删除部分重复的记录，保留重复记录的第一条.

方法:

select   identity ( int , 1 , 1 )  as  autoID,  *   into  Tmp  from  aa
select   min (autoID)  as  autoID  into  Tmp2  from  Tmp  group   by  b
select   *   from  Tmp  where  autoID  in ( select  autoID  from  tmp2)