oracle

select * from  gcfr_t_vch_delete E WHERE E.rowid < (SELECT max(X.rowid) 
FROM gcfr_t_vch_delete X WHERE X.vchno = E.vchno and e.type = x.type and e.curragency = x.curragency);
--或者
select * from  gcfr_t_vch_delete E WHERE E.rowid > (SELECT min(X.rowid) 
FROM gcfr_t_vch_delete X WHERE X.vchno =
 
 E.vchno and e.type = x.type and e.curragency = x.curragency);

以憑證的類型，編號，單位為判斷重復的依據。刪除重復的就是要刪除多余的並且只留下一條。上述sql的意思就是查出重復的記錄之後，以一個最大值為標準，把小於他的都刪掉。或者以最小的為標準，把大於他的都刪掉。這樣就正好滿足需求。

oracle利用rowid來保證數據的唯一性，其實用一個表字段也可以，但是要保證這個字段的唯一性，最好是主鍵。

mysql

select * from  gcfr_t_vch_delete E WHERE E.guid < (SELECT max
 
(X.guid) 
FROM gcfr_t_vch_delete X WHERE X.vchno = E.vchno and e.type = x.type and e.curragency = x.curragency);
--或者
select * from  gcfr_t_vch_delete E WHERE E.guid > (SELECT min(X.guid) 
FROM gcfr_t_vch_delete X WHERE X.vchno = E.vchno and e.type = x.type and e.curragency = x.curragency);

mysql這裏沒有rowid，以guid為例，guid是表的主鍵。

刪除數據庫表中的重復記錄