /src/ch/clickhouse/src/Disks/ObjectStorages/DiskObjectStorage.cpp:181: DB::DiskObjectStorage::exists(std::__1::basic_string, std::__1::allocator> const&) const @ 0x4c65aa in /src/ch/clickhouse/.cmake/src/libdbmsd.so 2022.10.28 14:08:41.704697 [ 9027 ] {} BaseDaemon: 9. /src/ch/clickhouse/src/Storages/MergeTree/DataPartStorageOnDisk.cpp:74: DB::DataPartStorageOnDisk::exists() const @ 0xda0d42 in /src/ch/clickhouse/.cmake/src/libclickhouse_storages_mergetreed.so 2022.10.28 14:08:43.032459 [ 9027 ] {} BaseDaemon: 10. /src/ch/clickhouse/src/Storages/MergeTree/MergeTask.cpp:147: DB::MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() @ 0x14e79ec in /src/ch/clickhouse/.cmake/src/libclickhouse_storages_mergetreed.so ... Segmentation fault (core dumped)
Signed-off-by: Azat Khuzhin --- .../Web/MetadataStorageFromStaticFilesWebServer.cpp | 3 +++ tests/integration/test_disk_over_web_server/test.py | 3 +++ 2 files changed, 6 insertions(+) diff --git a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp index 06e36a2ddd8..aa125e93dee 100644 --- a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp +++ b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp @@ -55,6 +55,9 @@ bool MetadataStorageFromStaticFilesWebServer::exists(const std::string & path) c path, [](const auto & file, const std::string & path_) { return file.first < path_; } ); + if (it == object_storage.files.end()) + return false; + if (startsWith(it->first, path) || (it != object_storage.files.begin() && startsWith(std::prev(it)->first, path))) return true; diff --git a/tests/integration/test_disk_over_web_server/test.py b/tests/integration/test_disk_over_web_server/test.py index ea6e407a18f..2ccc17db4f4 100644 --- a/tests/integration/test_disk_over_web_server/test.py +++ b/tests/integration/test_disk_over_web_server/test.py @@ -129,6 +129,9 @@ def test_incorrect_usage(cluster): result = node2.query_and_get_error("TRUNCATE TABLE test0") assert "Table is read-only" in result + result = node2.query_and_get_error("OPTIMIZE TABLE test0 FINAL") + assert "Only read-only operations are supported" in result + node2.query("DROP TABLE test0 SYNC") From 665fcf55aa66c0a19d8c7e4c8b9d7fe6de852aee Mon Sep 17 00:00:00 2001 From: Vitalii S Date: Fri, 28 Oct 2022 10:08:04 -0400 Subject: [PATCH 157/239] Update database.md Minor text correction --- docs/en/sql-reference/statements/create/database.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/create/database.md b/docs/en/sql-reference/statements/create/database.md index 432f5975cc8..7954d1362f1 100644 --- a/docs/en/sql-reference/statements/create/database.md +++ b/docs/en/sql-reference/statements/create/database.md @@ -31,7 +31,7 @@ By default, ClickHouse uses its own [Atomic](../../../engines/database-engines/a ### COMMENT -You can add a comment to the database when you creating it. +You can add a comment to the database when you are creating it. The comment is supported for all database engines. From 4a4fb20731019f2b1a66c53b9e1b89bc823e93a9 Mon Sep 17 00:00:00 2001 From: Roman Vasin Date: Fri, 28 Oct 2022 14:08:28 +0000 Subject: [PATCH 158/239] Add Date32 to 01411_from_unixtime and improve 00718_format_datetime --- .../00718_format_datetime.reference | 49 +++++++++-------- .../0_stateless/00718_format_datetime.sql | 54 ++++++++++--------- .../0_stateless/01411_from_unixtime.reference | 34 ++++++------ .../0_stateless/01411_from_unixtime.sql | 34 ++++++------ 4 files changed, 86 insertions(+), 85 deletions(-) diff --git a/tests/queries/0_stateless/00718_format_datetime.reference b/tests/queries/0_stateless/00718_format_datetime.reference index 6a0fd451e73..bc98dd59d5f 100644 --- a/tests/queries/0_stateless/00718_format_datetime.reference +++ b/tests/queries/0_stateless/00718_format_datetime.reference @@ -1,33 +1,32 @@ -20 +20 20 +02 02 +01/02/18 01/02/18 + 2 2 +2018-01-02 2018-01-02 +22 00 02 -01/02/18 - 2 -2018-01-02 -22 -02 -10 +10 12 11 12 -001 -366 -01 -33 -\n -AM +001 001 +366 366 +01 01 +33 00 +\n \n +AM AM AM PM -22:33 -44 -\t -22:33:44 -1 7 -01 01 53 52 -1 0 -18 -2018 -% -no formatting pattern -2018-01-01 00:00:00 +22:33 00:00 +44 00 +\t \t +22:33:44 00:00:00 +1 7 1 7 +01 01 53 52 01 01 53 52 +1 0 1 0 +18 18 +2018 2018 +% % +no formatting pattern no formatting pattern 2018-01-01 00:00:00 1927-01-01 00:00:00 2018-01-01 01:00:00 2018-01-01 04:00:00 diff --git a/tests/queries/0_stateless/00718_format_datetime.sql b/tests/queries/0_stateless/00718_format_datetime.sql index 532a0c02e6f..deb5fb96c6c 100644 --- a/tests/queries/0_stateless/00718_format_datetime.sql +++ b/tests/queries/0_stateless/00718_format_datetime.sql @@ -8,40 +8,42 @@ SELECT formatDateTime(now(), 'unescaped %'); -- { serverError 36 } SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%U'); -- { serverError 48 } SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%W'); -- { serverError 48 } -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%C'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%d'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%D'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%e'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%F'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%H'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%C'), formatDateTime(toDate32('2018-01-02'), '%C'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%d'), formatDateTime(toDate32('2018-01-02'), '%d'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%D'), formatDateTime(toDate32('2018-01-02'), '%D'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%e'), formatDateTime(toDate32('2018-01-02'), '%e'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%F'), formatDateTime(toDate32('2018-01-02'), '%F'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%H'), formatDateTime(toDate32('2018-01-02'), '%H'); SELECT formatDateTime(toDateTime('2018-01-02 02:33:44'), '%H'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%I'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%I'), formatDateTime(toDate32('2018-01-02'), '%I'); SELECT formatDateTime(toDateTime('2018-01-02 11:33:44'), '%I'); SELECT formatDateTime(toDateTime('2018-01-02 00:33:44'), '%I'); -SELECT formatDateTime(toDateTime('2018-01-01 00:33:44'), '%j'); -SELECT formatDateTime(toDateTime('2000-12-31 00:33:44'), '%j'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%m'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%M'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%n'); -SELECT formatDateTime(toDateTime('2018-01-02 00:33:44'), '%p'); +SELECT formatDateTime(toDateTime('2018-01-01 00:33:44'), '%j'), formatDateTime(toDate32('2018-01-01'), '%j'); +SELECT formatDateTime(toDateTime('2000-12-31 00:33:44'), '%j'), formatDateTime(toDate32('2000-12-31'), '%j'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%m'), formatDateTime(toDate32('2018-01-02'), '%m'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%M'), formatDateTime(toDate32('2018-01-02'), '%M'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%n'), formatDateTime(toDate32('2018-01-02'), '%n'); +SELECT formatDateTime(toDateTime('2018-01-02 00:33:44'), '%p'), formatDateTime(toDateTime('2018-01-02'), '%p'); SELECT formatDateTime(toDateTime('2018-01-02 11:33:44'), '%p'); SELECT formatDateTime(toDateTime('2018-01-02 12:33:44'), '%p'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%R'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%S'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%t'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%T'); -SELECT formatDateTime(toDateTime('2018-01-01 22:33:44'), '%u'), formatDateTime(toDateTime('2018-01-07 22:33:44'), '%u'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%R'), formatDateTime(toDate32('2018-01-02'), '%R'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%S'), formatDateTime(toDate32('2018-01-02'), '%S'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%t'), formatDateTime(toDate32('2018-01-02'), '%t'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%T'), formatDateTime(toDate32('2018-01-02'), '%T'); +SELECT formatDateTime(toDateTime('2018-01-01 22:33:44'), '%u'), formatDateTime(toDateTime('2018-01-07 22:33:44'), '%u'), + formatDateTime(toDate32('2018-01-01'), '%u'), formatDateTime(toDate32('2018-01-07'), '%u'); SELECT formatDateTime(toDateTime('1996-01-01 22:33:44'), '%V'), formatDateTime(toDateTime('1996-12-31 22:33:44'), '%V'), - formatDateTime(toDateTime('1999-01-01 22:33:44'), '%V'), formatDateTime(toDateTime('1999-12-31 22:33:44'), '%V'); -SELECT formatDateTime(toDateTime('2018-01-01 22:33:44'), '%w'), formatDateTime(toDateTime('2018-01-07 22:33:44'), '%w'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%y'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%Y'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%%'); -SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), 'no formatting pattern'); + formatDateTime(toDateTime('1999-01-01 22:33:44'), '%V'), formatDateTime(toDateTime('1999-12-31 22:33:44'), '%V'), + formatDateTime(toDate32('1996-01-01'), '%V'), formatDateTime(toDate32('1996-12-31'), '%V'), + formatDateTime(toDate32('1999-01-01'), '%V'), formatDateTime(toDate32('1999-12-31'), '%V'); +SELECT formatDateTime(toDateTime('2018-01-01 22:33:44'), '%w'), formatDateTime(toDateTime('2018-01-07 22:33:44'), '%w'), + formatDateTime(toDate32('2018-01-01'), '%w'), formatDateTime(toDate32('2018-01-07'), '%w'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%y'), formatDateTime(toDate32('2018-01-02'), '%y'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%Y'), formatDateTime(toDate32('2018-01-02'), '%Y'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), '%%'), formatDateTime(toDate32('2018-01-02'), '%%'); +SELECT formatDateTime(toDateTime('2018-01-02 22:33:44'), 'no formatting pattern'), formatDateTime(toDate32('2018-01-02'), 'no formatting pattern'); SELECT formatDateTime(toDate('2018-01-01'), '%F %T'); - -SELECT formatDateTime(toDate32('2018-01-01'), '%F %T'); SELECT formatDateTime(toDate32('1927-01-01'), '%F %T'); SELECT diff --git a/tests/queries/0_stateless/01411_from_unixtime.reference b/tests/queries/0_stateless/01411_from_unixtime.reference index 1bc7519e668..17086e8c58b 100644 --- a/tests/queries/0_stateless/01411_from_unixtime.reference +++ b/tests/queries/0_stateless/01411_from_unixtime.reference @@ -5,25 +5,25 @@ 11 1970-01-15 1970-01-15 06:52:36 -20 +20 20 +02 02 +01/02/18 01/02/18 + 2 2 +2018-01-02 2018-01-02 +22 00 02 -01/02/18 - 2 -2018-01-02 -22 -02 -10 +10 12 11 12 -001 -366 -01 -33 -\n -AM +001 001 +366 366 +01 01 +33 00 +\n \n +AM AM AM PM -22:33 -44 -\t -22:33:44 +22:33 00:00 +44 00 +\t \t +22:33:44 00:00:00 diff --git a/tests/queries/0_stateless/01411_from_unixtime.sql b/tests/queries/0_stateless/01411_from_unixtime.sql index ec7b4d65b57..9a6655768e0 100644 --- a/tests/queries/0_stateless/01411_from_unixtime.sql +++ b/tests/queries/0_stateless/01411_from_unixtime.sql @@ -5,25 +5,25 @@ SELECT FROM_UNIXTIME(5345345, '%C', 'UTC'); SELECT FROM_UNIXTIME(645123, '%H', 'UTC'); SELECT FROM_UNIXTIME(1232456, '%Y-%m-%d', 'UTC'); SELECT FROM_UNIXTIME(1234356, '%Y-%m-%d %R:%S', 'UTC'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%C'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%d'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%D'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%e'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%F'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%H'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%C'), FROM_UNIXTIME(toDate32('2018-01-02'), '%C'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%d'), FROM_UNIXTIME(toDate32('2018-01-02'), '%d'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%D'), FROM_UNIXTIME(toDate32('2018-01-02'), '%D'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%e'), FROM_UNIXTIME(toDate32('2018-01-02'), '%e'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%F'), FROM_UNIXTIME(toDate32('2018-01-02'), '%F'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%H'), FROM_UNIXTIME(toDate32('2018-01-02'), '%H'); SELECT FROM_UNIXTIME(toDateTime('2018-01-02 02:33:44'), '%H'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%I'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%I'), FROM_UNIXTIME(toDate32('2018-01-02'), '%I'); SELECT FROM_UNIXTIME(toDateTime('2018-01-02 11:33:44'), '%I'); SELECT FROM_UNIXTIME(toDateTime('2018-01-02 00:33:44'), '%I'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-01 00:33:44'), '%j'); -SELECT FROM_UNIXTIME(toDateTime('2000-12-31 00:33:44'), '%j'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%m'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%M'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%n'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 00:33:44'), '%p'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-01 00:33:44'), '%j'), FROM_UNIXTIME(toDate32('2018-01-01'), '%j'); +SELECT FROM_UNIXTIME(toDateTime('2000-12-31 00:33:44'), '%j'), FROM_UNIXTIME(toDate32('2000-12-31'), '%j'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%m'), FROM_UNIXTIME(toDate32('2018-01-02'), '%m'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%M'), FROM_UNIXTIME(toDate32('2018-01-02'), '%M'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%n'), FROM_UNIXTIME(toDate32('2018-01-02'), '%n'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 00:33:44'), '%p'), FROM_UNIXTIME(toDate32('2018-01-02'), '%p'); SELECT FROM_UNIXTIME(toDateTime('2018-01-02 11:33:44'), '%p'); SELECT FROM_UNIXTIME(toDateTime('2018-01-02 12:33:44'), '%p'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%R'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%S'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%t'); -SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%T'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%R'), FROM_UNIXTIME(toDate32('2018-01-02'), '%R'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%S'), FROM_UNIXTIME(toDate32('2018-01-02'), '%S'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%t'), FROM_UNIXTIME(toDate32('2018-01-02'), '%t'); +SELECT FROM_UNIXTIME(toDateTime('2018-01-02 22:33:44'), '%T'), FROM_UNIXTIME(toDate32('2018-01-02'), '%T'); From d4c62688069b5d6baad52d328702973860e17ec1 Mon Sep 17 00:00:00 2001 From: UnamedRus Date: Fri, 28 Oct 2022 18:47:32 +0400 Subject: [PATCH 159/239] Fix Polygon dict xml config Missing tag --- .../external-dictionaries/external-dicts-dict-polygon.md | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-polygon.md b/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-polygon.md index 912af5b5bce..e5ee48c9166 100644 --- a/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-polygon.md +++ b/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-polygon.md @@ -14,8 +14,10 @@ Example of a polygon dictionary configuration: - key - Array(Array(Array(Array(Float64)))) + + key + Array(Array(Array(Array(Float64)))) + From 0872d5c440ada1cb9966f8e69b832b2c8eb1a026 Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Fri, 28 Oct 2022 13:33:12 -0400 Subject: [PATCH 160/239] style Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 834f74822a1..eb6e131886c 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -141,7 +141,7 @@ jobs: - name: Download and set up sonar-scanner env: SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip - run: | + "run": | mkdir -p $HOME/.sonar curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ @@ -149,7 +149,7 @@ jobs: - name: Download and set up build-wrapper env: BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip - run: | + "run": | curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH From 144648a633e90e5f40b196bb4af668afe74c4e2d Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Fri, 28 Oct 2022 13:47:37 -0400 Subject: [PATCH 161/239] disable sc2086 Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index eb6e131886c..527e3b1f2af 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -141,7 +141,8 @@ jobs: - name: Download and set up sonar-scanner env: SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip - "run": | + # shellcheck disable=SC2086 + run: | mkdir -p $HOME/.sonar curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ @@ -149,7 +150,8 @@ jobs: - name: Download and set up build-wrapper env: BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip - "run": | + # shellcheck disable=SC2086 + run: | curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH From 10bee23c584a2effb509e71d4c7aa09b781e853d Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Fri, 28 Oct 2022 13:56:12 -0400 Subject: [PATCH 162/239] no newline at end of file. Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 527e3b1f2af..1fa6603c960 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -176,4 +176,4 @@ jobs: --define sonar.host.url="${{ env.SONAR_SERVER_URL }}" \ --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ --define sonar.projectKey="ClickHouse_ClickHouse" \ - --define sonar.organization="clickhouse-java" \ No newline at end of file + --define sonar.organization="clickhouse-java" From dc10c7a446d2d18823fe3e1c07d8996aaeb99787 Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Fri, 28 Oct 2022 13:57:47 -0400 Subject: [PATCH 163/239] no newline at end of file. Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 3 --- 1 file changed, 3 deletions(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 1fa6603c960..0bada9eb913 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -123,7 +123,6 @@ jobs: docker ps --all --quiet | xargs --no-run-if-empty docker rm -f ||: sudo rm -fr "$TEMP_PATH" "$CACHES_PATH" SonarCloud: - name: Sonar Cloud runs-on: [self-hosted, builder] env: SONAR_SCANNER_VERSION: 4.7.0.2747 @@ -141,7 +140,6 @@ jobs: - name: Download and set up sonar-scanner env: SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip - # shellcheck disable=SC2086 run: | mkdir -p $HOME/.sonar curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} @@ -150,7 +148,6 @@ jobs: - name: Download and set up build-wrapper env: BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip - # shellcheck disable=SC2086 run: | curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ From 51f2d6046c1b40edc6d07d3b63abe2d5d4e1697f Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Fri, 28 Oct 2022 14:10:23 -0400 Subject: [PATCH 164/239] sc2086 Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 0bada9eb913..81878ccd16f 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -142,14 +142,14 @@ jobs: SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip run: | mkdir -p $HOME/.sonar - curl -sSLo $HOME/.sonar/sonar-scanner.zip ${{ env.SONAR_SCANNER_DOWNLOAD_URL }} + curl -sSLo $HOME/.sonar/sonar-scanner.zip "${{ env.SONAR_SCANNER_DOWNLOAD_URL }}" unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> $GITHUB_PATH - name: Download and set up build-wrapper env: BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip run: | - curl -sSLo $HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }} + curl -sSLo "$HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }}" unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH - name: Set Up Build Tools From 53acc23856f83727b53eda5c39e979038f80c52e Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Fri, 28 Oct 2022 14:17:29 -0400 Subject: [PATCH 165/239] sc2086 Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 81878ccd16f..e6da4df7200 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -141,17 +141,17 @@ jobs: env: SONAR_SCANNER_DOWNLOAD_URL: https://binaries.sonarsource.com/Distribution/sonar-scanner-cli/sonar-scanner-cli-${{ env.SONAR_SCANNER_VERSION }}-linux.zip run: | - mkdir -p $HOME/.sonar - curl -sSLo $HOME/.sonar/sonar-scanner.zip "${{ env.SONAR_SCANNER_DOWNLOAD_URL }}" - unzip -o $HOME/.sonar/sonar-scanner.zip -d $HOME/.sonar/ - echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> $GITHUB_PATH + mkdir -p "$HOME/.sonar" + curl -sSLo "$HOME/.sonar/sonar-scanner.zip" "${{ env.SONAR_SCANNER_DOWNLOAD_URL }}" + unzip -o "$HOME/.sonar/sonar-scanner.zip" -d "$HOME/.sonar/" + echo "$HOME/.sonar/sonar-scanner-${{ env.SONAR_SCANNER_VERSION }}-linux/bin" >> "$GITHUB_PATH" - name: Download and set up build-wrapper env: BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip run: | curl -sSLo "$HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }}" - unzip -o $HOME/.sonar/build-wrapper-linux-x86.zip -d $HOME/.sonar/ - echo "$HOME/.sonar/build-wrapper-linux-x86" >> $GITHUB_PATH + unzip -o "$HOME/.sonar/build-wrapper-linux-x86.zip" -d "$HOME/.sonar/" + echo "$HOME/.sonar/build-wrapper-linux-x86" >> "$GITHUB_PATH" - name: Set Up Build Tools run: | sudo apt-get update From e903efda0a71c9d47d0217d6cbf4e38fa0125a16 Mon Sep 17 00:00:00 2001 From: Denny Crane Date: Fri, 28 Oct 2022 16:00:37 -0300 Subject: [PATCH 166/239] Update column.md --- docs/en/sql-reference/statements/alter/column.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/alter/column.md b/docs/en/sql-reference/statements/alter/column.md index 067a350dca7..4c0b79dd8f7 100644 --- a/docs/en/sql-reference/statements/alter/column.md +++ b/docs/en/sql-reference/statements/alter/column.md @@ -204,8 +204,9 @@ It is used if it is necessary to add or update a column with a complicated expre Syntax: ```sql -ALTER TABLE table MATERIALIZE COLUMN col; +ALTER TABLE table MATERIALIZE COLUMN col [IN PARTITION partition | IN PARTITION ID 'partition_id']; ``` +- If you specify a PARTITION, a column will be materialized with only the specified partition. **Example** From 7fa71d8c04e20436e232e1f8c55f6fb46a10f279 Mon Sep 17 00:00:00 2001 From: Denny Crane Date: Fri, 28 Oct 2022 16:04:00 -0300 Subject: [PATCH 167/239] Update column.md --- docs/en/sql-reference/statements/alter/column.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/alter/column.md b/docs/en/sql-reference/statements/alter/column.md index 4c0b79dd8f7..f36aa1357f4 100644 --- a/docs/en/sql-reference/statements/alter/column.md +++ b/docs/en/sql-reference/statements/alter/column.md @@ -204,7 +204,7 @@ It is used if it is necessary to add or update a column with a complicated expre Syntax: ```sql -ALTER TABLE table MATERIALIZE COLUMN col [IN PARTITION partition | IN PARTITION ID 'partition_id']; +ALTER TABLE [db.]table [ON CLUSTER cluster] MATERIALIZE COLUMN col [IN PARTITION partition | IN PARTITION ID 'partition_id']; ``` - If you specify a PARTITION, a column will be materialized with only the specified partition. From 004ca99712c0037f4fc3e162eb451775965dfaba Mon Sep 17 00:00:00 2001 From: Yuko Takagi <70714860+yukotakagi@users.noreply.github.com> Date: Fri, 28 Oct 2022 13:58:14 -0600 Subject: [PATCH 168/239] Update README.md --- README.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 003b78a3cbb..f90df9686c2 100644 --- a/README.md +++ b/README.md @@ -5,7 +5,7 @@ ClickHouse® is an open-source column-oriented database management system that a ## Useful Links * [Official website](https://clickhouse.com/) has a quick high-level overview of ClickHouse on the main page. -* [ClickHouse Cloud](https://clickhouse.com/cloud) ClickHouse as a service, built by the creators and maintainers. +* [ClickHouse Cloud](https://clickhouse.cloud) ClickHouse as a service, built by the creators and maintainers. * [Tutorial](https://clickhouse.com/docs/en/getting_started/tutorial/) shows how to set up and query a small ClickHouse cluster. * [Documentation](https://clickhouse.com/docs/en/) provides more in-depth information. * [YouTube channel](https://www.youtube.com/c/ClickHouseDB) has a lot of content about ClickHouse in video format. @@ -16,5 +16,6 @@ ClickHouse® is an open-source column-oriented database management system that a * [Contacts](https://clickhouse.com/company/contact) can help to get your questions answered if there are any. ## Upcoming events -* [**v22.10 Release Webinar**](https://clickhouse.com/company/events/v22-10-release-webinar) Original creator, co-founder, and CTO of ClickHouse Alexey Milovidov will walk us through the highlights of the release, provide live demos, and share vision into what is coming in the roadmap. -* [**Introducing ClickHouse Cloud**](https://clickhouse.com/company/events/cloud-beta) Introducing ClickHouse as a service, built by creators and maintainers of the fastest OLAP database on earth. Join Tanya Bragin for a detailed walkthrough of ClickHouse Cloud capabilities, as well as a peek behind the curtain to understand the unique architecture that makes our service tick. +* [**v22.11 Release Webinar**](https://clickhouse.com/company/events/v22-11-release-webinar) Original creator, co-founder, and CTO of ClickHouse Alexey Milovidov will walk us through the highlights of the release, provide live demos, and share vision into what is coming in the roadmap. +* [**ClickHouse Meetup at the Deutsche Bank office in Berlin**](https://www.meetup.com/clickhouse-berlin-user-group/events/289311596/) Hear from Deutsche Bank on why they chose ClickHouse for big sensitive data in a regulated environment. The ClickHouse team will then present how ClickHouse is used for real time financial data analytics, including tick data, trade analytics and risk management. +* [**AWS re:Invent**](https://clickhouse.com/company/events/aws-reinvent) Core members of the ClickHouse team -- including 2 of our founders -- will be at re:Invent from November 29 to December 3. We are available on the show floor, but are also determining interest in holding an event during the time there. From fe0aea2e3a2011ed803894c8e59c83db357e9e26 Mon Sep 17 00:00:00 2001 From: avogar Date: Fri, 28 Oct 2022 18:44:36 +0000 Subject: [PATCH 169/239] Support parallel parsing for LineAsString input format --- src/Formats/newLineSegmentationEngine.cpp | 50 +++++++++++++++++++ src/Formats/newLineSegmentationEngine.h | 9 ++++ src/Formats/registerFormats.cpp | 2 + .../Impl/LineAsStringRowInputFormat.cpp | 8 ++- .../Formats/Impl/RegexpRowInputFormat.cpp | 40 +-------------- tests/performance/line_as_string_parsing.xml | 9 ++++ 6 files changed, 79 insertions(+), 39 deletions(-) create mode 100644 src/Formats/newLineSegmentationEngine.cpp create mode 100644 src/Formats/newLineSegmentationEngine.h create mode 100644 tests/performance/line_as_string_parsing.xml diff --git a/src/Formats/newLineSegmentationEngine.cpp b/src/Formats/newLineSegmentationEngine.cpp new file mode 100644 index 00000000000..a605bba7e5b --- /dev/null +++ b/src/Formats/newLineSegmentationEngine.cpp @@ -0,0 +1,50 @@ +#include +#include +#include + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int LOGICAL_ERROR; +} + +std::pair newLineFileSegmentationEngine(ReadBuffer & in, DB::Memory<> & memory, size_t min_bytes, size_t max_rows) +{ + char * pos = in.position(); + bool need_more_data = true; + size_t number_of_rows = 0; + + while (loadAtPosition(in, memory, pos) && need_more_data) + { + pos = find_first_symbols<'\r', '\n'>(pos, in.buffer().end()); + if (pos > in.buffer().end()) + throw Exception("Position in buffer is out of bounds. There must be a bug.", ErrorCodes::LOGICAL_ERROR); + else if (pos == in.buffer().end()) + continue; + + ++number_of_rows; + if ((memory.size() + static_cast(pos - in.position()) >= min_bytes) || (number_of_rows == max_rows)) + need_more_data = false; + + if (*pos == '\n') + { + ++pos; + if (loadAtPosition(in, memory, pos) && *pos == '\r') + ++pos; + } + else if (*pos == '\r') + { + ++pos; + if (loadAtPosition(in, memory, pos) && *pos == '\n') + ++pos; + } + } + + saveUpToPosition(in, memory, pos); + + return {loadAtPosition(in, memory, pos), number_of_rows}; +} + +} diff --git a/src/Formats/newLineSegmentationEngine.h b/src/Formats/newLineSegmentationEngine.h new file mode 100644 index 00000000000..598f808b798 --- /dev/null +++ b/src/Formats/newLineSegmentationEngine.h @@ -0,0 +1,9 @@ +#pragma once + +#include +#include + +namespace DB +{ + std::pair newLineFileSegmentationEngine(ReadBuffer & in, DB::Memory<> & memory, size_t min_bytes, size_t max_rows); +} diff --git a/src/Formats/registerFormats.cpp b/src/Formats/registerFormats.cpp index 593e4568be1..ba40fe442ab 100644 --- a/src/Formats/registerFormats.cpp +++ b/src/Formats/registerFormats.cpp @@ -18,6 +18,7 @@ void registerFileSegmentationEngineJSONCompactEachRow(FormatFactory & factory); #if USE_HIVE void registerFileSegmentationEngineHiveText(FormatFactory & factory); #endif +void registerFileSegmentationEngineLineAsString(FormatFactory & factory); /// Formats for both input/output. @@ -153,6 +154,7 @@ void registerFormats() #if USE_HIVE registerFileSegmentationEngineHiveText(factory); #endif + registerFileSegmentationEngineLineAsString(factory); registerInputFormatNative(factory); diff --git a/src/Processors/Formats/Impl/LineAsStringRowInputFormat.cpp b/src/Processors/Formats/Impl/LineAsStringRowInputFormat.cpp index 30084804d92..677f8bb28ec 100644 --- a/src/Processors/Formats/Impl/LineAsStringRowInputFormat.cpp +++ b/src/Processors/Formats/Impl/LineAsStringRowInputFormat.cpp @@ -1,5 +1,5 @@ #include -#include +#include #include #include #include @@ -63,6 +63,12 @@ void registerInputFormatLineAsString(FormatFactory & factory) }); } +void registerFileSegmentationEngineLineAsString(FormatFactory & factory) +{ + factory.registerFileSegmentationEngine("LineAsString", &newLineFileSegmentationEngine); +} + + void registerLineAsStringSchemaReader(FormatFactory & factory) { factory.registerExternalSchemaReader("LineAsString", []( diff --git a/src/Processors/Formats/Impl/RegexpRowInputFormat.cpp b/src/Processors/Formats/Impl/RegexpRowInputFormat.cpp index 6eacfe621e1..2ad2ad6f7a3 100644 --- a/src/Processors/Formats/Impl/RegexpRowInputFormat.cpp +++ b/src/Processors/Formats/Impl/RegexpRowInputFormat.cpp @@ -3,6 +3,7 @@ #include #include #include +#include #include namespace DB @@ -178,46 +179,9 @@ void registerInputFormatRegexp(FormatFactory & factory) }); } -static std::pair fileSegmentationEngineRegexpImpl(ReadBuffer & in, DB::Memory<> & memory, size_t min_bytes, size_t max_rows) -{ - char * pos = in.position(); - bool need_more_data = true; - size_t number_of_rows = 0; - - while (loadAtPosition(in, memory, pos) && need_more_data) - { - pos = find_first_symbols<'\r', '\n'>(pos, in.buffer().end()); - if (pos > in.buffer().end()) - throw Exception("Position in buffer is out of bounds. There must be a bug.", ErrorCodes::LOGICAL_ERROR); - else if (pos == in.buffer().end()) - continue; - - ++number_of_rows; - if ((memory.size() + static_cast(pos - in.position()) >= min_bytes) || (number_of_rows == max_rows)) - need_more_data = false; - - if (*pos == '\n') - { - ++pos; - if (loadAtPosition(in, memory, pos) && *pos == '\r') - ++pos; - } - else if (*pos == '\r') - { - ++pos; - if (loadAtPosition(in, memory, pos) && *pos == '\n') - ++pos; - } - } - - saveUpToPosition(in, memory, pos); - - return {loadAtPosition(in, memory, pos), number_of_rows}; -} - void registerFileSegmentationEngineRegexp(FormatFactory & factory) { - factory.registerFileSegmentationEngine("Regexp", &fileSegmentationEngineRegexpImpl); + factory.registerFileSegmentationEngine("Regexp", &newLineFileSegmentationEngine); } void registerRegexpSchemaReader(FormatFactory & factory) diff --git a/tests/performance/line_as_string_parsing.xml b/tests/performance/line_as_string_parsing.xml new file mode 100644 index 00000000000..d9fa1d4fa6e --- /dev/null +++ b/tests/performance/line_as_string_parsing.xml @@ -0,0 +1,9 @@ + + +INSERT INTO FUNCTION file(test_line_as_string.tsv) SELECT randomString(1000) FROM numbers(1000000) SETTINGS engine_file_truncate_on_insert=1 + +SELECT * FROM file(test_line_as_string.tsv, LineAsString) FORMAT Null + +INSERT INTO FUNCTION file(test_line_as_string.tsv) SELECT * FROM numbers(0) SETTINGS engine_file_truncate_on_insert=1 + + From ca5dbe88dbbde7bc139b6bfc864a8a1b52167ca7 Mon Sep 17 00:00:00 2001 From: clarkcaoliu Date: Wed, 19 Oct 2022 03:18:12 +0800 Subject: [PATCH 170/239] match function can use index if prefix --- src/Storages/MergeTree/KeyCondition.cpp | 105 +++++++++++++++++- .../02462_match_regexp_pk.reference | 5 + .../0_stateless/02462_match_regexp_pk.sql | 9 ++ 3 files changed, 118 insertions(+), 1 deletion(-) create mode 100644 tests/queries/0_stateless/02462_match_regexp_pk.reference create mode 100644 tests/queries/0_stateless/02462_match_regexp_pk.sql diff --git a/src/Storages/MergeTree/KeyCondition.cpp b/src/Storages/MergeTree/KeyCondition.cpp index 927ca98cc9a..8fc5f13d27d 100644 --- a/src/Storages/MergeTree/KeyCondition.cpp +++ b/src/Storages/MergeTree/KeyCondition.cpp @@ -86,6 +86,88 @@ String extractFixedPrefixFromLikePattern(const String & like_pattern) return fixed_prefix; } +/// for "^prefix..." string it returns "prefix" +static String extractFixedPrefixFromRegularExpression(const String & regexp) +{ + if (regexp.size() <= 1 || regexp[0] != '^') + return {}; + + String fixed_prefix; + const char * begin = regexp.data() + 1; + const char * pos = begin; + const char * end = regexp.data() + regexp.size(); + + while (pos != end) + { + switch (*pos) + { + case '\0': + pos = end; + break; + + case '\\': + { + ++pos; + if (pos == end) + break; + + switch (*pos) + { + case '|': + case '(': + case ')': + case '^': + case '$': + case '.': + case '[': + case '?': + case '*': + case '+': + case '{': + fixed_prefix += *pos; + break; + default: + /// all other escape sequences are not supported + pos = end; + break; + } + + ++pos; + break; + } + + /// non-trivial cases + case '|': + fixed_prefix.clear(); + [[fallthrough]]; + case '(': + case '[': + case '^': + case '$': + case '.': + case '+': + pos = end; + break; + + /// Quantifiers that allow a zero number of occurrences. + case '{': + case '?': + case '*': + if (!fixed_prefix.empty()) + fixed_prefix.pop_back(); + + pos = end; + break; + default: + fixed_prefix += *pos; + pos++; + break; + } + } + + return fixed_prefix; +} + /** For a given string, get a minimum string that is strictly greater than all strings with this prefix, * or return an empty string if there are no such strings. @@ -581,6 +663,27 @@ const KeyCondition::AtomMap KeyCondition::atom_map return true; } }, + { + "match", + [] (RPNElement & out, const Field & value) + { + if (value.getType() != Field::Types::String) + return false; + + String prefix = extractFixedPrefixFromRegularExpression(value.get()); + if (prefix.empty()) + return false; + + String right_bound = firstStringThatIsGreaterThanAllStringsWithPrefix(prefix); + + out.function = RPNElement::FUNCTION_IN_RANGE; + out.range = !right_bound.empty() + ? Range(prefix, true, right_bound, false) + : Range::createLeftBounded(prefix, true); + + return true; + } + }, { "isNotNull", [] (RPNElement & out, const Field &) @@ -1738,7 +1841,7 @@ bool KeyCondition::tryParseAtomFromAST(const Tree & node, ContextPtr context, Bl else if (func_name == "in" || func_name == "notIn" || func_name == "like" || func_name == "notLike" || func_name == "ilike" || func_name == "notIlike" || - func_name == "startsWith") + func_name == "startsWith" || func_name == "match") { /// "const IN data_column" doesn't make sense (unlike "data_column IN const") return false; diff --git a/tests/queries/0_stateless/02462_match_regexp_pk.reference b/tests/queries/0_stateless/02462_match_regexp_pk.reference new file mode 100644 index 00000000000..428d6556f4c --- /dev/null +++ b/tests/queries/0_stateless/02462_match_regexp_pk.reference @@ -0,0 +1,5 @@ +4 +1 +3 +4 +4 diff --git a/tests/queries/0_stateless/02462_match_regexp_pk.sql b/tests/queries/0_stateless/02462_match_regexp_pk.sql new file mode 100644 index 00000000000..1a944b96196 --- /dev/null +++ b/tests/queries/0_stateless/02462_match_regexp_pk.sql @@ -0,0 +1,9 @@ +CREATE TABLE mt_match_pk (v String) ENGINE = MergeTree ORDER BY v SETTINGS index_granularity = 1; +INSERT INTO mt_match_pk VALUES ('a'), ('aaa'), ('aba'), ('bac'), ('acccca'); + +SET force_primary_key = 1; +SELECT count() FROM mt_match_pk WHERE match(v, '^a'); +SELECT count() FROM mt_match_pk WHERE match(v, '^ab'); +SELECT count() FROM mt_match_pk WHERE match(v, '^a.'); +SELECT count() FROM mt_match_pk WHERE match(v, '^ab*'); +SELECT count() FROM mt_match_pk WHERE match(v, '^ac?'); From 8de75bf29c57d58ee430227ecf230c2e186438ba Mon Sep 17 00:00:00 2001 From: Anton Popov Date: Sat, 29 Oct 2022 14:26:34 +0000 Subject: [PATCH 171/239] better code with projections --- src/Storages/MergeTree/DataPartsExchange.cpp | 199 +++++++----------- src/Storages/MergeTree/DataPartsExchange.h | 13 +- .../MergeTree/MergeTreeDataWriter.cpp | 113 ++++------ src/Storages/MergeTree/MergeTreeDataWriter.h | 15 +- .../MergeTree/MergeTreeWriteAheadLog.cpp | 2 +- src/Storages/MergeTree/checkDataPart.cpp | 133 ++++-------- 6 files changed, 173 insertions(+), 302 deletions(-) diff --git a/src/Storages/MergeTree/DataPartsExchange.cpp b/src/Storages/MergeTree/DataPartsExchange.cpp index afef354a2e6..4f9c9ffd596 100644 --- a/src/Storages/MergeTree/DataPartsExchange.cpp +++ b/src/Storages/MergeTree/DataPartsExchange.cpp @@ -180,6 +180,13 @@ void Service::processQuery(const HTMLForm & params, ReadBuffer & /*body*/, Write std::sregex_token_iterator(remote_fs_metadata.begin(), remote_fs_metadata.end(), re, -1), std::sregex_token_iterator()); + bool send_projections = client_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_PROJECTION; + if (send_projections) + { + const auto & projections = part->getProjectionParts(); + writeBinary(projections.size(), out); + } + if (data_settings->allow_remote_fs_zero_copy_replication && /// In memory data part does not have metadata yet. !isInMemoryPart(part) && @@ -190,33 +197,15 @@ void Service::processQuery(const HTMLForm & params, ReadBuffer & /*body*/, Write { /// Send metadata if the receiver's capability covers the source disk type. response.addCookie({"remote_fs_metadata", disk_type}); - if (client_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_PROJECTION) - { - const auto & projections = part->getProjectionParts(); - writeBinary(projections.size(), out); - } - - sendPartFromDiskRemoteMeta(part, out, true, part->getProjectionParts()); + sendPartFromDiskRemoteMeta(part, out, true, send_projections); return; } } - if (client_protocol_version >= REPLICATION_PROTOCOL_VERSION_WITH_PARTS_PROJECTION) - { - const auto & projections = part->getProjectionParts(); - writeBinary(projections.size(), out); - if (isInMemoryPart(part)) - sendPartFromMemory(part, out, projections); - else - sendPartFromDisk(part, out, client_protocol_version, projections); - } + if (isInMemoryPart(part)) + sendPartFromMemory(part, out, send_projections); else - { - if (isInMemoryPart(part)) - sendPartFromMemory(part, out); - else - sendPartFromDisk(part, out, client_protocol_version); - } + sendPartFromDisk(part, out, client_protocol_version, send_projections); } catch (const NetException &) { @@ -238,20 +227,23 @@ void Service::processQuery(const HTMLForm & params, ReadBuffer & /*body*/, Write } void Service::sendPartFromMemory( - const MergeTreeData::DataPartPtr & part, WriteBuffer & out, const std::map> & projections) + const MergeTreeData::DataPartPtr & part, WriteBuffer & out, bool send_projections) { auto metadata_snapshot = data.getInMemoryMetadataPtr(); - for (const auto & [name, projection] : projections) + if (send_projections) { - auto projection_sample_block = metadata_snapshot->projections.get(name).sample_block; - auto part_in_memory = asInMemoryPart(projection); - if (!part_in_memory) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Projection {} of part {} is not stored in memory", name, part->name); + for (const auto & [name, projection] : part->getProjectionParts()) + { + auto projection_sample_block = metadata_snapshot->projections.get(name).sample_block; + auto part_in_memory = asInMemoryPart(projection); + if (!part_in_memory) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Projection {} of part {} is not stored in memory", name, part->name); - writeStringBinary(name, out); - projection->checksums.write(out); - NativeWriter block_out(out, 0, projection_sample_block); - block_out.write(part_in_memory->block); + writeStringBinary(name, out); + projection->checksums.write(out); + NativeWriter block_out(out, 0, projection_sample_block); + block_out.write(part_in_memory->block); + } } auto part_in_memory = asInMemoryPart(part); @@ -269,7 +261,7 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDisk( const MergeTreeData::DataPartPtr & part, WriteBuffer & out, int client_protocol_version, - const std::map> & projections) + bool send_projections) { /// We'll take a list of files from the list of checksums. MergeTreeData::DataPart::Checksums checksums = part->checksums; @@ -277,7 +269,8 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDisk( auto file_names_without_checksums = part->getFileNamesWithoutChecksums(); for (const auto & file_name : file_names_without_checksums) { - if (client_protocol_version < REPLICATION_PROTOCOL_VERSION_WITH_PARTS_DEFAULT_COMPRESSION && file_name == IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME) + if (client_protocol_version < REPLICATION_PROTOCOL_VERSION_WITH_PARTS_DEFAULT_COMPRESSION + && file_name == IMergeTreeDataPart::DEFAULT_COMPRESSION_CODEC_FILE_NAME) continue; checksums.files[file_name] = {}; @@ -288,11 +281,10 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDisk( { // Get rid of projection files checksums.files.erase(name + ".proj"); - auto it = projections.find(name); - if (it != projections.end()) + if (send_projections) { writeStringBinary(name, out); - MergeTreeData::DataPart::Checksums projection_checksum = sendPartFromDisk(it->second, out, client_protocol_version); + MergeTreeData::DataPart::Checksums projection_checksum = sendPartFromDisk(projection, out, client_protocol_version, false); data_checksums.addFile(name + ".proj", projection_checksum.getTotalSizeOnDisk(), projection_checksum.getTotalChecksumUInt128()); } else if (part->checksums.has(name + ".proj")) @@ -337,18 +329,15 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDisk( return data_checksums; } -MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( +void Service::sendPartFromDiskRemoteMeta( const MergeTreeData::DataPartPtr & part, WriteBuffer & out, bool send_part_id, - const std::map> & projections) + bool send_projections) { - const auto * data_part_storage_on_disk = dynamic_cast(&part->getDataPartStorage()); - if (!data_part_storage_on_disk) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Storage '{}' doesn't support zero-copy replication", part->getDataPartStorage().getDiskName()); - - if (!data_part_storage_on_disk->supportZeroCopyReplication()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "Disk '{}' doesn't support zero-copy replication", data_part_storage_on_disk->getDiskName()); + auto data_part_storage = part->getDataPartStoragePtr(); + if (!data_part_storage->supportZeroCopyReplication()) + throw Exception(ErrorCodes::LOGICAL_ERROR, "Disk '{}' doesn't support zero-copy replication", data_part_storage->getDiskName()); /// We'll take a list of files from the list of checksums. MergeTreeData::DataPart::Checksums checksums = part->checksums; @@ -369,30 +358,20 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( paths.push_back(fs::path(part->getDataPartStorage().getRelativePath()) / it.first); /// Serialized metadatadatas with zero ref counts. - auto metadatas = data_part_storage_on_disk->getSerializedMetadata(paths); + auto metadatas = data_part_storage->getSerializedMetadata(paths); if (send_part_id) { - String part_id = data_part_storage_on_disk->getUniqueId(); + String part_id = data_part_storage->getUniqueId(); writeStringBinary(part_id, out); } - MergeTreeData::DataPart::Checksums data_checksums; - for (const auto & [name, projection] : part->getProjectionParts()) + if (send_projections) { - auto it = projections.find(name); - if (it != projections.end()) + for (const auto & [name, projection] : part->getProjectionParts()) { - writeStringBinary(name, out); - MergeTreeData::DataPart::Checksums projection_checksum = sendPartFromDiskRemoteMeta(it->second, out, false); - data_checksums.addFile(name + ".proj", projection_checksum.getTotalSizeOnDisk(), projection_checksum.getTotalChecksumUInt128()); - } - else if (part->checksums.has(name + ".proj")) - { - // We don't send this projection, just add out checksum to bypass the following check - const auto & our_checksum = part->checksums.files.find(name + ".proj")->second; - data_checksums.addFile(name + ".proj", our_checksum.file_size, our_checksum.file_hash); + sendPartFromDiskRemoteMeta(projection, out, false, false); } } @@ -403,7 +382,7 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( String file_path_prefix = fs::path(part->getDataPartStorage().getRelativePath()) / file_name; /// Just some additional checks - String metadata_file_path = fs::path(data_part_storage_on_disk->getDiskPath()) / file_path_prefix; + String metadata_file_path = fs::path(data_part_storage->getDiskPath()) / file_path_prefix; fs::path metadata(metadata_file_path); if (!fs::exists(metadata)) throw Exception(ErrorCodes::CORRUPTED_DATA, "Remote metadata '{}' is not exists", file_name); @@ -427,12 +406,7 @@ MergeTreeData::DataPart::Checksums Service::sendPartFromDiskRemoteMeta( throw Exception(ErrorCodes::BAD_SIZE_OF_FILE_IN_DATA_PART, "Unexpected size of file {}", metadata_file_path); writePODBinary(hashing_out.getHash(), out); - - if (!file_names_without_checksums.contains(file_name)) - data_checksums.addFile(file_name, hashing_out.count(), hashing_out.getHash()); } - - return data_checksums; } MergeTreeData::DataPartPtr Service::findPart(const String & name) @@ -707,68 +681,54 @@ MergeTreeData::MutableDataPartPtr Fetcher::fetchSelectedPart( in->setNextCallback(ReplicatedFetchReadCallback(*entry)); - return part_type == "InMemory" - ? downloadPartToMemory(part_name, part_uuid, metadata_snapshot, context, disk, *in, projections, throttler) - : downloadPartToDisk(part_name, replica_path, to_detached, tmp_prefix, sync, disk, *in, projections, checksums, throttler); + if (part_type == "InMemory") + { + auto volume = std::make_shared("volume_" + part_name, disk, 0); + + auto data_part_storage = std::make_shared( + volume, + data.getRelativeDataPath(), + part_name); + + return downloadPartToMemory( + data_part_storage, part_name, + MergeTreePartInfo::fromPartName(part_name, data.format_version), + part_uuid, metadata_snapshot, context, *in, + projections, false, throttler); + } + + return downloadPartToDisk( + part_name, replica_path, to_detached, tmp_prefix, + sync, disk, *in, projections, checksums, throttler); } MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( + MutableDataPartStoragePtr data_part_storage, const String & part_name, + const MergeTreePartInfo & part_info, const UUID & part_uuid, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, - DiskPtr disk, PooledReadWriteBufferFromHTTP & in, size_t projections, + bool is_projection, ThrottlerPtr throttler) { - auto volume = std::make_shared("volume_" + part_name, disk, 0); + auto new_data_part = std::make_shared(data, part_name, part_info, data_part_storage); - auto data_part_storage = std::make_shared( - volume, - data.getRelativeDataPath(), - part_name); - - MergeTreeData::MutableDataPartPtr new_data_part = - std::make_shared(data, part_name, data_part_storage); - - new_data_part->version.setCreationTID(Tx::PrehistoricTID, nullptr); - - for (auto i = 0ul; i < projections; ++i) + for (size_t i = 0; i < projections; ++i) { String projection_name; readStringBinary(projection_name, in); - MergeTreeData::DataPart::Checksums checksums; - if (!checksums.read(in)) - throw Exception("Cannot deserialize checksums", ErrorCodes::CORRUPTED_DATA); - - NativeReader block_in(in, 0); - auto block = block_in.read(); - throttler->add(block.bytes()); - - auto projection_part_storage = data_part_storage->getProjection(projection_name + ".proj"); MergeTreePartInfo new_part_info("all", 0, 0, 0); - MergeTreeData::MutableDataPartPtr new_projection_part = - std::make_shared(data, projection_name, new_part_info, projection_part_storage, new_data_part.get()); + auto projection_part_storage = data_part_storage->getProjection(projection_name + ".proj"); - new_projection_part->is_temp = false; - new_projection_part->setColumns(block.getNamesAndTypesList(), {}); - MergeTreePartition partition{}; - new_projection_part->partition = std::move(partition); - new_projection_part->minmax_idx = std::make_shared(); + auto new_projection_part = downloadPartToMemory( + projection_part_storage, projection_name, + new_part_info, part_uuid, metadata_snapshot, + context, in, 0, true, throttler); - MergedBlockOutputStream part_out( - new_projection_part, - metadata_snapshot->projections.get(projection_name).metadata, - block.getNamesAndTypesList(), - {}, - CompressionCodecFactory::instance().get("NONE", {}), - NO_TRANSACTION_PTR); - - part_out.write(block); - part_out.finalizePart(new_projection_part, false); - new_projection_part->checksums.checkEqual(checksums, /* have_uncompressed = */ true); new_data_part->addProjectionPart(projection_name, std::move(new_projection_part)); } @@ -780,11 +740,16 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToMemory( auto block = block_in.read(); throttler->add(block.bytes()); - new_data_part->uuid = part_uuid; - new_data_part->is_temp = true; new_data_part->setColumns(block.getNamesAndTypesList(), {}); - new_data_part->minmax_idx->update(block, data.getMinMaxColumnsNames(metadata_snapshot->getPartitionKey())); - new_data_part->partition.create(metadata_snapshot, block, 0, context); + + if (!is_projection) + { + new_data_part->version.setCreationTID(Tx::PrehistoricTID, nullptr); + new_data_part->uuid = part_uuid; + new_data_part->is_temp = true; + new_data_part->minmax_idx->update(block, data.getMinMaxColumnsNames(metadata_snapshot->getPartitionKey())); + new_data_part->partition.create(metadata_snapshot, block, 0, context); + } MergedBlockOutputStream part_out( new_data_part, metadata_snapshot, block.getNamesAndTypesList(), {}, @@ -850,7 +815,6 @@ void Fetcher::downloadBasePartOrProjectionPartToDiskRemoteMeta( checksums.addFile(file_name, file_size, expected_hash); } } - } @@ -966,11 +930,11 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDisk( SyncGuardPtr sync_guard; if (data.getSettings()->fsync_part_directory) - sync_guard = disk->getDirectorySyncGuard(data_part_storage->getRelativePath()); + sync_guard = data_part_storage->getDirectorySyncGuard(); CurrentMetrics::Increment metric_increment{CurrentMetrics::ReplicatedFetch}; - for (auto i = 0ul; i < projections; ++i) + for (size_t i = 0; i < projections; ++i) { String projection_name; readStringBinary(projection_name, in); @@ -1043,7 +1007,7 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDiskRemoteMeta( data_part_storage->createDirectories(); - for (auto i = 0ul; i < projections; ++i) + for (size_t i = 0; i < projections; ++i) { String projection_name; readStringBinary(projection_name, in); @@ -1071,7 +1035,6 @@ MergeTreeData::MutableDataPartPtr Fetcher::downloadPartToDiskRemoteMeta( new_data_part->version.setCreationTID(Tx::PrehistoricTID, nullptr); new_data_part->is_temp = true; new_data_part->modification_time = time(nullptr); - new_data_part->loadColumnsChecksumsIndexes(true, false); } #if USE_AWS_S3 diff --git a/src/Storages/MergeTree/DataPartsExchange.h b/src/Storages/MergeTree/DataPartsExchange.h index 59800756c34..6c92fad4092 100644 --- a/src/Storages/MergeTree/DataPartsExchange.h +++ b/src/Storages/MergeTree/DataPartsExchange.h @@ -1,5 +1,6 @@ #pragma once +#include "Storages/MergeTree/MergeTreePartInfo.h" #include #include #include @@ -42,19 +43,19 @@ private: void sendPartFromMemory( const MergeTreeData::DataPartPtr & part, WriteBuffer & out, - const std::map> & projections = {}); + bool send_projections); MergeTreeData::DataPart::Checksums sendPartFromDisk( const MergeTreeData::DataPartPtr & part, WriteBuffer & out, int client_protocol_version, - const std::map> & projections = {}); + bool send_projections); - MergeTreeData::DataPart::Checksums sendPartFromDiskRemoteMeta( + void sendPartFromDiskRemoteMeta( const MergeTreeData::DataPartPtr & part, WriteBuffer & out, bool send_part_id, - const std::map> & projections = {}); + bool send_projections); /// StorageReplicatedMergeTree::shutdown() waits for all parts exchange handlers to finish, /// so Service will never access dangling reference to storage @@ -120,13 +121,15 @@ private: ThrottlerPtr throttler); MergeTreeData::MutableDataPartPtr downloadPartToMemory( + MutableDataPartStoragePtr data_part_storage, const String & part_name, + const MergeTreePartInfo & part_info, const UUID & part_uuid, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, - DiskPtr disk, PooledReadWriteBufferFromHTTP & in, size_t projections, + bool is_projection, ThrottlerPtr throttler); MergeTreeData::MutableDataPartPtr downloadPartToDiskRemoteMeta( diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.cpp b/src/Storages/MergeTree/MergeTreeDataWriter.cpp index 55404324b75..7e064f5164f 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.cpp +++ b/src/Storages/MergeTree/MergeTreeDataWriter.cpp @@ -480,8 +480,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempPart( MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( const String & part_name, - MergeTreeDataPartType part_type, - const String & relative_path, bool is_temp, IMergeTreeDataPart * parent_part, const MergeTreeData & data, @@ -493,6 +491,21 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( const StorageMetadataPtr & metadata_snapshot = projection.metadata; MergeTreePartInfo new_part_info("all", 0, 0, 0); + MergeTreeDataPartType part_type; + if (parent_part->getType() == MergeTreeDataPartType::InMemory) + { + part_type = MergeTreeDataPartType::InMemory; + } + else + { + /// Size of part would not be greater than block.bytes() + epsilon + size_t expected_size = block.bytes(); + // just check if there is enough space on parent volume + data.reserveSpace(expected_size, parent_part->getDataPartStorage()); + part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); + } + + auto relative_path = part_name + (is_temp ? ".tmp_proj" : ".proj"); auto projection_part_storage = parent_part->getDataPartStorage().getProjection(relative_path); auto new_data_part = data.createPart( part_name, @@ -583,77 +596,6 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPartImpl( } MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeProjectionPart( - MergeTreeData & data, - Poco::Logger * log, - Block block, - const ProjectionDescription & projection, - IMergeTreeDataPart * parent_part) -{ - String part_name = projection.name; - MergeTreeDataPartType part_type; - if (parent_part->getType() == MergeTreeDataPartType::InMemory) - { - part_type = MergeTreeDataPartType::InMemory; - } - else - { - /// Size of part would not be greater than block.bytes() + epsilon - size_t expected_size = block.bytes(); - // just check if there is enough space on parent volume - data.reserveSpace(expected_size, parent_part->getDataPartStorage()); - part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); - } - - return writeProjectionPartImpl( - part_name, - part_type, - part_name + ".proj" /* relative_path */, - false /* is_temp */, - parent_part, - data, - log, - block, - projection); -} - -/// This is used for projection materialization process which may contain multiple stages of -/// projection part merges. -MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( - MergeTreeData & data, - Poco::Logger * log, - Block block, - const ProjectionDescription & projection, - IMergeTreeDataPart * parent_part, - size_t block_num) -{ - String part_name = fmt::format("{}_{}", projection.name, block_num); - MergeTreeDataPartType part_type; - if (parent_part->getType() == MergeTreeDataPartType::InMemory) - { - part_type = MergeTreeDataPartType::InMemory; - } - else - { - /// Size of part would not be greater than block.bytes() + epsilon - size_t expected_size = block.bytes(); - // just check if there is enough space on parent volume - data.reserveSpace(expected_size, parent_part->getDataPartStorage()); - part_type = data.choosePartTypeOnDisk(expected_size, block.rows()); - } - - return writeProjectionPartImpl( - part_name, - part_type, - part_name + ".tmp_proj" /* relative_path */, - true /* is_temp */, - parent_part, - data, - log, - block, - projection); -} - -MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeInMemoryProjectionPart( const MergeTreeData & data, Poco::Logger * log, Block block, @@ -662,13 +604,32 @@ MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeInMemoryProjectionP { return writeProjectionPartImpl( projection.name, - MergeTreeDataPartType::InMemory, - projection.name + ".proj" /* relative_path */, false /* is_temp */, parent_part, data, log, - block, + std::move(block), + projection); +} + +/// This is used for projection materialization process which may contain multiple stages of +/// projection part merges. +MergeTreeDataWriter::TemporaryPart MergeTreeDataWriter::writeTempProjectionPart( + const MergeTreeData & data, + Poco::Logger * log, + Block block, + const ProjectionDescription & projection, + IMergeTreeDataPart * parent_part, + size_t block_num) +{ + String part_name = fmt::format("{}_{}", projection.name, block_num); + return writeProjectionPartImpl( + part_name, + true /* is_temp */, + parent_part, + data, + log, + std::move(block), projection); } diff --git a/src/Storages/MergeTree/MergeTreeDataWriter.h b/src/Storages/MergeTree/MergeTreeDataWriter.h index 8e405016cde..8c2bf66e8f8 100644 --- a/src/Storages/MergeTree/MergeTreeDataWriter.h +++ b/src/Storages/MergeTree/MergeTreeDataWriter.h @@ -73,7 +73,7 @@ public: /// For insertion. static TemporaryPart writeProjectionPart( - MergeTreeData & data, + const MergeTreeData & data, Poco::Logger * log, Block block, const ProjectionDescription & projection, @@ -81,21 +81,13 @@ public: /// For mutation: MATERIALIZE PROJECTION. static TemporaryPart writeTempProjectionPart( - MergeTreeData & data, + const MergeTreeData & data, Poco::Logger * log, Block block, const ProjectionDescription & projection, IMergeTreeDataPart * parent_part, size_t block_num); - /// For WriteAheadLog AddPart. - static TemporaryPart writeInMemoryProjectionPart( - const MergeTreeData & data, - Poco::Logger * log, - Block block, - const ProjectionDescription & projection, - IMergeTreeDataPart * parent_part); - static Block mergeBlock( const Block & block, SortDescription sort_description, @@ -106,8 +98,6 @@ public: private: static TemporaryPart writeProjectionPartImpl( const String & part_name, - MergeTreeDataPartType part_type, - const String & relative_path, bool is_temp, IMergeTreeDataPart * parent_part, const MergeTreeData & data, @@ -116,7 +106,6 @@ private: const ProjectionDescription & projection); MergeTreeData & data; - Poco::Logger * log; }; diff --git a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp index 85b41f6dd83..b3625ba8e93 100644 --- a/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp +++ b/src/Storages/MergeTree/MergeTreeWriteAheadLog.cpp @@ -237,7 +237,7 @@ MergeTreeData::MutableDataPartsVector MergeTreeWriteAheadLog::restore( for (const auto & projection : metadata_snapshot->getProjections()) { auto projection_block = projection.calculate(block, context); - auto temp_part = MergeTreeDataWriter::writeInMemoryProjectionPart(storage, log, projection_block, projection, part.get()); + auto temp_part = MergeTreeDataWriter::writeProjectionPart(storage, log, projection_block, projection, part.get()); temp_part.finalize(); if (projection_block.rows()) part->addProjectionPart(projection.name, std::move(temp_part.part)); diff --git a/src/Storages/MergeTree/checkDataPart.cpp b/src/Storages/MergeTree/checkDataPart.cpp index 4758ccb201a..6f9f16b6155 100644 --- a/src/Storages/MergeTree/checkDataPart.cpp +++ b/src/Storages/MergeTree/checkDataPart.cpp @@ -112,98 +112,17 @@ IMergeTreeDataPart::Checksums checkDataPart( }; /// This function calculates only checksum of file content (compressed or uncompressed). - /// It also calculates checksum of projections. auto checksum_file = [&](const String & file_name) { - if (data_part_storage.isDirectory(file_name) && endsWith(file_name, ".proj")) - { - auto projection_name = file_name.substr(0, file_name.size() - sizeof(".proj") + 1); - auto pit = data_part->getProjectionParts().find(projection_name); - if (pit == data_part->getProjectionParts().end()) - { - if (require_checksums) - throw Exception("Unexpected file " + file_name + " in data part", ErrorCodes::UNEXPECTED_FILE_IN_DATA_PART); - else - return; - } - - const auto & projection = pit->second; - IMergeTreeDataPart::Checksums projection_checksums_data; - - auto projection_part_storage = data_part_storage.getProjection(file_name); - - if (projection->getType() == MergeTreeDataPartType::Compact) - { - auto file_buf = projection_part_storage->readFile(MergeTreeDataPartCompact::DATA_FILE_NAME_WITH_EXTENSION, {}, std::nullopt, std::nullopt); - HashingReadBuffer hashing_buf(*file_buf); - hashing_buf.ignoreAll(); - projection_checksums_data.files[MergeTreeDataPartCompact::DATA_FILE_NAME_WITH_EXTENSION] - = IMergeTreeDataPart::Checksums::Checksum(hashing_buf.count(), hashing_buf.getHash()); - } - else - { - const NamesAndTypesList & projection_columns_list = projection->getColumns(); - for (const auto & projection_column : projection_columns_list) - { - get_serialization(projection_column)->enumerateStreams( - [&](const ISerialization::SubstreamPath & substream_path) - { - String projection_file_name = ISerialization::getFileNameForStream(projection_column, substream_path) + ".bin"; - projection_checksums_data.files[projection_file_name] = checksum_compressed_file(*projection_part_storage, projection_file_name); - }); - } - } - - IMergeTreeDataPart::Checksums projection_checksums_txt; - - if (require_checksums || projection_part_storage->exists("checksums.txt")) - { - auto buf = projection_part_storage->readFile("checksums.txt", {}, std::nullopt, std::nullopt); - projection_checksums_txt.read(*buf); - assertEOF(*buf); - } - - const auto & projection_checksum_files_txt = projection_checksums_txt.files; - for (auto projection_it = projection_part_storage->iterate(); projection_it->isValid(); projection_it->next()) - { - const String & projection_file_name = projection_it->name(); - auto projection_checksum_it = projection_checksums_data.files.find(projection_file_name); - - /// Skip files that we already calculated. Also skip metadata files that are not checksummed. - if (projection_checksum_it == projection_checksums_data.files.end() && !files_without_checksums.contains(projection_file_name)) - { - auto projection_txt_checksum_it = projection_checksum_files_txt.find(file_name); - if (projection_txt_checksum_it == projection_checksum_files_txt.end() - || projection_txt_checksum_it->second.uncompressed_size == 0) - { - auto projection_file_buf = projection_part_storage->readFile(projection_file_name, {}, std::nullopt, std::nullopt); - HashingReadBuffer projection_hashing_buf(*projection_file_buf); - projection_hashing_buf.ignoreAll(); - projection_checksums_data.files[projection_file_name] = IMergeTreeDataPart::Checksums::Checksum( - projection_hashing_buf.count(), projection_hashing_buf.getHash()); - } - else - { - projection_checksums_data.files[projection_file_name] = checksum_compressed_file(*projection_part_storage, projection_file_name); - } - } - } - checksums_data.files[file_name] = IMergeTreeDataPart::Checksums::Checksum( - projection_checksums_data.getTotalSizeOnDisk(), projection_checksums_data.getTotalChecksumUInt128()); - - if (require_checksums || !projection_checksums_txt.files.empty()) - projection_checksums_txt.checkEqual(projection_checksums_data, false); - } - else - { - auto file_buf = data_part_storage.readFile(file_name, {}, std::nullopt, std::nullopt); - HashingReadBuffer hashing_buf(*file_buf); - hashing_buf.ignoreAll(); - checksums_data.files[file_name] = IMergeTreeDataPart::Checksums::Checksum(hashing_buf.count(), hashing_buf.getHash()); - } + auto file_buf = data_part_storage.readFile(file_name, {}, std::nullopt, std::nullopt); + HashingReadBuffer hashing_buf(*file_buf); + hashing_buf.ignoreAll(); + checksums_data.files[file_name] = IMergeTreeDataPart::Checksums::Checksum(hashing_buf.count(), hashing_buf.getHash()); }; - bool check_uncompressed = true; + /// Do not check uncompressed for projections. But why? + bool check_uncompressed = !data_part->isProjectionPart(); + /// First calculate checksums for columns data if (part_type == MergeTreeDataPartType::Compact) { @@ -238,10 +157,19 @@ IMergeTreeDataPart::Checksums checkDataPart( assertEOF(*buf); } + NameSet projections_on_disk; const auto & checksum_files_txt = checksums_txt.files; for (auto it = data_part_storage.iterate(); it->isValid(); it->next()) { - const String & file_name = it->name(); + auto file_name = it->name(); + + /// We will check projections later. + if (data_part_storage.isDirectory(file_name) && endsWith(file_name, ".proj")) + { + projections_on_disk.insert(file_name); + continue; + } + auto checksum_it = checksums_data.files.find(file_name); /// Skip files that we already calculated. Also skip metadata files that are not checksummed. @@ -260,11 +188,38 @@ IMergeTreeDataPart::Checksums checkDataPart( } } + for (const auto & [name, projection] : data_part->getProjectionParts()) + { + if (is_cancelled()) + return {}; + + auto projection_file = name + ".proj"; + auto projection_checksums = checkDataPart( + projection, *data_part_storage.getProjection(projection_file), + projection->getColumns(), projection->getType(), + projection->getFileNamesWithoutChecksums(), + require_checksums, is_cancelled); + + checksums_data.files[projection_file] = IMergeTreeDataPart::Checksums::Checksum( + projection_checksums.getTotalSizeOnDisk(), + projection_checksums.getTotalChecksumUInt128()); + + projections_on_disk.erase(projection_file); + } + + if (require_checksums && !projections_on_disk.empty()) + { + throw Exception(ErrorCodes::UNEXPECTED_FILE_IN_DATA_PART, + "Found unexpected projection directories: {}", + fmt::join(projections_on_disk, ",")); + } + if (is_cancelled()) return {}; if (require_checksums || !checksums_txt.files.empty()) checksums_txt.checkEqual(checksums_data, check_uncompressed); + return checksums_data; } From 25d35a97f9a9f19b00cb76834fc28d26e72c8dfd Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Sat, 29 Oct 2022 18:10:33 +0200 Subject: [PATCH 172/239] Update CCTZ to 2022f --- contrib/cctz | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/cctz b/contrib/cctz index 7a454c25c7d..5c8528fb35e 160000 --- a/contrib/cctz +++ b/contrib/cctz @@ -1 +1 @@ -Subproject commit 7a454c25c7d16053bcd327cdd16329212a08fa4a +Subproject commit 5c8528fb35e89ee0b3a7157490423fba0d4dd7b5 From c68ab231f917bc41b354f7d2a1fa769a7fbb8d06 Mon Sep 17 00:00:00 2001 From: Alexander Tokmakov Date: Sun, 30 Oct 2022 17:30:51 +0100 Subject: [PATCH 173/239] fix accessing part in Deleting state --- src/Storages/MergeTree/MergeTreeData.cpp | 11 ++++++--- src/Storages/MergeTree/MergeTreeData.h | 25 ++++++++++++++++++++- src/Storages/StorageReplicatedMergeTree.cpp | 23 ++++++++++--------- src/Storages/StorageReplicatedMergeTree.h | 2 +- 4 files changed, 46 insertions(+), 15 