删除语句中未使用聚集索引
tol*_*ahC 13 index sql-server delete clustered-index sql-server-2012
我有一个 SQL Server 表定义如下
CREATE TABLE [dbo].[Production_Detail]
(
[Id] [bigint] NOT NULL DEFAULT (NEXT VALUE FOR [dbo].[Production_Detail_Seq]),
[Meta_Data_ID] INT NOT NULL ,
[Production_Detail_Time] DATETIME NOT NULL,
[Production_Detail_Time_Local] DATETIME NOT NULL,
[Production_Detail_Value] FLOAT NULL,
[IntegratedDM] BIT NOT NULL DEFAULT 0,
[DailyIntegratedDM] BIT NOT NULL DEFAULT 0,
[InsertedDate] DateTime NOT NULL,
[ModifiedDate] DateTime NOT NULL,
CONSTRAINT [PK_Production_Detail] PRIMARY KEY ([Id]),
CONSTRAINT [UK_Production_Detail] UNIQUE ([Meta_Data_Id], [Production_Detail_Time]),
CONSTRAINT [FK_Production_Detail_To_Meta_Data] FOREIGN KEY ([Meta_Data_ID]) REFERENCES [Meta_Data]([Meta_Data_ID])
)
GO
CREATE NONCLUSTERED INDEX [IX_Production_Detail_Production_Detail_Time_Local]
ON [dbo].[Production_Detail] ([Production_Detail_Time_Local])
INCLUDE ([Meta_Data_ID], [Production_Detail_Time], [Production_Detail_Value])
GO
CREATE UNIQUE NONCLUSTERED INDEX [IX_Production_Detail_Meta_Data_ID_Production_Detail_Time]
ON [dbo].[Production_Detail] ([Meta_Data_ID] ASC, [Production_Detail_Time] ASC)
GO
CREATE NONCLUSTERED INDEX [IX_Production_Detail_IntegratedDM_MetaDataId]
ON [dbo].[Production_Detail] ([IntegratedDM], [Meta_Data_ID])
INCLUDE ([Production_Detail_Time])
GO
此表包含实时数据，因此会始终插入记录。我正在尝试删除超过 5 天的数据。但是当删除运行时，我有几个会引发死锁的插入语句。
这是删除查询。该#ToPurge表包含需要删除的 ID。该#ToPurge表没有任何索引。行数可能会有所不同，但对于我的测试用例，它已超过 300 万：
Declare @RowCount int;
set @RowCount = @@ROWCOUNT;
while @RowCount > 0
begin
delete top(20000) pd
from #ToPurge tp
inner join [DataWarehouse].[dbo].[Production_Detail] pd on (tp.ID = pd.ID)
set @RowCount = @@ROWCOUNT;
end
这是删除查询的执行计划
<?xml version="1.0" encoding="utf-16"?>
<ShowPlanXML xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" Version="1.2" Build="11.0.3128.0" xmlns="http://schemas.microsoft.com/sqlserver/2004/07/showplan">
<BatchSequence>
<Batch>
<Statements>
<StmtSimple StatementCompId="1" StatementEstRows="20000" StatementId="1" StatementOptmLevel="FULL" StatementSubTreeCost="34.7996" StatementText=" delete top(20000) pd
 from #ToPurge tp
 inner join [DataWarehouse].[dbo].[Production_Detail] pd on (tp.ID = pd.ID)" StatementType="DELETE" QueryHash="0xF7A9CF8E140CDB4E" QueryPlanHash="0x2CC61A780737A398" RetrievedFromCache="false">
<StatementSetOptions ANSI_NULLS="true" ANSI_PADDING="true" ANSI_WARNINGS="true" ARITHABORT="true" CONCAT_NULL_YIELDS_NULL="true" NUMERIC_ROUNDABORT="false" QUOTED_IDENTIFIER="true" />
<QueryPlan CachedPlanSize="88" CompileTime="65" CompileCPU="65" CompileMemory="592">
<ThreadStat Branches="2" />
<MemoryGrantInfo SerialRequiredMemory="1024" SerialDesiredMemory="477448" />
<OptimizerHardwareDependentProperties EstimatedAvailableMemoryGrant="209715" EstimatedPagesCached="419430" EstimatedAvailableDegreeOfParallelism="4" />
<RelOp AvgRowSize="9" EstimateCPU="0.08" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Sequence" NodeId="0" Parallel="false" PhysicalOp="Sequence" EstimatedTotalSubtreeCost="34.7996">
<OutputList />
<Sequence>
<RelOp AvgRowSize="9" EstimateCPU="0.02" EstimateIO="0.458155" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Delete" NodeId="1" Parallel="false" PhysicalOp="Index Delete" EstimatedTotalSubtreeCost="8.56272">
<OutputList />
<Update WithUnorderedPrefetch="true" DMLRequestSort="false">
<Object Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Index="[UK_Production_Detail]" Alias="[pd]" IndexKind="NonClustered" />
<RelOp AvgRowSize="36" EstimateCPU="0.00182568" EstimateIO="0.00328238" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Eager Spool" NodeId="3" Parallel="false" PhysicalOp="Table Spool" EstimatedTotalSubtreeCost="8.08456">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Meta_Data_ID" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Production_Detail_Time" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Production_Detail_Time_Local" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="IntegratedDM" />
</OutputList>
<Spool>
<RelOp AvgRowSize="36" EstimateCPU="0.02" EstimateIO="1.31642" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Delete" NodeId="4" Parallel="false" PhysicalOp="Clustered Index Delete" EstimatedTotalSubtreeCost="32.3067">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Meta_Data_ID" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Production_Detail_Time" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Production_Detail_Time_Local" />
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="IntegratedDM" />
</OutputList>
<Update WithUnorderedPrefetch="true" DMLRequestSort="false">
<Object Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Index="[PK_Production_Detail]" Alias="[pd]" IndexKind="Clustered" />
<RelOp AvgRowSize="15" EstimateCPU="0.002" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Top" NodeId="6" Parallel="false" PhysicalOp="Top" EstimatedTotalSubtreeCost="30.9703">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</OutputList>
<Top RowCount="false" IsPercent="false" WithTies="false">
<TopExpression>
<ScalarOperator ScalarString="(20000)">
<Const ConstValue="(20000)" />
</ScalarOperator>
</TopExpression>
<RelOp AvgRowSize="15" EstimateCPU="0.0392491" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Gather Streams" NodeId="7" Parallel="true" PhysicalOp="Parallelism" EstimatedTotalSubtreeCost="30.9683">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</OutputList>
<Parallelism>
<RelOp AvgRowSize="15" EstimateCPU="276.841" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="20000" LogicalOp="Right Semi Join" NodeId="8" Parallel="true" PhysicalOp="Hash Match" EstimatedTotalSubtreeCost="30.929">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</OutputList>
<MemoryFractions Input="1" Output="1" />
<Hash>
<DefinedValues />
<HashKeysBuild>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</HashKeysBuild>
<HashKeysProbe>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</HashKeysProbe>
<ProbeResidual>
<ScalarOperator ScalarString="[DataWarehouse].[dbo].[Production_Detail].[Id] as [pd].[Id]=[tempdb].[dbo].[#ToPurge].[Id] as [tp].[Id]">
<Compare CompareOp="EQ">
<ScalarOperator>
<Identifier>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</Identifier>
</ScalarOperator>
<ScalarOperator>
<Identifier>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</Identifier>
</ScalarOperator>
</Compare>
</ScalarOperator>
</ProbeResidual>
<RelOp AvgRowSize="15" EstimateCPU="2.83429" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="3910180" LogicalOp="Bitmap Create" NodeId="9" Parallel="true" PhysicalOp="Bitmap" EstimatedTotalSubtreeCost="9.99726">
<OutputList>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</OutputList>
<Bitmap>
<DefinedValues>
<DefinedValue>
<ColumnReference Column="Bitmap1007" />
</DefinedValue>
</DefinedValues>
<HashKeys>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</HashKeys>
<RelOp AvgRowSize="15" EstimateCPU="2.83429" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="3910180" LogicalOp="Repartition Streams" NodeId="10" Parallel="true" PhysicalOp="Parallelism" EstimatedTotalSubtreeCost="9.99726">
<OutputList>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</OutputList>
<Parallelism PartitioningType="Hash">
<PartitionColumns>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</PartitionColumns>
