具有以下查询,在联接情况下,如果没有此子查询,该查询的效果很好。关于更有效的方法有什么想法?目前尚不清楚是否发生这种情况,因为子查询正在针对项中的所有记录运行,并且不遵守where = cms.orderno ...谢谢您的帮助!
子查询:
select count(distinct(ship_to))
from items
where cms.orderno = items.orderno and ship_to <> 0
完整查询:
select
cms.custnum,
cms.odr_date,
cms.orderno,
internetid,
ltrim(rtrim(cust.firstname))+' '+ltrim(rtrim(cust.lastname)) as recipient,
ord_total,
ordertype,
order_st2,
websitestatdesc,
holdcode
from
cms
join statdesc on cms.order_st2 = statdesc.statcode
join items on cms.orderno = items.orderno
join cust on
case
when (
select count(distinct(ship_to))
from items
where cms.orderno = items.orderno and ship_to <> 0
) = 1 then items.ship_to
when cms.shipnum <> 0 then cms.shipnum
else cms.custnum
end = cust.custnum
where
cms.custnum = 3300
and statdesc.stattype = 'O'
group by
cms.orderno,
cms.custnum,
cms.odr_date,
internetid,
firstname,
lastname,
ord_total,
ordertype,
order_st2,
websitestatdesc,
holdcode
order by cms.orderno desc
答案 0 :(得分:1)
首先,您在表orderno
的列items
上有索引吗?我想您不会返回很多行...因此缺少索引可能是一个杀手.。
然后,我将倾向于修改查询以避免出现如下group by
(假设我当然理解您的意图)的情况:
select
cms.custnum
, cms.odr_date
, cms.orderno
, internetid
, ltrim(rtrim(cust.firstname))+' '+ltrim(rtrim(cust.lastname)) as recipient
, ord_total
, ordertype
, order_st2
, websitestatdesc
, holdcode
from (
select *, (
select count(distinct(ship_to))
from items
where cms.orderno = items.orderno and ship_to <> 0
) distinct_ship_to_count
, (
select top 1 ship_to
from items
where cms.orderno = items.orderno and ship_to <> 0
) distinct_ship_to
from cms
) cms
join statdesc on cms.order_st2 = statdesc.statcode
--join items on cms.orderno = items.orderno
join cust on
case
when cms.distinct_ship_to_count = 1 then cms.distinct_ship_to
when cms.shipnum <> 0 then cms.shipnum
else cms.custnum
end = cust.custnum
where cms.custnum = 3300
and statdesc.stattype = 'O'
--group by cms.orderno, cms.custnum, cms.odr_date, internetid, firstname, lastname, ord_total, ordertype, order_st2, websitestatdesc, holdcode
order by cms.orderno desc