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index bb589161b57..d38c9c465d5 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -3142,7 +3142,7 @@ void MergeTreeData::removePartsInRangeFromWorkingSet(MergeTreeTransaction * txn, removePartsInRangeFromWorkingSetAndGetPartsToRemoveFromZooKeeper(txn, drop_range, lock); } -MergeTreeData::DataPartsVector MergeTreeData::removePartsInRangeFromWorkingSetAndGetPartsToRemoveFromZooKeeper( +MergeTreeData::PartsToRemoveFromZooKeeper MergeTreeData::removePartsInRangeFromWorkingSetAndGetPartsToRemoveFromZooKeeper( MergeTreeTransaction * txn, const MergeTreePartInfo & drop_range, DataPartsLock & lock) { DataPartsVector parts_to_remove; @@ -3220,15 +3220,20 @@ MergeTreeData::DataPartsVector MergeTreeData::removePartsInRangeFromWorkingSetAn /// FIXME refactor removePartsFromWorkingSet(...), do not remove parts twice removePartsFromWorkingSet(txn, parts_to_remove, clear_without_timeout, lock); + /// Since we can return parts in Deleting state, we have to use a wrapper that restricts access to such parts. + PartsToRemoveFromZooKeeper parts_to_remove_from_zookeeper; + for (auto & part : parts_to_remove) + parts_to_remove_from_zookeeper.emplace_back(std::move(part)); + for (auto & part : inactive_parts_to_remove_immediately) { if (!drop_range.contains(part->info)) continue; part->remove_time.store(0, std::memory_order_relaxed); - parts_to_remove.push_back(std::move(part)); + parts_to_remove_from_zookeeper.emplace_back(std::move(part), /* was_active */ false); } - return parts_to_remove; + return parts_to_remove_from_zookeeper; } void MergeTreeData::restoreAndActivatePart(const DataPartPtr & part, DataPartsLock * acquired_lock) diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 2b67face570..60255ce1b16 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -584,10 +584,33 @@ public: /// Used in REPLACE PARTITION command. void removePartsInRangeFromWorkingSet(MergeTreeTransaction * txn, const MergeTreePartInfo & drop_range, DataPartsLock & lock); + /// This wrapper is required to restrict access to parts in Deleting state + class PartToRemoveFromZooKeeper + { + DataPartPtr part; + bool was_active; + + public: + PartToRemoveFromZooKeeper(DataPartPtr && part_, bool was_active_ = true) + : part(std::move(part_)), was_active(was_active_) + { + } + + /// It's s to get name of any part + const String & getPartName() const { return part->name; } + + DataPartPtr getPartIfItWasActive() const + { + return was_active ? part : nullptr; + } + }; + + using PartsToRemoveFromZooKeeper = std::vector; + /// Same as above, but also returns list of parts to remove from ZooKeeper. /// It includes parts that have been just removed by these method /// and Outdated parts covered by drop_range that were removed earlier for any reason. - DataPartsVector removePartsInRangeFromWorkingSetAndGetPartsToRemoveFromZooKeeper( + PartsToRemoveFromZooKeeper removePartsInRangeFromWorkingSetAndGetPartsToRemoveFromZooKeeper( MergeTreeTransaction * txn, const MergeTreePartInfo & drop_range, DataPartsLock & lock); /// Restores Outdated part and adds it to working set diff --git a/src/Storages/StorageReplicatedMergeTree.cpp b/src/Storages/StorageReplicatedMergeTree.cpp index 8dad5755dab..3c0fbb162bc 100644 --- a/src/Storages/StorageReplicatedMergeTree.cpp +++ b/src/Storages/StorageReplicatedMergeTree.cpp @@ -1827,7 +1827,7 @@ void StorageReplicatedMergeTree::executeDropRange(const LogEntry & entry) /// Therefore, we use all data parts. auto metadata_snapshot = getInMemoryMetadataPtr(); - DataPartsVector parts_to_remove; + PartsToRemoveFromZooKeeper parts_to_remove; { auto data_parts_lock = lockParts(); parts_to_remove = removePartsInRangeFromWorkingSetAndGetPartsToRemoveFromZooKeeper(NO_TRANSACTION_RAW, drop_range_info, data_parts_lock); @@ -1849,8 +1849,11 @@ void StorageReplicatedMergeTree::executeDropRange(const LogEntry & entry) /// If DETACH clone parts to detached/ directory for (const auto & part : parts_to_remove) { - LOG_INFO(log, "Detaching {}", part->getDataPartStorage().getPartDirectory()); - part->makeCloneInDetached("", metadata_snapshot); + if (auto part_to_detach = part.getPartIfItWasActive()) + { + LOG_INFO(log, "Detaching {}", part_to_detach->getDataPartStorage().getPartDirectory()); + part_to_detach->makeCloneInDetached("", metadata_snapshot); + } } } @@ -1941,7 +1944,7 @@ bool StorageReplicatedMergeTree::executeReplaceRange(const LogEntry & entry) PartDescriptions all_parts; PartDescriptions parts_to_add; - DataPartsVector parts_to_remove; + PartsToRemoveFromZooKeeper parts_to_remove; auto table_lock_holder_dst_table = lockForShare( RWLockImpl::NO_QUERY, getSettings()->lock_acquire_timeout_for_background_operations); @@ -1972,7 +1975,7 @@ bool StorageReplicatedMergeTree::executeReplaceRange(const LogEntry & entry) String parts_to_remove_str; for (const auto & part : parts_to_remove) { - parts_to_remove_str += part->name; + parts_to_remove_str += part.getPartName(); parts_to_remove_str += " "; } LOG_TRACE(log, "Replacing {} parts {}with empty set", parts_to_remove.size(), parts_to_remove_str); @@ -2248,7 +2251,7 @@ bool StorageReplicatedMergeTree::executeReplaceRange(const LogEntry & entry) String parts_to_remove_str; for (const auto & part : parts_to_remove) { - parts_to_remove_str += part->name; + parts_to_remove_str += part.getPartName(); parts_to_remove_str += " "; } LOG_TRACE(log, "Replacing {} parts {}with {} parts {}", parts_to_remove.size(), parts_to_remove_str, @@ -6230,11 +6233,11 @@ void StorageReplicatedMergeTree::clearOldPartsAndRemoveFromZK() } -void StorageReplicatedMergeTree::removePartsFromZooKeeperWithRetries(DataPartsVector & parts, size_t max_retries) +void StorageReplicatedMergeTree::removePartsFromZooKeeperWithRetries(PartsToRemoveFromZooKeeper & parts, size_t max_retries) { Strings part_names_to_remove; for (const auto & part : parts) - part_names_to_remove.emplace_back(part->name); + part_names_to_remove.emplace_back(part.getPartName()); return removePartsFromZooKeeperWithRetries(part_names_to_remove, max_retries); } @@ -6561,7 +6564,7 @@ void StorageReplicatedMergeTree::replacePartitionFrom( if (replace) clearBlocksInPartition(*zookeeper, drop_range.partition_id, drop_range.max_block, drop_range.max_block); - DataPartsVector parts_to_remove; + PartsToRemoveFromZooKeeper parts_to_remove; Coordination::Responses op_results; try @@ -6797,7 +6800,7 @@ void StorageReplicatedMergeTree::movePartitionToTable(const StoragePtr & dest_ta clearBlocksInPartition(*zookeeper, drop_range.partition_id, drop_range.max_block, drop_range.max_block); - DataPartsVector parts_to_remove; + PartsToRemoveFromZooKeeper parts_to_remove; Coordination::Responses op_results; try diff --git a/src/Storages/StorageReplicatedMergeTree.h b/src/Storages/StorageReplicatedMergeTree.h index 5213f963fdf..323b1ce02bf 100644 --- a/src/Storages/StorageReplicatedMergeTree.h +++ b/src/Storages/StorageReplicatedMergeTree.h @@ -549,7 +549,7 @@ private: /// Remove parts from ZooKeeper, throw exception if unable to do so after max_retries. void removePartsFromZooKeeperWithRetries(const Strings & part_names, size_t max_retries = 5); - void removePartsFromZooKeeperWithRetries(DataPartsVector & parts, size_t max_retries = 5); + void removePartsFromZooKeeperWithRetries(PartsToRemoveFromZooKeeper & parts, size_t max_retries = 5); /// Removes a part from ZooKeeper and adds a task to the queue to download it. It is supposed to do this with broken parts. void removePartAndEnqueueFetch(const String & part_name); From 978aa16e29d67cfeed8abead1e99802983448699 Mon Sep 17 00:00:00 2001 From: Nikolay Degterinsky Date: Sun, 30 Oct 2022 16:42:57 +0000 Subject: [PATCH 174/239] Fix a bug in ParserCreateUserQuery --- src/Parsers/Access/ParserCreateUserQuery.cpp | 5 ++--- .../0_stateless/02474_create_user_query_fuzzer_bug.reference | 0 .../0_stateless/02474_create_user_query_fuzzer_bug.sql | 1 + 3 files changed, 3 insertions(+), 3 deletions(-) create mode 100644 tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.reference create mode 100644 tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.sql diff --git a/src/Parsers/Access/ParserCreateUserQuery.cpp b/src/Parsers/Access/ParserCreateUserQuery.cpp index 9e32b3c4618..ed6ecb62667 100644 --- a/src/Parsers/Access/ParserCreateUserQuery.cpp +++ b/src/Parsers/Access/ParserCreateUserQuery.cpp @@ -295,11 +295,11 @@ namespace } - bool parseHosts(IParserBase::Pos & pos, Expected & expected, const String & prefix, AllowedClientHosts & hosts) + bool parseHosts(IParserBase::Pos & pos, Expected & expected, std::string_view prefix, AllowedClientHosts & hosts) { return IParserBase::wrapParseImpl(pos, [&] { - if (!prefix.empty() && !ParserKeyword{prefix.c_str()}.ignore(pos, expected)) + if (!prefix.empty() && !ParserKeyword{prefix}.ignore(pos, expected)) return false; if (!ParserKeyword{"HOST"}.ignore(pos, expected)) @@ -492,7 +492,6 @@ bool ParserCreateUserQuery::parseImpl(Pos & pos, ASTPtr & node, Expected & expec if (alter) { - String maybe_new_name; if (!new_name && (names->size() == 1) && parseRenameTo(pos, expected, new_name)) continue; diff --git a/tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.reference b/tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.reference new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.sql b/tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.sql new file mode 100644 index 00000000000..3ef1469cf1b --- /dev/null +++ b/tests/queries/0_stateless/02474_create_user_query_fuzzer_bug.sql @@ -0,0 +1 @@ +EXPLAIN AST ALTER user WITH a; -- { clientError SYNTAX_ERROR } From f53df7870c50dab252969a7bf37e3f776686888d Mon Sep 17 00:00:00 2001 From: Alexander Tokmakov Date: Sun, 30 Oct 2022 17:59:47 +0100 Subject: [PATCH 175/239] fix race between drop and failed insert --- src/Storages/MergeTree/MergeTreeData.cpp | 25 +++++++++++++++++++++--- src/Storages/MergeTree/MergeTreeData.h | 3 +++ 2 files changed, 25 insertions(+), 3 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp index d38c9c465d5..a9e726f25fd 100644 --- a/src/Storages/MergeTree/MergeTreeData.cpp +++ b/src/Storages/MergeTree/MergeTreeData.cpp @@ -2199,6 +2199,7 @@ void MergeTreeData::dropAllData() LOG_TRACE(log, "dropAllData: removing all data parts from memory."); data_parts_indexes.clear(); + all_data_dropped = true; } catch (...) { @@ -5181,9 +5182,27 @@ void MergeTreeData::Transaction::rollback() buf << "."; LOG_DEBUG(data.log, "Undoing transaction.{}", buf.str()); - data.removePartsFromWorkingSet(txn, - DataPartsVector(precommitted_parts.begin(), precommitted_parts.end()), - /* clear_without_timeout = */ true); + auto lock = data.lockParts(); + + if (data.data_parts_indexes.empty()) + { + /// Table was dropped concurrently and all parts (including PreActive parts) were cleared, so there's nothing to rollback + if (!data.all_data_dropped) + { + Strings part_names; + for (const auto & part : precommitted_parts) + part_names.emplace_back(part->name); + throw Exception(ErrorCodes::LOGICAL_ERROR, "There are some PreActive parts ({}) to rollback, " + "but data parts set is empty and table {} was not dropped. It's a bug", + fmt::join(part_names, ", "), data.getStorageID().getNameForLogs()); + } + } + else + { + data.removePartsFromWorkingSet(txn, + DataPartsVector(precommitted_parts.begin(), precommitted_parts.end()), + /* clear_without_timeout = */ true, &lock); + } } clear(); diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 60255ce1b16..40eaa679845 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -663,6 +663,9 @@ public: /// Deletes the data directory and flushes the uncompressed blocks cache and the marks cache. void dropAllData(); + /// This flag is for hardening and assertions. + bool all_data_dropped = false; + /// Drop data directories if they are empty. It is safe to call this method if table creation was unsuccessful. void dropIfEmpty(); From 620caeb07c9a09f855c3deb65e30de2674b4504f Mon Sep 17 00:00:00 2001 From: Gabriel Date: Mon, 31 Oct 2022 13:20:58 +0800 Subject: [PATCH 176/239] Fix typo in comments --- src/Processors/Executors/ExecutingGraph.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Processors/Executors/ExecutingGraph.cpp b/src/Processors/Executors/ExecutingGraph.cpp index 651ede10cfd..c7e89d7547c 100644 --- a/src/Processors/Executors/ExecutingGraph.cpp +++ b/src/Processors/Executors/ExecutingGraph.cpp @@ -71,7 +71,7 @@ bool ExecutingGraph::addEdges(uint64_t node) } } - /// Add direct edges form output ports. + /// Add direct edges from output ports. auto & outputs = from->getOutputs(); auto from_output = nodes[node]->direct_edges.size(); From 32febf5155d55bb63b46b2f04d2ea4a91b85c5d2 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 31 Oct 2022 07:56:59 +0100 Subject: [PATCH 177/239] Remove dead code in LowerUpperUTF8Impl::array() AFAICS it was there before since it was possible to overrun the expected_end, since utf8.convert() was called with "src_end - src" not "expected_end - src". Refs: 5a21f3908b054a0efc90c65a12fbe151c74d90dc:dbms/include/DB/Functions/FunctionsString.h Signed-off-by: Azat Khuzhin --- src/Functions/LowerUpperUTF8Impl.h | 10 ---------- 1 file changed, 10 deletions(-) diff --git a/src/Functions/LowerUpperUTF8Impl.h b/src/Functions/LowerUpperUTF8Impl.h index 3a050e2bd6f..1f45772c84a 100644 --- a/src/Functions/LowerUpperUTF8Impl.h +++ b/src/Functions/LowerUpperUTF8Impl.h @@ -230,16 +230,6 @@ private: while (src < expected_end) toCase(src, expected_end, dst); - - /// adjust src_end_sse by pushing it forward or backward - const auto diff = src - expected_end; - if (diff != 0) - { - if (src_end_sse + diff < src_end) - src_end_sse += diff; - else - src_end_sse -= bytes_sse - diff; - } } } From 5fe44f27368fb27e40fc46ce957ae0da5f37e8f4 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Sun, 30 Oct 2022 08:50:05 +0100 Subject: [PATCH 178/239] Fix lowerUTF8()/upperUTF8() in case of symbol was in between 16-byte boundary MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In lowerUTF8()/upperUTF8() there is an SSE optimization that handles 16 byte at a time, but only for ASCII, for UTF8 symbols converion will be done by symbol. Consider the following example: КВ АМ И СЖ ^ - offset is 15, length of sequence is 2 so first byte of a symbol is in first 16 bytes second byte of a symbol is not ther And in this case it will be handled incorrectly because it will try to process oly these 16 bytes w/o looking forward. This had been broken by #41286, before this patch it does not looks at the row boundaries but only at the string end and so this sutation wasn't possible. Fixes: #42756 Signed-off-by: Azat Khuzhin --- src/Functions/LowerUpperUTF8Impl.h | 22 +++++++++++++++---- ...71_lower_upper_utf8_row_overlaps.reference | 4 ++++ .../02071_lower_upper_utf8_row_overlaps.sql | 4 ++++ 3 files changed, 26 insertions(+), 4 deletions(-) diff --git a/src/Functions/LowerUpperUTF8Impl.h b/src/Functions/LowerUpperUTF8Impl.h index 1f45772c84a..b8fd20d56de 100644 --- a/src/Functions/LowerUpperUTF8Impl.h +++ b/src/Functions/LowerUpperUTF8Impl.h @@ -104,7 +104,7 @@ struct LowerUpperUTF8Impl /** Converts a single code point starting at `src` to desired case, storing result starting at `dst`. * `src` and `dst` are incremented by corresponding sequence lengths. */ - static void toCase(const UInt8 *& src, const UInt8 * src_end, UInt8 *& dst) + static bool toCase(const UInt8 *& src, const UInt8 * src_end, UInt8 *& dst, bool partial) { if (src[0] <= ascii_upper_bound) { @@ -136,6 +136,11 @@ struct LowerUpperUTF8Impl static const Poco::UTF8Encoding utf8; size_t src_sequence_length = UTF8::seqLength(*src); + /// In case partial buffer was passed (due to SSE optimization) + /// we cannot convert it with current src_end, but we may have more + /// bytes to convert and eventually got correct symbol. + if (partial && src_sequence_length > static_cast(src_end-src)) + return false; auto src_code_point = UTF8::convertUTF8ToCodePoint(src, src_end - src); if (src_code_point) @@ -152,7 +157,7 @@ struct LowerUpperUTF8Impl { src += dst_sequence_length; dst += dst_sequence_length; - return; + return true; } } } @@ -161,6 +166,8 @@ struct LowerUpperUTF8Impl ++dst; ++src; } + + return true; } private: @@ -229,7 +236,14 @@ private: const UInt8 * expected_end = std::min(src + bytes_sse, row_end); while (src < expected_end) - toCase(src, expected_end, dst); + { + if (!toCase(src, expected_end, dst, /* partial= */ true)) + { + /// Fallback to handling byte by byte. + src_end_sse = src; + break; + } + } } } @@ -245,7 +259,7 @@ private: chassert(row_end >= src); while (src < row_end) - toCase(src, row_end, dst); + toCase(src, row_end, dst, /* partial= */ false); ++offset_it; } } diff --git a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference index 2b3f8138c2b..190c7895011 100644 --- a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference +++ b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference @@ -9,3 +9,7 @@ select length(str), if(l_ == '\xe2', h_, l_), if(u_ == '\xe2', h_, u_) from utf8 15,"foo⚊barbazbam","FOO⚊BARBAZBAM" 1,"0xE2","0xE2" 15,"foo⚊barbazbam","FOO⚊BARBAZBAM" +-- NOTE: regression test for introduced bug +-- https://github.com/ClickHouse/ClickHouse/issues/42756 +SELECT lowerUTF8('КВ АМ И СЖ'); +кв ам и сж diff --git a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql index ee0d29be699..161d8fb512e 100644 --- a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql +++ b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql @@ -8,3 +8,7 @@ insert into utf8_overlap values ('\xe2'), ('Foo⚊BarBazBam'), ('\xe2'), ('Foo -- MONOGRAM FOR YANG with lowerUTF8(str) as l_, upperUTF8(str) as u_, '0x' || hex(str) as h_ select length(str), if(l_ == '\xe2', h_, l_), if(u_ == '\xe2', h_, u_) from utf8_overlap format CSV; + +-- NOTE: regression test for introduced bug +-- https://github.com/ClickHouse/ClickHouse/issues/42756 +SELECT lowerUTF8('КВ АМ И СЖ'); From b62170426a43702bc5f9ad8854a9b468fe4788b3 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 31 Oct 2022 10:48:37 +0100 Subject: [PATCH 179/239] Fix compilation of LLVM with cmake cache Simple reproducer: $ cmake $ ninja contrib/llvm-project/llvm/lib/MC/MCParser/CMakeFiles/LLVMMCParser.dir/MasmParser.cpp.o # will have -std=c++14 $ touch CMakeLists.txt $ cmake $ ninja contrib/llvm-project/llvm/lib/MC/MCParser/CMakeFiles/LLVMMCParser.dir/MasmParser.cpp.o # will have -std=c++20 and fail (fails because std::vector cannot work with opaque types anymore) Fixes: #42249 (cc @rschu1ze) --- contrib/llvm-project-cmake/CMakeLists.txt | 3 +++ 1 file changed, 3 insertions(+) diff --git a/contrib/llvm-project-cmake/CMakeLists.txt b/contrib/llvm-project-cmake/CMakeLists.txt index 6a73ae0f0c6..7af4a23bc9d 100644 --- a/contrib/llvm-project-cmake/CMakeLists.txt +++ b/contrib/llvm-project-cmake/CMakeLists.txt @@ -21,6 +21,9 @@ set (LLVM_INCLUDE_DIRS "${ClickHouse_BINARY_DIR}/contrib/llvm-project/llvm/include" ) set (LLVM_LIBRARY_DIRS "${ClickHouse_BINARY_DIR}/contrib/llvm-project/llvm") +# NOTE: You should not remove this line since otherwise it will use default 20, +# and llvm cannot be compiled with bundled libcxx and 20 standard. +set (CMAKE_CXX_STANDARD 14) # This list was generated by listing all LLVM libraries, compiling the binary and removing all libraries while it still compiles. set (REQUIRED_LLVM_LIBRARIES From 115fcaffc558c89cb95bf7fceaee1322f2bb0707 Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Mon, 31 Oct 2022 11:30:56 +0100 Subject: [PATCH 180/239] Fixed tests --- .../MergeTree/MergeTreeIndexConditionBloomFilter.cpp | 7 +++++++ src/Storages/MergeTree/RPNBuilder.cpp | 2 +- 2 files changed, 8 insertions(+), 1 deletion(-) diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp index 1413f084536..9856500c65c 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp @@ -280,6 +280,13 @@ bool MergeTreeIndexConditionBloomFilter::traverseFunction(const RPNBuilderTreeNo auto arguments_size = function.getArgumentsSize(); auto function_name = function.getFunctionName(); + for (size_t i = 0; i < arguments_size; ++i) + { + auto argument = function.getArgumentAt(i); + if (traverseFunction(argument, out, &node)) + maybe_useful = true; + } + if (arguments_size != 2) return false; diff --git a/src/Storages/MergeTree/RPNBuilder.cpp b/src/Storages/MergeTree/RPNBuilder.cpp index 4cc311b8b9e..395daab76ba 100644 --- a/src/Storages/MergeTree/RPNBuilder.cpp +++ b/src/Storages/MergeTree/RPNBuilder.cpp @@ -361,7 +361,7 @@ ConstSetPtr RPNBuilderTreeNode::tryGetPreparedSet( RPNBuilderFunctionTreeNode RPNBuilderTreeNode::toFunctionNode() const { if (!isFunction()) - throw Exception(ErrorCodes::LOGICAL_ERROR, "RPNBuilderTree node is not a constant"); + throw Exception(ErrorCodes::LOGICAL_ERROR, "RPNBuilderTree node is not a function"); if (this->ast_node) return RPNBuilderFunctionTreeNode(this->ast_node, tree_context); From d760c78754d8e6ff133496dfa1a48a6c6a7683ae Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Mon, 31 Oct 2022 11:45:12 +0100 Subject: [PATCH 181/239] Fixed code review issues --- src/Analyzer/Passes/QueryAnalysisPass.cpp | 16 +-------------- src/Analyzer/QueryTreeBuilder.cpp | 16 +-------------- src/Analyzer/TableExpressionModifiers.cpp | 24 +++++++++++++++++++++++ src/Analyzer/TableExpressionModifiers.h | 3 +++ src/Analyzer/UnionNode.cpp | 2 +- 5 files changed, 30 insertions(+), 31 deletions(-) diff --git a/src/Analyzer/Passes/QueryAnalysisPass.cpp b/src/Analyzer/Passes/QueryAnalysisPass.cpp index 019b002c527..9db2d66d99d 100644 --- a/src/Analyzer/Passes/QueryAnalysisPass.cpp +++ b/src/Analyzer/Passes/QueryAnalysisPass.cpp @@ -4646,23 +4646,9 @@ void QueryAnalyzer::initializeQueryJoinTreeNode(QueryTreeNodePtr & join_tree_nod if (table_expression_modifiers.has_value()) { - String table_expression_modifiers_error_message; - - if (table_expression_modifiers->hasFinal()) - { - table_expression_modifiers_error_message += "FINAL"; - - if (table_expression_modifiers->hasSampleSizeRatio()) - table_expression_modifiers_error_message += ", SAMPLE"; - } - else if (table_expression_modifiers->hasSampleSizeRatio()) - { - table_expression_modifiers_error_message += "SAMPLE"; - } - throw Exception(ErrorCodes::UNSUPPORTED_METHOD, "Table expression modifiers {} are not supported for subquery {}", - table_expression_modifiers_error_message, + table_expression_modifiers->formatForErrorMessage(), resolved_identifier->formatASTForErrorMessage()); } } diff --git a/src/Analyzer/QueryTreeBuilder.cpp b/src/Analyzer/QueryTreeBuilder.cpp index c5bb3ae7074..51745d820e7 100644 --- a/src/Analyzer/QueryTreeBuilder.cpp +++ b/src/Analyzer/QueryTreeBuilder.cpp @@ -673,23 +673,9 @@ QueryTreeNodePtr QueryTreeBuilder::buildJoinTree(const ASTPtr & tables_in_select if (table_expression_modifiers) { - String table_expression_modifiers_error_message; - - if (table_expression_modifiers->hasFinal()) - { - table_expression_modifiers_error_message += "FINAL"; - - if (table_expression_modifiers->hasSampleSizeRatio()) - table_expression_modifiers_error_message += ", SAMPLE"; - } - else if (table_expression_modifiers->hasSampleSizeRatio()) - { - table_expression_modifiers_error_message += "SAMPLE"; - } - throw Exception(ErrorCodes::UNSUPPORTED_METHOD, "Table expression modifiers {} are not supported for subquery {}", - table_expression_modifiers_error_message, + table_expression_modifiers->formatForErrorMessage(), node->formatASTForErrorMessage()); } diff --git a/src/Analyzer/TableExpressionModifiers.cpp b/src/Analyzer/TableExpressionModifiers.cpp index 79b5a8dba41..c8002f44c97 100644 --- a/src/Analyzer/TableExpressionModifiers.cpp +++ b/src/Analyzer/TableExpressionModifiers.cpp @@ -5,6 +5,7 @@ #include #include #include +#include namespace DB { @@ -39,4 +40,27 @@ void TableExpressionModifiers::updateTreeHash(SipHash & hash_state) const } } +String TableExpressionModifiers::formatForErrorMessage() const +{ + WriteBufferFromOwnString buffer; + if (has_final) + buffer << "FINAL"; + + if (sample_size_ratio) + { + if (has_final) + buffer << ' '; + buffer << "SAMPLE " << ASTSampleRatio::toString(*sample_size_ratio); + } + + if (sample_offset_ratio) + { + if (has_final || sample_size_ratio) + buffer << ' '; + buffer << "OFFSET " << ASTSampleRatio::toString(*sample_offset_ratio); + } + + return buffer.str(); +} + } diff --git a/src/Analyzer/TableExpressionModifiers.h b/src/Analyzer/TableExpressionModifiers.h index cc5ac3948bf..f61c2a61610 100644 --- a/src/Analyzer/TableExpressionModifiers.h +++ b/src/Analyzer/TableExpressionModifiers.h @@ -58,6 +58,9 @@ public: /// Update tree hash void updateTreeHash(SipHash & hash_state) const; + /// Format for error message + String formatForErrorMessage() const; + private: bool has_final = false; std::optional sample_size_ratio; diff --git a/src/Analyzer/UnionNode.cpp b/src/Analyzer/UnionNode.cpp index 52ad0102d7c..67860438335 100644 --- a/src/Analyzer/UnionNode.cpp +++ b/src/Analyzer/UnionNode.cpp @@ -40,7 +40,7 @@ UnionNode::UnionNode(SelectUnionMode union_mode_) if (union_mode == SelectUnionMode::UNION_DEFAULT || union_mode == SelectUnionMode::EXCEPT_DEFAULT || union_mode == SelectUnionMode::INTERSECT_DEFAULT) - throw Exception(ErrorCodes::BAD_ARGUMENTS, "UNION mode must be normalized"); + throw Exception(ErrorCodes::BAD_ARGUMENTS, "UNION mode {} must be normalized", toString(union_mode)); children[queries_child_index] = std::make_shared(); } From f6d0c03ee5dcdf703ca63ebb781f99a22d566a9e Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 19:39:02 +0200 Subject: [PATCH 182/239] Fix path to files in MetadataStorageFromPlainObjectStorage Signed-off-by: Azat Khuzhin --- .../MetadataStorageFromPlainObjectStorage.cpp | 21 ++++++++++--------- .../MetadataStorageFromPlainObjectStorage.h | 3 +++ 2 files changed, 14 insertions(+), 10 deletions(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index 35cd3be15d2..f5b1b4d287a 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -33,10 +33,14 @@ const std::string & MetadataStorageFromPlainObjectStorage::getPath() const { return object_storage_root_path; } +std::filesystem::path MetadataStorageFromPlainObjectStorage::getAbsolutePath(const std::string & path) const +{ + return fs::path(object_storage_root_path) / path; +} bool MetadataStorageFromPlainObjectStorage::exists(const std::string & path) const { - auto object = StoredObject::create(*object_storage, fs::path(object_storage_root_path) / path); + auto object = StoredObject::create(*object_storage, getAbsolutePath(path)); return object_storage->exists(object); } @@ -48,7 +52,7 @@ bool MetadataStorageFromPlainObjectStorage::isFile(const std::string & path) con bool MetadataStorageFromPlainObjectStorage::isDirectory(const std::string & path) const { - std::string directory = path; + std::string directory = getAbsolutePath(path); trimRight(directory); directory += "/"; @@ -77,7 +81,7 @@ time_t MetadataStorageFromPlainObjectStorage::getLastChanged(const std::string & uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) const { RelativePathsWithSize children; - object_storage->listPrefix(path, children); + object_storage->listPrefix(getAbsolutePath(path), children); if (children.empty()) return 0; if (children.size() != 1) @@ -88,7 +92,7 @@ uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) std::vector MetadataStorageFromPlainObjectStorage::listDirectory(const std::string & path) const { RelativePathsWithSize children; - object_storage->listPrefix(path, children); + object_storage->listPrefix(getAbsolutePath(path), children); std::vector result; for (const auto & path_size : children) @@ -120,11 +124,8 @@ std::unordered_map MetadataStorageFromPlainObjectStorage::getSer StoredObjects MetadataStorageFromPlainObjectStorage::getStorageObjects(const std::string & path) const { std::string blob_name = object_storage->generateBlobNameForPath(path); - - std::string object_path = fs::path(object_storage_root_path) / blob_name; - size_t object_size = getFileSize(object_path); - - auto object = StoredObject::create(*object_storage, object_path, object_size, /* exists */true); + size_t object_size = getFileSize(blob_name); + auto object = StoredObject::create(*object_storage, getAbsolutePath(blob_name), object_size, /* exists */true); return {std::move(object)}; } @@ -150,7 +151,7 @@ void MetadataStorageFromPlainObjectStorageTransaction::setLastModified(const std void MetadataStorageFromPlainObjectStorageTransaction::unlinkFile(const std::string & path) { - auto object = StoredObject::create(*metadata_storage.object_storage, fs::path(metadata_storage.object_storage_root_path) / path); + auto object = StoredObject::create(*metadata_storage.object_storage, metadata_storage.getAbsolutePath(path)); metadata_storage.object_storage->removeObject(object); } diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h index bd993918413..c58ee17b495 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h @@ -70,6 +70,9 @@ public: StoredObjects getStorageObjects(const std::string & path) const override; std::string getObjectStorageRootPath() const override { return object_storage_root_path; } + +private: + std::filesystem::path getAbsolutePath(const std::string & path) const; }; class MetadataStorageFromPlainObjectStorageTransaction final : public IMetadataTransaction From b34ffda272fe9be0a014eb71ff6053f537911150 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 19:39:34 +0200 Subject: [PATCH 183/239] Implement MetadataStorageFromPlainObjectStorage::getLastModified() (used by MergeTree) Signed-off-by: Azat Khuzhin --- .../ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index f5b1b4d287a..f99f4eb21f3 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -65,7 +65,8 @@ bool MetadataStorageFromPlainObjectStorage::isDirectory(const std::string & path Poco::Timestamp MetadataStorageFromPlainObjectStorage::getLastModified(const std::string &) const { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "getLastModified is not implemented for MetadataStorageFromPlainObjectStorage"); + /// NOTE: This is required for MergeTree + return {}; } struct stat MetadataStorageFromPlainObjectStorage::stat(const std::string &) const From 94d9600fb813ea25f892a461f8ffc3c632123318 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 19:39:50 +0200 Subject: [PATCH 184/239] Implement MetadataStorageFromPlainObjectStorage::getLastChanged() (as for web) Signed-off-by: Azat Khuzhin --- .../ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index f99f4eb21f3..dda7ed98837 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -76,7 +76,8 @@ struct stat MetadataStorageFromPlainObjectStorage::stat(const std::string &) con time_t MetadataStorageFromPlainObjectStorage::getLastChanged(const std::string &) const { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "getLastChanged is not implemented for MetadataStorageFromPlainObjectStorage"); + /// NOTE: by analogy with MetadataStorageFromStaticFilesWebServer::getLastChanged() + return {}; } uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) const From 88db8ae7fac9c09cf4e92fc3a9a55d45887682b9 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 20:05:24 +0200 Subject: [PATCH 185/239] Add a comment for IObjectStorage::listPrefix() Signed-off-by: Azat Khuzhin --- src/Disks/ObjectStorages/IObjectStorage.h | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/src/Disks/ObjectStorages/IObjectStorage.h b/src/Disks/ObjectStorages/IObjectStorage.h index 52e1a2cb270..ec696726ed2 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.h +++ b/src/Disks/ObjectStorages/IObjectStorage.h @@ -66,6 +66,13 @@ public: virtual bool exists(const StoredObject & object) const = 0; /// List on prefix, return children (relative paths) with their sizes. + /// So it should return all "objects" with their sizes. + /// + /// For example if you do this over filesystem, you should skip folders and return files only. + /// + /// NOTE: It makes sense only for real object storages (S3, Azure), since + /// this is used for restoring metadata (see "send_metadata" and + /// DiskObjectStorage::restoreMetadataIfNeeded()). virtual void listPrefix(const std::string & path, RelativePathsWithSize & children) const = 0; /// Get object metadata if supported. It should be possible to receive From ee18140c4843fddd067255b8848a2ef8dc5c5041 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 20:06:18 +0200 Subject: [PATCH 186/239] Remove listPrefix() implementation for disks that does not support send_metadata The reason for removing is it because not compatible with restoring (with send_metadata set) anyway: - HDFS - is not compatible with send_metadata, and besides it's implementaion is not correct, since it is simply `ls -l`, while the following is required: `find . -maxdepth 1 -type f` - Web - is not compatible with send_metadata anyway - Local - is not compatible with send_metadata anyway Signed-off-by: Azat Khuzhin --- src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.cpp | 12 ------------ src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.h | 2 -- src/Disks/ObjectStorages/IObjectStorage.cpp | 5 +++++ src/Disks/ObjectStorages/IObjectStorage.h | 13 +++++++++---- src/Disks/ObjectStorages/LocalObjectStorage.cpp | 7 ------- src/Disks/ObjectStorages/LocalObjectStorage.h | 2 -- src/Disks/ObjectStorages/Web/WebObjectStorage.cpp | 11 ----------- src/Disks/ObjectStorages/Web/WebObjectStorage.h | 2 -- 8 files changed, 14 insertions(+), 40 deletions(-) diff --git a/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.cpp b/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.cpp index 2f82458ecd8..80c4bb2bc64 100644 --- a/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.cpp +++ b/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.cpp @@ -101,18 +101,6 @@ std::unique_ptr HDFSObjectStorage::writeObject( /// NOL } -void HDFSObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize & children) const -{ - const size_t begin_of_path = path.find('/', path.find("//") + 2); - int32_t num_entries; - auto * files_list = hdfsListDirectory(hdfs_fs.get(), path.substr(begin_of_path).c_str(), &num_entries); - if (num_entries == -1) - throw Exception(ErrorCodes::HDFS_ERROR, "HDFSDelete failed with path: " + path); - - for (int32_t i = 0; i < num_entries; ++i) - children.emplace_back(files_list[i].mName, files_list[i].mSize); -} - /// Remove file. Throws exception if file doesn't exists or it's a directory. void HDFSObjectStorage::removeObject(const StoredObject & object) { diff --git a/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.h b/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.h index 82cddfb9122..4064a5c5b7f 100644 --- a/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.h +++ b/src/Disks/ObjectStorages/HDFS/HDFSObjectStorage.h @@ -85,8 +85,6 @@ public: size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE, const WriteSettings & write_settings = {}) override; - void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; - /// Remove file. Throws exception if file doesn't exists or it's a directory. void removeObject(const StoredObject & object) override; diff --git a/src/Disks/ObjectStorages/IObjectStorage.cpp b/src/Disks/ObjectStorages/IObjectStorage.cpp index 9d6610ee326..7f1cdcbfa59 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.cpp +++ b/src/Disks/ObjectStorages/IObjectStorage.cpp @@ -14,6 +14,11 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } +void IObjectStorage::listPrefix(const std::string &, RelativePathsWithSize &) const +{ + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "listPrefix() is not supported"); +} + IAsynchronousReader & IObjectStorage::getThreadPoolReader() { auto context = Context::getGlobalContextInstance(); diff --git a/src/Disks/ObjectStorages/IObjectStorage.h b/src/Disks/ObjectStorages/IObjectStorage.h index ec696726ed2..39819e2e58c 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.h +++ b/src/Disks/ObjectStorages/IObjectStorage.h @@ -68,12 +68,17 @@ public: /// List on prefix, return children (relative paths) with their sizes. /// So it should return all "objects" with their sizes. /// - /// For example if you do this over filesystem, you should skip folders and return files only. + /// For example if you do this over filesystem, you should skip folders and + /// return files only, so something like on local filesystem: + /// + /// find . -type f /// /// NOTE: It makes sense only for real object storages (S3, Azure), since - /// this is used for restoring metadata (see "send_metadata" and - /// DiskObjectStorage::restoreMetadataIfNeeded()). - virtual void listPrefix(const std::string & path, RelativePathsWithSize & children) const = 0; + /// it is used only for one of the following: + /// - send_metadata (to restore metadata) + /// - see DiskObjectStorage::restoreMetadataIfNeeded() + /// - MetadataStorageFromPlainObjectStorage - only for s3_plain disk + virtual void listPrefix(const std::string & path, RelativePathsWithSize & children) const; /// Get object metadata if supported. It should be possible to receive /// at least size of object diff --git a/src/Disks/ObjectStorages/LocalObjectStorage.cpp b/src/Disks/ObjectStorages/LocalObjectStorage.cpp index dbb3a7c2aba..67e2cc2d74b 100644 --- a/src/Disks/ObjectStorages/LocalObjectStorage.cpp +++ b/src/Disks/ObjectStorages/LocalObjectStorage.cpp @@ -104,13 +104,6 @@ std::unique_ptr LocalObjectStorage::writeObject( /// NO return std::make_unique(path, buf_size, flags); } -void LocalObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize & children) const -{ - fs::directory_iterator end_it; - for (auto it = fs::directory_iterator(path); it != end_it; ++it) - children.emplace_back(it->path().filename(), it->file_size()); -} - void LocalObjectStorage::removeObject(const StoredObject & object) { /// For local object storage files are actually removed when "metadata" is removed. diff --git a/src/Disks/ObjectStorages/LocalObjectStorage.h b/src/Disks/ObjectStorages/LocalObjectStorage.h index 0e4c71b4a47..b04e3fa6285 100644 --- a/src/Disks/ObjectStorages/LocalObjectStorage.h +++ b/src/Disks/ObjectStorages/LocalObjectStorage.h @@ -45,8 +45,6 @@ public: size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE, const WriteSettings & write_settings = {}) override; - void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; - void removeObject(const StoredObject & object) override; void removeObjects(const StoredObjects & objects) override; diff --git a/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp b/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp index 71bde110fa6..f97409cfc6c 100644 --- a/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp +++ b/src/Disks/ObjectStorages/Web/WebObjectStorage.cpp @@ -178,17 +178,6 @@ std::unique_ptr WebObjectStorage::readObject( /// NOLINT } } -void WebObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize & children) const -{ - for (const auto & [file_path, file_info] : files) - { - if (file_info.type == FileType::File && file_path.starts_with(path)) - { - children.emplace_back(file_path, file_info.size); - } - } -} - void WebObjectStorage::throwNotAllowed() { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Only read-only operations are supported"); diff --git a/src/Disks/ObjectStorages/Web/WebObjectStorage.h b/src/Disks/ObjectStorages/Web/WebObjectStorage.h index 2fda5e576aa..2dab8fdb62d 100644 --- a/src/Disks/ObjectStorages/Web/WebObjectStorage.h +++ b/src/Disks/ObjectStorages/Web/WebObjectStorage.h @@ -55,8 +55,6 @@ public: size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE, const WriteSettings & write_settings = {}) override; - void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; - void removeObject(const StoredObject & object) override; void removeObjects(const StoredObjects & objects) override; From 18e4fdf40f47cb5ad5cd608e4d0dabdfd6529bdb Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 20:32:18 +0200 Subject: [PATCH 187/239] Introduce IObjectStorage::listPrefixInPath() Signed-off-by: Azat Khuzhin --- src/Disks/ObjectStorages/IObjectStorage.cpp | 6 ++++++ src/Disks/ObjectStorages/IObjectStorage.h | 12 ++++++++++++ 2 files changed, 18 insertions(+) diff --git a/src/Disks/ObjectStorages/IObjectStorage.cpp b/src/Disks/ObjectStorages/IObjectStorage.cpp index 7f1cdcbfa59..7c187e3dd2f 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.cpp +++ b/src/Disks/ObjectStorages/IObjectStorage.cpp @@ -18,6 +18,12 @@ void IObjectStorage::listPrefix(const std::string &, RelativePathsWithSize &) co { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "listPrefix() is not supported"); } +void IObjectStorage::listPrefixInPath(const std::string &, + RelativePathsWithSize &, + std::vector &) const +{ + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "listPrefixInPath() is not supported"); +} IAsynchronousReader & IObjectStorage::getThreadPoolReader() { diff --git a/src/Disks/ObjectStorages/IObjectStorage.h b/src/Disks/ObjectStorages/IObjectStorage.h index 39819e2e58c..a7ad4e5b515 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.h +++ b/src/Disks/ObjectStorages/IObjectStorage.h @@ -80,6 +80,18 @@ public: /// - MetadataStorageFromPlainObjectStorage - only for s3_plain disk virtual void listPrefix(const std::string & path, RelativePathsWithSize & children) const; + /// List on prefix, but it is limited only to current path, something like + /// on local filesystem: + /// + /// find . -maxdepth 1 + /// + /// @param path - path to list prefix in + /// @param children - list of child nodes (for unix filesystem - files goes here) + /// @param common_prefixes - list of common prefixes in this path, (for unix filesystem - subfolders) + virtual void listPrefixInPath(const std::string & path, + RelativePathsWithSize & children, + std::vector & common_prefixes) const; + /// Get object metadata if supported. It should be possible to receive /// at least size of object virtual ObjectMetadata getObjectMetadata(const std::string & path) const = 0; From f8ba24f04046f06af42e756560b10cb8884299d3 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 20:33:12 +0200 Subject: [PATCH 188/239] Implement S3ObjectStorage::listPrefixInPath() Signed-off-by: Azat Khuzhin --- .../ObjectStorages/S3/S3ObjectStorage.cpp | 45 ++++++++++++++++++- src/Disks/ObjectStorages/S3/S3ObjectStorage.h | 3 ++ 2 files changed, 47 insertions(+), 1 deletion(-) diff --git a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp index 83908d02f48..af3374f8d3c 100644 --- a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp +++ b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp @@ -28,7 +28,7 @@ #include #include - +#include #include #include @@ -279,6 +279,49 @@ void S3ObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize } while (outcome.GetResult().GetIsTruncated()); } +void S3ObjectStorage::listPrefixInPath(const std::string & path, + RelativePathsWithSize & children, + std::vector & common_prefixes) const +{ + auto settings_ptr = s3_settings.get(); + auto client_ptr = client.get(); + + Aws::S3::Model::ListObjectsV2Request request; + request.SetBucket(bucket); + request.SetPrefix(path); + request.SetMaxKeys(settings_ptr->list_object_keys_size); + request.SetDelimiter("/"); + + Aws::S3::Model::ListObjectsV2Outcome outcome; + do + { + ProfileEvents::increment(ProfileEvents::S3ListObjects); + ProfileEvents::increment(ProfileEvents::DiskS3ListObjects); + outcome = client_ptr->ListObjectsV2(request); + throwIfError(outcome); + + auto result = outcome.GetResult(); + auto result_objects = result.GetContents(); + auto result_common_prefixes = result.GetCommonPrefixes(); + + if (result_objects.empty() && result_common_prefixes.empty()) + break; + + for (const auto & object : result_objects) + children.emplace_back(object.GetKey(), object.GetSize()); + + for (const auto & common_prefix : result_common_prefixes) + { + std::string subfolder = common_prefix.GetPrefix(); + /// Make it compatible with std::filesystem::path::filename() + trimRight(subfolder, '/'); + common_prefixes.emplace_back(subfolder); + } + + request.SetContinuationToken(outcome.GetResult().GetNextContinuationToken()); + } while (outcome.GetResult().GetIsTruncated()); +} + void S3ObjectStorage::removeObjectImpl(const StoredObject & object, bool if_exists) { auto client_ptr = client.get(); diff --git a/src/Disks/ObjectStorages/S3/S3ObjectStorage.h b/src/Disks/ObjectStorages/S3/S3ObjectStorage.h index a193653db9a..9e90efcb9dc 100644 --- a/src/Disks/ObjectStorages/S3/S3ObjectStorage.h +++ b/src/Disks/ObjectStorages/S3/S3ObjectStorage.h @@ -106,6 +106,9 @@ public: const WriteSettings & write_settings = {}) override; void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; + void listPrefixInPath(const std::string & path, + RelativePathsWithSize & children, + std::vector & common_prefixes) const override; /// Uses `DeleteObjectRequest`. void removeObject(const StoredObject & object) override; From 4e42521f44116ca5c9275e7f78259611b8405488 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 21:48:27 +0200 Subject: [PATCH 189/239] Introduce ReadOnlyMetadataStorage And use it for: - MetadataStorageFromPlainObjectStorage - MetadataStorageFromStaticFilesWebServer This will allow to reduce ~100-200 lines of duplicated code, and plus make the code less error prone. Note, for now I tried to make this without behaviour changes. Signed-off-by: Azat Khuzhin --- .../MetadataStorageFromPlainObjectStorage.cpp | 97 ---------- .../MetadataStorageFromPlainObjectStorage.h | 54 +----- .../ObjectStorages/ReadOnlyMetadataStorage.h | 165 ++++++++++++++++++ ...etadataStorageFromStaticFilesWebServer.cpp | 107 ------------ .../MetadataStorageFromStaticFilesWebServer.h | 55 +----- 5 files changed, 176 insertions(+), 302 deletions(-) create mode 100644 src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index dda7ed98837..0722c3622b9 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -12,7 +12,6 @@ namespace DB namespace ErrorCodes { - extern const int NOT_IMPLEMENTED; extern const int LOGICAL_ERROR; } @@ -63,23 +62,6 @@ bool MetadataStorageFromPlainObjectStorage::isDirectory(const std::string & path return !children.empty(); } -Poco::Timestamp MetadataStorageFromPlainObjectStorage::getLastModified(const std::string &) const -{ - /// NOTE: This is required for MergeTree - return {}; -} - -struct stat MetadataStorageFromPlainObjectStorage::stat(const std::string &) const -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "stat is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -time_t MetadataStorageFromPlainObjectStorage::getLastChanged(const std::string &) const -{ - /// NOTE: by analogy with MetadataStorageFromStaticFilesWebServer::getLastChanged() - return {}; -} - uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) const { RelativePathsWithSize children; @@ -113,16 +95,6 @@ DirectoryIteratorPtr MetadataStorageFromPlainObjectStorage::iterateDirectory(con return std::make_unique(std::move(fs_paths)); } -std::string MetadataStorageFromPlainObjectStorage::readFileToString(const std::string &) const -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "readFileToString is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -std::unordered_map MetadataStorageFromPlainObjectStorage::getSerializedMetadata(const std::vector &) const -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "getSerializedMetadata is not implemented for MetadataStorageFromPlainObjectStorage"); -} - StoredObjects MetadataStorageFromPlainObjectStorage::getStorageObjects(const std::string & path) const { std::string blob_name = object_storage->generateBlobNameForPath(path); @@ -131,99 +103,30 @@ StoredObjects MetadataStorageFromPlainObjectStorage::getStorageObjects(const std return {std::move(object)}; } -uint32_t MetadataStorageFromPlainObjectStorage::getHardlinkCount(const std::string &) const -{ - return 1; -} - const IMetadataStorage & MetadataStorageFromPlainObjectStorageTransaction::getStorageForNonTransactionalReads() const { return metadata_storage; } -void MetadataStorageFromPlainObjectStorageTransaction::writeStringToFile(const std::string &, const std::string & /* data */) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "writeStringToFile is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::setLastModified(const std::string &, const Poco::Timestamp &) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "setLastModified is not implemented for MetadataStorageFromPlainObjectStorage"); -} - void MetadataStorageFromPlainObjectStorageTransaction::unlinkFile(const std::string & path) { auto object = StoredObject::create(*metadata_storage.object_storage, metadata_storage.getAbsolutePath(path)); metadata_storage.object_storage->removeObject(object); } -void MetadataStorageFromPlainObjectStorageTransaction::removeRecursive(const std::string &) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "removeRecursive is not implemented for MetadataStorageFromPlainObjectStorage"); -} - void MetadataStorageFromPlainObjectStorageTransaction::createDirectory(const std::string &) { /// Noop. It is an Object Storage not a filesystem. } - void MetadataStorageFromPlainObjectStorageTransaction::createDirectoryRecursive(const std::string &) { /// Noop. It is an Object Storage not a filesystem. } - -void MetadataStorageFromPlainObjectStorageTransaction::removeDirectory(const std::string &) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "removeDirectory is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::moveFile(const std::string & /* path_from */, const std::string & /* path_to */) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "moveFile is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::moveDirectory(const std::string & /* path_from */, const std::string & /* path_to */) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "moveDirectory is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::replaceFile(const std::string & /* path_from */, const std::string & /* path_to */) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "replaceFile is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::chmod(const String &, mode_t) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "chmod is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::setReadOnly(const std::string &) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "setReadOnly is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::createHardLink(const std::string & /* path_from */, const std::string & /* path_to */) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "createHardLink is not implemented for MetadataStorageFromPlainObjectStorage"); -} - -void MetadataStorageFromPlainObjectStorageTransaction::createEmptyMetadataFile(const std::string &) -{ - /// Noop, no separate metadata. -} - -void MetadataStorageFromPlainObjectStorageTransaction::createMetadataFile( - const std::string &, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) -{ - /// Noop, no separate metadata. -} - void MetadataStorageFromPlainObjectStorageTransaction::addBlobToMetadata( const std::string &, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) { /// Noop, local metadata files is only one file, it is the metadata file itself. } - void MetadataStorageFromPlainObjectStorageTransaction::unlinkMetadata(const std::string &) { /// Noop, no separate metadata. diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h index c58ee17b495..2cfa123ecc9 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h @@ -2,6 +2,7 @@ #include #include +#include #include #include @@ -20,7 +21,10 @@ namespace DB /// It is used to allow BACKUP/RESTORE to ObjectStorage (S3/...) with the same /// structure as on disk MergeTree, and does not requires metadata from local /// disk to restore. -class MetadataStorageFromPlainObjectStorage final : public IMetadataStorage +/// +/// NOTE: Inheritance from ReadOnlyMetadataStorage is used here to throw +/// NOT_IMPLEMENTED error for lost of unsupported methods (mtime/move/stat/...) +class MetadataStorageFromPlainObjectStorage final : public ReadOnlyMetadataStorage { private: friend class MetadataStorageFromPlainObjectStorageTransaction; @@ -45,26 +49,10 @@ public: uint64_t getFileSize(const String & path) const override; - Poco::Timestamp getLastModified(const std::string & path) const override; - - time_t getLastChanged(const std::string & path) const override; - - bool supportsChmod() const override { return false; } - - bool supportsStat() const override { return false; } - - struct stat stat(const String & path) const override; - std::vector listDirectory(const std::string & path) const override; DirectoryIteratorPtr iterateDirectory(const std::string & path) const override; - std::string readFileToString(const std::string & path) const override; - - std::unordered_map getSerializedMetadata(const std::vector & file_paths) const override; - - uint32_t getHardlinkCount(const std::string & path) const override; - DiskPtr getDisk() const { return {}; } StoredObjects getStorageObjects(const std::string & path) const override; @@ -75,7 +63,7 @@ private: std::filesystem::path getAbsolutePath(const std::string & path) const; }; -class MetadataStorageFromPlainObjectStorageTransaction final : public IMetadataTransaction +class MetadataStorageFromPlainObjectStorageTransaction final : public ReadOnlyMetadataTransaction { private: const MetadataStorageFromPlainObjectStorage & metadata_storage; @@ -90,41 +78,13 @@ public: const IMetadataStorage & getStorageForNonTransactionalReads() const final; - void commit() final {} - - void writeStringToFile(const std::string & path, const std::string & data) override; - - void createEmptyMetadataFile(const std::string & path) override; - - void createMetadataFile(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) override; - void addBlobToMetadata(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) override; - void setLastModified(const std::string & path, const Poco::Timestamp & timestamp) override; - - bool supportsChmod() const override { return false; } - - void chmod(const String & path, mode_t mode) override; - - void setReadOnly(const std::string & path) override; - - void unlinkFile(const std::string & path) override; - void createDirectory(const std::string & path) override; void createDirectoryRecursive(const std::string & path) override; - void removeDirectory(const std::string & path) override; - - void removeRecursive(const std::string & path) override; - - void createHardLink(const std::string & path_from, const std::string & path_to) override; - - void moveFile(const std::string & path_from, const std::string & path_to) override; - - void moveDirectory(const std::string & path_from, const std::string & path_to) override; - - void replaceFile(const std::string & path_from, const std::string & path_to) override; + void unlinkFile(const std::string & path) override; void unlinkMetadata(const std::string & path) override; }; diff --git a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h new file mode 100644 index 00000000000..e136b0580ea --- /dev/null +++ b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h @@ -0,0 +1,165 @@ +#pragma once + +#include +#include + +namespace DB +{ + +namespace ErrorCodes +{ + extern const int NOT_IMPLEMENTED; +} + +class ReadOnlyMetadataStorage; + +/// Readonly, throws NOT_IMPLEMENTED error. +/// Can be used to add limited read-only support of MergeTree. +class ReadOnlyMetadataTransaction : public IMetadataTransaction +{ +public: + /// + /// Noop + /// + void commit() override + { + /// Noop, nothing to commit. + } + void createEmptyMetadataFile(const std::string & /* path */) override + { + /// No metadata, no need to create anything. + } + void createMetadataFile(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override + { + /// Noop + } + + /// + /// Throws + /// + void writeStringToFile(const std::string & /* path */, const std::string & /* data */) override + { + throwNotAllowed(); + } + void setLastModified(const std::string & /* path */, const Poco::Timestamp & /* timestamp */) override + { + throwNotAllowed(); + } + void chmod(const String & /* path */, mode_t /* mode */) override + { + throwNotAllowed(); + } + void setReadOnly(const std::string & /* path */) override + { + throwNotAllowed(); + } + void unlinkFile(const std::string & /* path */) override + { + throwNotAllowed(); + } + void removeDirectory(const std::string & /* path */) override + { + throwNotAllowed(); + } + void removeRecursive(const std::string & /* path */) override + { + throwNotAllowed(); + } + void createHardLink(const std::string & /* path_from */, const std::string & /* path_to */) override + { + throwNotAllowed(); + } + void moveFile(const std::string & /* path_from */, const std::string & /* path_to */) override + { + throwNotAllowed(); + } + void moveDirectory(const std::string & /* path_from */, const std::string & /* path_to */) override + { + throwNotAllowed(); + } + void replaceFile(const std::string & /* path_from */, const std::string & /* path_to */) override + { + throwNotAllowed(); + } + void createDirectory(const std::string & /* path */) override + { + throwNotAllowed(); + } + void createDirectoryRecursive(const std::string & /* path */) override + { + throwNotAllowed(); + } + void addBlobToMetadata(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override + { + throwNotAllowed(); + } + void unlinkMetadata(const std::string & /* path */) override + { + throwNotAllowed(); + } + + + /// + /// Others + /// + bool supportsChmod() const override { return false; } + +private: + [[noreturn]] static void throwNotAllowed() + { + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Only read-only transaction operations are supported"); + } +}; + +/// Readonly, throws NOT_IMPLEMENTED error. +/// Can be used to add limited read-only support of MergeTree. +class ReadOnlyMetadataStorage : public IMetadataStorage +{ +public: + /// + /// Noop + /// + Poco::Timestamp getLastModified(const std::string & /* path */) const override + { + /// Required by MergeTree + return {}; + } + time_t getLastChanged(const std::string & /* path */) const override + { + return {}; + } + + struct stat stat(const String & /* path */) const override + { + throwNotAllowed(); + } + + uint32_t getHardlinkCount(const std::string & /* path */) const override + { + return 1; + } + + std::string readFileToString(const std::string & /* path */) const override + { + throwNotAllowed(); + } + + std::unordered_map getSerializedMetadata(const std::vector & /* file_paths */) const override + { + throwNotAllowed(); + } + + /// + /// Others + /// + bool supportsChmod() const override { return false; } + bool supportsStat() const override { return false; } + +private: + [[noreturn]] static void throwNotAllowed() + { + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Only read-only metadata operations are supported"); + } +}; + +} diff --git a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp index aa125e93dee..12c2cd16a9f 100644 --- a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp +++ b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp @@ -12,7 +12,6 @@ namespace DB namespace ErrorCodes { - extern const int NOT_IMPLEMENTED; extern const int FILE_DOESNT_EXIST; extern const int NETWORK_ERROR; } @@ -168,91 +167,11 @@ DirectoryIteratorPtr MetadataStorageFromStaticFilesWebServer::iterateDirectory(c return std::make_unique(std::move(dir_file_paths)); } -std::string MetadataStorageFromStaticFilesWebServer::readFileToString(const std::string &) const -{ - WebObjectStorage::throwNotAllowed(); -} - -Poco::Timestamp MetadataStorageFromStaticFilesWebServer::getLastModified(const std::string &) const -{ - return {}; -} - -time_t MetadataStorageFromStaticFilesWebServer::getLastChanged(const std::string &) const -{ - return {}; -} - -uint32_t MetadataStorageFromStaticFilesWebServer::getHardlinkCount(const std::string &) const -{ - return 1; -} - const IMetadataStorage & MetadataStorageFromStaticFilesWebServerTransaction::getStorageForNonTransactionalReads() const { return metadata_storage; } -void MetadataStorageFromStaticFilesWebServerTransaction::writeStringToFile(const std::string &, const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::setLastModified(const std::string &, const Poco::Timestamp &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::unlinkFile(const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::removeRecursive(const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::removeDirectory(const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::moveFile(const std::string &, const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::moveDirectory(const std::string &, const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::replaceFile(const std::string &, const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::setReadOnly(const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::createHardLink(const std::string &, const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::addBlobToMetadata(const std::string &, const std::string &, uint64_t) -{ - WebObjectStorage::throwNotAllowed(); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::unlinkMetadata(const std::string &) -{ - WebObjectStorage::throwNotAllowed(); -} - void MetadataStorageFromStaticFilesWebServerTransaction::createDirectory(const std::string &) { /// Noop. @@ -263,30 +182,4 @@ void MetadataStorageFromStaticFilesWebServerTransaction::createDirectoryRecursiv /// Noop. } -void MetadataStorageFromStaticFilesWebServerTransaction::createEmptyMetadataFile(const std::string & /* path */) -{ - /// Noop. -} - -void MetadataStorageFromStaticFilesWebServerTransaction::createMetadataFile( - const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) -{ - /// Noop. -} - -void MetadataStorageFromStaticFilesWebServerTransaction::commit() -{ - /// Noop. -} - -std::unordered_map MetadataStorageFromStaticFilesWebServer::getSerializedMetadata(const std::vector &) const -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "getSerializedMetadata is not implemented for MetadataStorageFromStaticFilesWebServer"); -} - -void MetadataStorageFromStaticFilesWebServerTransaction::chmod(const String &, mode_t) -{ - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "chmod is not implemented for MetadataStorageFromStaticFilesWebServer"); -} - } diff --git a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h index 27a1ae8b8fa..6ef44defad0 100644 --- a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h +++ b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h @@ -1,6 +1,7 @@ #pragma once #include +#include #include #include #include @@ -9,7 +10,7 @@ namespace DB { -class MetadataStorageFromStaticFilesWebServer final : public IMetadataStorage +class MetadataStorageFromStaticFilesWebServer final : public ReadOnlyMetadataStorage { private: friend class MetadataStorageFromStaticFilesWebServerTransaction; @@ -36,32 +37,18 @@ public: uint64_t getFileSize(const String & path) const override; - Poco::Timestamp getLastModified(const std::string & path) const override; - - time_t getLastChanged(const std::string & path) const override; - std::vector listDirectory(const std::string & path) const override; DirectoryIteratorPtr iterateDirectory(const std::string & path) const override; - std::string readFileToString(const std::string & path) const override; - - std::unordered_map getSerializedMetadata(const std::vector & file_paths) const override; - - uint32_t getHardlinkCount(const std::string & path) const override; - StoredObjects getStorageObjects(const std::string & path) const override; std::string getObjectStorageRootPath() const override { return ""; } - bool supportsChmod() const override { return false; } - - bool supportsStat() const override { return false; } - - struct stat stat(const String &) const override { return {}; } + struct stat stat(const String & /* path */) const override { return {}; } }; -class MetadataStorageFromStaticFilesWebServerTransaction final : public IMetadataTransaction +class MetadataStorageFromStaticFilesWebServerTransaction final : public ReadOnlyMetadataTransaction { private: DiskPtr disk; @@ -77,43 +64,9 @@ public: const IMetadataStorage & getStorageForNonTransactionalReads() const override; - void commit() override; - - void writeStringToFile(const std::string & path, const std::string & data) override; - - void createEmptyMetadataFile(const std::string & path) override; - - void createMetadataFile(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) override; - - void addBlobToMetadata(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) override; - - void setLastModified(const std::string & path, const Poco::Timestamp & timestamp) override; - - void setReadOnly(const std::string & path) override; - - void unlinkFile(const std::string & path) override; - void createDirectory(const std::string & path) override; void createDirectoryRecursive(const std::string & path) override; - - void removeDirectory(const std::string & path) override; - - void removeRecursive(const std::string & path) override; - - void createHardLink(const std::string & path_from, const std::string & path_to) override; - - void moveFile(const std::string & path_from, const std::string & path_to) override; - - void moveDirectory(const std::string & path_from, const std::string & path_to) override; - - void replaceFile(const std::string & path_from, const std::string & path_to) override; - - void unlinkMetadata(const std::string & path) override; - - bool supportsChmod() const override { return false; } - - void chmod(const String &, mode_t) override; }; } From 95fb2ad3cf3118dfbd9f5183f6fdcecae59fba2f Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 22:02:29 +0200 Subject: [PATCH 190/239] Implement ATTACH of MergeTree table for s3_plain disk Signed-off-by: Azat Khuzhin --- .../MetadataStorageFromPlainObjectStorage.cpp | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index 0722c3622b9..19f7f107b35 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -76,20 +76,20 @@ uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) std::vector MetadataStorageFromPlainObjectStorage::listDirectory(const std::string & path) const { RelativePathsWithSize children; - object_storage->listPrefix(getAbsolutePath(path), children); + std::vector common_prefixes; + object_storage->listPrefixInPath(getAbsolutePath(path), children, common_prefixes); std::vector result; for (const auto & path_size : children) - { result.push_back(path_size.relative_path); - } + for (const auto & common_prefix : common_prefixes) + result.push_back(common_prefix); return result; } DirectoryIteratorPtr MetadataStorageFromPlainObjectStorage::iterateDirectory(const std::string & path) const { - /// NOTE: this is not required for BACKUP/RESTORE, but this is a first step - /// towards MergeTree on plain S3. + /// Required for MergeTree auto paths = listDirectory(path); std::vector fs_paths(paths.begin(), paths.end()); return std::make_unique(std::move(fs_paths)); From 82ea67eb51d878d697476385fe2feed5573cb81f Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 22:04:13 +0200 Subject: [PATCH 191/239] Slightly better MetadataStorageFromPlainObjectStorage::isFile() Signed-off-by: Azat Khuzhin --- .../ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index 19f7f107b35..de97203d771 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -46,7 +46,7 @@ bool MetadataStorageFromPlainObjectStorage::exists(const std::string & path) con bool MetadataStorageFromPlainObjectStorage::isFile(const std::string & path) const { /// NOTE: This check is inaccurate and has excessive API calls - return !isDirectory(path) && exists(path); + return exists(path) && !isDirectory(path); } bool MetadataStorageFromPlainObjectStorage::isDirectory(const std::string & path) const From fa7535c90d22f514dac6d8c85c79beab88ebce7e Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 22:05:26 +0200 Subject: [PATCH 192/239] Slightly optimize MetadataStorageFromPlainObjectStorage::isDirectory() Signed-off-by: Azat Khuzhin --- .../ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index de97203d771..aa83effc481 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -58,8 +58,9 @@ bool MetadataStorageFromPlainObjectStorage::isDirectory(const std::string & path /// NOTE: This check is far from ideal, since it work only if the directory /// really has files, and has excessive API calls RelativePathsWithSize children; - object_storage->listPrefix(directory, children); - return !children.empty(); + std::vector common_prefixes; + object_storage->listPrefixInPath(directory, children, common_prefixes); + return !children.empty() || !common_prefixes.empty(); } uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) const From 8345c5f8812d3a8a1a2d9ea0eed03d12e94e270d Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 24 Oct 2022 22:07:26 +0200 Subject: [PATCH 193/239] tests: cover ATTACH of BACKUP from s3_plain disk Signed-off-by: Azat Khuzhin --- .../__init__.py | 0 .../configs/disk_s3.xml | 34 ++++++++++++++++ .../test_attach_backup_from_s3_plain/test.py | 40 +++++++++++++++++++ 3 files changed, 74 insertions(+) create mode 100644 tests/integration/test_attach_backup_from_s3_plain/__init__.py create mode 100644 tests/integration/test_attach_backup_from_s3_plain/configs/disk_s3.xml create mode 100644 tests/integration/test_attach_backup_from_s3_plain/test.py diff --git a/tests/integration/test_attach_backup_from_s3_plain/__init__.py b/tests/integration/test_attach_backup_from_s3_plain/__init__.py new file mode 100644 index 00000000000..e69de29bb2d diff --git a/tests/integration/test_attach_backup_from_s3_plain/configs/disk_s3.xml b/tests/integration/test_attach_backup_from_s3_plain/configs/disk_s3.xml new file mode 100644 index 00000000000..67278694d39 --- /dev/null +++ b/tests/integration/test_attach_backup_from_s3_plain/configs/disk_s3.xml @@ -0,0 +1,34 @@ + + + + + + s3_plain + http://minio1:9001/root/data/disks/disk_s3_plain/ + minio + minio123 + 33554432 + + + s3_plain + + http://minio1:9001/root/data/disks/disk_s3_plain/backup/ + minio + minio123 + 33554432 + + + + + +
