各位大侠meger into 是否可以开分枝? 解决了膜拜您!!!!!~~~

[usernamezero]

描述：有三个表A01,B01,C01 现以C来比较两个表，C01表中存有A01.id,b01.id....
C01结构是这样的 ID,A01ID,B01ID,A01NAME,A01IDCARD,A01SEX,B01NAME,BO1IDCARD,BO1SEX.....就不一一列举了.
下边是我写的meger into
meger into C01 C
using A01 A on(c.A1ID=A.ID)
when matched then
update set C.A1NAME=A.NAME,C.A1SEX=A.SEX,C.A1IDCARD=A.IDCARD
when not matched then
insert (C.A1id,C.A1NAME,C.A1IDCARD,C.A1SEX)
values (A.ID,A.NAME,A.IDCARD,A.SEX);
commit;
以上相同的更新成功了,如果插入了一条叫张三的记录(ID为:6676$5675,身份证号:1234567,姓名:张三,性别:男)
问当与B01这个表在做比较的时候,B01表中也有个叫张三的身份证号为1234567,性别男这个时候怎么办:
我的理解是这个时候B01这个表中的张三应该和A01这个表中的张三合成一条记录不应该在单独插入,这种meger into 怎么写谢谢.

[usernamezero]

谢谢二楼三楼的回复可是我这个报错了,你们能给看一下吧
ORA-30926:unable to get a stable set of rows in the source tables

[yf520gn]

ORA-30926: 无法在源表中获得一组稳定的行这个错误是使用MERGE最常见的错误，造成这个错误的原因是由于通过连接条件得到的T的记录不唯一。

[yf520gn]

意思就是通过关联条件，得到的记录不唯一~
贴你的代码出来看看~

[usernamezero]

merge into sqfmdm m
using (select * from sqfhr union select * from sqfsb) h
on (m.hr_psn_id=h.hr_id)
when matched then
update set m.hrname=h.name
when not matched then
insert (m.id,m.hr_psn_id,m.hrname)
values (h.id,h.hr_id,h.name);
commit;

[yf520gn]

select * from sqfhr union select * from sqfsb
这一句产生的数据集无法通过你的关联条件(m.hr_psn_id=h.hr_id) 产生唯一的记录。

PS:我感觉你对MERGE的理解还是不够深入，给你个链接去看看
http://xsb.itpub.net/post/419/30766

[usernamezero]

这样交流太费劲了,不过还得谢谢你

[Andy__Huang]

merge into sqfmdm m
using (select id,hr_id,name from sqfhr
union select id,hr_id,name from sqfsb a where not exists(select 1 from sqfhr b where a.name=b.name)
) h
on (m.hr_psn_id=h.hr_id)
when matched then
update set m.hrname=h.name
when not matched then
insert (m.id,m.hr_psn_id,m.hrname)
values (h.id,h.hr_id,h.name);
commit;

[Andy__Huang]

如果上面还不行,那么再加条件
merge into sqfmdm m
using (select id,hr_id,name from sqfhr
union select id,hr_id,name from sqfsb a where not exists(select 1 from sqfhr b where a.name=b.name and a.hr_id=b.hr_id)
) h
on (m.hr_psn_id=h.hr_id)
when matched then
update set m.hrname=h.name
when not matched then
insert (m.id,m.hr_psn_id,m.hrname)
values (h.id,h.hr_id,h.name);
commit;

[usernamezero]

谢谢楼上的哥们儿

[Andy__Huang]

meger into C01 C
using (select * from A01 union select * from B01) A
on(c.A1ID=A.ID)
when matched then
update set C.A1NAME=A.NAME,C.A1SEX=A.SEX,C.A1IDCARD=A.IDCARD
when not matched then
insert (C.A1id,C.A1NAME,C.A1IDCARD,C.A1SEX)
values (A.ID,A.NAME,A.IDCARD,A.SEX);
commit;

用union连接两个查询结果,它自动把相同的记录屏蔽掉;
记住,不能用union all连接

[yf520gn]

USING后面不光只能跟一个表，还可以跟一个数据集，LZ应该灵活应用~·