mysql group by 组内排序方法

mysql的group by语法可以根据指定的规则对数据进行分组，分组就是将一个数据集划分成若干个小区域，然后再针对若干个小区域进行数据处理。本文将介绍mysql使用group by分组时，实现组内排序的方法。

mysql的group by语法可以对数据进行分组，但是分组后的数据并不能进行组内排序。


例如一个评论表有多个用户评论，需要获取每个用户最后评论的内容。

创建测试数据表及数据

CREATE TABLE `comment` (
  `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `user_id` int(10) unsigned NOT NULL,
  `content` varchar(200) NOT NULL,
  `addtime` datetime NOT NULL,
  `lastmodify` datetime NOT NULL,
  PRIMARY KEY (`id`),
  KEY `user_id` (`user_id`),
  KEY `addtime` (`addtime`),
  KEY `uid_addtime` (`user_id`,`addtime`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

INSERT INTO `comment` (`id`, `user_id`, `content`, `addtime`, `lastmodify`) VALUES
(1, 1, '评论1', '2017-05-17 00:00:00', '2017-05-17 00:00:00'),
(2, 1, '评论2', '2017-05-17 00:00:01', '2017-05-17 00:00:01'),
(3, 2, '评论1', '2017-05-17 00:00:02', '2017-05-17 00:00:02'),
(4, 2, '评论2', '2017-05-17 00:00:03', '2017-05-17 00:00:03'),
(5, 3, '评论1', '2017-05-17 00:00:04', '2017-05-17 00:00:04'),
(6, 1, '评论3', '2017-05-17 00:00:05', '2017-05-17 00:00:05'),
(7, 4, '评论1', '2017-05-17 00:00:06', '2017-05-17 00:00:06'),
(8, 4, '评论2', '2017-05-17 00:00:07', '2017-05-17 00:00:07'),
(9, 4, '评论3', '2017-05-17 00:00:08', '2017-05-17 00:00:08'),
(10, 4, '评论4', '2017-05-17 00:00:09', '2017-05-17 00:00:09'),
(11, 3, '评论2', '2017-05-17 00:00:10', '2017-05-17 00:00:10');

select * from comment;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime             | lastmodify          |
+----+---------+---------+---------------------+---------------------+
|  1 |       1 | 评论1   | 2017-05-17 00:00:00 | 2017-05-17 00:00:00 |
|  2 |       1 | 评论2   | 2017-05-17 00:00:01 | 2017-05-17 00:00:01 |
|  3 |       2 | 评论1   | 2017-05-17 00:00:02 | 2017-05-17 00:00:02 |
|  4 |       2 | 评论2   | 2017-05-17 00:00:03 | 2017-05-17 00:00:03 |
|  5 |       3 | 评论1   | 2017-05-17 00:00:04 | 2017-05-17 00:00:04 |
|  6 |       1 | 评论3   | 2017-05-17 00:00:05 | 2017-05-17 00:00:05 |
|  7 |       4 | 评论1   | 2017-05-17 00:00:06 | 2017-05-17 00:00:06 |
|  8 |       4 | 评论2   | 2017-05-17 00:00:07 | 2017-05-17 00:00:07 |
|  9 |       4 | 评论3   | 2017-05-17 00:00:08 | 2017-05-17 00:00:08 |
| 10 |       4 | 评论4   | 2017-05-17 00:00:09 | 2017-05-17 00:00:09 |
| 11 |       3 | 评论2   | 2017-05-17 00:00:10 | 2017-05-17 00:00:10 |
+----+---------+---------+---------------------+---------------------+

在comment表中，每个用户最后评论的内容就是id为6,4,11,10的记录。

使用group by查询

select * from comment group by user_id;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime             | lastmodify          |
+----+---------+---------+---------------------+---------------------+
|  1 |       1 | 评论1   | 2017-05-17 00:00:00 | 2017-05-17 00:00:00 |
|  3 |       2 | 评论1   | 2017-05-17 00:00:02 | 2017-05-17 00:00:02 |
|  5 |       3 | 评论1   | 2017-05-17 00:00:04 | 2017-05-17 00:00:04 |
|  7 |       4 | 评论1   | 2017-05-17 00:00:06 | 2017-05-17 00:00:06 |
+----+---------+---------+---------------------+---------------------+