+ attach_disk_s3_plain +
+ + + + + + backup_disk_s3_plain + + diff --git a/tests/integration/test_attach_backup_from_s3_plain/test.py b/tests/integration/test_attach_backup_from_s3_plain/test.py new file mode 100644 index 00000000000..35d53d5b8bd --- /dev/null +++ b/tests/integration/test_attach_backup_from_s3_plain/test.py @@ -0,0 +1,40 @@ +# pylint: disable=global-statement +# pylint: disable=line-too-long + +import pytest +from helpers.cluster import ClickHouseCluster + +cluster = ClickHouseCluster(__file__) +node = cluster.add_instance( + "node", + main_configs=["configs/disk_s3.xml"], + with_minio=True, +) + + +@pytest.fixture(scope="module", autouse=True) +def start_cluster(): + try: + cluster.start() + yield + finally: + cluster.shutdown() + + +def test_attach_backup(): + node.query( + f""" + -- BACKUP writes Ordinary like structure + set allow_deprecated_database_ordinary=1; + create database ordinary engine=Ordinary; + + create table ordinary.test_backup_attach engine=MergeTree() order by tuple() as select * from numbers(100); + -- NOTE: name of backup ("backup") is significant. + backup table ordinary.test_backup_attach TO Disk('backup_disk_s3_plain', 'backup'); + + drop table ordinary.test_backup_attach; + attach table ordinary.test_backup_attach (number UInt64) engine=MergeTree() order by tuple() settings storage_policy='attach_policy_s3_plain'; + """ + ) + + assert int(node.query("select count() from ordinary.test_backup_attach")) == 100 From 3b7abbbff437cd6c38ae35bc5c1e04a419f279fb Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Fri, 28 Oct 2022 16:17:52 +0200 Subject: [PATCH 194/239] Rename s/listPrefix/findAllFiles, s/listPrefixInPath/getDirectoryContents/ Signed-off-by: Azat Khuzhin --- .../AzureBlobStorage/AzureObjectStorage.cpp | 2 +- .../AzureBlobStorage/AzureObjectStorage.h | 2 +- .../Cached/CachedObjectStorage.cpp | 4 ++-- .../Cached/CachedObjectStorage.h | 2 +- ...jectStorageRemoteMetadataRestoreHelper.cpp | 4 ++-- src/Disks/ObjectStorages/IObjectStorage.cpp | 8 +++---- src/Disks/ObjectStorages/IObjectStorage.h | 24 +++++++++---------- .../MetadataStorageFromPlainObjectStorage.cpp | 24 +++++++++---------- .../ObjectStorages/S3/S3ObjectStorage.cpp | 16 ++++++------- src/Disks/ObjectStorages/S3/S3ObjectStorage.h | 8 +++---- 10 files changed, 47 insertions(+), 47 deletions(-) diff --git a/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp b/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp index b3dcfdafa9e..c3549701ec1 100644 --- a/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp +++ b/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp @@ -141,7 +141,7 @@ std::unique_ptr AzureObjectStorage::writeObject( /// NO return std::make_unique(std::move(buffer), std::move(finalize_callback), object.absolute_path); } -void AzureObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize & children) const +void AzureObjectStorage::findAllFiles(const std::string & path, RelativePathsWithSize & children) const { auto client_ptr = client.get(); diff --git a/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.h b/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.h index 47ac0d6badd..6fd41dae2ec 100644 --- a/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.h +++ b/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.h @@ -84,7 +84,7 @@ public: size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE, const WriteSettings & write_settings = {}) override; - void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; + void findAllFiles(const std::string & path, RelativePathsWithSize & children) const override; /// Remove file. Throws exception if file doesn't exists or it's a directory. void removeObject(const StoredObject & object) override; diff --git a/src/Disks/ObjectStorages/Cached/CachedObjectStorage.cpp b/src/Disks/ObjectStorages/Cached/CachedObjectStorage.cpp index 7e829847846..f3d3f049dc1 100644 --- a/src/Disks/ObjectStorages/Cached/CachedObjectStorage.cpp +++ b/src/Disks/ObjectStorages/Cached/CachedObjectStorage.cpp @@ -282,9 +282,9 @@ std::unique_ptr CachedObjectStorage::cloneObjectStorage( return object_storage->cloneObjectStorage(new_namespace, config, config_prefix, context); } -void CachedObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize & children) const +void CachedObjectStorage::findAllFiles(const std::string & path, RelativePathsWithSize & children) const { - object_storage->listPrefix(path, children); + object_storage->findAllFiles(path, children); } ObjectMetadata CachedObjectStorage::getObjectMetadata(const std::string & path) const diff --git a/src/Disks/ObjectStorages/Cached/CachedObjectStorage.h b/src/Disks/ObjectStorages/Cached/CachedObjectStorage.h index 68ded61a9f1..64e6eed45bb 100644 --- a/src/Disks/ObjectStorages/Cached/CachedObjectStorage.h +++ b/src/Disks/ObjectStorages/Cached/CachedObjectStorage.h @@ -72,7 +72,7 @@ public: const std::string & config_prefix, ContextPtr context) override; - void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; + void findAllFiles(const std::string & path, RelativePathsWithSize & children) const override; ObjectMetadata getObjectMetadata(const std::string & path) const override; diff --git a/src/Disks/ObjectStorages/DiskObjectStorageRemoteMetadataRestoreHelper.cpp b/src/Disks/ObjectStorages/DiskObjectStorageRemoteMetadataRestoreHelper.cpp index 65f0d24035a..4ea42616ba2 100644 --- a/src/Disks/ObjectStorages/DiskObjectStorageRemoteMetadataRestoreHelper.cpp +++ b/src/Disks/ObjectStorages/DiskObjectStorageRemoteMetadataRestoreHelper.cpp @@ -390,7 +390,7 @@ void DiskObjectStorageRemoteMetadataRestoreHelper::restoreFiles(IObjectStorage * }; RelativePathsWithSize children; - source_object_storage->listPrefix(restore_information.source_path, children); + source_object_storage->findAllFiles(restore_information.source_path, children); restore_files(children); @@ -540,7 +540,7 @@ void DiskObjectStorageRemoteMetadataRestoreHelper::restoreFileOperations(IObject }; RelativePathsWithSize children; - source_object_storage->listPrefix(restore_information.source_path + "operations/", children); + source_object_storage->findAllFiles(restore_information.source_path + "operations/", children); restore_file_operations(children); if (restore_information.detached) diff --git a/src/Disks/ObjectStorages/IObjectStorage.cpp b/src/Disks/ObjectStorages/IObjectStorage.cpp index 7c187e3dd2f..3f8ac566603 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.cpp +++ b/src/Disks/ObjectStorages/IObjectStorage.cpp @@ -14,15 +14,15 @@ namespace ErrorCodes extern const int LOGICAL_ERROR; } -void IObjectStorage::listPrefix(const std::string &, RelativePathsWithSize &) const +void IObjectStorage::findAllFiles(const std::string &, RelativePathsWithSize &) const { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "listPrefix() is not supported"); + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "findAllFiles() is not supported"); } -void IObjectStorage::listPrefixInPath(const std::string &, +void IObjectStorage::getDirectoryContents(const std::string &, RelativePathsWithSize &, std::vector &) const { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "listPrefixInPath() is not supported"); + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "getDirectoryContents() is not supported"); } IAsynchronousReader & IObjectStorage::getThreadPoolReader() diff --git a/src/Disks/ObjectStorages/IObjectStorage.h b/src/Disks/ObjectStorages/IObjectStorage.h index a7ad4e5b515..9451ae31b07 100644 --- a/src/Disks/ObjectStorages/IObjectStorage.h +++ b/src/Disks/ObjectStorages/IObjectStorage.h @@ -65,32 +65,32 @@ public: /// Object exists or not virtual bool exists(const StoredObject & object) const = 0; - /// List on prefix, return children (relative paths) with their sizes. - /// So it should return all "objects" with their sizes. + /// List all objects with specific prefix. /// /// For example if you do this over filesystem, you should skip folders and /// return files only, so something like on local filesystem: /// /// find . -type f /// + /// @param children - out files (relative paths) with their sizes. + /// /// NOTE: It makes sense only for real object storages (S3, Azure), since /// it is used only for one of the following: /// - send_metadata (to restore metadata) /// - see DiskObjectStorage::restoreMetadataIfNeeded() /// - MetadataStorageFromPlainObjectStorage - only for s3_plain disk - virtual void listPrefix(const std::string & path, RelativePathsWithSize & children) const; + virtual void findAllFiles(const std::string & path, RelativePathsWithSize & children) const; - /// List on prefix, but it is limited only to current path, something like - /// on local filesystem: + /// Analog of directory content for object storage (object storage does not + /// have "directory" definition, but it can be emulated with usage of + /// "delimiter"), so this is analog of: /// - /// find . -maxdepth 1 + /// find . -maxdepth 1 $path /// - /// @param path - path to list prefix in - /// @param children - list of child nodes (for unix filesystem - files goes here) - /// @param common_prefixes - list of common prefixes in this path, (for unix filesystem - subfolders) - virtual void listPrefixInPath(const std::string & path, - RelativePathsWithSize & children, - std::vector & common_prefixes) const; + /// Return files in @files and directories in @directories + virtual void getDirectoryContents(const std::string & path, + RelativePathsWithSize & files, + std::vector & directories) const; /// Get object metadata if supported. It should be possible to receive /// at least size of object diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp index aa83effc481..259f6e01fd7 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.cpp @@ -57,34 +57,34 @@ bool MetadataStorageFromPlainObjectStorage::isDirectory(const std::string & path /// NOTE: This check is far from ideal, since it work only if the directory /// really has files, and has excessive API calls - RelativePathsWithSize children; - std::vector common_prefixes; - object_storage->listPrefixInPath(directory, children, common_prefixes); - return !children.empty() || !common_prefixes.empty(); + RelativePathsWithSize files; + std::vector directories; + object_storage->getDirectoryContents(directory, files, directories); + return !files.empty() || !directories.empty(); } uint64_t MetadataStorageFromPlainObjectStorage::getFileSize(const String & path) const { RelativePathsWithSize children; - object_storage->listPrefix(getAbsolutePath(path), children); + object_storage->findAllFiles(getAbsolutePath(path), children); if (children.empty()) return 0; if (children.size() != 1) - throw Exception(ErrorCodes::LOGICAL_ERROR, "listPrefix() return multiple paths ({}) for {}", children.size(), path); + throw Exception(ErrorCodes::LOGICAL_ERROR, "findAllFiles() return multiple paths ({}) for {}", children.size(), path); return children.front().bytes_size; } std::vector MetadataStorageFromPlainObjectStorage::listDirectory(const std::string & path) const { - RelativePathsWithSize children; - std::vector common_prefixes; - object_storage->listPrefixInPath(getAbsolutePath(path), children, common_prefixes); + RelativePathsWithSize files; + std::vector directories; + object_storage->getDirectoryContents(getAbsolutePath(path), files, directories); std::vector result; - for (const auto & path_size : children) + for (const auto & path_size : files) result.push_back(path_size.relative_path); - for (const auto & common_prefix : common_prefixes) - result.push_back(common_prefix); + for (const auto & directory : directories) + result.push_back(directory); return result; } diff --git a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp index af3374f8d3c..0c421ee03d7 100644 --- a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp +++ b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp @@ -248,7 +248,7 @@ std::unique_ptr S3ObjectStorage::writeObject( /// NOLIN std::move(s3_buffer), std::move(finalize_callback), object.absolute_path); } -void S3ObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize & children) const +void S3ObjectStorage::findAllFiles(const std::string & path, RelativePathsWithSize & children) const { auto settings_ptr = s3_settings.get(); auto client_ptr = client.get(); @@ -279,9 +279,9 @@ void S3ObjectStorage::listPrefix(const std::string & path, RelativePathsWithSize } while (outcome.GetResult().GetIsTruncated()); } -void S3ObjectStorage::listPrefixInPath(const std::string & path, - RelativePathsWithSize & children, - std::vector & common_prefixes) const +void S3ObjectStorage::getDirectoryContents(const std::string & path, + RelativePathsWithSize & files, + std::vector & directories) const { auto settings_ptr = s3_settings.get(); auto client_ptr = client.get(); @@ -308,14 +308,14 @@ void S3ObjectStorage::listPrefixInPath(const std::string & path, break; for (const auto & object : result_objects) - children.emplace_back(object.GetKey(), object.GetSize()); + files.emplace_back(object.GetKey(), object.GetSize()); for (const auto & common_prefix : result_common_prefixes) { - std::string subfolder = common_prefix.GetPrefix(); + std::string directory = common_prefix.GetPrefix(); /// Make it compatible with std::filesystem::path::filename() - trimRight(subfolder, '/'); - common_prefixes.emplace_back(subfolder); + trimRight(directory, '/'); + directories.emplace_back(directory); } request.SetContinuationToken(outcome.GetResult().GetNextContinuationToken()); diff --git a/src/Disks/ObjectStorages/S3/S3ObjectStorage.h b/src/Disks/ObjectStorages/S3/S3ObjectStorage.h index 9e90efcb9dc..6b1e8289b15 100644 --- a/src/Disks/ObjectStorages/S3/S3ObjectStorage.h +++ b/src/Disks/ObjectStorages/S3/S3ObjectStorage.h @@ -105,10 +105,10 @@ public: size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE, const WriteSettings & write_settings = {}) override; - void listPrefix(const std::string & path, RelativePathsWithSize & children) const override; - void listPrefixInPath(const std::string & path, - RelativePathsWithSize & children, - std::vector & common_prefixes) const override; + void findAllFiles(const std::string & path, RelativePathsWithSize & children) const override; + void getDirectoryContents(const std::string & path, + RelativePathsWithSize & files, + std::vector & directories) const override; /// Uses `DeleteObjectRequest`. void removeObject(const StoredObject & object) override; From bceca73f6f362577ef6f9f8b05ac1a1ae65da1c1 Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Fri, 28 Oct 2022 18:10:00 +0200 Subject: [PATCH 195/239] Throw NOT_IMPLEMENTED form ReadOnlyMetadataStorage::getLastChanged() Signed-off-by: Azat Khuzhin --- .../ObjectStorages/ReadOnlyMetadataStorage.h | 21 ++++++++++--------- 1 file changed, 11 insertions(+), 10 deletions(-) diff --git a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h index e136b0580ea..b21d2c3a28b 100644 --- a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h +++ b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h @@ -124,26 +124,27 @@ public: /// Required by MergeTree return {}; } - time_t getLastChanged(const std::string & /* path */) const override - { - return {}; - } - - struct stat stat(const String & /* path */) const override - { - throwNotAllowed(); - } uint32_t getHardlinkCount(const std::string & /* path */) const override { return 1; } + /// + /// Throw + /// + struct stat stat(const String & /* path */) const override + { + throwNotAllowed(); + } + time_t getLastChanged(const std::string & /* path */) const override + { + throwNotAllowed(); + } std::string readFileToString(const std::string & /* path */) const override { throwNotAllowed(); } - std::unordered_map getSerializedMetadata(const std::vector & /* file_paths */) const override { throwNotAllowed(); From 51bd0c2ac11f5b36aab2824938d60b60066359cc Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Sat, 29 Oct 2022 11:56:12 +0200 Subject: [PATCH 196/239] Make ReadOnlyMetadataStorage really readonly Signed-off-by: Azat Khuzhin --- .../MetadataStorageFromPlainObjectStorage.h | 10 +++++ .../ObjectStorages/ReadOnlyMetadataStorage.h | 44 +++++-------------- .../MetadataStorageFromStaticFilesWebServer.h | 10 +++++ 3 files changed, 32 insertions(+), 32 deletions(-) diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h index 2cfa123ecc9..e4c0fb90fb1 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h @@ -80,6 +80,16 @@ public: void addBlobToMetadata(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) override; + void createEmptyMetadataFile(const std::string & /* path */) override + { + /// No metadata, no need to create anything. + } + + void createMetadataFile(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override + { + /// Noop + } + void createDirectory(const std::string & path) override; void createDirectoryRecursive(const std::string & path) override; diff --git a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h index b21d2c3a28b..f59bc0ad77f 100644 --- a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h +++ b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h @@ -13,30 +13,24 @@ namespace ErrorCodes class ReadOnlyMetadataStorage; -/// Readonly, throws NOT_IMPLEMENTED error. +/// Transaction for read-only storage, throws NOT_IMPLEMENTED error. /// Can be used to add limited read-only support of MergeTree. class ReadOnlyMetadataTransaction : public IMetadataTransaction { public: - /// - /// Noop - /// void commit() override { /// Noop, nothing to commit. } + void createEmptyMetadataFile(const std::string & /* path */) override { - /// No metadata, no need to create anything. + throwNotAllowed(); } void createMetadataFile(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override { - /// Noop + throwNotAllowed(); } - - /// - /// Throws - /// void writeStringToFile(const std::string & /* path */, const std::string & /* data */) override { throwNotAllowed(); @@ -98,68 +92,54 @@ public: throwNotAllowed(); } - - /// - /// Others - /// bool supportsChmod() const override { return false; } private: [[noreturn]] static void throwNotAllowed() { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Only read-only transaction operations are supported"); + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Transaction for read-only storage is not supported"); } }; -/// Readonly, throws NOT_IMPLEMENTED error. +/// Readonly storage, throws NOT_IMPLEMENTED error. /// Can be used to add limited read-only support of MergeTree. class ReadOnlyMetadataStorage : public IMetadataStorage { public: - /// - /// Noop - /// Poco::Timestamp getLastModified(const std::string & /* path */) const override { /// Required by MergeTree return {}; } - uint32_t getHardlinkCount(const std::string & /* path */) const override { return 1; } - /// - /// Throw - /// struct stat stat(const String & /* path */) const override { - throwNotAllowed(); + throwNotImplemented(); } time_t getLastChanged(const std::string & /* path */) const override { - throwNotAllowed(); + throwNotImplemented(); } std::string readFileToString(const std::string & /* path */) const override { - throwNotAllowed(); + throwNotImplemented(); } std::unordered_map getSerializedMetadata(const std::vector & /* file_paths */) const override { - throwNotAllowed(); + throwNotImplemented(); } - /// - /// Others - /// bool supportsChmod() const override { return false; } bool supportsStat() const override { return false; } private: - [[noreturn]] static void throwNotAllowed() + [[noreturn]] static void throwNotImplemented() { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Only read-only metadata operations are supported"); + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Operation is ont implemented"); } }; diff --git a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h index 6ef44defad0..eb70b9c8108 100644 --- a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h +++ b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h @@ -64,6 +64,16 @@ public: const IMetadataStorage & getStorageForNonTransactionalReads() const override; + void createEmptyMetadataFile(const std::string & /* path */) override + { + /// No metadata, no need to create anything. + } + + void createMetadataFile(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override + { + /// Noop + } + void createDirectory(const std::string & path) override; void createDirectoryRecursive(const std::string & path) override; From c93262170df0c0fb09ccbb56f5f73b7a41bc58fc Mon Sep 17 00:00:00 2001 From: Azat Khuzhin Date: Mon, 31 Oct 2022 11:02:13 +0100 Subject: [PATCH 197/239] Remove ReadOnlyMetadataStorage Throw exceptions from IMetadataStorage instead to avoid introducing extra abstractions. Signed-off-by: Azat Khuzhin --- src/Disks/ObjectStorages/IMetadataStorage.h | 108 ++++++++++--- .../MetadataStorageFromPlainObjectStorage.h | 33 ++-- .../ObjectStorages/ReadOnlyMetadataStorage.h | 146 ------------------ .../MetadataStorageFromStaticFilesWebServer.h | 27 +++- 4 files changed, 136 insertions(+), 178 deletions(-) delete mode 100644 src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h diff --git a/src/Disks/ObjectStorages/IMetadataStorage.h b/src/Disks/ObjectStorages/IMetadataStorage.h index 3d6c772157d..597d7744c78 100644 --- a/src/Disks/ObjectStorages/IMetadataStorage.h +++ b/src/Disks/ObjectStorages/IMetadataStorage.h @@ -11,10 +11,16 @@ #include #include #include +#include namespace DB { +namespace ErrorCodes +{ + extern const int NOT_IMPLEMENTED; +} + class IMetadataStorage; /// Tries to provide some "transactions" interface, which allow @@ -33,32 +39,71 @@ public: /// General purpose methods /// Write metadata string to file - virtual void writeStringToFile(const std::string & path, const std::string & data) = 0; + virtual void writeStringToFile(const std::string & /* path */, const std::string & /* data */) + { + throwNotImplemented(); + } - virtual void setLastModified(const std::string & path, const Poco::Timestamp & timestamp) = 0; + virtual void setLastModified(const std::string & /* path */, const Poco::Timestamp & /* timestamp */) + { + throwNotImplemented(); + } virtual bool supportsChmod() const = 0; - virtual void chmod(const String & path, mode_t mode) = 0; + virtual void chmod(const String & /* path */, mode_t /* mode */) + { + throwNotImplemented(); + } - virtual void setReadOnly(const std::string & path) = 0; + virtual void setReadOnly(const std::string & /* path */) + { + throwNotImplemented(); + } - virtual void unlinkFile(const std::string & path) = 0; + virtual void unlinkFile(const std::string & /* path */) + { + throwNotImplemented(); + } - virtual void createDirectory(const std::string & path) = 0; + virtual void createDirectory(const std::string & /* path */) + { + throwNotImplemented(); + } - virtual void createDirectoryRecursive(const std::string & path) = 0; + virtual void createDirectoryRecursive(const std::string & /* path */) + { + throwNotImplemented(); + } - virtual void removeDirectory(const std::string & path) = 0; + virtual void removeDirectory(const std::string & /* path */) + { + throwNotImplemented(); + } - virtual void removeRecursive(const std::string & path) = 0; + virtual void removeRecursive(const std::string & /* path */) + { + throwNotImplemented(); + } - virtual void createHardLink(const std::string & path_from, const std::string & path_to) = 0; + virtual void createHardLink(const std::string & /* path_from */, const std::string & /* path_to */) + { + throwNotImplemented(); + } - virtual void moveFile(const std::string & path_from, const std::string & path_to) = 0; + virtual void moveFile(const std::string & /* path_from */, const std::string & /* path_to */) + { + throwNotImplemented(); + } - virtual void moveDirectory(const std::string & path_from, const std::string & path_to) = 0; + virtual void moveDirectory(const std::string & /* path_from */, const std::string & /* path_to */) + { + throwNotImplemented(); + } - virtual void replaceFile(const std::string & path_from, const std::string & path_to) = 0; + virtual void replaceFile(const std::string & /* path_from */, const std::string & /* path_to */) + { + throwNotImplemented(); + } /// Metadata related methods @@ -69,7 +114,10 @@ public: virtual void createMetadataFile(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) = 0; /// Add to new blob to metadata file (way to implement appends) - virtual void addBlobToMetadata(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) = 0; + virtual void addBlobToMetadata(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) + { + throwNotImplemented(); + } /// Unlink metadata file and do something special if required /// By default just remove file (unlink file). @@ -79,6 +127,12 @@ public: } virtual ~IMetadataTransaction() = default; + +private: + [[noreturn]] static void throwNotImplemented() + { + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Operation is not implemented"); + } }; using MetadataTransactionPtr = std::shared_ptr; @@ -106,12 +160,18 @@ public: virtual Poco::Timestamp getLastModified(const std::string & path) const = 0; - virtual time_t getLastChanged(const std::string & path) const = 0; + virtual time_t getLastChanged(const std::string & /* path */) const + { + throwNotImplemented(); + } virtual bool supportsChmod() const = 0; virtual bool supportsStat() const = 0; - virtual struct stat stat(const String & path) const = 0; + virtual struct stat stat(const String & /* path */) const + { + throwNotImplemented(); + } virtual std::vector listDirectory(const std::string & path) const = 0; @@ -120,20 +180,32 @@ public: virtual uint32_t getHardlinkCount(const std::string & path) const = 0; /// Read metadata file to string from path - virtual std::string readFileToString(const std::string & path) const = 0; + virtual std::string readFileToString(const std::string & /* path */) const + { + throwNotImplemented(); + } virtual ~IMetadataStorage() = default; /// ==== More specific methods. Previous were almost general purpose. ==== /// Read multiple metadata files into strings and return mapping from file_path -> metadata - virtual std::unordered_map getSerializedMetadata(const std::vector & file_paths) const = 0; + virtual std::unordered_map getSerializedMetadata(const std::vector & /* file_paths */) const + { + throwNotImplemented(); + } /// Return object information (absolute_path, bytes_size, ...) for metadata path. /// object_storage_path is absolute. virtual StoredObjects getStorageObjects(const std::string & path) const = 0; virtual std::string getObjectStorageRootPath() const = 0; + +private: + [[noreturn]] static void throwNotImplemented() + { + throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Operation is not implemented"); + } }; using MetadataStoragePtr = std::shared_ptr; diff --git a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h index e4c0fb90fb1..99cc960b9e4 100644 --- a/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h +++ b/src/Disks/ObjectStorages/MetadataStorageFromPlainObjectStorage.h @@ -2,7 +2,6 @@ #include #include -#include #include #include @@ -21,10 +20,7 @@ namespace DB /// It is used to allow BACKUP/RESTORE to ObjectStorage (S3/...) with the same /// structure as on disk MergeTree, and does not requires metadata from local /// disk to restore. -/// -/// NOTE: Inheritance from ReadOnlyMetadataStorage is used here to throw -/// NOT_IMPLEMENTED error for lost of unsupported methods (mtime/move/stat/...) -class MetadataStorageFromPlainObjectStorage final : public ReadOnlyMetadataStorage +class MetadataStorageFromPlainObjectStorage final : public IMetadataStorage { private: friend class MetadataStorageFromPlainObjectStorageTransaction; @@ -59,11 +55,25 @@ public: std::string getObjectStorageRootPath() const override { return object_storage_root_path; } + Poco::Timestamp getLastModified(const std::string & /* path */) const override + { + /// Required by MergeTree + return {}; + } + + uint32_t getHardlinkCount(const std::string & /* path */) const override + { + return 1; + } + + bool supportsChmod() const override { return false; } + bool supportsStat() const override { return false; } + private: std::filesystem::path getAbsolutePath(const std::string & path) const; }; -class MetadataStorageFromPlainObjectStorageTransaction final : public ReadOnlyMetadataTransaction +class MetadataStorageFromPlainObjectStorageTransaction final : public IMetadataTransaction { private: const MetadataStorageFromPlainObjectStorage & metadata_storage; @@ -74,9 +84,7 @@ public: : metadata_storage(metadata_storage_) {} - ~MetadataStorageFromPlainObjectStorageTransaction() override = default; - - const IMetadataStorage & getStorageForNonTransactionalReads() const final; + const IMetadataStorage & getStorageForNonTransactionalReads() const override; void addBlobToMetadata(const std::string & path, const std::string & blob_name, uint64_t size_in_bytes) override; @@ -97,6 +105,13 @@ public: void unlinkFile(const std::string & path) override; void unlinkMetadata(const std::string & path) override; + + void commit() override + { + /// Nothing to commit. + } + + bool supportsChmod() const override { return false; } }; } diff --git a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h b/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h deleted file mode 100644 index f59bc0ad77f..00000000000 --- a/src/Disks/ObjectStorages/ReadOnlyMetadataStorage.h +++ /dev/null @@ -1,146 +0,0 @@ -#pragma once - -#include -#include - -namespace DB -{ - -namespace ErrorCodes -{ - extern const int NOT_IMPLEMENTED; -} - -class ReadOnlyMetadataStorage; - -/// Transaction for read-only storage, throws NOT_IMPLEMENTED error. -/// Can be used to add limited read-only support of MergeTree. -class ReadOnlyMetadataTransaction : public IMetadataTransaction -{ -public: - void commit() override - { - /// Noop, nothing to commit. - } - - void createEmptyMetadataFile(const std::string & /* path */) override - { - throwNotAllowed(); - } - void createMetadataFile(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override - { - throwNotAllowed(); - } - void writeStringToFile(const std::string & /* path */, const std::string & /* data */) override - { - throwNotAllowed(); - } - void setLastModified(const std::string & /* path */, const Poco::Timestamp & /* timestamp */) override - { - throwNotAllowed(); - } - void chmod(const String & /* path */, mode_t /* mode */) override - { - throwNotAllowed(); - } - void setReadOnly(const std::string & /* path */) override - { - throwNotAllowed(); - } - void unlinkFile(const std::string & /* path */) override - { - throwNotAllowed(); - } - void removeDirectory(const std::string & /* path */) override - { - throwNotAllowed(); - } - void removeRecursive(const std::string & /* path */) override - { - throwNotAllowed(); - } - void createHardLink(const std::string & /* path_from */, const std::string & /* path_to */) override - { - throwNotAllowed(); - } - void moveFile(const std::string & /* path_from */, const std::string & /* path_to */) override - { - throwNotAllowed(); - } - void moveDirectory(const std::string & /* path_from */, const std::string & /* path_to */) override - { - throwNotAllowed(); - } - void replaceFile(const std::string & /* path_from */, const std::string & /* path_to */) override - { - throwNotAllowed(); - } - void createDirectory(const std::string & /* path */) override - { - throwNotAllowed(); - } - void createDirectoryRecursive(const std::string & /* path */) override - { - throwNotAllowed(); - } - void addBlobToMetadata(const std::string & /* path */, const std::string & /* blob_name */, uint64_t /* size_in_bytes */) override - { - throwNotAllowed(); - } - void unlinkMetadata(const std::string & /* path */) override - { - throwNotAllowed(); - } - - bool supportsChmod() const override { return false; } - -private: - [[noreturn]] static void throwNotAllowed() - { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Transaction for read-only storage is not supported"); - } -}; - -/// Readonly storage, throws NOT_IMPLEMENTED error. -/// Can be used to add limited read-only support of MergeTree. -class ReadOnlyMetadataStorage : public IMetadataStorage -{ -public: - Poco::Timestamp getLastModified(const std::string & /* path */) const override - { - /// Required by MergeTree - return {}; - } - uint32_t getHardlinkCount(const std::string & /* path */) const override - { - return 1; - } - - struct stat stat(const String & /* path */) const override - { - throwNotImplemented(); - } - time_t getLastChanged(const std::string & /* path */) const override - { - throwNotImplemented(); - } - std::string readFileToString(const std::string & /* path */) const override - { - throwNotImplemented(); - } - std::unordered_map getSerializedMetadata(const std::vector & /* file_paths */) const override - { - throwNotImplemented(); - } - - bool supportsChmod() const override { return false; } - bool supportsStat() const override { return false; } - -private: - [[noreturn]] static void throwNotImplemented() - { - throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Operation is ont implemented"); - } -}; - -} diff --git a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h index eb70b9c8108..338a2690b8f 100644 --- a/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h +++ b/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.h @@ -1,7 +1,6 @@ #pragma once #include -#include #include #include #include @@ -10,7 +9,7 @@ namespace DB { -class MetadataStorageFromStaticFilesWebServer final : public ReadOnlyMetadataStorage +class MetadataStorageFromStaticFilesWebServer final : public IMetadataStorage { private: friend class MetadataStorageFromStaticFilesWebServerTransaction; @@ -46,9 +45,22 @@ public: std::string getObjectStorageRootPath() const override { return ""; } struct stat stat(const String & /* path */) const override { return {}; } + + Poco::Timestamp getLastModified(const std::string & /* path */) const override + { + /// Required by MergeTree + return {}; + } + uint32_t getHardlinkCount(const std::string & /* path */) const override + { + return 1; + } + + bool supportsChmod() const override { return false; } + bool supportsStat() const override { return false; } }; -class MetadataStorageFromStaticFilesWebServerTransaction final : public ReadOnlyMetadataTransaction +class MetadataStorageFromStaticFilesWebServerTransaction final : public IMetadataTransaction { private: DiskPtr disk; @@ -60,8 +72,6 @@ public: : metadata_storage(metadata_storage_) {} - ~MetadataStorageFromStaticFilesWebServerTransaction() override = default; - const IMetadataStorage & getStorageForNonTransactionalReads() const override; void createEmptyMetadataFile(const std::string & /* path */) override @@ -77,6 +87,13 @@ public: void createDirectory(const std::string & path) override; void createDirectoryRecursive(const std::string & path) override; + + void commit() override + { + /// Nothing to commit. + } + + bool supportsChmod() const override { return false; } }; } From bca22ec5f50b05e4e578da816278379d86ab3a6d Mon Sep 17 00:00:00 2001 From: Maksim Kita Date: Mon, 31 Oct 2022 12:30:00 +0100 Subject: [PATCH 198/239] Fixed code review issues --- .../MergeTreeIndexConditionBloomFilter.cpp | 19 +++++------ .../MergeTree/MergeTreeIndexFullText.cpp | 25 +++++++------- src/Storages/MergeTree/MergeTreeIndexUtils.h | 2 +- src/Storages/MergeTree/RPNBuilder.cpp | 33 ++++++++++++------- src/Storages/MergeTree/RPNBuilder.h | 18 ++++++---- 5 files changed, 54 insertions(+), 43 deletions(-) diff --git a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp index 9856500c65c..be7118066bb 100644 --- a/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexConditionBloomFilter.cpp @@ -293,8 +293,6 @@ bool MergeTreeIndexConditionBloomFilter::traverseFunction(const RPNBuilderTreeNo auto lhs_argument = function.getArgumentAt(0); auto rhs_argument = function.getArgumentAt(1); - auto lhs_argument_column_name = lhs_argument.getColumnName(); - if (functionIsInOrGlobalInOperator(function_name)) { ConstSetPtr prepared_set = rhs_argument.tryGetPreparedSet(); @@ -407,20 +405,21 @@ bool MergeTreeIndexConditionBloomFilter::traverseTreeIn( if (set_contain_default_value) return false; - const auto & col_name = key_node_function.getArgumentAt(0).getColumnName(); - auto map_keys_index_column_name = fmt::format("mapKeys({})", col_name); - auto map_values_index_column_name = fmt::format("mapValues({})", col_name); + auto first_argument = key_node_function.getArgumentAt(0); + const auto column_name = first_argument.getColumnName(); + auto map_keys_index_column_name = fmt::format("mapKeys({})", column_name); + auto map_values_index_column_name = fmt::format("mapValues({})", column_name); if (header.has(map_keys_index_column_name)) { /// For mapKeys we serialize key argument with bloom filter - auto first_argument = key_node_function.getArgumentAt(1); + auto second_argument = key_node_function.getArgumentAt(1); Field constant_value; DataTypePtr constant_type; - if (first_argument.tryGetConstant(constant_value, constant_type)) + if (second_argument.tryGetConstant(constant_value, constant_type)) { size_t position = header.getPositionByName(map_keys_index_column_name); const DataTypePtr & index_type = header.getByPosition(position).type; @@ -688,10 +687,10 @@ bool MergeTreeIndexConditionBloomFilter::traverseTreeEquals( return false; auto first_argument = key_node_function.getArgumentAt(0); - const auto col_name = first_argument.getColumnName(); + const auto column_name = first_argument.getColumnName(); - auto map_keys_index_column_name = fmt::format("mapKeys({})", col_name); - auto map_values_index_column_name = fmt::format("mapValues({})", col_name); + auto map_keys_index_column_name = fmt::format("mapKeys({})", column_name); + auto map_values_index_column_name = fmt::format("mapValues({})", column_name); size_t position = 0; Field const_value = value_field; diff --git a/src/Storages/MergeTree/MergeTreeIndexFullText.cpp b/src/Storages/MergeTree/MergeTreeIndexFullText.cpp index 9074485a020..b96d40f5759 100644 --- a/src/Storages/MergeTree/MergeTreeIndexFullText.cpp +++ b/src/Storages/MergeTree/MergeTreeIndexFullText.cpp @@ -434,7 +434,7 @@ bool MergeTreeConditionFullText::traverseTreeEquals( return false; auto first_argument = key_function_node.getArgumentAt(0); - const auto & map_column_name = first_argument.getColumnName(); + const auto map_column_name = first_argument.getColumnName(); size_t map_keys_key_column_num = 0; auto map_keys_index_column_name = fmt::format("mapKeys({})", map_column_name); @@ -588,23 +588,20 @@ bool MergeTreeConditionFullText::tryPrepareSetBloomFilter( std::vector key_tuple_mapping; DataTypes data_types; - if (left_argument.isFunction()) + auto left_argument_function_node_optional = left_argument.toFunctionNodeOrNull(); + + if (left_argument_function_node_optional && left_argument_function_node_optional->getFunctionName() == "tuple") { - auto left_argument_function_node = left_argument.toFunctionNode(); - auto left_argument_function_node_name = left_argument_function_node.getFunctionName(); + const auto & left_argument_function_node = *left_argument_function_node_optional; + size_t left_argument_function_node_arguments_size = left_argument_function_node.getArgumentsSize(); - if (left_argument_function_node_name == "tuple") + for (size_t i = 0; i < left_argument_function_node_arguments_size; ++i) { - size_t left_argument_function_node_arguments_size = left_argument_function_node.getArgumentsSize(); - - for (size_t i = 0; i < left_argument_function_node_arguments_size; ++i) + size_t key = 0; + if (getKey(left_argument_function_node.getArgumentAt(i).getColumnName(), key)) { - size_t key = 0; - if (getKey(left_argument_function_node.getArgumentAt(i).getColumnName(), key)) - { - key_tuple_mapping.emplace_back(i, key); - data_types.push_back(index_data_types[key]); - } + key_tuple_mapping.emplace_back(i, key); + data_types.push_back(index_data_types[key]); } } } diff --git a/src/Storages/MergeTree/MergeTreeIndexUtils.h b/src/Storages/MergeTree/MergeTreeIndexUtils.h index 23cd92f3d99..6ba9725b564 100644 --- a/src/Storages/MergeTree/MergeTreeIndexUtils.h +++ b/src/Storages/MergeTree/MergeTreeIndexUtils.h @@ -6,7 +6,7 @@ namespace DB { /** Build AST filter node for index analysis from WHERE and PREWHERE sections of select query and additional filters. - * If select query does not have WHERE or PREWHERE and additional filters are empty null is returned. + * If select query does not have WHERE and PREWHERE and additional filters are empty null is returned. */ ASTPtr buildFilterNode(const ASTPtr & select_query, ASTs additional_filters = {}); diff --git a/src/Storages/MergeTree/RPNBuilder.cpp b/src/Storages/MergeTree/RPNBuilder.cpp index 395daab76ba..d7ea68e7d64 100644 --- a/src/Storages/MergeTree/RPNBuilder.cpp +++ b/src/Storages/MergeTree/RPNBuilder.cpp @@ -129,7 +129,7 @@ std::string RPNBuilderTreeNode::getColumnNameWithModuloLegacy() const } else { - return getColumnNameWithoutAlias(*dag_node, true); + return getColumnNameWithoutAlias(*dag_node, true /*legacy*/); } } @@ -203,7 +203,7 @@ bool RPNBuilderTreeNode::tryGetConstant(Field & output_value, DataTypePtr & outp String column_name = ast_node->getColumnName(); const auto & block_with_constants = tree_context.getBlockWithConstants(); - if (const auto * lit = ast_node->as()) + if (const auto * literal = ast_node->as()) { /// By default block_with_constants has only one column named "_dummy". /// If block contains only constants it's may not be preprocessed by @@ -212,7 +212,7 @@ bool RPNBuilderTreeNode::tryGetConstant(Field & output_value, DataTypePtr & outp column_name = "_dummy"; /// Simple literal - output_value = lit->value; + output_value = literal->value; output_type = block_with_constants.getByName(column_name).type; /// If constant is not Null, we can assume it's type is not Nullable as well. @@ -225,9 +225,9 @@ bool RPNBuilderTreeNode::tryGetConstant(Field & output_value, DataTypePtr & outp isColumnConst(*block_with_constants.getByName(column_name).column)) { /// An expression which is dependent on constants only - const auto & expr_info = block_with_constants.getByName(column_name); - output_value = (*expr_info.column)[0]; - output_type = expr_info.type; + const auto & constant_column = block_with_constants.getByName(column_name); + output_value = (*constant_column.column)[0]; + output_type = constant_column.type; if (!output_value.isNull()) output_type = removeNullable(output_type); @@ -260,13 +260,13 @@ ConstSetPtr tryGetSetFromDAGNode(const ActionsDAG::Node * dag_node) if (!dag_node->column) return {}; - const IColumn * col = dag_node->column.get(); - if (const auto * col_const = typeid_cast(col)) - col = &col_const->getDataColumn(); + const IColumn * column = dag_node->column.get(); + if (const auto * column_const = typeid_cast(column)) + column = &column_const->getDataColumn(); - if (const auto * col_set = typeid_cast(col)) + if (const auto * column_set = typeid_cast(column)) { - auto set = col_set->getData(); + auto set = column_set->getData(); if (set->isCreated()) return set; @@ -369,6 +369,17 @@ RPNBuilderFunctionTreeNode RPNBuilderTreeNode::toFunctionNode() const return RPNBuilderFunctionTreeNode(this->dag_node, tree_context); } +std::optional RPNBuilderTreeNode::toFunctionNodeOrNull() const +{ + if (!isFunction()) + return {}; + + if (this->ast_node) + return RPNBuilderFunctionTreeNode(this->ast_node, tree_context); + else + return RPNBuilderFunctionTreeNode(this->dag_node, tree_context); +} + std::string RPNBuilderFunctionTreeNode::getFunctionName() const { if (ast_node) diff --git a/src/Storages/MergeTree/RPNBuilder.h b/src/Storages/MergeTree/RPNBuilder.h index f6fc4a70102..132d3aa44e8 100644 --- a/src/Storages/MergeTree/RPNBuilder.h +++ b/src/Storages/MergeTree/RPNBuilder.h @@ -12,7 +12,7 @@ namespace DB /** Context of RPNBuilderTree. * - * For AST tree context, precalculated block with constansts and prepared sets are required for index analysis. + * For AST tree context, precalculated block with constants and prepared sets are required for index analysis. * For DAG tree precalculated block with constants and prepared sets are not required, because constants and sets already * calculated inside COLUMN actions dag node. */ @@ -112,6 +112,9 @@ public: */ RPNBuilderFunctionTreeNode toFunctionNode() const; + /// Convert node to function node or null optional + std::optional toFunctionNodeOrNull() const; + /// Get tree context const RPNBuilderTreeContext & getTreeContext() const { @@ -136,16 +139,16 @@ protected: class RPNBuilderFunctionTreeNode : public RPNBuilderTreeNode { public: + using RPNBuilderTreeNode::RPNBuilderTreeNode; + /// Get function name std::string getFunctionName() const; /// Get function arguments size size_t getArgumentsSize() const; - /// Get argument at index + /// Get function argument at index RPNBuilderTreeNode getArgumentAt(size_t index) const; - - using RPNBuilderTreeNode::RPNBuilderTreeNode; }; /** RPN Builder build stack of reverse polish notation elements (RPNElements) required for index analysis. @@ -240,9 +243,10 @@ private: bool extractLogicalOperatorFromTree(const RPNBuilderFunctionTreeNode & function_node, RPNElement & out) { - /// Functions AND, OR, NOT. - /// Also a special function `indexHint` - works as if instead of calling a function there are just parentheses - /// (or, the same thing - calling the function `and` from one argument). + /** Functions AND, OR, NOT. + * Also a special function `indexHint` - works as if instead of calling a function there are just parentheses + * (or, the same thing - calling the function `and` from one argument). + */ auto function_name = function_node.getFunctionName(); if (function_name == "not") From b677e68c4eadcc7b59e6a6094506cfe23cc19707 Mon Sep 17 00:00:00 2001 From: Sergei Trifonov Date: Mon, 31 Oct 2022 12:46:14 +0100 Subject: [PATCH 199/239] Update column.md --- docs/en/sql-reference/statements/alter/column.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/alter/column.md b/docs/en/sql-reference/statements/alter/column.md index f36aa1357f4..cc278465437 100644 --- a/docs/en/sql-reference/statements/alter/column.md +++ b/docs/en/sql-reference/statements/alter/column.md @@ -107,7 +107,7 @@ ALTER TABLE visits RENAME COLUMN webBrowser TO browser CLEAR COLUMN [IF EXISTS] name IN PARTITION partition_name ``` -Resets all data in a column for a specified partition. Read more about setting the partition name in the section [How to specify the partition expression](#alter-how-to-specify-part-expr). +Resets all data in a column for a specified partition. Read more about setting the partition name in the section [How to set the partition expression](partition.md#how-to-set-partition-expression). If the `IF EXISTS` clause is specified, the query won’t return an error if the column does not exist. From e43ecf9ca0f9c66661b706a2f80896d43264daa1 Mon Sep 17 00:00:00 2001 From: Sergei Trifonov Date: Mon, 31 Oct 2022 12:52:31 +0100 Subject: [PATCH 200/239] Link to proper place in docs --- docs/en/sql-reference/statements/optimize.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/optimize.md b/docs/en/sql-reference/statements/optimize.md index 680ff773992..036d3f0599a 100644 --- a/docs/en/sql-reference/statements/optimize.md +++ b/docs/en/sql-reference/statements/optimize.md @@ -22,7 +22,7 @@ The `OPTIMIZE` query is supported for [MergeTree](../../engines/table-engines/me When `OPTIMIZE` is used with the [ReplicatedMergeTree](../../engines/table-engines/mergetree-family/replication.md) family of table engines, ClickHouse creates a task for merging and waits for execution on all replicas (if the [replication_alter_partitions_sync](../../operations/settings/settings.md#replication-alter-partitions-sync) setting is set to `2`) or on current replica (if the [replication_alter_partitions_sync](../../operations/settings/settings.md#replication-alter-partitions-sync) setting is set to `1`). - If `OPTIMIZE` does not perform a merge for any reason, it does not notify the client. To enable notifications, use the [optimize_throw_if_noop](../../operations/settings/settings.md#setting-optimize_throw_if_noop) setting. -- If you specify a `PARTITION`, only the specified partition is optimized. [How to set partition expression](../../sql-reference/statements/alter/index.md#alter-how-to-specify-part-expr). +- If you specify a `PARTITION`, only the specified partition is optimized. [How to set partition expression](alter/partition.md#how-to-set-partition-expression). - If you specify `FINAL`, optimization is performed even when all the data is already in one part. Also merge is forced even if concurrent merges are performed. - If you specify `DEDUPLICATE`, then completely identical rows (unless by-clause is specified) will be deduplicated (all columns are compared), it makes sense only for the MergeTree engine. From 89a155d95cecf91031633f020851e1fe0b52858e Mon Sep 17 00:00:00 2001 From: Denny Crane Date: Mon, 31 Oct 2022 12:00:49 +0000 Subject: [PATCH 201/239] test for #28083 --- ...02343_analyzer_lambdas_issue_28083.reference | 2 ++ .../02343_analyzer_lambdas_issue_28083.sql | 17 +++++++++++++++++ 2 files changed, 19 insertions(+) create mode 100644 tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.reference create mode 100644 tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.sql diff --git a/tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.reference b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.reference new file mode 100644 index 00000000000..a2ed8e55d62 --- /dev/null +++ b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.reference @@ -0,0 +1,2 @@ +n [('n',1)] +[('n',1)] diff --git a/tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.sql b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.sql new file mode 100644 index 00000000000..3b780e1dec3 --- /dev/null +++ b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_28083.sql @@ -0,0 +1,17 @@ +SET allow_experimental_analyzer = 1; + +select so, + r +from + (select [('y',0),('n',1)] as cg, + if( arrayMap( x -> x.1, cg ) != ['y', 'n'], 'y', 'n') as so, + arrayFilter( x -> x.1 = so , cg) as r + ); + +select + r +from + (select [('y',0),('n',1)] as cg, + if( arrayMap( x -> x.1, cg ) != ['y', 'n'], 'y', 'n') as so, + arrayFilter( x -> x.1 = so , cg) as r + ); From 1c9fd1d0c777cc05d4650bb6900bb9f8efbd942e Mon Sep 17 00:00:00 2001 From: Denny Crane Date: Mon, 31 Oct 2022 12:01:04 +0000 Subject: [PATCH 202/239] test for #36677 --- .../02343_analyzer_lambdas_issue_36677.reference | 2 ++ .../02343_analyzer_lambdas_issue_36677.sql | 14 ++++++++++++++ 2 files changed, 16 insertions(+) create mode 100644 tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.reference create mode 100644 tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.sql diff --git a/tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.reference b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.reference new file mode 100644 index 00000000000..bec52aa46b6 --- /dev/null +++ b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.reference @@ -0,0 +1,2 @@ +2.1999999999999997 289.99999999999994 [1,2,3,4] [0.1,0.2,0.1,0.2] +2.1999999999999997 289.99999999999994 [1,2,3,4] [0.1,0.2,0.1,0.2] diff --git a/tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.sql b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.sql new file mode 100644 index 00000000000..b07f3f33ac3 --- /dev/null +++ b/tests/queries/0_stateless/02343_analyzer_lambdas_issue_36677.sql @@ -0,0 +1,14 @@ +SET allow_experimental_analyzer = 1; + +SELECT + arraySum(x -> ((x.1) / ((x.2) * (x.2))), arrayZip(mag, magerr)) / arraySum(x -> (1. / (x * x)), magerr) AS weightedmeanmag, + arraySum(x -> ((((x.1) - weightedmeanmag) * ((x.1) - weightedmeanmag)) / ((x.2) * (x.2))), arrayZip(mag, magerr)) AS chi2, + [1, 2, 3, 4] AS mag, + [0.1, 0.2, 0.1, 0.2] AS magerr; + +SELECT + arraySum(x -> ((x.1) / ((x.2) * (x.2))), arrayZip(mag, magerr)) / arraySum(x -> (1. / (x * x)), magerr) AS weightedmeanmag, + arraySum(x -> ((((x.1) - weightedmeanmag) * ((x.1) - weightedmeanmag)) / ((x.2) * (x.2))), arrayZip(mag, magerr)) AS chi2, + [1, 2, 3, 4] AS mag, + [0.1, 0.2, 0.1, 0.2] AS magerr +WHERE isFinite(chi2) From b2b9479afa810ce7e69f820ecc4a3e0b5db4e6a7 Mon Sep 17 00:00:00 2001 From: Alexander Tokmakov Date: Mon, 31 Oct 2022 16:15:35 +0300 Subject: [PATCH 203/239] Update src/Storages/MergeTree/MergeTreeData.h Co-authored-by: Sergei Trifonov --- src/Storages/MergeTree/MergeTreeData.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index 40eaa679845..cfad11b8d36 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -591,7 +591,7 @@ public: bool was_active; public: - PartToRemoveFromZooKeeper(DataPartPtr && part_, bool was_active_ = true) + explicit PartToRemoveFromZooKeeper(DataPartPtr && part_, bool was_active_ = true) : part(std::move(part_)), was_active(was_active_) { } From 599ccb99396a84151b280bdf809ff54d04e8c38a Mon Sep 17 00:00:00 2001 From: Alexander Tokmakov Date: Mon, 31 Oct 2022 16:18:17 +0300 Subject: [PATCH 204/239] Update MergeTreeData.h --- src/Storages/MergeTree/MergeTreeData.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/Storages/MergeTree/MergeTreeData.h b/src/Storages/MergeTree/MergeTreeData.h index cfad11b8d36..99ba6991e43 100644 --- a/src/Storages/MergeTree/MergeTreeData.h +++ b/src/Storages/MergeTree/MergeTreeData.h @@ -596,7 +596,7 @@ public: { } - /// It's s to get name of any part + /// It's safe to get name of any part const String & getPartName() const { return part->name; } DataPartPtr getPartIfItWasActive() const From 399d024c2c58f928c14058f887dd2f635eb06aca Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?M=C3=A1rcio=20Martins?= Date: Mon, 31 Oct 2022 15:02:08 +0100 Subject: [PATCH 205/239] Fix array index argument type check for JSONExtract* functions --- src/Functions/FunctionsJSON.cpp | 2 +- tests/queries/0_stateless/00918_json_functions.reference | 1 + tests/queries/0_stateless/00918_json_functions.sql | 3 +++ 3 files changed, 5 insertions(+), 1 deletion(-) diff --git a/src/Functions/FunctionsJSON.cpp b/src/Functions/FunctionsJSON.cpp index 7c221209071..c856419c9e8 100644 --- a/src/Functions/FunctionsJSON.cpp +++ b/src/Functions/FunctionsJSON.cpp @@ -191,7 +191,7 @@ private: for (const auto i : collections::range(first_index_argument, first_index_argument + num_index_arguments)) { const auto & column = columns[i]; - if (!isString(column.type) && !isInteger(column.type)) + if (!isString(column.type) && !isNativeInteger(column.type)) throw Exception{"The argument " + std::to_string(i + 1) + " of function " + String(function_name) + " should be a string specifying key or an integer specifying index, illegal type: " + column.type->getName(), ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT}; diff --git a/tests/queries/0_stateless/00918_json_functions.reference b/tests/queries/0_stateless/00918_json_functions.reference index 7d925c1236d..fc03457c677 100644 --- a/tests/queries/0_stateless/00918_json_functions.reference +++ b/tests/queries/0_stateless/00918_json_functions.reference @@ -253,3 +253,4 @@ e u v --show error: type should be const string +--show error: index type should be integer diff --git a/tests/queries/0_stateless/00918_json_functions.sql b/tests/queries/0_stateless/00918_json_functions.sql index d614d507dda..3105994ce20 100644 --- a/tests/queries/0_stateless/00918_json_functions.sql +++ b/tests/queries/0_stateless/00918_json_functions.sql @@ -280,3 +280,6 @@ SELECT JSONExtractString(json, 's') FROM (SELECT arrayJoin(['{"s":"u"}', '{"s":" SELECT '--show error: type should be const string'; SELECT JSONExtractKeysAndValues([], JSONLength('^?V{LSwp')); -- { serverError 44 } WITH '{"i": 1, "f": 1.2}' AS json SELECT JSONExtract(json, 'i', JSONType(json, 'i')); -- { serverError 44 } + +SELECT '--show error: index type should be integer'; +SELECT JSONExtract('[]', JSONExtract('0', 'UInt256'), 'UInt256'); -- { serverError 43 } From 3742fd848a11642325bc345ee84f4d8574b3887b Mon Sep 17 00:00:00 2001 From: Alexander Gololobov <440544+davenger@users.noreply.github.com> Date: Mon, 31 Oct 2022 15:31:14 +0100 Subject: [PATCH 206/239] Tests for upperUTF8() and for 32 and 64 byte boundaries --- .../02071_lower_upper_utf8_row_overlaps.reference | 15 +++++++++++++++ .../02071_lower_upper_utf8_row_overlaps.sql | 8 ++++++++ 2 files changed, 23 insertions(+) diff --git a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference index 190c7895011..a3bac432482 100644 --- a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference +++ b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.reference @@ -13,3 +13,18 @@ select length(str), if(l_ == '\xe2', h_, l_), if(u_ == '\xe2', h_, u_) from utf8 -- https://github.com/ClickHouse/ClickHouse/issues/42756 SELECT lowerUTF8('КВ АМ И СЖ'); кв ам и сж +SELECT upperUTF8('кв ам и сж'); +КВ АМ И СЖ +SELECT lowerUTF8('КВ АМ И СЖ КВ АМ И СЖ'); +кв ам и сж кв ам и сж +SELECT upperUTF8('кв ам и сж кв ам и сж'); +КВ АМ И СЖ КВ АМ И СЖ +-- Test at 32 and 64 byte boundaries +SELECT lowerUTF8(repeat('0', 16) || 'КВ АМ И СЖ'); +0000000000000000кв ам и сж +SELECT upperUTF8(repeat('0', 16) || 'кв ам и сж'); +0000000000000000КВ АМ И СЖ +SELECT lowerUTF8(repeat('0', 48) || 'КВ АМ И СЖ'); +000000000000000000000000000000000000000000000000кв ам и сж +SELECT upperUTF8(repeat('0', 48) || 'кв ам и сж'); +000000000000000000000000000000000000000000000000КВ АМ И СЖ diff --git a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql index 161d8fb512e..8ca0a3f5f75 100644 --- a/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql +++ b/tests/queries/0_stateless/02071_lower_upper_utf8_row_overlaps.sql @@ -12,3 +12,11 @@ select length(str), if(l_ == '\xe2', h_, l_), if(u_ == '\xe2', h_, u_) from utf8 -- NOTE: regression test for introduced bug -- https://github.com/ClickHouse/ClickHouse/issues/42756 SELECT lowerUTF8('КВ АМ И СЖ'); +SELECT upperUTF8('кв ам и сж'); +SELECT lowerUTF8('КВ АМ И СЖ КВ АМ И СЖ'); +SELECT upperUTF8('кв ам и сж кв ам и сж'); +-- Test at 32 and 64 byte boundaries +SELECT lowerUTF8(repeat('0', 16) || 'КВ АМ И СЖ'); +SELECT upperUTF8(repeat('0', 16) || 'кв ам и сж'); +SELECT lowerUTF8(repeat('0', 48) || 'КВ АМ И СЖ'); +SELECT upperUTF8(repeat('0', 48) || 'кв ам и сж'); From 9fb6f5228675add4662c52cacb70d38dd571776d Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Mon, 31 Oct 2022 11:02:08 -0400 Subject: [PATCH 207/239] Fix Missing Quotes - Sonar Nightly Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index e6da4df7200..5a208807c81 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -149,7 +149,7 @@ jobs: env: BUILD_WRAPPER_DOWNLOAD_URL: ${{ env.SONAR_SERVER_URL }}/static/cpp/build-wrapper-linux-x86.zip run: | - curl -sSLo "$HOME/.sonar/build-wrapper-linux-x86.zip ${{ env.BUILD_WRAPPER_DOWNLOAD_URL }}" + curl -sSLo "$HOME/.sonar/build-wrapper-linux-x86.zip" "${{ env.BUILD_WRAPPER_DOWNLOAD_URL }}" unzip -o "$HOME/.sonar/build-wrapper-linux-x86.zip" -d "$HOME/.sonar/" echo "$HOME/.sonar/build-wrapper-linux-x86" >> "$GITHUB_PATH" - name: Set Up Build Tools From f8091424629714ff43094e224f6cf95f79b3f715 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Mon, 31 Oct 2022 17:47:02 +0100 Subject: [PATCH 208/239] Add exclusions from the Snyk scan --- .snyk | 4 ++++ 1 file changed, 4 insertions(+) create mode 100644 .snyk diff --git a/.snyk b/.snyk new file mode 100644 index 00000000000..7acc6b9fbf5 --- /dev/null +++ b/.snyk @@ -0,0 +1,4 @@ +# Snyk (https://snyk.io) policy file +exclude: + global: + - tests/** From 440cc51a7e7aa6df4bde96c895965048a710f3c1 Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Mon, 31 Oct 2022 14:11:52 -0400 Subject: [PATCH 209/239] Fix Missing Env Vars - Sonar Nightly Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 5a208807c81..904c0dffe57 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -128,6 +128,8 @@ jobs: SONAR_SCANNER_VERSION: 4.7.0.2747 SONAR_SERVER_URL: "https://sonarcloud.io" BUILD_WRAPPER_OUT_DIR: build_wrapper_output_directory # Directory where build-wrapper output will be placed + CC: clang-15 + CXX: clang++-15 steps: - uses: actions/checkout@v2 with: From d18d08bcc78431977abe735e3bc6793e6e2308c2 Mon Sep 17 00:00:00 2001 From: Zhiguo Zhou Date: Tue, 1 Nov 2022 04:36:15 +0800 Subject: [PATCH 210/239] Remove short-circuit evaluation in AssociativeApplierImpl::apply (#42214) The short-circuit evaluation was implemented when applying the saturable operators (and, or) on a vector of ColumnUInt8. However, its control flow would be compiled as a series of conditional branch instructions which are hard to predict by the hardware and at the same time hinder the vectorization optimization by the compiler. This commit removes the short-circuit and evaluates the whole expression. --- src/Functions/FunctionsLogical.cpp | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/src/Functions/FunctionsLogical.cpp b/src/Functions/FunctionsLogical.cpp index 2ac7688737f..7e52c55e5b0 100644 --- a/src/Functions/FunctionsLogical.cpp +++ b/src/Functions/FunctionsLogical.cpp @@ -168,10 +168,7 @@ public: inline ResultValueType apply(const size_t i) const { const auto a = !!vec[i]; - if constexpr (Op::isSaturable()) - return Op::isSaturatedValue(a) ? a : Op::apply(a, next.apply(i)); - else - return Op::apply(a, next.apply(i)); + return Op::apply(a, next.apply(i)); } private: From 439ddc2bf7d07122cd15d8054b7c9400ad29578e Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Mon, 31 Oct 2022 16:53:58 -0400 Subject: [PATCH 211/239] exclude java Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 904c0dffe57..94fe404ea97 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -175,4 +175,5 @@ jobs: --define sonar.host.url="${{ env.SONAR_SERVER_URL }}" \ --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ --define sonar.projectKey="ClickHouse_ClickHouse" \ - --define sonar.organization="clickhouse-java" + --define sonar.organization="clickhouse-java" \ + --define sonar.exclusions="**/*.java" From baf0e92e60185ade936ec0caa4ff0cd2960d6751 Mon Sep 17 00:00:00 2001 From: Julio Jimenez Date: Mon, 31 Oct 2022 17:58:43 -0400 Subject: [PATCH 212/239] exclude more stuff Signed-off-by: Julio Jimenez --- .github/workflows/nightly.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/nightly.yml b/.github/workflows/nightly.yml index 94fe404ea97..612bb1f8f9b 100644 --- a/.github/workflows/nightly.yml +++ b/.github/workflows/nightly.yml @@ -176,4 +176,4 @@ jobs: --define sonar.cfamily.build-wrapper-output="${{ env.BUILD_WRAPPER_OUT_DIR }}" \ --define sonar.projectKey="ClickHouse_ClickHouse" \ --define sonar.organization="clickhouse-java" \ - --define sonar.exclusions="**/*.java" + --define sonar.exclusions="**/*.java,**/*.ts,**/*.js,**/*.css,**/*.sql" From 70eaf69df176a3cc2dee84f88fbb22fc1185478b Mon Sep 17 00:00:00 2001 From: Alexander Yakovlev <33040934+AlexJameson@users.noreply.github.com> Date: Tue, 1 Nov 2022 02:57:55 +0300 Subject: [PATCH 213/239] Fix a typo in table-engines/integrations/s3.md --- docs/en/engines/table-engines/integrations/s3.