日期:2014-05-17  浏览次数:20708 次

请问我这个视图能优化吗?
create or replace view v_biz_hadhandled as
select f1.*,f3.flowstepname as nextworkname from
(select ff.*,j.cetf_id,j.live_address,j.rpr_live_jd
  from t_biz_hadhandled ff,v_jz_app p,jz_basic_person j
  where ff.biztableid=p.pid(+)
  and j.pid=p.person_id
  and ff.person_name is not null
union all
select ff.*,null,null,null from t_biz_hadhandled ff where ff.person_name is null
)f1,(
  select f.biztableid,MAX(F.PCTIME)
  AS max_pctime,max(f.flowstep)
  as max_flowstep from t_biz_hadhandled f

  group by f.biztableid,f.senderid
  ) f2,t_bizwaithandle f3
  where f1.biztableid=f2.biztableid
  and f1.pctime=f2.max_pctime
  and f1.flowstep=max_flowstep
  and f3.biztableid(+)=f1.biztableid
;


我已经建立了索引,但感觉速度还是不理想,想问问各位大哥看这个语句能不能修改下,使查询更快。

------解决方案--------------------
SQL code
create or replace view v_biz_hadhandled as
select ff.*,
       j.cetf_id,
       j.live_address,
       j.rpr_live_jd,
       f3.flowstepname as nextworkname
  from t_biz_hadhandled ff
  left join v_jz_app p on ff.biztableid = p.pid
                      and ff.person_name is not null
  left join jz_basic_person j on j.pid = p.person_id
  left join t_bizwaithandle f3 on f3.biztableid = ff.biztableid
 where exists
 (select 1
          from t_biz_hadhandled
         where biztableid = ff.biztableid
         group by senderid
        having MAX(PCTIME) = ff.pctime and max(flowstep) = ff.flowstep)