ClickHouse/src/Storages/StorageJoin.h

117 lines
4.5 KiB
C++
Raw Normal View History

#pragma once
#include <Common/RWLock.h>
#include <Storages/StorageSet.h>
#include <Storages/TableLockHolder.h>
#include <Parsers/ASTTablesInSelectQuery.h>
2022-08-04 15:20:19 +00:00
#include <Interpreters/JoinUtils.h>
namespace DB
{
class TableJoin;
class HashJoin;
using HashJoinPtr = std::shared_ptr<HashJoin>;
Squashed commit of the following: commit e712f469a55ff34ad34b482b15cc4153b7ad7233 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:59:13 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 2a002823084e3a79bffcc17d479620a68eb0644b Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:58:30 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 9e06f407c8ee781ed8ddf98bdfcc31846bf2a0fe Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:55:14 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 9581620f1e839f456fa7894aa1f996d5162ac6cd Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:54:22 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 2a8564c68cb6cc3649fafaf401256d43c9a2e777 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:47:34 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit cf60632d78ec656be3304ef4565e859bb6ce80ba Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:40:09 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit ee3d1dc6e0c4ca60e3ac1e0c30d4b3ed1e66eca0 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:22:49 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 65592ef7116a90104fcd524b53ef8b7cf22640f2 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:18:17 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 37972c257320d3b7e7b294e0fdeffff218647bfd Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:17:06 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit dd909d149974ce5bed2456de1261aa5a368fd3ff Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:16:28 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 3cf43266ca7e30adf01212b1a739ba5fe43639fd Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:15:42 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 6731a3df96d1609286e2536b6432916af7743f0f Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:13:35 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 1b5727e0d56415b7add4cb76110105358663602c Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:11:18 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit bbcf726a55685b8e72f5b40ba0bf1904bd1c0407 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:09:04 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit c03b477d5e2e65014e8906ecfa2efb67ee295af1 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:06:30 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 2986e2fb0466bc18d73693dcdded28fccc0dc66b Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:05:44 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit 5d6cdef13d2e02bd5c4954983334e9162ab2635b Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:04:53 2017 +0300 Less dependencies [#CLICKHOUSE-2] commit f2b819b25ce8b2ccdcb201eefb03e1e6f5aab590 Author: Alexey Milovidov <milovidov@yandex-team.ru> Date: Sat Jan 14 11:01:47 2017 +0300 Less dependencies [#CLICKHOUSE-2]
2017-01-14 09:00:19 +00:00
2017-04-16 15:00:33 +00:00
/** Allows you save the state for later use on the right side of the JOIN.
* When inserted into a table, the data will be inserted into the state,
* and also written to the backup file, to restore after the restart.
* Reading from the table is not possible directly - only specifying on the right side of JOIN is possible.
*
2017-04-16 15:00:33 +00:00
* When using, JOIN must be of the appropriate type (ANY|ALL LEFT|INNER ...).
*/
class StorageJoin final : public StorageSetOrJoinBase
{
public:
StorageJoin(
DiskPtr disk_,
const String & relative_path_,
const StorageID & table_id_,
const Names & key_names_,
bool use_nulls_,
SizeLimits limits_,
2022-07-29 16:30:50 +00:00
JoinKind kind_,
JoinStrictness strictness_,
const ColumnsDescription & columns_,
const ConstraintsDescription & constraints_,
const String & comment,
bool overwrite,
bool persistent_);
String getName() const override { return "Join"; }
void truncate(const ASTPtr &, const StorageMetadataPtr & metadata_snapshot, ContextPtr, TableExclusiveLockHolder &) override;
2018-06-09 16:09:37 +00:00
2021-05-31 23:22:05 +00:00
/// Only delete is supported.
2021-04-18 09:38:50 +00:00
void checkMutationIsPossible(const MutationCommands & commands, const Settings & settings) const override;
2023-01-30 17:38:28 +00:00
void mutate(const MutationCommands & commands, ContextPtr context) override;
2021-04-18 09:38:50 +00:00
/// Return instance of HashJoin holding lock that protects from insertions to StorageJoin.
/// HashJoin relies on structure of hash table that's why we need to return it with locked mutex.
HashJoinPtr getJoinLocked(std::shared_ptr<TableJoin> analyzed_join, ContextPtr context, const Names & required_columns_names) const;
/// Get result type for function "joinGet(OrNull)"
DataTypePtr joinGetCheckAndGetReturnType(const DataTypes & data_types, const String & column_name, bool or_null) const;
/// Execute function "joinGet(OrNull)" on data block.
/// Takes rwlock for read to prevent parallel StorageJoin updates during processing data block
/// (but not during processing whole query, it's safe for joinGet that doesn't involve `used_flags` from HashJoin)
ColumnWithTypeAndName joinGet(const Block & block, const Block & block_with_columns_to_add, ContextPtr context) const;
SinkToStoragePtr write(const ASTPtr & query, const StorageMetadataPtr & metadata_snapshot, ContextPtr context, bool async_insert) override;
2021-05-31 23:22:05 +00:00
2020-08-03 13:54:14 +00:00
Pipe read(
const Names & column_names,
const StorageSnapshotPtr & storage_snapshot,
SelectQueryInfo & query_info,
ContextPtr context,
QueryProcessingStage::Enum processed_stage,
size_t max_block_size,
size_t num_streams) override;
2020-11-25 13:47:32 +00:00
std::optional<UInt64> totalRows(const Settings & settings) const override;
std::optional<UInt64> totalBytes(const Settings & settings) const override;
Block getRightSampleBlock() const
{
auto metadata_snapshot = getInMemoryMetadataPtr();
Block block = metadata_snapshot->getSampleBlock();
convertRightBlock(block);
return block;
}
bool useNulls() const { return use_nulls; }
2022-12-07 15:02:37 +00:00
const Names & getKeyNames() const { return key_names; }
private:
Block sample_block;
const Names key_names;
bool use_nulls;
SizeLimits limits;
2022-07-29 16:30:50 +00:00
JoinKind kind; /// LEFT | INNER ...
JoinStrictness strictness; /// ANY | ALL
bool overwrite;
std::shared_ptr<TableJoin> table_join;
2019-09-09 19:43:37 +00:00
HashJoinPtr join;
/// Protect state for concurrent use in insertFromBlock and joinBlock.
/// Lock is stored in HashJoin instance during query and blocks concurrent insertions.
mutable RWLock rwlock = RWLockImpl::create();
2021-05-31 23:22:05 +00:00
mutable std::mutex mutate_mutex;
2021-04-18 09:38:50 +00:00
2021-09-30 08:47:15 +00:00
void insertBlock(const Block & block, ContextPtr context) override;
2019-11-01 11:03:35 +00:00
void finishInsert() override {}
size_t getSize(ContextPtr context) const override;
2021-09-30 08:47:15 +00:00
RWLockImpl::LockHolder tryLockTimedWithContext(const RWLock & lock, RWLockImpl::Type type, ContextPtr context) const;
/// Same as tryLockTimedWithContext, but returns `nullptr` if lock is already acquired by current query.
2023-03-13 10:49:51 +00:00
static RWLockImpl::LockHolder tryLockForCurrentQueryTimedWithContext(const RWLock & lock, RWLockImpl::Type type, ContextPtr context);
void convertRightBlock(Block & block) const;
};
}