add virtual column _data_version

ClickHouse · Mar 1, 2024 · dd400de · dd400de
1 parent dcff866
commit dd400de
Show file tree

Hide file tree

Showing 6 changed files with 31 additions and 3 deletions.
diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp
@@ -7,7 +7,6 @@
 
 #include <Common/logger_useful.h>
 #include <Common/ActionBlocker.h>
-#include "Storages/MergeTreeVirtualColumns.h"
 #include <Processors/Transforms/CheckSortedTransform.h>
 #include <Storages/MergeTree/DataPartStorageOnDiskFull.h>
 

diff --git a/src/Storages/MergeTree/MergeTreeData.cpp b/src/Storages/MergeTree/MergeTreeData.cpp
@@ -440,6 +440,7 @@ VirtualColumnsDescription MergeTreeData::createVirtuals(const StorageInMemoryMet
     desc.addEphemeral("_partition_id", std::make_shared<DataTypeLowCardinality>(std::make_shared<DataTypeString>()), "Name of partition");
     desc.addEphemeral("_sample_factor", std::make_shared<DataTypeFloat64>(), "Sample factor (from the query)");
     desc.addEphemeral("_part_offset", std::make_shared<DataTypeUInt64>(), "Number of row in the part");
+    desc.addEphemeral("_data_version", std::make_shared<DataTypeUInt64>(), "Data version of mutation of part");
 
     if (metadata.hasPartitionKey())
     {
@@ -1027,7 +1028,7 @@ void MergeTreeData::MergingParams::check(const StorageInMemoryMetadata & metadat
     /// TODO Checks for Graphite mode.
 }
 
-const Names MergeTreeData::virtuals_useful_for_filter = {"_part", "_partition_id", "_part_uuid", "_partition_value"};
+const Names MergeTreeData::virtuals_useful_for_filter = {"_part", "_partition_id", "_part_uuid", "_partition_value", "_data_version"};
 
 Block MergeTreeData::getHeaderWithVirtualsForFilter() const
 {

diff --git a/src/Storages/MergeTree/MergeTreeRangeReader.cpp b/src/Storages/MergeTree/MergeTreeRangeReader.cpp
@@ -1,11 +1,11 @@
 #include <Storages/MergeTree/MergeTreeRangeReader.h>
 #include <Storages/MergeTree/IMergeTreeReader.h>
+#include <Storages/MergeTreeVirtualColumns.h>
 #include <Columns/FilterDescription.h>
 #include <Columns/ColumnConst.h>
 #include <Columns/ColumnsCommon.h>
 #include <Common/TargetSpecific.h>
 #include <Common/logger_useful.h>
-#include "Storages/MergeTreeVirtualColumns.h"
 #include <Core/UUID.h>
 #include <IO/WriteBufferFromString.h>
 #include <IO/Operators.h>

diff --git a/src/Storages/MergeTreeVirtualColumns.cpp b/src/Storages/MergeTreeVirtualColumns.cpp
@@ -47,6 +47,9 @@ Field getFieldForConstVirtualColumn(const String & column_name, const IMergeTree
     if (column_name == "_partition_id")
         return part.info.partition_id;
 
+    if (column_name == "_data_version")
+        return part.info.getDataVersion();
+
     if (column_name == "_partition_value")
         return Tuple(part.partition.value.begin(), part.partition.value.end());
 

diff --git a/tests/queries/0_stateless/03001_data_version_column.reference b/tests/queries/0_stateless/03001_data_version_column.reference
@@ -0,0 +1,5 @@
+all_1_1_0	1	1	1
+all_2_2_0	2	2	2
+all_1_1_0_3	3	1	100
+all_2_2_0_3	3	2	200
+all_4_4_0	4	3	3
diff --git a/tests/queries/0_stateless/03001_data_version_column.sql b/tests/queries/0_stateless/03001_data_version_column.sql
@@ -0,0 +1,20 @@
+DROP TABLE IF EXISTS t_data_version;
+
+CREATE TABLE t_data_version (a UInt64, b UInt64) ENGINE = MergeTree ORDER BY a;
+
+INSERT INTO t_data_version VALUES (1, 1);
+INSERT INTO t_data_version VALUES (2, 2);
+
+SELECT _part, _data_version, * FROM t_data_version ORDER BY a;
+
+ALTER TABLE t_data_version UPDATE b = a * 100 WHERE 1 SETTINGS mutations_sync = 2;
+
+SELECT _part, _data_version, * FROM t_data_version ORDER BY a;
+
+INSERT INTO t_data_version VALUES (3, 3);
+
+-- Check part pruning.
+SELECT _part, _data_version, * FROM t_data_version WHERE _data_version = 4 ORDER BY a SETTINGS max_rows_to_read = 1;
+
+DROP TABLE t_data_version;
+