Question

我正在使用Merge语句在我的存储过程中插入/更新数据。

MERGE [dbo].[tReserveData_4541] AS Target
USING (SELECT * FROM  [dbo].[tblcangrowhitesh]) AS Source
ON ( [Source].[PK_ID] =  [Target].[PK_ID])

WHEN MATCHED THEN
    UPDATE SET [Target].[mgender] = Source.[mgender]

WHEN NOT MATCHED BY TARGET THEN
    INSERT ([Granularity], [PK_ID], [ROWID], 
            [mgender], [mma1], [mma2], [mma3], 
            [mmadmincost], [mmcumulativevolume], 
            [mmcurrency], [mmdate],
            [mmfileimporteddate], [mmfilename])
    VALUES ([Source].[Granularity], [Source].[PK_ID], [Source].[ROWID],
            [Source].[mgender], [Source].[mma1], [Source].[mma2], [Source].[mma3], 
            [Source].[mmadmincost], [Source].[mmcumulativevolume],
            [Source].[mmcurrency], [Source].[mmdate],
            [Source].[mmfileimporteddate], [Source].[mmfilename])

如您所见，我将在[PK_ID]语句中插入标识列，即MERGE。但我无法这样做。

Answer 1

我不熟悉MERGE命令，但是当您直接插入标识列时，您是否不必使用IDENTITY_INSERT命令？ e.g。

SET IDENTITY_INSERT [dbo].[tReserveData_4541] on

MERGE...

SET IDENTITY_INSERT [dbo].[tReserveData_4541] off

Answer 2

如果您不需要在目标上生成值，请删除PK_ID上的IDENTITY属性。

为此，您可以添加PK_ID_temp列UPDATE [dbo].[tReserveData_4541] SET PK_ID_temp = PK_ID。删除PK_ID列，然后将PK_ID_temp重命名为PK_ID。

这可能还需要删除主键，删除所有关联索引并重建所有非聚簇索引（如果它是聚簇索引）。

合并声明和身份插入

2 个答案: