temp1: 2499条数据
temp6: 969596条数据
//注意时间单位ms和s

//其中temp1和temp2已经给eid加上索引

1.in和 exists

//外表大于子表的时候，使用in
//外表小于字表的时候，使用exists
select * from temp1 where eid in (select eid from temp6)//1.92s
select * from temp1 where exists (select eid from temp6 where eid = temp1.eid)//66.48ms

select * from temp6 where eid in (select eid from temp1)//98.25ms
select * from temp6 where exists (select eid from temp1 where eid = temp6.eid)//2.40s

2.limit

//offset 越大，执行时间越长。
select * from temp6 limit 900000,100//1.25s
select * from temp6 limit 1,100//7.12ms

//下面两句并没有网上说的效果
select * from temp6 where eid >= (select eid from temp6 order by eid limit 800000,1) limit 100//515.14ms
select * from temp6 limit 800000,100//474.13ms

3.union 和 union all

//在不考虑重复数据的时候，union all 比union效率高。
//union all 两张表的时候，应该将大表放在左边
select * from temp6 union select * from temp1//187.17s
select * from temp1 union select * from temp6//190.30s

select * from temp6 union all select * from temp1//8.39s
select * from temp1 union all select * from temp6//42.03s

4.join

//join两张表的时候，应该把小表放在左边 select * from temp1 a join temp6 b on a.eid = b.eid//54.21ms select * from temp6 a join temp1 b on a.eid = b.eid//1.12s