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/engines/table-engines/integrations/s3.md b/docs/en/engines/table-engines/integrations/s3.md index 986a29b8307..db983ab9c68 100644 --- a/docs/en/engines/table-engines/integrations/s3.md +++ b/docs/en/engines/table-engines/integrations/s3.md @@ -139,7 +139,7 @@ The following settings can be specified in configuration file for given endpoint - `use_environment_credentials` — If set to `true`, S3 client will try to obtain credentials from environment variables and [Amazon EC2](https://en.wikipedia.org/wiki/Amazon_Elastic_Compute_Cloud) metadata for given endpoint. Optional, default value is `false`. - `region` — Specifies S3 region name. Optional. - `use_insecure_imds_request` — If set to `true`, S3 client will use insecure IMDS request while obtaining credentials from Amazon EC2 metadata. Optional, default value is `false`. -- `header` — Adds specified HTTP header to a request to given endpoint. Optional, can be speficied multiple times. +- `header` — Adds specified HTTP header to a request to given endpoint. Optional, can be specified multiple times. - `server_side_encryption_customer_key_base64` — If specified, required headers for accessing S3 objects with SSE-C encryption will be set. Optional. - `max_single_read_retries` — The maximum number of attempts during single read. Default value is `4`. Optional. From 8b849d4b374db3527d3857b9e88908a5a32179a7 Mon Sep 17 00:00:00 2001 From: 94rain <21193371+94rain@users.noreply.github.com> Date: Mon, 31 Oct 2022 20:45:01 +0000 Subject: [PATCH 214/239] Fix broken links in system tables docs Signed-off-by: 94rain <21193371+94rain@users.noreply.github.com> --- docs/en/operations/system-tables/information_schema.md | 2 +- docs/en/operations/system-tables/replicated_fetches.md | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/en/operations/system-tables/information_schema.md b/docs/en/operations/system-tables/information_schema.md index a573491282a..811e0dc3c3f 100644 --- a/docs/en/operations/system-tables/information_schema.md +++ b/docs/en/operations/system-tables/information_schema.md @@ -178,7 +178,7 @@ Columns: - `view_definition` ([String](../../sql-reference/data-types/string.md)) — `SELECT` query for view. - `check_option` ([String](../../sql-reference/data-types/string.md)) — `NONE`, no checking. - `is_updatable` ([Enum8](../../sql-reference/data-types/enum.md)) — `NO`, the view is not updated. -- `is_insertable_into` ([Enum8](../../sql-reference/data-types/enum.md)) — Shows whether the created view is [materialized](../../sql-reference/statements/create/view/#materialized). Possible values: +- `is_insertable_into` ([Enum8](../../sql-reference/data-types/enum.md)) — Shows whether the created view is [materialized](../../sql-reference/statements/create/view.md#materialized-view). Possible values: - `NO` — The created view is not materialized. - `YES` — The created view is materialized. - `is_trigger_updatable` ([Enum8](../../sql-reference/data-types/enum.md)) — `NO`, the trigger is not updated. diff --git a/docs/en/operations/system-tables/replicated_fetches.md b/docs/en/operations/system-tables/replicated_fetches.md index 3536bbaff4d..734da63a600 100644 --- a/docs/en/operations/system-tables/replicated_fetches.md +++ b/docs/en/operations/system-tables/replicated_fetches.md @@ -68,6 +68,6 @@ thread_id: 54 **See Also** -- [Managing ReplicatedMergeTree Tables](../../sql-reference/statements/system/#query-language-system-replicated) +- [Managing ReplicatedMergeTree Tables](../../sql-reference/statements/system.md#managing-replicatedmergetree-tables) -[Original article](https://clickhouse.com/docs/en/operations/system_tables/replicated_fetches) +[Original article](https://clickhouse.com/docs/en/operations/system-tables/replicated_fetches/) From 7234600ec31f511146bd795df91d1936c5a42200 Mon Sep 17 00:00:00 2001 From: shuyang <21193371+94rain@users.noreply.github.com> Date: Mon, 31 Oct 2022 19:14:58 -0500 Subject: [PATCH 215/239] Apply suggestions from code review Co-authored-by: Dan Roscigno --- docs/en/operations/system-tables/information_schema.md | 2 +- docs/en/operations/system-tables/replicated_fetches.md | 3 +-- 2 files changed, 2 insertions(+), 3 deletions(-) diff --git a/docs/en/operations/system-tables/information_schema.md b/docs/en/operations/system-tables/information_schema.md index 811e0dc3c3f..a8e516f02a3 100644 --- a/docs/en/operations/system-tables/information_schema.md +++ b/docs/en/operations/system-tables/information_schema.md @@ -178,7 +178,7 @@ Columns: - `view_definition` ([String](../../sql-reference/data-types/string.md)) — `SELECT` query for view. - `check_option` ([String](../../sql-reference/data-types/string.md)) — `NONE`, no checking. - `is_updatable` ([Enum8](../../sql-reference/data-types/enum.md)) — `NO`, the view is not updated. -- `is_insertable_into` ([Enum8](../../sql-reference/data-types/enum.md)) — Shows whether the created view is [materialized](../../sql-reference/statements/create/view.md#materialized-view). Possible values: +- `is_insertable_into` ([Enum8](../../sql-reference/data-types/enum.md)) — Shows whether the created view is [materialized](../../sql-reference/statements/create/view.md/#materialized-view). Possible values: - `NO` — The created view is not materialized. - `YES` — The created view is materialized. - `is_trigger_updatable` ([Enum8](../../sql-reference/data-types/enum.md)) — `NO`, the trigger is not updated. diff --git a/docs/en/operations/system-tables/replicated_fetches.md b/docs/en/operations/system-tables/replicated_fetches.md index 734da63a600..74888fd2f13 100644 --- a/docs/en/operations/system-tables/replicated_fetches.md +++ b/docs/en/operations/system-tables/replicated_fetches.md @@ -68,6 +68,5 @@ thread_id: 54 **See Also** -- [Managing ReplicatedMergeTree Tables](../../sql-reference/statements/system.md#managing-replicatedmergetree-tables) +- [Managing ReplicatedMergeTree Tables](../../sql-reference/statements/system.md/#managing-replicatedmergetree-tables) -[Original article](https://clickhouse.com/docs/en/operations/system-tables/replicated_fetches/) From 2b0cde9ef9f9d77eac3b8045f1683ed693a1fb01 Mon Sep 17 00:00:00 2001 From: GoGoWen2021 Date: Tue, 1 Nov 2022 17:46:44 +0800 Subject: [PATCH 216/239] fix typo --- src/Columns/IColumn.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/Columns/IColumn.h b/src/Columns/IColumn.h index 19f3dea4f82..eda4ae62b03 100644 --- a/src/Columns/IColumn.h +++ b/src/Columns/IColumn.h @@ -85,8 +85,8 @@ public: [[nodiscard]] virtual MutablePtr cloneEmpty() const { return cloneResized(0); } /// Creates column with the same type and specified size. - /// If size is less current size, then data is cut. - /// If size is greater, than default values are appended. + /// If size is less than current size, then data is cut. + /// If size is greater, then default values are appended. [[nodiscard]] virtual MutablePtr cloneResized(size_t /*size*/) const { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot cloneResized() column {}", getName()); } /// Returns number of values in column. From 29318a1e133a29ff99602f53b3376bd0fa296465 Mon Sep 17 00:00:00 2001 From: Derek Chia Date: Tue, 1 Nov 2022 18:03:17 +0800 Subject: [PATCH 217/239] Update nyc-taxi.md - Add enum 'UNK' = 5 to `payment_type` to avoid `Unknown element 'UNK' for enum` error when loading trips_*.gz - Change url table function to s3 table function as described in the earlier paragraph --- docs/en/getting-started/example-datasets/nyc-taxi.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/en/getting-started/example-datasets/nyc-taxi.md b/docs/en/getting-started/example-datasets/nyc-taxi.md index e24fb4b01a7..69098f63037 100644 --- a/docs/en/getting-started/example-datasets/nyc-taxi.md +++ b/docs/en/getting-started/example-datasets/nyc-taxi.md @@ -33,7 +33,7 @@ CREATE TABLE trips ( tip_amount Float32, tolls_amount Float32, total_amount Float32, - payment_type Enum('CSH' = 1, 'CRE' = 2, 'NOC' = 3, 'DIS' = 4), + payment_type Enum('CSH' = 1, 'CRE' = 2, 'NOC' = 3, 'DIS' = 4, 'UNK' = 5), pickup_ntaname LowCardinality(String), dropoff_ntaname LowCardinality(String) ) @@ -63,7 +63,7 @@ SELECT payment_type, pickup_ntaname, dropoff_ntaname -FROM url( +FROM s3( 'https://datasets-documentation.s3.eu-west-3.amazonaws.com/nyc-taxi/trips_{0..2}.gz', 'TabSeparatedWithNames' ) From e0735e1e67da3c74dd43dabd7147cce2d85d49ed Mon Sep 17 00:00:00 2001 From: vdimir Date: Wed, 26 Oct 2022 14:02:04 +0000 Subject: [PATCH 218/239] Debug TestBugfixCheck --- tests/ci/functional_test_check.py | 13 ++++++++++--- .../0_stateless/01744_fuse_sum_count_aggregate.sql | 1 + 2 files changed, 11 insertions(+), 3 deletions(-) diff --git a/tests/ci/functional_test_check.py b/tests/ci/functional_test_check.py index 388f93f34ec..b4601f766f8 100644 --- a/tests/ci/functional_test_check.py +++ b/tests/ci/functional_test_check.py @@ -7,6 +7,7 @@ import os import subprocess import sys import atexit +import time from github import Github @@ -221,12 +222,18 @@ if __name__ == "__main__": if args.post_commit_status == "file": post_commit_status_to_file( os.path.join(temp_path, "post_commit_status.tsv"), - "Skipped (no pr-bugfix)", + f"Skipped (no pr-bugfix in {pr_info.labels})", "success", "null", ) - logging.info("Skipping '%s' (no pr-bugfix)", check_name) - sys.exit(0) + logging.info("Skipping '%s' (no pr-bugfix in %s)", check_name, pr_info.labels) + + # >>>> + time.sleep(10) + pr_info = PRInfo(need_changed_files=run_changed_tests) + logging.info("pr_info '%s' (%s)", pr_info, pr_info.labels) + sys.exit(1) + # <<<< if "RUN_BY_HASH_NUM" in os.environ: run_by_hash_num = int(os.getenv("RUN_BY_HASH_NUM")) diff --git a/tests/queries/0_stateless/01744_fuse_sum_count_aggregate.sql b/tests/queries/0_stateless/01744_fuse_sum_count_aggregate.sql index 5b6ed440ba4..375662eb405 100644 --- a/tests/queries/0_stateless/01744_fuse_sum_count_aggregate.sql +++ b/tests/queries/0_stateless/01744_fuse_sum_count_aggregate.sql @@ -10,4 +10,5 @@ EXPLAIN SYNTAX SELECT sum(a), sum(b), count(b) from fuse_tbl; SELECT '---------NOT trigger fuse--------'; SELECT sum(a), avg(b) from fuse_tbl; EXPLAIN SYNTAX SELECT sum(a), avg(b) from fuse_tbl; + DROP TABLE fuse_tbl; From 9f5dcd14af323c9c0063cc585f6186556559d691 Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 10:49:02 +0000 Subject: [PATCH 219/239] Request labels from gh api for validate bugfix check --- tests/ci/functional_test_check.py | 11 +++-------- tests/ci/integration_test_check.py | 8 +++++--- 2 files changed, 8 insertions(+), 11 deletions(-) diff --git a/tests/ci/functional_test_check.py b/tests/ci/functional_test_check.py index b4601f766f8..9b40d7facbe 100644 --- a/tests/ci/functional_test_check.py +++ b/tests/ci/functional_test_check.py @@ -211,7 +211,8 @@ if __name__ == "__main__": run_changed_tests = flaky_check or validate_bugix_check gh = Github(get_best_robot_token(), per_page=100) - pr_info = PRInfo(need_changed_files=run_changed_tests) + # For validate_bugix_check we need up to date information about labels, so pr_event_from_api is used + pr_info = PRInfo(need_changed_files=run_changed_tests, pr_event_from_api=validate_bugix_check) atexit.register(update_mergeable_check, gh, pr_info, check_name) @@ -227,13 +228,7 @@ if __name__ == "__main__": "null", ) logging.info("Skipping '%s' (no pr-bugfix in %s)", check_name, pr_info.labels) - - # >>>> - time.sleep(10) - pr_info = PRInfo(need_changed_files=run_changed_tests) - logging.info("pr_info '%s' (%s)", pr_info, pr_info.labels) - sys.exit(1) - # <<<< + sys.exit(0) if "RUN_BY_HASH_NUM" in os.environ: run_by_hash_num = int(os.getenv("RUN_BY_HASH_NUM")) diff --git a/tests/ci/integration_test_check.py b/tests/ci/integration_test_check.py index 3709a7271d7..2cdbd553144 100644 --- a/tests/ci/integration_test_check.py +++ b/tests/ci/integration_test_check.py @@ -167,17 +167,19 @@ if __name__ == "__main__": os.makedirs(temp_path) is_flaky_check = "flaky" in check_name - pr_info = PRInfo(need_changed_files=is_flaky_check or validate_bugix_check) + + # For validate_bugix_check we need up to date information about labels, so pr_event_from_api is used + pr_info = PRInfo(need_changed_files=is_flaky_check or validate_bugix_check, pr_event_from_api=validate_bugix_check) if validate_bugix_check and "pr-bugfix" not in pr_info.labels: if args.post_commit_status == "file": post_commit_status_to_file( os.path.join(temp_path, "post_commit_status.tsv"), - "Skipped (no pr-bugfix)", + f"Skipped (no pr-bugfix in {pr_info.labels})", "success", "null", ) - logging.info("Skipping '%s' (no pr-bugfix)", check_name) + logging.info("Skipping '%s' (no pr-bugfix in '%s')", check_name, pr_info.labels) sys.exit(0) gh = Github(get_best_robot_token(), per_page=100) From c18838715da399605df9de0a7d59561f89281dcf Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 11:32:42 +0000 Subject: [PATCH 220/239] Post status for test bugfix check --- tests/ci/bugfix_validate_check.py | 52 ++++++++++++++++++++++++------ tests/ci/functional_test_check.py | 6 ++-- tests/ci/integration_test_check.py | 7 ++-- 3 files changed, 51 insertions(+), 14 deletions(-) diff --git a/tests/ci/bugfix_validate_check.py b/tests/ci/bugfix_validate_check.py index 4e6001aaa74..6ab47392dca 100644 --- a/tests/ci/bugfix_validate_check.py +++ b/tests/ci/bugfix_validate_check.py @@ -7,10 +7,18 @@ import os import sys +from github import Github + +from s3_helper import S3Helper +from get_robot_token import get_best_robot_token +from pr_info import PRInfo +from upload_result_helper import upload_results +from commit_status_helper import post_commit_status + + def parse_args(): parser = argparse.ArgumentParser() - parser.add_argument("report1") - parser.add_argument("report2") + parser.add_argument("status", nargs="+", help="Path to status file") return parser.parse_args() @@ -27,19 +35,43 @@ def post_commit_status_from_file(file_path): def process_results(file_path): + test_results = [] state, report_url, description = post_commit_status_from_file(file_path) prefix = os.path.basename(os.path.dirname(file_path)) - print( - f"::notice:: bugfix check: {prefix} - {state}: {description} Report url: {report_url}" - ) - return state == "success" + + test_results.append([f"{prefix}: {description}", state, report_url]) + return state == "success", test_results def main(args): - is_ok = False - is_ok = process_results(args.report1) or is_ok - is_ok = process_results(args.report2) or is_ok - sys.exit(0 if is_ok else 1) + all_ok = False + all_results = [] + for status_path in args.status: + is_ok, test_results = process_results(status_path) + all_ok = all_ok or is_ok + all_results.extend(test_results) + + check_name_with_group = "Bugfix validate check" + + pr_info = PRInfo() + report_url = upload_results( + S3Helper(), + pr_info.number, + pr_info.sha, + all_results, + [], + check_name_with_group, + ) + + gh = Github(get_best_robot_token(), per_page=100) + post_commit_status( + gh, + pr_info.sha, + check_name_with_group, + "", + "success" if is_ok else "error", + report_url, + ) if __name__ == "__main__": diff --git a/tests/ci/functional_test_check.py b/tests/ci/functional_test_check.py index 9b40d7facbe..1d215920cb5 100644 --- a/tests/ci/functional_test_check.py +++ b/tests/ci/functional_test_check.py @@ -212,7 +212,9 @@ if __name__ == "__main__": gh = Github(get_best_robot_token(), per_page=100) # For validate_bugix_check we need up to date information about labels, so pr_event_from_api is used - pr_info = PRInfo(need_changed_files=run_changed_tests, pr_event_from_api=validate_bugix_check) + pr_info = PRInfo( + need_changed_files=run_changed_tests, pr_event_from_api=validate_bugix_check + ) atexit.register(update_mergeable_check, gh, pr_info, check_name) @@ -322,7 +324,7 @@ if __name__ == "__main__": state, description, test_results, additional_logs = process_results( result_path, server_log_path ) - state = override_status(state, check_name, validate_bugix_check) + state = override_status(state, check_name, invert=validate_bugix_check) ch_helper = ClickHouseHelper() mark_flaky_tests(ch_helper, check_name, test_results) diff --git a/tests/ci/integration_test_check.py b/tests/ci/integration_test_check.py index 2cdbd553144..cba428cbcf5 100644 --- a/tests/ci/integration_test_check.py +++ b/tests/ci/integration_test_check.py @@ -169,7 +169,10 @@ if __name__ == "__main__": is_flaky_check = "flaky" in check_name # For validate_bugix_check we need up to date information about labels, so pr_event_from_api is used - pr_info = PRInfo(need_changed_files=is_flaky_check or validate_bugix_check, pr_event_from_api=validate_bugix_check) + pr_info = PRInfo( + need_changed_files=is_flaky_check or validate_bugix_check, + pr_event_from_api=validate_bugix_check, + ) if validate_bugix_check and "pr-bugfix" not in pr_info.labels: if args.post_commit_status == "file": @@ -246,7 +249,7 @@ if __name__ == "__main__": subprocess.check_call(f"sudo chown -R ubuntu:ubuntu {temp_path}", shell=True) state, description, test_results, additional_logs = process_results(result_path) - state = override_status(state, check_name, validate_bugix_check) + state = override_status(state, check_name, invert=validate_bugix_check) ch_helper = ClickHouseHelper() mark_flaky_tests(ch_helper, check_name, test_results) From 11a933b0465bedea333ff48dfb90078869d5822f Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 11:34:45 +0000 Subject: [PATCH 221/239] broke style in src/Interpreters/InterpreterSelectQuery.cpp --- src/Interpreters/InterpreterSelectQuery.cpp | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/Interpreters/InterpreterSelectQuery.cpp b/src/Interpreters/InterpreterSelectQuery.cpp index d8ac263e3d1..225dfd1f607 100644 --- a/src/Interpreters/InterpreterSelectQuery.cpp +++ b/src/Interpreters/InterpreterSelectQuery.cpp @@ -138,8 +138,7 @@ FilterDAGInfoPtr generateFilterActions( expr_list->children.push_back(row_policy_filter->clone()); /// Keep columns that are required after the filter actions. - for (const auto & column_str : prerequisite_columns) - { + for (const auto & column_str : prerequisite_columns) { ParserExpression expr_parser; expr_list->children.push_back(parseQuery(expr_parser, column_str, 0, context->getSettingsRef().max_parser_depth)); } From 1b32c7d21277887b3475ef15a53c15a4c13c16a8 Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 13:06:37 +0000 Subject: [PATCH 222/239] Update bugfix_validate_check.py --- tests/ci/bugfix_validate_check.py | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/tests/ci/bugfix_validate_check.py b/tests/ci/bugfix_validate_check.py index 6ab47392dca..975c3eba2bb 100644 --- a/tests/ci/bugfix_validate_check.py +++ b/tests/ci/bugfix_validate_check.py @@ -38,9 +38,16 @@ def process_results(file_path): test_results = [] state, report_url, description = post_commit_status_from_file(file_path) prefix = os.path.basename(os.path.dirname(file_path)) + is_ok = state == "success" - test_results.append([f"{prefix}: {description}", state, report_url]) - return state == "success", test_results + test_results.append( + [ + f"{prefix}: {description}", + "Bug reproduced" if is_ok else "Bug is not reproduced", + report_url, + ] + ) + return is_ok, test_results def main(args): @@ -68,7 +75,7 @@ def main(args): gh, pr_info.sha, check_name_with_group, - "", + "" if is_ok else "Changed tests doesn't reproduce the bug", "success" if is_ok else "error", report_url, ) From bc78810f2c73ef91d178c0814805f7ec222c4769 Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 15:08:05 +0000 Subject: [PATCH 223/239] Add 02474_dummy.sql --- tests/queries/0_stateless/02474_dummy.reference | 1 + tests/queries/0_stateless/02474_dummy.sql | 2 ++ 2 files changed, 3 insertions(+) create mode 100644 tests/queries/0_stateless/02474_dummy.reference create mode 100644 tests/queries/0_stateless/02474_dummy.sql diff --git a/tests/queries/0_stateless/02474_dummy.reference b/tests/queries/0_stateless/02474_dummy.reference new file mode 100644 index 00000000000..d00491fd7e5 --- /dev/null +++ b/tests/queries/0_stateless/02474_dummy.reference @@ -0,0 +1 @@ +1 diff --git a/tests/queries/0_stateless/02474_dummy.sql b/tests/queries/0_stateless/02474_dummy.sql new file mode 100644 index 00000000000..bf2485aeb69 --- /dev/null +++ b/tests/queries/0_stateless/02474_dummy.sql @@ -0,0 +1,2 @@ +-- Tags: no-fasttest +SELECT 0; From 63b63da214340704269ee881d8129e09725b31af Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 15:11:16 +0000 Subject: [PATCH 224/239] Update bugfix_validate_check.py --- tests/ci/bugfix_validate_check.py | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) diff --git a/tests/ci/bugfix_validate_check.py b/tests/ci/bugfix_validate_check.py index 975c3eba2bb..97294152941 100644 --- a/tests/ci/bugfix_validate_check.py +++ b/tests/ci/bugfix_validate_check.py @@ -34,7 +34,7 @@ def post_commit_status_from_file(file_path): return res[0] -def process_results(file_path): +def process_result(file_path): test_results = [] state, report_url, description = post_commit_status_from_file(file_path) prefix = os.path.basename(os.path.dirname(file_path)) @@ -50,22 +50,27 @@ def process_results(file_path): return is_ok, test_results -def main(args): - all_ok = False +def process_all_results(file_paths): + all_ok = True all_results = [] - for status_path in args.status: - is_ok, test_results = process_results(status_path) - all_ok = all_ok or is_ok + for status_path in file_paths: + is_ok, test_results = process_result(status_path) + all_ok = all_ok and is_ok all_results.extend(test_results) + return all_ok, all_results + +def main(args): check_name_with_group = "Bugfix validate check" + is_ok, test_results = process_all_results(args.status) + pr_info = PRInfo() report_url = upload_results( S3Helper(), pr_info.number, pr_info.sha, - all_results, + test_results, [], check_name_with_group, ) From e83d61749a7748b347f05cd4c99c9fd69d59c81e Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 16:51:11 +0000 Subject: [PATCH 225/239] update test bugfix report --- tests/ci/bugfix_validate_check.py | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) diff --git a/tests/ci/bugfix_validate_check.py b/tests/ci/bugfix_validate_check.py index 97294152941..c9d3886b73b 100644 --- a/tests/ci/bugfix_validate_check.py +++ b/tests/ci/bugfix_validate_check.py @@ -40,24 +40,21 @@ def process_result(file_path): prefix = os.path.basename(os.path.dirname(file_path)) is_ok = state == "success" - test_results.append( - [ - f"{prefix}: {description}", - "Bug reproduced" if is_ok else "Bug is not reproduced", - report_url, - ] - ) + status = f'OK: Bug reproduced (Report' + if not is_ok: + status = f'Bug is not reproduced (Report)' + test_results.append([f"{prefix}: {description}", status]) return is_ok, test_results def process_all_results(file_paths): - all_ok = True + any_ok = False all_results = [] for status_path in file_paths: is_ok, test_results = process_result(status_path) - all_ok = all_ok and is_ok + any_ok = any_ok or is_ok all_results.extend(test_results) - return all_ok, all_results + return any_ok, all_results def main(args): From e7e223ffc635d3f2ac4eec590237a0c486108d09 Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 16:51:34 +0000 Subject: [PATCH 226/239] Revert "Add 02474_dummy.sql" This reverts commit c2a8176716746364e559d61d8475407610cb3d3f. --- tests/queries/0_stateless/02474_dummy.reference | 1 - tests/queries/0_stateless/02474_dummy.sql | 2 -- 2 files changed, 3 deletions(-) delete mode 100644 tests/queries/0_stateless/02474_dummy.reference delete mode 100644 tests/queries/0_stateless/02474_dummy.sql diff --git a/tests/queries/0_stateless/02474_dummy.reference b/tests/queries/0_stateless/02474_dummy.reference deleted file mode 100644 index d00491fd7e5..00000000000 --- a/tests/queries/0_stateless/02474_dummy.reference +++ /dev/null @@ -1 +0,0 @@ -1 diff --git a/tests/queries/0_stateless/02474_dummy.sql b/tests/queries/0_stateless/02474_dummy.sql deleted file mode 100644 index bf2485aeb69..00000000000 --- a/tests/queries/0_stateless/02474_dummy.sql +++ /dev/null @@ -1,2 +0,0 @@ --- Tags: no-fasttest -SELECT 0; From 16032fd0b2b120343c0f229984c5beec43fcd843 Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 16:51:43 +0000 Subject: [PATCH 227/239] Revert "broke style in src/Interpreters/InterpreterSelectQuery.cpp" This reverts commit 60e625918cd646c8b143d957709c56a0380fa6ab. --- src/Interpreters/InterpreterSelectQuery.cpp | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/Interpreters/InterpreterSelectQuery.cpp b/src/Interpreters/InterpreterSelectQuery.cpp index 225dfd1f607..d8ac263e3d1 100644 --- a/src/Interpreters/InterpreterSelectQuery.cpp +++ b/src/Interpreters/InterpreterSelectQuery.cpp @@ -138,7 +138,8 @@ FilterDAGInfoPtr generateFilterActions( expr_list->children.push_back(row_policy_filter->clone()); /// Keep columns that are required after the filter actions. - for (const auto & column_str : prerequisite_columns) { + for (const auto & column_str : prerequisite_columns) + { ParserExpression expr_parser; expr_list->children.push_back(parseQuery(expr_parser, column_str, 0, context->getSettingsRef().max_parser_depth)); } From fba8686db1eaeab90905004a7643fc3a674ba5b5 Mon Sep 17 00:00:00 2001 From: vdimir Date: Thu, 27 Oct 2022 16:52:16 +0000 Subject: [PATCH 228/239] CheckLabels, StyleCheck for TestsBugfixCheck --- .github/workflows/pull_request.yml | 1 + 1 file changed, 1 insertion(+) diff --git a/.github/workflows/pull_request.yml b/.github/workflows/pull_request.yml index 3951f99b16b..23245c16374 100644 --- a/.github/workflows/pull_request.yml +++ b/.github/workflows/pull_request.yml @@ -2023,6 +2023,7 @@ jobs: docker ps --all --quiet | xargs --no-run-if-empty docker rm -f ||: sudo rm -fr "$TEMP_PATH" TestsBugfixCheck: + needs: [CheckLabels, StyleCheck] runs-on: [self-hosted, stress-tester] steps: - name: Set envs From 2060c2d9ac6133017a4477b9cd0d61183d27aa18 Mon Sep 17 00:00:00 2001 From: vdimir Date: Fri, 28 Oct 2022 11:23:44 +0000 Subject: [PATCH 229/239] Fix style --- tests/ci/bugfix_validate_check.py | 2 -- tests/ci/functional_test_check.py | 1 - 2 files changed, 3 deletions(-) diff --git a/tests/ci/bugfix_validate_check.py b/tests/ci/bugfix_validate_check.py index c9d3886b73b..9ecada816b6 100644 --- a/tests/ci/bugfix_validate_check.py +++ b/tests/ci/bugfix_validate_check.py @@ -4,8 +4,6 @@ import argparse import csv import itertools import os -import sys - from github import Github diff --git a/tests/ci/functional_test_check.py b/tests/ci/functional_test_check.py index 1d215920cb5..f7d3288c316 100644 --- a/tests/ci/functional_test_check.py +++ b/tests/ci/functional_test_check.py @@ -7,7 +7,6 @@ import os import subprocess import sys import atexit -import time from github import Github From 3f424de0229f944f20204797ac24b8970c88f45c Mon Sep 17 00:00:00 2001 From: "Mikhail f. Shiryaev" Date: Tue, 1 Nov 2022 14:26:06 +0100 Subject: [PATCH 230/239] Adjust cancel and rerun workflow names to the actual --- .github/workflows/cancel.yml | 2 +- tests/ci/cancel_and_rerun_workflow_lambda/app.py | 2 +- tests/ci/workflow_approve_rerun_lambda/app.py | 6 +++--- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/.github/workflows/cancel.yml b/.github/workflows/cancel.yml index cb06d853219..3c2be767ad2 100644 --- a/.github/workflows/cancel.yml +++ b/.github/workflows/cancel.yml @@ -6,7 +6,7 @@ env: on: # yamllint disable-line rule:truthy workflow_run: - workflows: ["PullRequestCI", "ReleaseCI", "DocsCheck", "BackportPR"] + workflows: ["PullRequestCI", "ReleaseBranchCI", "DocsCheck", "BackportPR"] types: - requested jobs: diff --git a/tests/ci/cancel_and_rerun_workflow_lambda/app.py b/tests/ci/cancel_and_rerun_workflow_lambda/app.py index 813ee9d1ab7..21a5ce517f6 100644 --- a/tests/ci/cancel_and_rerun_workflow_lambda/app.py +++ b/tests/ci/cancel_and_rerun_workflow_lambda/app.py @@ -15,7 +15,7 @@ import boto3 # type: ignore NEED_RERUN_OR_CANCELL_WORKFLOWS = { "PullRequestCI", "DocsCheck", - "DocsRelease", + "DocsReleaseChecks", "BackportPR", } diff --git a/tests/ci/workflow_approve_rerun_lambda/app.py b/tests/ci/workflow_approve_rerun_lambda/app.py index 39bd9cfb283..f2b785840d8 100644 --- a/tests/ci/workflow_approve_rerun_lambda/app.py +++ b/tests/ci/workflow_approve_rerun_lambda/app.py @@ -61,11 +61,11 @@ TRUSTED_WORKFLOW_IDS = { NEED_RERUN_WORKFLOWS = { "BackportPR", - "Docs", - "DocsRelease", + "DocsCheck", + "DocsReleaseChecks", "MasterCI", "PullRequestCI", - "ReleaseCI", + "ReleaseBranchCI", } # Individual trusted contirbutors who are not in any trusted organization. From fc1ae1d99dbae2fe7bd15d3004b1e2551b8a040a Mon Sep 17 00:00:00 2001 From: robot-clickhouse Date: Tue, 1 Nov 2022 15:50:38 +0000 Subject: [PATCH 231/239] Update version_date.tsv and changelogs after v22.10.2.11-stable --- docker/server/Dockerfile.alpine | 2 +- docker/server/Dockerfile.ubuntu | 2 +- docs/changelogs/v22.10.2.11-stable.md | 18 ++++++++++++++++++ utils/list-versions/version_date.tsv | 1 + 4 files changed, 21 insertions(+), 2 deletions(-) create mode 100644 docs/changelogs/v22.10.2.11-stable.md diff --git a/docker/server/Dockerfile.alpine b/docker/server/Dockerfile.alpine index cf4eb3fe645..8f1cf6ee98b 100644 --- a/docker/server/Dockerfile.alpine +++ b/docker/server/Dockerfile.alpine @@ -33,7 +33,7 @@ RUN arch=${TARGETARCH:-amd64} \ # lts / testing / prestable / etc ARG REPO_CHANNEL="stable" ARG REPOSITORY="https://packages.clickhouse.com/tgz/${REPO_CHANNEL}" -ARG VERSION="22.10.1.1877" +ARG VERSION="22.10.2.11" ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static" # user/group precreated explicitly with fixed uid/gid on purpose. diff --git a/docker/server/Dockerfile.ubuntu b/docker/server/Dockerfile.ubuntu index d26657a7979..d5fc5d8e0d3 100644 --- a/docker/server/Dockerfile.ubuntu +++ b/docker/server/Dockerfile.ubuntu @@ -21,7 +21,7 @@ RUN sed -i "s|http://archive.ubuntu.com|${apt_archive}|g" /etc/apt/sources.list ARG REPO_CHANNEL="stable" ARG REPOSITORY="deb https://packages.clickhouse.com/deb ${REPO_CHANNEL} main" -ARG VERSION="22.10.1.1877" +ARG VERSION="22.10.2.11" ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static" # set non-empty deb_location_url url to create a docker image diff --git a/docs/changelogs/v22.10.2.11-stable.md b/docs/changelogs/v22.10.2.11-stable.md new file mode 100644 index 00000000000..e4507f4e745 --- /dev/null +++ b/docs/changelogs/v22.10.2.11-stable.md @@ -0,0 +1,18 @@ +--- +sidebar_position: 1 +sidebar_label: 2022 +--- + +# 2022 Changelog + +### ClickHouse release v22.10.2.11-stable (d2bfcaba002) FIXME as compared to v22.10.1.1877-stable (98ab5a3c189) + +#### Bug Fix (user-visible misbehavior in official stable or prestable release) + +* Backported in [#42750](https://github.com/ClickHouse/ClickHouse/issues/42750): A segmentation fault related to DNS & c-ares has been reported. The below error ocurred in multiple threads: ``` 2022-09-28 15:41:19.008,2022.09.28 15:41:19.008088 [ 356 ] {} BaseDaemon: ######################################## 2022-09-28 15:41:19.008,"2022.09.28 15:41:19.008147 [ 356 ] {} BaseDaemon: (version 22.8.5.29 (official build), build id: 92504ACA0B8E2267) (from thread 353) (no query) Received signal Segmentation fault (11)" 2022-09-28 15:41:19.008,2022.09.28 15:41:19.008196 [ 356 ] {} BaseDaemon: Address: 0xf Access: write. Address not mapped to object. 2022-09-28 15:41:19.008,2022.09.28 15:41:19.008216 [ 356 ] {} BaseDaemon: Stack trace: 0x188f8212 0x1626851b 0x1626a69e 0x16269b3f 0x16267eab 0x13cf8284 0x13d24afc 0x13c5217e 0x14ec2495 0x15ba440f 0x15b9d13b 0x15bb2699 0x1891ccb3 0x1891e00d 0x18ae0769 0x18ade022 0x7f76aa985609 0x7f76aa8aa133 2022-09-28 15:41:19.008,2022.09.28 15:41:19.008274 [ 356 ] {} BaseDaemon: 2. Poco::Net::IPAddress::family() const @ 0x188f8212 in /usr/bin/clickhouse 2022-09-28 15:41:19.008,2022.09.28 15:41:19.008297 [ 356 ] {} BaseDaemon: 3. ? @ 0x1626851b in /usr/bin/clickhouse 2022-09-28 15:41:19.008,2022.09.28 15:41:19.008309 [ 356 ] {} BaseDaemon: 4. ? @ 0x1626a69e in /usr/bin/clickhouse ```. [#42234](https://github.com/ClickHouse/ClickHouse/pull/42234) ([Arthur Passos](https://github.com/arthurpassos)). +* Backported in [#42793](https://github.com/ClickHouse/ClickHouse/issues/42793): Fix a bug in ParserFunction that could have led to a segmentation fault. [#42724](https://github.com/ClickHouse/ClickHouse/pull/42724) ([Nikolay Degterinsky](https://github.com/evillique)). + +#### NOT FOR CHANGELOG / INSIGNIFICANT + +* Always run `BuilderReport` and `BuilderSpecialReport` in all CI types [#42684](https://github.com/ClickHouse/ClickHouse/pull/42684) ([Mikhail f. Shiryaev](https://github.com/Felixoid)). + diff --git a/utils/list-versions/version_date.tsv b/utils/list-versions/version_date.tsv index 21197713694..47dbec5a5f8 100644 --- a/utils/list-versions/version_date.tsv +++ b/utils/list-versions/version_date.tsv @@ -1,3 +1,4 @@ +v22.10.2.11-stable 2022-11-01 v22.10.1.1877-stable 2022-10-26 v22.9.4.32-stable 2022-10-26 v22.9.3.18-stable 2022-09-30 From e5f9c6559016ff26bbd00e528d5c0ca7bc41cecd Mon Sep 17 00:00:00 2001 From: DanRoscigno Date: Tue, 1 Nov 2022 12:56:41 -0400 Subject: [PATCH 232/239] add to page title for search --- docs/en/sql-reference/statements/create/function.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/create/function.md b/docs/en/sql-reference/statements/create/function.md index 63c006b1e3e..90be007bf43 100644 --- a/docs/en/sql-reference/statements/create/function.md +++ b/docs/en/sql-reference/statements/create/function.md @@ -4,7 +4,7 @@ sidebar_position: 38 sidebar_label: FUNCTION --- -# CREATE FUNCTION +# CREATE FUNCTION — user defined function (UDF) Creates a user defined function from a lambda expression. The expression must consist of function parameters, constants, operators, or other function calls. From 62428989b4024eef45a5437a319f016064d18131 Mon Sep 17 00:00:00 2001 From: Dan Roscigno Date: Tue, 1 Nov 2022 14:06:42 -0400 Subject: [PATCH 233/239] Add link to docs repo contrib writing guide --- docs/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/README.md b/docs/README.md index fa8b6bed85c..3ca87dc03c3 100644 --- a/docs/README.md +++ b/docs/README.md @@ -212,4 +212,4 @@ Templates: ## How to Build Documentation -You can build your documentation manually by following the instructions in [docs/tools/README.md](../docs/tools/README.md). Also, our CI runs the documentation build after the `documentation` label is added to PR. You can see the results of a build in the GitHub interface. If you have no permissions to add labels, a reviewer of your PR will add it. +You can build your documentation manually by following the instructions in the docs repo [contrib-writing-guide](https://github.com/ClickHouse/clickhouse-docs/blob/main/contrib-writing-guide.md). Also, our CI runs the documentation build after the `documentation` label is added to PR. You can see the results of a build in the GitHub interface. If you have no permissions to add labels, a reviewer of your PR will add it. From 32e1c3b1e648795e05e3494093035f79ad213fb4 Mon Sep 17 00:00:00 2001 From: DanRoscigno Date: Tue, 1 Nov 2022 16:25:54 -0400 Subject: [PATCH 234/239] rename update file as it is included in Manage > Updates --- docs/en/operations/{update.md => _update.md} | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) rename docs/en/operations/{update.md => _update.md} (88%) diff --git a/docs/en/operations/update.md b/docs/en/operations/_update.md similarity index 88% rename from docs/en/operations/update.md rename to docs/en/operations/_update.md index 24f7efecc7b..86981da2be6 100644 --- a/docs/en/operations/update.md +++ b/docs/en/operations/_update.md @@ -1,10 +1,7 @@ ---- -slug: /en/operations/update -sidebar_position: 47 -sidebar_label: ClickHouse Upgrade ---- -# ClickHouse Upgrade +[//]: # (This file is included in Manage > Updates) + +## Self-managed ClickHouse Upgrade If ClickHouse was installed from `deb` packages, execute the following commands on the server: From ba34ad37d1f3ab7d5becbf4c3ec699ff02f60009 Mon Sep 17 00:00:00 2001 From: Habibullah Oladepo Date: Wed, 2 Nov 2022 08:02:01 +0100 Subject: [PATCH 235/239] Fix minor typos in intersect statement documentation --- docs/en/sql-reference/statements/select/intersect.