结果如下：

mysql> SELECT * FROM (SELECT cid,author,COUNT(*) AS number FROM test GROUP BY cid,author) a 
    -> WHERE  
    -> 3>(
    ->     SELECT COUNT(*) 
    ->     FROM (SELECT cid,author,COUNT(*) AS number FROM test GROUP BY cid,author) b
    ->     WHERE a.cid=b.cid AND a.number<b.number
    -> )ORDER BY cid,number DESC;
+------+----------------+--------+
| cid  | author         | number |
+------+----------------+--------+
|    1 | test3          |      4 |
|    1 | test2          |      3 |
|    1 | test1          |      2 |
|    2 | test33         |      4 |
|    2 | test22         |      3 |
|    2 | test11         |      2 |
|  200 | 200test_nagios |      1 |
+------+----------------+--------+
7 rows in set (0.00 sec)

N就是取分组之后的最前面几个判断，N=3就是取前3个

 -- 生成自动数字序列
SET @ROW=0;