--测试数据
/*-----------------------------
select * from tt
-----------------------------*/
id pid
----------- -----------
1 1
1 1
2 2
3 3
3 3
3 3
(所影响的行数为 6 行)
首先,如何查询table中有重复记录
select *,count(1) as rownum
from tt
group by id, pid
having count(1) > 1
id pid rownum
----------- ----------- -----------
1 1 2
3 3 3
(所影响的行数为 2 行)
方法一:使用distinct和临时表
if object_id('tempdb..#tmp') is not null
drop table #tmp
select distinct * into #tmp from tt
truncate table tt
insert into tt select * from #tmp
方法二:添加标识列
alter table tt add NewID int identity(1,1)
go
delete from tt where exists(select 1 from tt a where a.newid>tt.newid and tt.id=a.id and tt.pid=a.pid)
go
alter table tt drop column NewID
go
--测试结果
/*-----------------------------
select * from tt
-----------------------------*/
id pid
----------- -----------
1 1
2 2
3 3
(所影响的行数为 3 行)
本文地址:http://com.8s8s.com/it/it19579.htm