delete from table_name a where rowid < (select max(rowid) from table_name b where a.col1 = b.col1 and a.col2 = b.col2 -- 遍历所以数据,以确定重复纪录 )