使用EXISTS(或NOT EXISTS)通常将提高查询的效率，由于NOT IN子句将对子查询中的表执行了一个全表遍历。
oracle在执行IN子查询过程中，先执行子查询结果放入临时表再进行主查询；
而exists先运行主查询，再执行子查询查到第一个匹配项
如：查询 A表中没有和B表或C表相连的数据
select A.id,A.reads,A.addtime
from A
where A.person_id not in (select user_id from B )
or A.worksite_id not in (select id from C)
order by A.addtime desc
程序执行时间：40109.38毫秒
select A.id,A.reads,A.addtime
from A
where not exists (select id FROM B where B.user_id=A.person_id)
or not exists (select id FROM C where C.id=A.worksite_id)
order by Sendorder.addtime desc
程序执行时间：8531.25毫秒
ms ssql 实例：
select
subscriber_id
from
NewsLetterSystem_CampaignCategorySubscriber ccs WITH ( NOLOCK )
where NOT EXISTS(
SELECT distinct subscriber_id
FROM NewsLetterSystem_OpenTracking ot WITH ( NOLOCK )
WHERE ot.subscriber_id=ccs.subscriber_id and ot.campaign_id = 52801)
and ccs.campaign_id = 52801 and ccs.status_id = 1