相关插入语句

Question

我有一个表，其中包含有关客户的数据，客户（姓名，地址），其中包含“John Doe”、“Some Street 123”等行。对于表中的每一行，我想在 Person(id, name) 表中插入一行，并在 Address(id, person_id, address) 表中插入一行。

我可以通过为 Customer 中的每一行运行两个插入语句来完成此操作：

insert into Person(name) values (@name);
insert into Address(person_id, address) values (scope_identity(), @address);

但这效率很低。我想批量插入，如下所示：

-- This works, the problem is with the Address table...
insert into Person(name)
select name from Customer

-- This looks good but does not work because name is not unique.
insert into Address(person_id, address)
select p.person_id, c.address
from Customer c join Person p on c.name = p.name

Answer 1

把这个留给像我一样找到这篇文章的谷歌旅行者。

我找到了这个解决方案，它似乎工作得很好，并且不需要任何时髦的模式更改：
https://dba.stackexchange.com/questions/160210/将数据一次性拆分为两个表

他们使用 MERGE 语句执行对第一个表（生成要生成的标识的表）的初始插入。在其他地方使用）。它使用 MERGE 语句的原因是它允许您使用 OUTPUT 语句，您可以使用该语句输出新的标识值以及来自源表（与在标准 INSERT 上使用 OUTPUT 语句相反，后者不允许您输出源表标识）。您可以将此输出数据插入到映射表中，并使用该映射表执行第二次插入。

这是我的解决方案的示例代码：

------------------------------------------------------------------------------

------------------------------------------------------------------------------
-- Set up sample schema and data
------------------------------------------------------------------------------
--Source Data
IF OBJECT_ID('dbo.tmp1') IS NOT NULL DROP TABLE dbo.tmp1 --SELECT * FROM dbo.tmp1
CREATE TABLE dbo.tmp1 (tmp1ID INT IDENTITY(1,1), Col1 CHAR(1) NOT NULL, Col2 CHAR(1) NOT NULL, Col3 CHAR(1) NOT NULL, Col4 CHAR(1) NOT NULL, Col5 CHAR(1) NOT NULL, Col6 CHAR(1) NOT NULL)

INSERT INTO dbo.tmp1 (Col1, Col2, Col3, Col4, Col5, Col6)
SELECT x.c1, x.c2, x.c3, x.c4, x.c5, x.c6
FROM (VALUES ('A','B','C','D','E','F'),
             ('G','H','I','J','K','L'),
             ('M','N','O','P','Q','R')
) x(c1,c2,c3,c4,c5,c6)

IF OBJECT_ID('dbo.tmp3') IS NOT NULL DROP TABLE dbo.tmp3 --SELECT * FROM dbo.tmp3
IF OBJECT_ID('dbo.tmp2') IS NOT NULL DROP TABLE dbo.tmp2 --SELECT * FROM dbo.tmp2

--Taget tables to split into
CREATE TABLE dbo.tmp2 (
      tmp2ID INT IDENTITY(1,1) NOT NULL CONSTRAINT PK_tmp2 PRIMARY KEY CLUSTERED (tmp2ID ASC)
    , Col1 CHAR(1) NOT NULL
    , Col2 CHAR(1) NOT NULL
    , Col3 CHAR(1) NOT NULL
)

CREATE TABLE dbo.tmp3 (
      tmp2ID INT NOT NULL
    , Col4 CHAR(1) NOT NULL
    , Col5 CHAR(1) NOT NULL
    , Col6 CHAR(1) NOT NULL
    , CONSTRAINT FK_tmp3_tmp2ID FOREIGN KEY(tmp2ID) REFERENCES dbo.tmp2 (tmp2ID)
)
------------------------------------------------------------------------------

------------------------------------------------------------------------------
-- Split data into two tables
------------------------------------------------------------------------------
DECLARE @Mapping TABLE (tmp1ID INT NOT NULL, tmp2ID INT NOT NULL);

--Use merge statment to output the source data PK as well as the newly inserted identity to generate a mapping table
MERGE INTO dbo.tmp2 AS tgt
USING dbo.tmp1 AS src ON (1=0)
WHEN NOT MATCHED THEN
    INSERT (    Col1,     Col2,     Col3)
    VALUES (src.Col1, src.Col2, src.Col3)
OUTPUT src.tmp1ID, Inserted.tmp2ID INTO @Mapping (tmp1ID, tmp2ID);

--Use the mapping table to insert the split data into the second table
INSERT INTO dbo.tmp3 (tmp2ID, Col4, Col5, Col6)
SELECT t2.tmp2ID, t1.Col4, t1.Col5, t1.Col6
FROM dbo.tmp2 t2
    JOIN @Mapping m ON m.tmp2ID = t2.tmp2ID
    JOIN dbo.tmp1 t1 ON t1.tmp1ID = m.tmp1ID

SELECT tmp2ID, Col1, Col2, Col3 FROM dbo.tmp2
SELECT tmp2ID, Col4, Col5, Col6 FROM dbo.tmp3
------------------------------------------------------------------------------

------------------------------------------------------------------------------
-- Clean up
------------------------------------------------------------------------------
DROP TABLE dbo.tmp1
DROP TABLE dbo.tmp3
DROP TABLE dbo.tmp2
------------------------------------------------------------------------------

------------------------------------------------------------------------------
GO

Answer 2

正如您所解释的那样，无法执行此操作，因为您丢失了第一次插入的每行的scope_identity() 值。

解决方法可能是将客户主键字段添加到人员表中，然后将第二次插入与此字段连接：

在插入之前，在人员上创建 customerID 字段，

alter table Person add customerID int null;

然后批量插入：

-- inserting customerID
insert into Person(name, customerID)
select name, customerID from Customer

-- joining on customerID.
insert into Address(person_id, address)
select p.person_id, c.address
  from Customer c 
  join Person p on c.customerID = p.customerID

之后，您可以从人员表中删除 customerID 字段：

alter table Person drop column customerID

Answer 3

最好在两个表中创建一些与它们相关的唯一类型的字段。否则您需要连接，因为您没有条件的唯一字段