可以看到结果，分组后只会返回分组内的第一条数据。因为group by语法没有进行组内排序的功能，只会按mysql默认的排序显示。


如何才能对group by分组内的数据进行排序了，这个需要根据不同的需求处理。


1.id最大的，评论时间肯定最新

这种情况我们可以使用id代替时间去搜寻并组内排序，使用max(id)就可以获取到每个分组中最大的评论id(即最新的评论)

select * from comment where id in(select max(id) from comment group by user_id) order by user_id;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime             | lastmodify          |
+----+---------+---------+---------------------+---------------------+
|  6 |       1 | 评论3   | 2017-05-17 00:00:05 | 2017-05-17 00:00:05 |
|  4 |       2 | 评论2   | 2017-05-17 00:00:03 | 2017-05-17 00:00:03 |
| 11 |       3 | 评论2   | 2017-05-17 00:00:10 | 2017-05-17 00:00:10 |
| 10 |       4 | 评论4   | 2017-05-17 00:00:09 | 2017-05-17 00:00:09 |
+----+---------+---------+---------------------+---------------------+

2.id与评论时间没有关系，id大的评论时间可能不是最新

这种情况我们就需要使用max(addtime)来获取最新的评论，但因为不同用户的评论时间有可能相同，因此还需要加多user_id这个条件去查询。

重新创建测试数据

truncate table comment;

INSERT INTO `comment` (`id`, `user_id`, `content`, `addtime`, `lastmodify`) VALUES
(1, 1, '评论1', '2017-05-17 00:00:00', '2017-05-17 00:00:00'),
(2, 1, '评论2', '2017-05-17 00:10:01', '2017-05-17 00:10:01'),
(3, 2, '评论1', '2017-05-17 00:10:02', '2017-05-17 00:10:02'),
(4, 2, '评论2', '2017-05-17 00:00:03', '2017-05-17 00:00:03'),
(5, 3, '评论1', '2017-05-17 00:10:04', '2017-05-17 00:10:04'),
(6, 1, '评论3', '2017-05-17 00:00:05', '2017-05-17 00:00:05'),
(7, 4, '评论1', '2017-05-17 00:00:06', '2017-05-17 00:00:06'),
(8, 4, '评论2', '2017-05-17 00:10:07', '2017-05-17 00:10:07'),
(9, 4, '评论3', '2017-05-17 00:00:08', '2017-05-17 00:00:08'),
(10, 4, '评论4', '2017-05-17 00:00:09', '2017-05-17 00:00:09'),
(11, 3, '评论2', '2017-05-17 00:00:10', '2017-05-17 00:00:10');

select * from comment;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime             | lastmodify          |
+----+---------+---------+---------------------+---------------------+
|  1 |       1 | 评论1   | 2017-05-17 00:00:00 | 2017-05-17 00:00:00 |
|  2 |       1 | 评论2   | 2017-05-17 00:10:01 | 2017-05-17 00:10:01 |
|  3 |       2 | 评论1   | 2017-05-17 00:10:02 | 2017-05-17 00:10:02 |
|  4 |       2 | 评论2   | 2017-05-17 00:00:03 | 2017-05-17 00:00:03 |
|  5 |       3 | 评论1   | 2017-05-17 00:10:04 | 2017-05-17 00:10:04 |
|  6 |       1 | 评论3   | 2017-05-17 00:00:05 | 2017-05-17 00:00:05 |
|  7 |       4 | 评论1   | 2017-05-17 00:00:06 | 2017-05-17 00:00:06 |
|  8 |       4 | 评论2   | 2017-05-17 00:10:07 | 2017-05-17 00:10:07 |
|  9 |       4 | 评论3   | 2017-05-17 00:00:08 | 2017-05-17 00:00:08 |
| 10 |       4 | 评论4   | 2017-05-17 00:00:09 | 2017-05-17 00:00:09 |
| 11 |       3 | 评论2   | 2017-05-17 00:00:10 | 2017-05-17 00:00:10 |
+----+---------+---------+---------------------+---------------------+

符合条件的应该是id为2,3,5,8的记录

select a.* from comment as a right join 
(select user_id, max(addtime) as maxtime from comment where user_id is not null group by user_id) as b 
on a.user_id=b.user_id and a.addtime=b.maxtime order by a.user_id asc;

+------+---------+---------+---------------------+---------------------+
| id   | user_id | content | addtime             | lastmodify          |
+------+---------+---------+---------------------+---------------------+
|    2 |       1 | 评论2   | 2017-05-17 00:10:01 | 2017-05-17 00:10:01 |
|    3 |       2 | 评论1   | 2017-05-17 00:10:02 | 2017-05-17 00:10:02 |
|    5 |       3 | 评论1   | 2017-05-17 00:10:04 | 2017-05-17 00:10:04 |
|    8 |       4 | 评论2   | 2017-05-17 00:10:07 | 2017-05-17 00:10:07 |
+------+---------+---------+---------------------+---------------------+

使用right join可以减少外层的数据集。
where user_id is not null 可以使group by user_id时使用索引。