这句SQL语句如何优化,现在执行很慢很慢......

[entice]

select column_oid,value from (
select a.column_oid,a.doc_oid, c.lable value
from f_item a,f_column b,f_refitem c
Where a.column_oid = b.column_oid
and b.ref_id != -1 and b.type !=5
and b.ref_id = c.ref_oid and a.value = c.value
and (a.doc_oid = ? or doc_oid in (select doc_oid from f_document where supdoc = ? ))
Union
select a.column_oid,a.doc_oid,a.value
from f_item a,f_column b
Where a.column_oid = b.column_oid And b.ref_id = -1
and (a.doc_oid = ? or a.doc_oid
in (select doc_oid from f_document where supdoc = ? ))
)
order by doc_oid

[beckhambobo]

order by doc_oid打乱了索引顺序，当然慢

[entice]

应该怎么办?请高手指点...

[wwl007]

索引应该没有问题的 因为是对于查询出来的记录再操作
已经没有很多数据了
我感觉你的 in 语法有点问题
可能数度很慢

[wwl007]

select a.column_oid,a.doc_oid, c.lable value
from f_item a,f_column b,f_refitem c，(select doc_oid from f_document
where supdoc = ?
union
select '?' from dual) d
Where a.column_oid = b.column_oid
and b.ref_id != -1 and b.type !=5
and b.ref_id = c.ref_oid and a.value = c.value
and a.doc_oid =d.doc_iod

[Michaelyfj]

建议不要用in，如果可以用exists来代替，尽量用exists来代替

[csusmart]

你的索引建了没有,是怎么建的?