在查询中使用exists来进行判断性能会高,因为exists有短路的效果,整个表不必查完,通常比IN效果来的好,现在SQL2008R2中用代码测试一下,看看结果:
首先是使用IN来查询:
DBCC FREEPROCCACHE CHECKPOINT DBCC DROPCLEANBUFFERS SELECT * FROM sales.salesorderheader AS soh WHERE contactid in ( SELECT contactid FROM person.contact WHERE firstname='carla' AND lastname='adams' )
逻辑读为:
(4 row(s) affected)
Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'SalesOrderHeader'. Scan count 1, logical reads 703, physical reads 19, read-ahead reads 699, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Contact'. Scan count 1, logical reads 366, physical reads 6, read-ahead reads 364, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
第二种是使用EXISTS写法:
DBCC FREEPROCCACHE CHECKPOINT DBCC DROPCLEANBUFFERS SELECT * FROM sales.salesorderheader AS soh WHERE EXISTS ( SELECT c.contactid FROM person.contact AS c WHERE firstname='carla' AND lastname='adams' AND c.contactid=soh.contactid )
逻辑读为:
(4 row(s) affected)
Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'SalesOrderHeader'. Scan count 1, logical reads 703, physical reads 19, read-ahead reads 699, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Contact'. Scan count 1, logical reads 366, physical reads 6, read-ahead reads 364, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
第三种写法使用连接查询:
DBCC FREEPROCCACHE CHECKPOINT DBCC DROPCLEANBUFFERS SELECT soh.* FROM sales.salesorderheader AS soh JOIN person.contact AS c ON soh.contactid=c.contactid AND c.firstname='carla' AND lastname='adams'
逻辑读为:
(4 row(s) affected)
Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'SalesOrderHeader'. Scan count 1, logical reads 703, physical reads 19, read-ahead reads 699, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Contact'. Scan count 1, logical reads 366, physical reads 6, read-ahead reads 364, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
最后来看看三种写法的执行计划:
可以看到执行计划一模一样,所以得出的结论如下:
1:IN,EXIST,连接查询性能孰高孰低,不能想当然,以上就是明证,以前脱口而出的答案,现在不管用了
2:我的感觉规则总有例外,尤其在SQL性能调优这块,以前最佳实践,现在只能通过不断的测试哪种写法最优了