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/sql-reference/statements/select/intersect.md b/docs/en/sql-reference/statements/select/intersect.md index d3b2b51b6be..f1eb4738543 100644 --- a/docs/en/sql-reference/statements/select/intersect.md +++ b/docs/en/sql-reference/statements/select/intersect.md @@ -7,7 +7,7 @@ sidebar_label: INTERSECT The `INTERSECT` clause returns only those rows that result from both the first and the second queries. The queries must match the number of columns, order, and type. The result of `INTERSECT` can contain duplicate rows. -Multiple `INTERSECT` statements are executes left to right if parenthesis are not specified. The `INTERSECT` operator has a higher priority than the `UNION` and `EXCEPT` clause. +Multiple `INTERSECT` statements are executed left to right if parentheses are not specified. The `INTERSECT` operator has a higher priority than the `UNION` and `EXCEPT` clauses. ``` sql From 40314e90303b366b2bcdf584170ff1aa2318926e Mon Sep 17 00:00:00 2001 From: vdimir Date: Tue, 1 Nov 2022 10:43:53 +0000 Subject: [PATCH 236/239] Add timezone to 02458_datediff_date32 --- .../02458_datediff_date32.reference | 154 +++++++++--------- .../0_stateless/02458_datediff_date32.sql | 154 +++++++++--------- 2 files changed, 154 insertions(+), 154 deletions(-) diff --git a/tests/queries/0_stateless/02458_datediff_date32.reference b/tests/queries/0_stateless/02458_datediff_date32.reference index 9e93af03896..fdb8273a74b 100644 --- a/tests/queries/0_stateless/02458_datediff_date32.reference +++ b/tests/queries/0_stateless/02458_datediff_date32.reference @@ -1,169 +1,169 @@ -- { echo } -- Date32 vs Date32 -SELECT dateDiff('second', toDate32('1927-01-01'), toDate32('1927-01-02')); +SELECT dateDiff('second', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDate32('1927-01-01'), toDate32('1927-01-02')); +SELECT dateDiff('minute', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDate32('1927-01-01'), toDate32('1927-01-02')); +SELECT dateDiff('hour', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDate32('1927-01-01'), toDate32('1927-01-02')); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDate32('1927-01-01'), toDate32('1927-01-08')); +SELECT dateDiff('week', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-08', 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDate32('1927-01-01'), toDate32('1927-02-01')); +SELECT dateDiff('month', toDate32('1927-01-01', 'UTC'), toDate32('1927-02-01', 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDate32('1927-01-01'), toDate32('1927-04-01')); +SELECT dateDiff('quarter', toDate32('1927-01-01', 'UTC'), toDate32('1927-04-01', 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDate32('1927-01-01'), toDate32('1928-01-01')); +SELECT dateDiff('year', toDate32('1927-01-01', 'UTC'), toDate32('1928-01-01', 'UTC'), 'UTC'); 1 -- With DateTime64 -- Date32 vs DateTime64 -SELECT dateDiff('second', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); +SELECT dateDiff('second', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); +SELECT dateDiff('minute', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); +SELECT dateDiff('hour', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDate32('1927-01-01'), toDateTime64('1927-01-08 00:00:00', 3)); +SELECT dateDiff('week', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-08 00:00:00', 3, 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDate32('1927-01-01'), toDateTime64('1927-02-01 00:00:00', 3)); +SELECT dateDiff('month', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-02-01 00:00:00', 3, 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDate32('1927-01-01'), toDateTime64('1927-04-01 00:00:00', 3)); +SELECT dateDiff('quarter', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-04-01 00:00:00', 3, 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDate32('1927-01-01'), toDateTime64('1928-01-01 00:00:00', 3)); +SELECT dateDiff('year', toDate32('1927-01-01', 'UTC'), toDateTime64('1928-01-01 00:00:00', 3, 'UTC'), 'UTC'); 1 -- DateTime64 vs Date32 -SELECT dateDiff('second', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); +SELECT dateDiff('second', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); +SELECT dateDiff('minute', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); +SELECT dateDiff('hour', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); +SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-08')); +SELECT dateDiff('week', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-08', 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-02-01')); +SELECT dateDiff('month', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-02-01', 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-04-01')); +SELECT dateDiff('quarter', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-04-01', 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1928-01-01')); +SELECT dateDiff('year', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1928-01-01', 'UTC'), 'UTC'); 1 -- With DateTime -- Date32 vs DateTime -SELECT dateDiff('second', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); +SELECT dateDiff('second', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); +SELECT dateDiff('minute', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); +SELECT dateDiff('hour', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDate32('2015-08-18'), toDateTime('2015-08-25 00:00:00')); +SELECT dateDiff('week', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-25 00:00:00', 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDate32('2015-08-18'), toDateTime('2015-09-18 00:00:00')); +SELECT dateDiff('month', toDate32('2015-08-18', 'UTC'), toDateTime('2015-09-18 00:00:00', 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDate32('2015-08-18'), toDateTime('2015-11-18 00:00:00')); +SELECT dateDiff('quarter', toDate32('2015-08-18', 'UTC'), toDateTime('2015-11-18 00:00:00', 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDate32('2015-08-18'), toDateTime('2016-08-18 00:00:00')); +SELECT dateDiff('year', toDate32('2015-08-18', 'UTC'), toDateTime('2016-08-18 00:00:00', 'UTC'), 'UTC'); 1 -- DateTime vs Date32 -SELECT dateDiff('second', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); +SELECT dateDiff('second', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); +SELECT dateDiff('minute', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); +SELECT dateDiff('hour', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); +SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-25')); +SELECT dateDiff('week', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-25', 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDateTime('2015-08-18 00:00:00'), toDate32('2015-09-18')); +SELECT dateDiff('month', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-09-18', 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDateTime('2015-08-18 00:00:00'), toDate32('2015-11-18')); +SELECT dateDiff('quarter', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-11-18', 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDateTime('2015-08-18 00:00:00'), toDate32('2016-08-18')); +SELECT dateDiff('year', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2016-08-18', 'UTC'), 'UTC'); 1 -- With Date -- Date32 vs Date -SELECT dateDiff('second', toDate32('2015-08-18'), toDate('2015-08-19')); +SELECT dateDiff('second', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDate32('2015-08-18'), toDate('2015-08-19')); +SELECT dateDiff('minute', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDate32('2015-08-18'), toDate('2015-08-19')); +SELECT dateDiff('hour', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDate32('2015-08-18'), toDate('2015-08-19')); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDate32('2015-08-18'), toDate('2015-08-25')); +SELECT dateDiff('week', toDate32('2015-08-18', 'UTC'), toDate('2015-08-25', 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDate32('2015-08-18'), toDate('2015-09-18')); +SELECT dateDiff('month', toDate32('2015-08-18', 'UTC'), toDate('2015-09-18', 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDate32('2015-08-18'), toDate('2015-11-18')); +SELECT dateDiff('quarter', toDate32('2015-08-18', 'UTC'), toDate('2015-11-18', 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDate32('2015-08-18'), toDate('2016-08-18')); +SELECT dateDiff('year', toDate32('2015-08-18', 'UTC'), toDate('2016-08-18', 'UTC'), 'UTC'); 1 -- Date vs Date32 -SELECT dateDiff('second', toDate('2015-08-18'), toDate32('2015-08-19')); +SELECT dateDiff('second', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 86400 -SELECT dateDiff('minute', toDate('2015-08-18'), toDate32('2015-08-19')); +SELECT dateDiff('minute', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 1440 -SELECT dateDiff('hour', toDate('2015-08-18'), toDate32('2015-08-19')); +SELECT dateDiff('hour', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 24 -SELECT dateDiff('day', toDate('2015-08-18'), toDate32('2015-08-19')); +SELECT dateDiff('day', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); 1 -SELECT dateDiff('week', toDate('2015-08-18'), toDate32('2015-08-25')); +SELECT dateDiff('week', toDate('2015-08-18', 'UTC'), toDate32('2015-08-25', 'UTC'), 'UTC'); 1 -SELECT dateDiff('month', toDate('2015-08-18'), toDate32('2015-09-18')); +SELECT dateDiff('month', toDate('2015-08-18', 'UTC'), toDate32('2015-09-18', 'UTC'), 'UTC'); 1 -SELECT dateDiff('quarter', toDate('2015-08-18'), toDate32('2015-11-18')); +SELECT dateDiff('quarter', toDate('2015-08-18', 'UTC'), toDate32('2015-11-18', 'UTC'), 'UTC'); 1 -SELECT dateDiff('year', toDate('2015-08-18'), toDate32('2016-08-18')); +SELECT dateDiff('year', toDate('2015-08-18', 'UTC'), toDate32('2016-08-18', 'UTC'), 'UTC'); 1 -- Const vs non-const columns -SELECT dateDiff('day', toDate32('1927-01-01'), materialize(toDate32('1927-01-02'))); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', toDate32('1927-01-01'), materialize(toDateTime64('1927-01-02 00:00:00', 3))); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), materialize(toDateTime64('1927-01-02 00:00:00', 3, 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3), materialize(toDate32('1927-01-02'))); +SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', toDate32('2015-08-18'), materialize(toDateTime('2015-08-19 00:00:00'))); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), materialize(toDateTime('2015-08-19 00:00:00', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00'), materialize(toDate32('2015-08-19'))); +SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00', 'UTC'), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', toDate32('2015-08-18'), materialize(toDate('2015-08-19'))); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), materialize(toDate('2015-08-19', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', toDate('2015-08-18'), materialize(toDate32('2015-08-19'))); +SELECT dateDiff('day', toDate('2015-08-18', 'UTC'), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); 1 -- Non-const vs const columns -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), toDate32('1927-01-02')); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), toDate32('1927-01-02', 'UTC'), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), toDateTime64('1927-01-02 00:00:00', 3)); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3)), toDate32('1927-01-02')); +SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3, 'UTC')), toDate32('1927-01-02', 'UTC'), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), toDateTime('2015-08-19 00:00:00')); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00')), toDate32('2015-08-19')); +SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00', 'UTC')), toDate32('2015-08-19', 'UTC'), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), toDate('2015-08-19')); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), toDate('2015-08-19', 'UTC'), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate('2015-08-18')), toDate32('2015-08-19')); +SELECT dateDiff('day', materialize(toDate('2015-08-18', 'UTC')), toDate32('2015-08-19', 'UTC'), 'UTC'); 1 -- Non-const vs non-const columns -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), materialize(toDate32('1927-01-02'))); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), materialize(toDateTime64('1927-01-02 00:00:00', 3))); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), materialize(toDateTime64('1927-01-02 00:00:00', 3, 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3)), materialize(toDate32('1927-01-02'))); +SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3, 'UTC')), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), materialize(toDateTime('2015-08-19 00:00:00'))); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), materialize(toDateTime('2015-08-19 00:00:00', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00')), materialize(toDate32('2015-08-19'))); +SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00', 'UTC')), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), materialize(toDate('2015-08-19'))); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), materialize(toDate('2015-08-19', 'UTC')), 'UTC'); 1 -SELECT dateDiff('day', materialize(toDate('2015-08-18')), materialize(toDate32('2015-08-19'))); +SELECT dateDiff('day', materialize(toDate('2015-08-18', 'UTC')), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); 1 diff --git a/tests/queries/0_stateless/02458_datediff_date32.sql b/tests/queries/0_stateless/02458_datediff_date32.sql index b4cb203080e..e41070e8146 100644 --- a/tests/queries/0_stateless/02458_datediff_date32.sql +++ b/tests/queries/0_stateless/02458_datediff_date32.sql @@ -1,101 +1,101 @@ -- { echo } -- Date32 vs Date32 -SELECT dateDiff('second', toDate32('1927-01-01'), toDate32('1927-01-02')); -SELECT dateDiff('minute', toDate32('1927-01-01'), toDate32('1927-01-02')); -SELECT dateDiff('hour', toDate32('1927-01-01'), toDate32('1927-01-02')); -SELECT dateDiff('day', toDate32('1927-01-01'), toDate32('1927-01-02')); -SELECT dateDiff('week', toDate32('1927-01-01'), toDate32('1927-01-08')); -SELECT dateDiff('month', toDate32('1927-01-01'), toDate32('1927-02-01')); -SELECT dateDiff('quarter', toDate32('1927-01-01'), toDate32('1927-04-01')); -SELECT dateDiff('year', toDate32('1927-01-01'), toDate32('1928-01-01')); +SELECT dateDiff('second', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('week', toDate32('1927-01-01', 'UTC'), toDate32('1927-01-08', 'UTC'), 'UTC'); +SELECT dateDiff('month', toDate32('1927-01-01', 'UTC'), toDate32('1927-02-01', 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDate32('1927-01-01', 'UTC'), toDate32('1927-04-01', 'UTC'), 'UTC'); +SELECT dateDiff('year', toDate32('1927-01-01', 'UTC'), toDate32('1928-01-01', 'UTC'), 'UTC'); -- With DateTime64 -- Date32 vs DateTime64 -SELECT dateDiff('second', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); -SELECT dateDiff('minute', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); -SELECT dateDiff('hour', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); -SELECT dateDiff('day', toDate32('1927-01-01'), toDateTime64('1927-01-02 00:00:00', 3)); -SELECT dateDiff('week', toDate32('1927-01-01'), toDateTime64('1927-01-08 00:00:00', 3)); -SELECT dateDiff('month', toDate32('1927-01-01'), toDateTime64('1927-02-01 00:00:00', 3)); -SELECT dateDiff('quarter', toDate32('1927-01-01'), toDateTime64('1927-04-01 00:00:00', 3)); -SELECT dateDiff('year', toDate32('1927-01-01'), toDateTime64('1928-01-01 00:00:00', 3)); +SELECT dateDiff('second', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('week', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-01-08 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('month', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-02-01 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDate32('1927-01-01', 'UTC'), toDateTime64('1927-04-01 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('year', toDate32('1927-01-01', 'UTC'), toDateTime64('1928-01-01 00:00:00', 3, 'UTC'), 'UTC'); -- DateTime64 vs Date32 -SELECT dateDiff('second', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); -SELECT dateDiff('minute', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); -SELECT dateDiff('hour', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); -SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-02')); -SELECT dateDiff('week', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-01-08')); -SELECT dateDiff('month', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-02-01')); -SELECT dateDiff('quarter', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1927-04-01')); -SELECT dateDiff('year', toDateTime64('1927-01-01 00:00:00', 3), toDate32('1928-01-01')); +SELECT dateDiff('second', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('week', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-01-08', 'UTC'), 'UTC'); +SELECT dateDiff('month', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-02-01', 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1927-04-01', 'UTC'), 'UTC'); +SELECT dateDiff('year', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), toDate32('1928-01-01', 'UTC'), 'UTC'); -- With DateTime -- Date32 vs DateTime -SELECT dateDiff('second', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); -SELECT dateDiff('minute', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); -SELECT dateDiff('hour', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); -SELECT dateDiff('day', toDate32('2015-08-18'), toDateTime('2015-08-19 00:00:00')); -SELECT dateDiff('week', toDate32('2015-08-18'), toDateTime('2015-08-25 00:00:00')); -SELECT dateDiff('month', toDate32('2015-08-18'), toDateTime('2015-09-18 00:00:00')); -SELECT dateDiff('quarter', toDate32('2015-08-18'), toDateTime('2015-11-18 00:00:00')); -SELECT dateDiff('year', toDate32('2015-08-18'), toDateTime('2016-08-18 00:00:00')); +SELECT dateDiff('second', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('week', toDate32('2015-08-18', 'UTC'), toDateTime('2015-08-25 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('month', toDate32('2015-08-18', 'UTC'), toDateTime('2015-09-18 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDate32('2015-08-18', 'UTC'), toDateTime('2015-11-18 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('year', toDate32('2015-08-18', 'UTC'), toDateTime('2016-08-18 00:00:00', 'UTC'), 'UTC'); -- DateTime vs Date32 -SELECT dateDiff('second', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); -SELECT dateDiff('minute', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); -SELECT dateDiff('hour', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); -SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-19')); -SELECT dateDiff('week', toDateTime('2015-08-18 00:00:00'), toDate32('2015-08-25')); -SELECT dateDiff('month', toDateTime('2015-08-18 00:00:00'), toDate32('2015-09-18')); -SELECT dateDiff('quarter', toDateTime('2015-08-18 00:00:00'), toDate32('2015-11-18')); -SELECT dateDiff('year', toDateTime('2015-08-18 00:00:00'), toDate32('2016-08-18')); +SELECT dateDiff('second', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('week', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-08-25', 'UTC'), 'UTC'); +SELECT dateDiff('month', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-09-18', 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2015-11-18', 'UTC'), 'UTC'); +SELECT dateDiff('year', toDateTime('2015-08-18 00:00:00', 'UTC'), toDate32('2016-08-18', 'UTC'), 'UTC'); -- With Date -- Date32 vs Date -SELECT dateDiff('second', toDate32('2015-08-18'), toDate('2015-08-19')); -SELECT dateDiff('minute', toDate32('2015-08-18'), toDate('2015-08-19')); -SELECT dateDiff('hour', toDate32('2015-08-18'), toDate('2015-08-19')); -SELECT dateDiff('day', toDate32('2015-08-18'), toDate('2015-08-19')); -SELECT dateDiff('week', toDate32('2015-08-18'), toDate('2015-08-25')); -SELECT dateDiff('month', toDate32('2015-08-18'), toDate('2015-09-18')); -SELECT dateDiff('quarter', toDate32('2015-08-18'), toDate('2015-11-18')); -SELECT dateDiff('year', toDate32('2015-08-18'), toDate('2016-08-18')); +SELECT dateDiff('second', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), toDate('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('week', toDate32('2015-08-18', 'UTC'), toDate('2015-08-25', 'UTC'), 'UTC'); +SELECT dateDiff('month', toDate32('2015-08-18', 'UTC'), toDate('2015-09-18', 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDate32('2015-08-18', 'UTC'), toDate('2015-11-18', 'UTC'), 'UTC'); +SELECT dateDiff('year', toDate32('2015-08-18', 'UTC'), toDate('2016-08-18', 'UTC'), 'UTC'); -- Date vs Date32 -SELECT dateDiff('second', toDate('2015-08-18'), toDate32('2015-08-19')); -SELECT dateDiff('minute', toDate('2015-08-18'), toDate32('2015-08-19')); -SELECT dateDiff('hour', toDate('2015-08-18'), toDate32('2015-08-19')); -SELECT dateDiff('day', toDate('2015-08-18'), toDate32('2015-08-19')); -SELECT dateDiff('week', toDate('2015-08-18'), toDate32('2015-08-25')); -SELECT dateDiff('month', toDate('2015-08-18'), toDate32('2015-09-18')); -SELECT dateDiff('quarter', toDate('2015-08-18'), toDate32('2015-11-18')); -SELECT dateDiff('year', toDate('2015-08-18'), toDate32('2016-08-18')); +SELECT dateDiff('second', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('minute', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('hour', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('day', toDate('2015-08-18', 'UTC'), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('week', toDate('2015-08-18', 'UTC'), toDate32('2015-08-25', 'UTC'), 'UTC'); +SELECT dateDiff('month', toDate('2015-08-18', 'UTC'), toDate32('2015-09-18', 'UTC'), 'UTC'); +SELECT dateDiff('quarter', toDate('2015-08-18', 'UTC'), toDate32('2015-11-18', 'UTC'), 'UTC'); +SELECT dateDiff('year', toDate('2015-08-18', 'UTC'), toDate32('2016-08-18', 'UTC'), 'UTC'); -- Const vs non-const columns -SELECT dateDiff('day', toDate32('1927-01-01'), materialize(toDate32('1927-01-02'))); -SELECT dateDiff('day', toDate32('1927-01-01'), materialize(toDateTime64('1927-01-02 00:00:00', 3))); -SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3), materialize(toDate32('1927-01-02'))); -SELECT dateDiff('day', toDate32('2015-08-18'), materialize(toDateTime('2015-08-19 00:00:00'))); -SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00'), materialize(toDate32('2015-08-19'))); -SELECT dateDiff('day', toDate32('2015-08-18'), materialize(toDate('2015-08-19'))); -SELECT dateDiff('day', toDate('2015-08-18'), materialize(toDate32('2015-08-19'))); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); +SELECT dateDiff('day', toDate32('1927-01-01', 'UTC'), materialize(toDateTime64('1927-01-02 00:00:00', 3, 'UTC')), 'UTC'); +SELECT dateDiff('day', toDateTime64('1927-01-01 00:00:00', 3, 'UTC'), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), materialize(toDateTime('2015-08-19 00:00:00', 'UTC')), 'UTC'); +SELECT dateDiff('day', toDateTime('2015-08-18 00:00:00', 'UTC'), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); +SELECT dateDiff('day', toDate32('2015-08-18', 'UTC'), materialize(toDate('2015-08-19', 'UTC')), 'UTC'); +SELECT dateDiff('day', toDate('2015-08-18', 'UTC'), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); -- Non-const vs const columns -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), toDate32('1927-01-02')); -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), toDateTime64('1927-01-02 00:00:00', 3)); -SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3)), toDate32('1927-01-02')); -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), toDateTime('2015-08-19 00:00:00')); -SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00')), toDate32('2015-08-19')); -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), toDate('2015-08-19')); -SELECT dateDiff('day', materialize(toDate('2015-08-18')), toDate32('2015-08-19')); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), toDateTime64('1927-01-02 00:00:00', 3, 'UTC'), 'UTC'); +SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3, 'UTC')), toDate32('1927-01-02', 'UTC'), 'UTC'); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), toDateTime('2015-08-19 00:00:00', 'UTC'), 'UTC'); +SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00', 'UTC')), toDate32('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), toDate('2015-08-19', 'UTC'), 'UTC'); +SELECT dateDiff('day', materialize(toDate('2015-08-18', 'UTC')), toDate32('2015-08-19', 'UTC'), 'UTC'); -- Non-const vs non-const columns -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), materialize(toDate32('1927-01-02'))); -SELECT dateDiff('day', materialize(toDate32('1927-01-01')), materialize(toDateTime64('1927-01-02 00:00:00', 3))); -SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3)), materialize(toDate32('1927-01-02'))); -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), materialize(toDateTime('2015-08-19 00:00:00'))); -SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00')), materialize(toDate32('2015-08-19'))); -SELECT dateDiff('day', materialize(toDate32('2015-08-18')), materialize(toDate('2015-08-19'))); -SELECT dateDiff('day', materialize(toDate('2015-08-18')), materialize(toDate32('2015-08-19'))); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); +SELECT dateDiff('day', materialize(toDate32('1927-01-01', 'UTC')), materialize(toDateTime64('1927-01-02 00:00:00', 3, 'UTC')), 'UTC'); +SELECT dateDiff('day', materialize(toDateTime64('1927-01-01 00:00:00', 3, 'UTC')), materialize(toDate32('1927-01-02', 'UTC')), 'UTC'); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), materialize(toDateTime('2015-08-19 00:00:00', 'UTC')), 'UTC'); +SELECT dateDiff('day', materialize(toDateTime('2015-08-18 00:00:00', 'UTC')), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); +SELECT dateDiff('day', materialize(toDate32('2015-08-18', 'UTC')), materialize(toDate('2015-08-19', 'UTC')), 'UTC'); +SELECT dateDiff('day', materialize(toDate('2015-08-18', 'UTC')), materialize(toDate32('2015-08-19', 'UTC')), 'UTC'); From 4c77987bf876706f8d0f579d8e33e95a94726239 Mon Sep 17 00:00:00 2001 From: Dan Roscigno Date: Wed, 2 Nov 2022 10:41:55 -0400 Subject: [PATCH 237/239] Update projection.md Experiment with Algolia search --- .../statements/alter/projection.md | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/docs/en/sql-reference/statements/alter/projection.md b/docs/en/sql-reference/statements/alter/projection.md index a1981b4a0bb..ff8ecf3a77f 100644 --- a/docs/en/sql-reference/statements/alter/projection.md +++ b/docs/en/sql-reference/statements/alter/projection.md @@ -7,18 +7,26 @@ title: "Manipulating Projections" The following operations with [projections](../../../engines/table-engines/mergetree-family/mergetree.md#projections) are available: -- `ALTER TABLE [db].name ADD PROJECTION name ( SELECT [GROUP BY] [ORDER BY] )` - Adds projection description to tables metadata. +## ADD PROJECTION -- `ALTER TABLE [db].name DROP PROJECTION name` - Removes projection description from tables metadata and deletes projection files from disk. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). +`ALTER TABLE [db].name ADD PROJECTION name ( SELECT [GROUP BY] [ORDER BY] )` - Adds projection description to tables metadata. -- `ALTER TABLE [db.]table MATERIALIZE PROJECTION name IN PARTITION partition_name` - The query rebuilds the projection `name` in the partition `partition_name`. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). +## DROP PROJECTION -- `ALTER TABLE [db.]table CLEAR PROJECTION name IN PARTITION partition_name` - Deletes projection files from disk without removing description. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). +`ALTER TABLE [db].name DROP PROJECTION name` - Removes projection description from tables metadata and deletes projection files from disk. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). + +## MATERIALIZE PROJECTION + +`ALTER TABLE [db.]table MATERIALIZE PROJECTION name IN PARTITION partition_name` - The query rebuilds the projection `name` in the partition `partition_name`. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). + +## CLEAR PROJECTION + +`ALTER TABLE [db.]table CLEAR PROJECTION name IN PARTITION partition_name` - Deletes projection files from disk without removing description. Implemented as a [mutation](../../../sql-reference/statements/alter/index.md#mutations). The commands `ADD`, `DROP` and `CLEAR` are lightweight in a sense that they only change metadata or remove files. -Also, they are replicated, syncing projections metadata via ZooKeeper. +Also, they are replicated, syncing projections metadata via ClickHouse Keeper or ZooKeeper. :::note Projection manipulation is supported only for tables with [`*MergeTree`](../../../engines/table-engines/mergetree-family/mergetree.md) engine (including [replicated](../../../engines/table-engines/mergetree-family/replication.md) variants). From b6d84ef5fa451d3905d09fd8bcc9b5220f70905e Mon Sep 17 00:00:00 2001 From: Camilo Sierra Date: Wed, 2 Nov 2022 15:43:57 +0100 Subject: [PATCH 238/239] range hashed missing parameters On the query we have missing two options: - `LIFETIME` Is not on the example if you copy paste you will have an Exception `DB::Exception: Cannot create dictionary with empty lifetime.` - `SOURCE` was not mentioned and it's important to link to the main/source table. - There was an error on the `dictGetT` function there was an additional T this function do not exist (I have tested and we need to use `dictGet`). - Also in the Dictionary example we have no extra attribute other than the id and the two dates, and for running the queries and the `dicGet` function you need an additional attribute this is why I have added `advertiser_id` (BTW I use advertiser_id as this was use in the example just before) and also add one example, without the example it was not easy to understand what was the 'attr_name' mentioned before. - I add an example as an user did not knew how to cast the date to a Uint64 (Because most of the time the original/raw dates are defined on the range as Date64, so this example will explain them how to cast when doing the query) --- .../external-dicts-dict-layout.md | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-layout.md b/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-layout.md index 198ff12f1d6..02a4ad57a3b 100644 --- a/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-layout.md +++ b/docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-layout.md @@ -303,17 +303,25 @@ or CREATE DICTIONARY somedict ( id UInt64, first Date, - last Date + last Date, + advertiser_id UInt64 ) PRIMARY KEY id +SOURCE(CLICKHOUSE(TABLE 'date_table')) +LIFETIME(MIN 1 MAX 1000) LAYOUT(RANGE_HASHED()) RANGE(MIN first MAX last) ``` -To work with these dictionaries, you need to pass an additional argument to the `dictGetT` function, for which a range is selected: +To work with these dictionaries, you need to pass an additional argument to the `dictGet` function, for which a range is selected: ``` sql -dictGetT('dict_name', 'attr_name', id, date) +dictGet('dict_name', 'attr_name', id, date) +``` +Query example: + +``` sql +SELECT dictGet('somedict', 'advertiser_id', 1, '2022-10-20 23:20:10.000'::DateTime64::UInt64); ``` This function returns the value for the specified `id`s and the date range that includes the passed date. From c3466fb52f635f8cbd5d0ae95781ac5a84b9ed87 Mon Sep 17 00:00:00 2001 From: Alexey Milovidov Date: Wed, 2 Nov 2022 22:22:16 +0300 Subject: [PATCH 239/239] Update clickhouse-keeper.md --- docs/en/operations/clickhouse-keeper.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/en/operations/clickhouse-keeper.md b/docs/en/operations/clickhouse-keeper.md index 82fa5c114ea..aad20da0010 100644 --- a/docs/en/operations/clickhouse-keeper.md +++ b/docs/en/operations/clickhouse-keeper.md @@ -309,7 +309,7 @@ Sessions with Ephemerals (1): /clickhouse/task_queue/ddl ``` -## [experimental] Migration from ZooKeeper {#migration-from-zookeeper} +## Migration from ZooKeeper {#migration-from-zookeeper} Seamlessly migration from ZooKeeper to ClickHouse Keeper is impossible you have to stop your ZooKeeper cluster, convert data and start ClickHouse Keeper. `clickhouse-keeper-converter` tool allows converting ZooKeeper logs and snapshots to ClickHouse Keeper snapshot. It works only with ZooKeeper > 3.4. Steps for migration: