Question

表A

name, street, city, state, zip
bill, 123 main, newcity, null, 77777
bill, 123 main, newcity, null, 77777
fred, 23 west, greattown, null, 12345
bob, 4 a St, nowhere, null, 34567

表B

name, street, city, st, zip
bill, 123 main, newcity, me, 77777
bill, 123 main, newcity, me, 77777
fred, 23 west, greattown, ny, 12345
bob, 4 a St, nowhere, wy, 34567

我想做

update table A
set state = tB.st
from 
table A tA inner join table B tB
on (tA.name = tB.name and tA.street = tB.street)

但我不想更新2条记录“bill，123 main，newcity，null，77777”。

如何从表格中排除这些行？

由于查尔斯

Answer 1

在SQL Server 2005+中，您可以执行以下操作：

;
WITH A_counted AS (
  SELECT
    *,
    cnt = COUNT(*) OVER (PARTITION BY name, street, city, zip)
  FROM TableA
),   B_counted AS (
  SELECT
    *,
    cnt = COUNT(*) OVER (PARTITION BY name, street, city, zip)
  FROM TableB
)
UPDATE A_counted
SET state = B.state
FROM B_counted B
WHERE A_counted.name   = B.name,
  AND A_counted.street = B.street,
  AND A_counted.city   = B.city,
  AND A_counted.zip    = B.zip
  AND A_counted.cnt    = 1
  AND B.cnt            = 1

Answer 2

我看到它的方式是：

1）创建一个临时表，其中A中的记录不止一次出现。

2）创建第二个临时表，其中B中的所有记录与步骤1中的临时表中的记录不匹配。

3）更新步骤2中的所有记录。

Answer 3

你应该只需要为你的加入增加1个条件：

update table A
set state = tB.st
from 
table A tA inner join table B tB
on (tA.name = tB.name and tA.street = tB.street and tA.name <> 'Bill')

这应该排除那两行甚至被更新语句看到。

更新表但只更新非重复记录

3 个答案: