什么是延迟索引？使用索引查询出来数据，之后把查询结果和同一张表中数据进行连接查询，进而提高查询速度!
分页是一个很常见功能，select ** from tablename limit ($page - 1 ) * $n ,$n
通过一个存储过程插入10000条数据进行测试：
create table smth1 ( id int auto_increment , ver int(11) default null, content varchar(1000) not null, intro varchar(1000) not null, primary key(id), key idver(id,ver) )engine = innodb default charset = utf8;create procedure smthtest1()begin declare num int default 100001; while num show profiles;+----------+------------+----------------------------------------------+| query_id | duration | query |+----------+------------+----------------------------------------------+| 1 | 0.002006 | select id ,content from smth1 limit 1000,10 || 2 | 0.030106 | select id ,content from smth1 limit 5000,10 || 3 | 0.042428 | select id ,content from smth1 limit 9000,10 || 4 | 0.01297225 | select id ,content from smth1 limit 10000,10 || 5 | 0.13077625 | select id ,content from smth1 limit 20000,10 |
可见随着查询$page 变大，时间会越来越大！
怎样避免这种情况？
一般我们数据库里面数据都不会直接删除，数据时很宝贵的，不舍得删除，另一方便能提高查询数据
先利用索引查询出来数据，再进行联合查询不就行了
select c.id,c.content from smth1 c inner join ( select id from smth1 where id > 1000 limit 10) as t on c.id = t.id ;select c.id,c.content from smth1 c inner join ( select id from smth1 where id > 5000 limit 10) as t on c.id = t.id ;select c.id,c.content from smth1 c inner join ( select id from smth1 where id > 9000 limit 10) as t on c.id = t.id ;select c.id,c.content from smth1 c inner join ( select id from smth1 where id > 10000 limit 10) as t on c.id = t.id ;select c.id,c.content from smth1 c inner join ( select id from smth1 where id > 20000 limit 10) as t on c.id = t.id ;进行执行计划分析，没有一个大于1s的11 | 0.04538625 | select c.id,c.content from smth1 c inner join( select id from smth1 where id > 5000 limit 10) as t on c.id = t.id || 12 | 0.023278 | select c.id,c.content from smth1 c inner join( select id from smth1 where id > 9000 limit 10) as t on c.id = t.id || 13 | 0.02320425 | select c.id,c.content from smth1 c inner join( select id from smth1 where id > 10000 limit 10) as t on c.id = t.id || 14 | 0.001938 | select c.id,c.content from smth1 c inner join( select id from smth1 where id > 20000 limit 10) as t on c.id = t.id |
此外，还会想到用in来查询而不是子查询，为什么不用in，使用in会先查询出来一条id，之后再去和下面进行匹配，会进行smth1进行全表扫描！