<RelOp AvgRowSize="15" EstimateCPU="1.07532" EstimateIO="6.08765" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="3910180" LogicalOp="Table Scan" NodeId="11" Parallel="true" PhysicalOp="Table Scan" EstimatedTotalSubtreeCost="7.16297" TableCardinality="3910180">
<OutputList>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</OutputList>
<TableScan Ordered="false" ForcedIndex="false" ForceScan="false" NoExpandHint="false">
<DefinedValues>
<DefinedValue>
<ColumnReference Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" Column="Id" />
</DefinedValue>
</DefinedValues>
<Object Database="[tempdb]" Schema="[dbo]" Table="[#ToPurge]" Alias="[tp]" IndexKind="Heap" />
</TableScan>
</RelOp>
</Parallelism>
</RelOp>
</Bitmap>
</RelOp>
<RelOp AvgRowSize="15" EstimateCPU="0.675882" EstimateIO="0" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="902193" LogicalOp="Repartition Streams" NodeId="12" Parallel="true" PhysicalOp="Parallelism" EstimatedTotalSubtreeCost="3.17684">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</OutputList>
<Parallelism PartitioningType="Hash">
<PartitionColumns>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</PartitionColumns>
<RelOp AvgRowSize="15" EstimateCPU="43.216" EstimateIO="391.867" EstimateRebinds="0" EstimateRewinds="0" EstimatedExecutionMode="Row" EstimateRows="902193" LogicalOp="Index Scan" NodeId="13" Parallel="true" PhysicalOp="Index Scan" EstimatedTotalSubtreeCost="2.50096" TableCardinality="157149000">
<OutputList>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</OutputList>
<IndexScan Ordered="true" ScanDirection="FORWARD" ForcedIndex="false" ForceSeek="false" ForceScan="false" NoExpandHint="false" Storage="RowStore">
<DefinedValues>
<DefinedValue>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</DefinedValue>
</DefinedValues>
<Object Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Index="[UK_Production_Detail]" Alias="[pd]" IndexKind="NonClustered" />
<Predicate>
<ScalarOperator ScalarString="PROBE([Bitmap1007],[DataWarehouse].[dbo].[Production_Detail].[Id] as [pd].[Id],N'[IN ROW]')">
<Intrinsic FunctionName="PROBE">
<ScalarOperator>
<Identifier>
<ColumnReference Column="Bitmap1007" />
</Identifier>
</ScalarOperator>
<ScalarOperator>
<Identifier>
<ColumnReference Database="[DataWarehouse]" Schema="[dbo]" Table="[Production_Detail]" Alias="[pd]" Column="Id" />
</Identifier>
</ScalarOperator>
<ScalarOperator>
<Const ConstValue="N'[IN ROW]'" />
</ScalarOperator>
</Intrinsic>
</ScalarOperator>
</Predicate>
</IndexScan>
</RelOp>
</Parallelism>
</RelOp>
</Hash>
</RelOp>
</Parallelism>
</RelOp>
Mar*_*ith 10
我想知道为什么优化器不使用聚集索引，而是使用非聚集索引？
这将是基于成本的优化器的决定。它估计完全扫描窄索引更便宜。
似乎您期望在聚集索引上有一个带有搜索的嵌套循环？
执行计划显示该表#ToPurge有 390 万行。
这将需要 195 个批次才能清除。假设所有行#ToPurge都是唯一的并且最初存在于其中，Production_Detail并且从#ToPurge每个批次输出的行的顺序一致（我忽略了并行性），那么每个后续批次最终将寻找之前已删除的所有行在最终到达尚未处理的行补丁之前，您的delete进程总共将为已删除的行执行 10 亿次索引搜索的三分之一以上。
0 + 20,000 + 40,000 + ... + 3,880,000 = 378,300,000
这不是基于成本的优化器使用（这将涉及两个表的直方图和逻辑rowgoal的TOP 20000），但它表明这不是一个好的策略反正。
您可以尝试以下方法将每次迭代的搜索次数限制在 20,000 次，看看您现在是否获得了所需的计划。
CREATE TABLE #ToPurge
(
[BatchId] [INT] NULL,
[Id] [BIGINT] NOT NULL,
UNIQUE CLUSTERED([BatchId], [Id])
);
/*Your insert statement to #ToPurge here*/
WITH CTE
AS (SELECT *,
1 + ( ROW_NUMBER() OVER (ORDER BY Id) / 20000 ) AS _BatchId
FROM #ToPurge)
UPDATE CTE
SET [BatchId] = [_BatchId];
DECLARE @BatchCounter INT = 1;
WHILE @BatchCounter <= (SELECT MAX([BatchId])
FROM #ToPurge)
BEGIN
DELETE pd
FROM #ToPurge tp
INNER JOIN [DataWarehouse].[dbo].[Production_Detail] pd
ON ( tp.ID = pd.ID )
WHERE tp.BatchId = @BatchCounter
SET @BatchCounter+=1;
END
| 归档时间： |
|
| 查看次数： |
975 次 |
| 最近记录： |