sql删除数据表中重复的记录，多字段匹配，只保留一条记录

在oracle中 可以直接使用in 来 匹配，sql server 中不可以使用in关键字多字段匹配，所以选择使用 and
连接。下面用Min()函数保留组前面一条记录，如果保留最后一条记录用Max()函数。

select * from test2 
where (t1 in (select 
t1 from test2 group by t1 having count(*) >1 )) and (t2 in 
(select t2 from test2 group by t2 having count(*) >1 
)) 
and tid not in (select min(tid) from test2 
group by t1 having count(*) >1 ) 
and tid not in (select 
min(tid) from test2 group by t2 having count(*) >1 ) 

在oracle中删除多余的重复项

select * from test2 

where (t1,t2) in (select t1,t2 from test2 group by 
t1,t2 having(count(*)>1 ))