Question

我正在使用SQL Server 2014.我查找了Joel Coehoorn为this question提供的解决方案，但它对我不起作用。

我有一个客户端表，我将其重命名为Aac_Client，其中包含客户端信息和地址信息作为列。我想将地址移到一个新表Address来组织事情，并将其余的事情保留在Ac_Client。我正在尝试执行以下操作：

请注意，地址与客户之间的关系是一对一的。

BEGIN TRANSACTION
   DECLARE @DataID int;

   -- Insert Address
    INSERT INTO Address ([StreetNumber],[StreetName] ,[StreetAddress2] ,[Unit] ,[City] ,[State] ,[Zip] ,[County])
      SELECT  [StreetNumber], [StreetName], [StreetAddress2], [Unit] ,[City] ,[State] ,[Zip] ,
      NULL AS [County]
      FROM Aac_Client

    -- Get Address Id
   SELECT @DataID = scope_identity();

    -- Insert Client    
    INSERT INTO Ac_Client ( AddressId, Name, Phone, Contact )
        SELECT  @DataID AS AddressId, Name , Phone, Contact FROM Aac_Client

COMMIT

但问题是INSERT是立即执行的，我在SELECT @DataID = scope_identity();

中获取了最后一个地址ID

Answer 1

BEGIN TRANSACTION

    IF OBJECT_ID('tempdb..#InsertedAddresses') IS NOT NULL
    BEGIN
        DROP TABLE #InsertedAddresses
    END

CREATE TABLE #InsertedAddresses (
    AddressId INT
    ,ClientId INT
)

   DECLARE @DataID int;

   -- add a unique client identifier
   -- Insert Address
    INSERT INTO Address ([StreetNumber],[StreetName] ,[StreetAddress2] ,[Unit] ,[City] ,[State] ,[Zip] ,[County], [ClientId])
    OUTPUT INSERTED.AddressId, INSERTED.ClientId (AddressId, ClientId)
      SELECT  [StreetNumber], [StreetName], [StreetAddress2], [Unit] ,[City] ,[State] ,[Zip] ,
      NULL AS [County]
      FROM Aac_Client

    -- Insert Client    
    INSERT INTO Ac_Client ( AddressId, Name, Phone, Contact, ClientId)
        SELECT i.AddressId, Name , Phone, Contact, c.ClientId
          FROM Aac_Client c
          INNER JON #InsertedAddresses i
          ON c.ClientId = i.ClientId

COMMIT

将第一个insert语句的output子句用于临时表。客户也可能有多个地址吗？或者在同一地址有多个客户？如果是这样，您可能需要考虑将AC_Client ID放在地址表中而不是客户端表中的address_id。此外，如果您不担心在同一地址的多个客户端的情况下独立更改地址，您可以考虑使用第3个表来关联地址和客户端，这样您就不必重复地址或客户端来创建多对多的关系。 / p>

我想扩展scope_identity是一个标量值，意味着只保留1个值，它是范围内更改的最后一行的标识。在您的情况下，您需要所有ID，因此您需要使用输出子句。

Answer 2

--Add the following Columns that you can later Drop 

ALTER TABLE [Address] ADD Col_GUID UNIQUEIDENTIFIER;
ALTER TABLE [Aac_Client] ADD Col_GUID UNIQUEIDENTIFIER;
GO

-- Give a Unique value to each row.
UPDATE [Aac_Client] SET Col_GUID = NEWID();

BEGIN TRANSACTION;

   -- Table variable to capture newly generated Identity values
   Declare @NewID TABLE (AddressId INT, Col_GUID UNIQUEIDENTIFIER) 

   -- Insert Address (with output clause to get the Identity values)
INSERT INTO [Address] ([StreetNumber],[StreetName] ,[StreetAddress2] 
                                      ,[Unit] ,[City] ,[State] ,[Zip] ,[County], Col_GUID)
OUTPUT Inserted.AddressId , Inserted.Col_GUID INTO @NewID (AddressId , Col_GUID)
SELECT  [StreetNumber], [StreetName], [StreetAddress2] 
                            ,[Unit] ,[City] ,[State] ,[Zip] , NULL AS [County] , Col_GUID
FROM Aac_Client


    -- Insert Client joining with the Table variable on the guid column to
    -- to get the new Identity values. 

INSERT INTO Ac_Client ( AddressId, Name, Phone, Contact )
SELECT  N.AddressId, a.Name , a.Phone, a.Contact 
FROM Aac_Client a
INNER JOIN @NewID N ON n.Col_GUID = a.Col_GUID

COMMIT TRANSACTION;

最后，您可以删除为此目的添加的列。

ALTER TABLE [Address] DROP COLUMN Col_GUID;
ALTER TABLE [Aac_Client] DROP COLUMN Col_GUID;

SQL - 将一个表拆分为两个和链接

2 个答案: