小编ahm*_*ary的帖子

我正在使用 SQL Server 2019，并面临一个问题，即在进行更新时，只需要 20 秒即可更新 162 行。

• 预计执行计划

• 实际执行计划

Update FT SET
    ft.ValueName=avo.name
FROM #package FT 
inner join parts.Nop_PackageAttribute PA with(nolock)
    on PA.PackageID=ft.PackageID
        and PA.[Key]=FT.ZfeatureId 
inner join Nop_AcceptedValuesOption AVO with(nolock)
    ON convert(varchar(20),AVO.AcceptedValuesOptionID)=PA.Value 
where FT.AcceptedValueID is not null

脚本示例

 create table #package
 (
 id int PRIMARY KEY IDENTITY(1,1),
 ZfeatureId INT NULL,
 AcceptedValueID INT NULL,
 PackageID INT NULL,
 ValueName NVARCHAR(2000) default ''
 )

我在 #package 表上的索引

 create nonclustered index IDX_PackageID on #package(PackageID) include (ZfeatureId,AcceptedValueID , ValueName)
 create index acceptedvaluesidpackage_idx on #package(AcceptedValueID) …

我在 SQL Server 中遇到一个问题，其中非聚集索引查找的性能很差。

下面是实际的执行计划 https://www.brentozar.com/pastetheplan/?id=Sk3-4JGAK

我怎样才能提高绩效？

下面是表定义

 CREATE TABLE [Parts].[ManufacturingData](
     [LeadFinishId] [int] IDENTITY(1,1) NOT FOR REPLICATION NOT NULL,
     [PartID] [int] NOT NULL,
     [LeadFinishMaterial] [varchar](50) NULL,
     [CreatedDate] [datetime] NULL,
     [CreatedBy] [int] NULL,
     [ModifiedDate] [datetime] NULL,
     [Modifiedby] [int] NULL,
     [DeletedDate] [datetime] NULL,
     [DeletedBy] [int] NULL,
     [Revision_Id] [int] NULL,
     [BaseMaterialID] [int] NULL,
     [MSLID] [int] NULL,
     [MSLSource_Revision_id] [int] NULL,
     [MaximumReflowTemperatureID] [int] NULL,
     [ReflowTemperatureSource_Revision_Id] [int] NULL,
     [MaximumWaveTemperatureID] [int] NULL,
     [WaveTemperatureSource_Revision_ID] [int] NULL,
     [ReflowSolderTimeID] [int] NULL,
     [WaveSolderTimeID] [int] NULL,
     [NumberOfReflowCycleID] [int] NULL,
     [LeadFinishPlatingID] [int] NULL, …