1、直接使用数据库提供的SQL语句 limit M ,N
SELECT * from user limit 0 , 10 ; -- 0.034s
SELECT * from user limit 100 , 10 ; -- 0.074s
SELECT * from user limit 1000 , 10 ; -- 0.087s
SELECT * from user limit 10000 , 10 ; -- 0.102s
SELECT * from user limit 100000 , 10 ; -- 0.236s
SELECT * from user limit 1000000 , 10 ; -- 1.684s
SELECT * from user limit 2000000 , 10 ; -- 2.971s
适用场景:可以看出,在数据量小时,使用limit M ,N 可以很快的查询出数据,但是当数据量达到十万级甚至百万级时,查询效率呈几何倍数递增;
该查询会导致全表扫描,速度会很慢,且有的数据库返回结果集不稳定(如某次返回1,2,3,另外一次返回2,1,3),limit限制的是从结果集的M位置取出N条输出,其余抛弃;
2、基于索引排序
SELECT * from user order by id asc limit 0 , 10 ; -- 0.075
SELECT * from user order by id asc limit 100 , 10 ; -- 0.084
SELECT * from user order by id asc limit 1000 , 10 ; -- 0.088
SELECT * from user order by id asc limit 10000 , 10 ; -- 0.099
SELECT * from user order by id asc limit 100000 , 10 ; -- 0.227
SELECT * from user order by id asc limit 1000000 , 10 ; -- 1.532
适用场景:数据量小时,当数据量达到十万到百万级时,查询效率与第一种相比,仍呈现几何倍数递增
在数据量小时,会根据索引进行选择,但是数据量大时,仍旧会导致全表扫描
3、基于所以筛选再排序
SELECT * from user order by id asc limit 10 ; -- 0.043
SELECT * from user where id > 100 order by id asc limit 10 ; -- 0.085
SELECT * from user where id > 1000 order by id asc limit 10 ; -- 0.086
SELECT * from user where id > 10000 order by id asc limit 10 ; -- 0.084
SELECT * from user where id > 100000 order by id asc limit 10 ; -- 0.083
SELECT * from user where id > 1000000 order by id asc limit 10 ; -- 0.083
SELECT * from user where id > 2000000 order by id asc limit 10 ; -- 0.085
SELECT * from user where id > 2000000 order by id desc limit 10 ; -- 0.085
可以看出,该查询效率稳定,适用于数据量多是情况,最后ORDER BY后的列对象是主键或者唯一索引,使得ORDER BY操作能利用索引被消除,但是结果集是稳定的