有时候我们想要获取每组的前n个记录，由于mysql中没有row_number() over 函数，之前部门大佬写了这个方法，觉得很实用，这里展示给大家。
student表：
s_id为学生id，s_name为学生姓名，s_birth为出生年月，s_sex为性别
score表：
s_id 为学生id，c_id为课程id，s_score为对应的成绩。
mysql> select * from student;
+------+--------+------------+-------+
| s_id | s_name | s_birth | s_sex |
+------+--------+------------+-------+
| 01 | 赵雷 | 1990-01-01 | 男 |
| 02 | 钱电 | 1990-12-21 | 男 |
| 03 | 孙风 | 1990-05-20 | 男 |
| 04 | 李云 | 1990-08-06 | 男 |
| 05 | 周梅 | 1991-12-01 | 女 |
| 06 | 吴兰 | 1992-03-01 | 女 |
| 07 | 郑竹 | 1989-07-01 | 女 |
| 08 | 王菊 | NULL | 女 |
+------+--------+------------+-------+
8 rows in set (0.00 sec)
mysql> select * from score;
+------+------+---------+
| s_id | c_id | s_score |
+------+------+---------+
| 01 | 01 | 80 |
| 01 | 02 | 90 |
| 01 | 03 | 99 |
| 02 | 01 | 70 |
| 02 | 02 | 60 |
| 02 | 03 | 80 |
| 03 | 01 | 80 |
| 03 | 02 | 80 |
| 03 | 03 | 80 |
| 04 | 01 | 50 |
| 04 | 02 | 30 |
| 04 | 03 | 20 |
| 05 | 01 | 76 |
| 05 | 02 | 87 |
| 06 | 01 | 31 |
| 06 | 03 | 34 |
| 07 | 02 | 89 |
| 07 | 03 | 98 |
+------+------+---------+
18 rows in set (0.00 sec)
现想要取出每个课程前3名的学生信息、课程id，成绩与对应课程内排名，如下所示：
+------+--------+------------+-------+------+---------+------+
| s_id | s_name | s_birth | s_sex | c_id | s_score | rank |
+------+--------+------------+-------+------+---------+------+
| 01 | 赵雷 | 1990-01-01 | 男 | 01 | 80 | 1 |
| 03 | 孙风 | 1990-05-20 | 男 | 01 | 80 | 2 |
| 05 | 周梅 | 1991-12-01 | 女 | 01 | 76 | 3 |
| 01 | 赵雷 | 1990-01-01 | 男 | 02 | 90 | 1 |
| 07 | 郑竹 | 1989-07-01 | 女 | 02 | 89 | 2 |
| 05 | 周梅 | 1991-12-01 | 女 | 02 | 87 | 3 |
| 01 | 赵雷 | 1990-01-01 | 男 | 03 | 99 | 1 |
| 07 | 郑竹 | 1989-07-01 | 女 | 03 | 98 | 2 |
| 02 | 钱电 | 1990-12-21 | 男 | 03 | 80 | 3 |
+------+--------+------------+-------+------+---------+------+
查询实现如下：
mysql> set @rank:=0;
Query OK, 0 rows affected (0.00 sec)
mysql> select a.*,b.c_id,b.s_score,b.rank from(
-> select *,@rank:=case when @current_id<>c_id then 1 else @rank+1 end as rank,@current_id:=c_id from score order by c_id,s_score desc)b
-> left join student a on a.s_id=b.s_id
-> having rank<=3;
首先看表b的部分,令@current_id=c_id，当c_id不是当前的课程时，rank重新从1开始计数，某则在在当前rank上加1。
需要注意的是，此时的排序需要现基于课程id，再基于成绩逆序。
mysql> select *,@rank:=case when @current_id<>c_id then 1 else @rank+1 end as rank,
-> @current_id:=c_id from score order by c_id,s_score desc;
+------+------+---------+------+-------------------+
| s_id | c_id | s_score | rank | @current_id:=c_id |
+------+------+---------+------+-------------------+
| 01 | 01 | 80 | 1 | 01 |
| 03 | 01 | 80 | 2 | 01 |
| 05 | 01 | 76 | 3 | 01 |
| 02 | 01 | 70 | 4 | 01 |
| 04 | 01 | 50 | 5 | 01 |
| 06 | 01 | 31 | 6 | 01 |
| 01 | 02 | 90 | 1 | 02 |
| 07 | 02 | 89 | 2 | 02 |
| 05 | 02 | 87 | 3 | 02 |
| 03 | 02 | 80 | 4 | 02 |
| 02 | 02 | 60 | 5 | 02 |
| 04 | 02 | 30 | 6 | 02 |
| 01 | 03 | 99 | 1 | 03 |
| 07 | 03 | 98 | 2 | 03 |
| 02 | 03 | 80 | 3 | 03 |
| 03 | 03 | 80 | 4 | 03 |
| 06 | 03 | 34 | 5 | 03 |
| 04 | 03 | 20 | 6 | 03 |
+------+------+---------+------+-------------------+
18 rows in set (0.00 sec)
在此基础上基于s_id连接student表即可，另外在最后的条件设定中需要用having不能用where，因为在原表中是不存在rank字段的，这是我们为了取数所构造的。
