这是我拥有的数据的子集:
|Id1|Id2|
---------
|100| 90|
|101| 90|
|102| 90|
|217|101|
|218|101|
|219|101|
但Id2应该匹配Id1每次单独出现的第一次出现,只要Id2发生变化,就像这样:
|Id1|Id2|
---------
|100|100|
|101|100|
|102|100|
|217|217|
|218|217|
|219|217|
如何使用更新语句在数千行中更改此内容?
感谢您的帮助,这很难解释!
答案 0 :(得分:9)
您可以CTE
使用FIRST_VALUE
:
;WITH ToUpdate AS (
SELECT Id1, Id2,
FIRST_VALUE(Id1) OVER (PARTITION BY Id2 ORDER BY Id1) AS newValue
FROM mytable
)
UPDATE ToUpdate
SET Id2 = newValue
嗯,遗憾的是,FIRST_VALUE
在SQL Server 2008中不可用。在这种情况下,您可以使用以下查询:
;WITH CTE_Rn AS (
SELECT Id1, Id2,
ROW_NUMBER() OVER (PARTITION BY Id2 ORDER BY Id1) AS rn
FROM mytable
), ToUpdate AS (
SELECT t1.Id1, t2.Id2, t2.Id1 AS newValue
FROM mytable AS t1
JOIN CTE_Rn AS t2 ON t1.Id2 = t2.Id2 AND t2.rn = 1
)
UPDATE ToUpdate
SET Id2 = newValue