4

我正在寻找正确的语法和方法来直接从 SQL 中执行以下操作：从具有相同复合主键TableMain的数据中插入或更新（如果内部已经存在数据）。TableA

两个表都定义为：

CREATE TABLE TableA (
[TID0] [int] NOT NULL,
[TID1] [int] NOT NULL,
[language] [nvarchar](2) NOT NULL,
[TID2] [nvarchar](200) NOT NULL,
[text] [nvarchar](max) NULL,
[updatedOn] [datetime] NOT NULL  DEFAULT (getdate())
PRIMARY KEY (
    [TID0],
    [TID1],
    [language],
    [TID2],
)
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]

TableA会定期删除和填充。

TableMain作为相同的定义，但将包含更多的数据行，我需要的是从TableAinto插入从未见过的值TableMain，并更新已经存在的行。

我曾经做过这种插入，但我不知道如何处理更新和复合主键：

INSERT INTO TableMain 
    SELECT * FROM TableA

编辑：我正在使用SQL Server 9.00.5000

编辑：受 MERGE 启发并模仿它的另一种方式

DECLARE @updatedIDs TABLE(
    [TID0] [int],
    [TID1] [int],
    [language] [nvarchar](2),
    [TID2] [nvarchar](200),
                PRIMARY KEY ([TID0], [TID1], [language], [TID2])  -- as stated by Nikola Markovinović above, thanks
);

-- First update records 
update TableMain 
   set [text]      = source.[text], 
       [updatedOn] = source.[updatedOn] 
       OUTPUT
         inserted.[TID0]
         inserted.[TID1]
         inserted.[language]
         inserted.[TID2]
       INTO @updatedIDs
  from 
       TableMain AS main
       , TableA AS source 
  WHERE    
   TableMain.[TID0]         = source.[TID0] 
   and TableMain.[TID1]     = source.[TID1] 
   and TableMain.[language] = source.[language] 
   and TableMain.[TID2]     = source.[TID2] 

-- And then insert 

insert into TableMain
select *
  from TableA AS source 
 where not exists 
       ( 
      select 1
        from @updatedIDs AS i
       where i.[TID0]     = source.[TID0] 
         and i.[TID1]     = source.[TID1] 
         and i.[language] = source.[language] 
         and i.[TID2]     = source.[TID2] 
       )
4

3 回答 3

5

你应该使用合并语句

像这样的东西：

merge TableMain  AS target
using TableA as source 
ON <join tables here>
WHEN MATCHED THEN <update>
WHEN NOT MATCHED BY TARGET <Insert>
WHEN NOT MATCHED BY SOURCE <delete>
于 2012-08-09T12:41:57.480 回答
4

这是一个您可以用来更新数据的脚本：

-- On error transaction is automatically rolled back
set xact_abort on

begin transaction
-- First update records
update TableMain
   set [text]      = source.[text],
       [updatedOn] = source.[updatedOn]
  from TableMain
 inner join TableA source
    on TableMain.[TID0]     = source.[TID0]
   and TableMain.[TID1]     = source.[TID1]
   and TableMain.[language] = source.[language]
   and TableMain.[TID2]     = source.[TID2]

-- And then insert

insert into TableMain ([TID0], [TID1], [language], [TID2], [text], [updatedOn])
select [TID0], [TID1], [language], [TID2], [text], [updatedOn]
  from TableA source
 where not exists
       (
          select *
            from TableMain
           where TableMain.[TID0]     = source.[TID0]
             and TableMain.[TID1]     = source.[TID1]
             and TableMain.[language] = source.[language]
             and TableMain.[TID2]     = source.[TID2]
       )

commit transaction

您可能会重写 not exists() ，left join ... where TableMain.TID0 is null好像性能不令人满意。

于 2012-08-09T13:16:36.723 回答
3

您可以使用 SQLServer 2008 中的Merge命令。它允许您将来自另一个数据源的数据合并到您的主数据源中，并在存在键匹配时定义特定行为（并且您可能希望更新您的表）或存在不匹配，您想插入新记录。

http://blog.sqlauthority.com/2010/06/08/sql-server-merge-operations-insert-update-delete-in-single-execution/

您可以访问此链接以获取一些代码示例。

于 2012-08-09T12:17:52.003 回答