Suppose the order business table structure is:
order(oid, date, uid, status, money, time, …)
in:
oid , order id , primary key
date , the date of the order, there is a common index , the management background often queries according to date
uid , user ID , there is a common index , users can query their own orders
status , the order status, there is a common index , the management background is often queried according to the status
money/time , order amount /time , queried field, no index
…
Suppose an order has three states: 0 Ordered, 1 Paid, 2 Completed
Business requirements, query outstanding orders, which SQL is faster?
select * from order where status!=2
select * from order where status=0 or status=1
select * from order where status IN (0,1)
select * from order where status=0
union all
select * from order where status=1
Conclusion: Scheme 1 is the slowest, schemes 2, 3, and 4 can all hit the index
but...
One: union all must be able to hit the index
select * from order where status=0
union all
select * from order where status=1
illustrate:
Tell MySQL directly what to do, MySQL consumes the least CPU
Programmers don't often write SQL like this (union all)
Two: simple in can hit the index
select * from order where status in (0,1)
illustrate:
Let MySQL think, query optimization consumes more CPU than union all , but it can be ignored
Programmers often write SQL(in) like this. In this example, it is most recommended to write like this
Three: For or , the new version of MySQL can hit the index
select * from order where status=0 or status=1
illustrate:
让MySQL思考,查询优化耗费的cpu比in多,别把负担交给MySQL
不建议程序员频繁用or,不是所有的or都命中索引
对于老版本的MySQL,建议查询分析下
四、对于!=,负向查询肯定不能命中索引
select * from order where status!=2
说明:
全表扫描,效率最低,所有方案中最慢
禁止使用负向查询
五、其他方案
select * from order where status < 2
这个具体的例子中,确实快,但是:
这个例子只举了3个状态,实际业务不止这3个状态,并且状态的“值”正好满足偏序关系,万一是查其他状态呢,SQL不宜依赖于枚举的值,方案不通用
这个SQL可读性差,可理解性差,可维护性差,强烈不推荐