嗨,我有这个触发器
create or replace TRIGGER trg_cust
after insert or update on nbty_customer
referencing old as old new as new
for each row
declare
pragma autonomous_transaction;
V_REQ VARCHAR2(10);
begin
IF(:NEW.cust_sal<1000)
THEN
select :NEW.cust_sal+:OLD.cust_sal INTO V_REQ FROM nbty_customer
where cust_id=:old.cust_id;
ELSE
SELECT :NEW.cust_sal-:OLD.cust_sal INTO V_REQ FROM nbty_customer where cust_id=:old.cust_id;
end if;
merge into nbty_cache_cust
using
(select distinct :new.cust_id cust_id, :new.cust_name cust_name,V_REQ V_REQ
from nbty_customer) b
on (cust_nbty_id = b.cust_id)
when matched then
update set cust_nbty_name = b.cust_name, cust_nbty_sal = b.V_REQ
when not matched then
insert (cust_nbty_id, cust_nbty_name, cust_nbty_sal)
values (b.cust_id, b.cust_name, b.V_REQ);
commit;
end;
哪个编译正确,但是当在表nbty_customer上进行插入时,例如
insert into nbty_customer values('2','abc','200')
它抛出ora-01403没有发现数据错误,执行触发器时出现ora-04088错误 请帮助,我无法弄清楚问题是什么?
答案 0 :(得分:2)
问题是由于select into
子句返回“未找到数据”所以我删除了select into
并直接使用了绑定变量:
IF(:new.cust_sal<1000)
then
v_req:=nvl(:new.cust_sal,0)+nvl(:old.cust_sal,0);
else
v_req:=nvl(:new.cust_sal,0)-nvl(:old.cust_sal,0);
end if;
这解决了这个问题。
答案 1 :(得分:1)
我刚评论过,然后我注意到了这一点:
select :NEW.cust_sal+:OLD.cust_sal INTO V_REQ FROM nbty_customer
where cust_id=:old.cust_id;
在insert
上,此select
将失败,因为没有:old.cust_id
。请改用:new.cust_id
。 (另外,:old.cust_sal
也将为空)
答案 2 :(得分:0)
在您的触发器正文中,您正在尝试选择新行的数据:
SELECT :NEW.cust_sal-:OLD.cust_sal INTO V_REQ FROM nbty_customer where cust_id=:old.cust_id;
但是您将触发器主体声明为PRAGMA AUTONOMOUS_TRANSACTION
。因此,您的触发器不会首先看到触发触发器执行的更改。
在触发器中使用AUTONOMOUS_TRANSACTION
从不一个好主意;请参阅Tom Kyte on Triggers,详细说明您为什么要避免这种情况。
如果出于性能原因需要缓存数据,我建议您使用实体化视图。如果需要,您可以将其设置为REFRESH ON COMMIT
:
create materialized view mv_cache
refresh on commit
as
select distinct cust_id cust_id, cust_name cust_name, V_REQ
from nbty_customer