mysql的group by語法可以根據指定的規則對數據進行分組，分組就是將一個數據集划分成若干個小區域，然后再針對若干個小區域進行數據處理。本文將介紹mysql使用group by分組時，實現組內排序的方法。
mysql的group by語法可以對數據進行分組，但是分組后的數據並不能進行組內排序。
例如一個評論表有多個用戶評論，需要獲取每個用戶最后評論的內容。
創建測試數據表及數據
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, '評論1', '2017-05-17 00:00:00', '2017-05-17 00:00:00'),
(2, 1, '評論2', '2017-05-17 00:00:01', '2017-05-17 00:00:01'),
(3, 2, '評論1', '2017-05-17 00:00:02', '2017-05-17 00:00:02'),
(4, 2, '評論2', '2017-05-17 00:00:03', '2017-05-17 00:00:03'),
(5, 3, '評論1', '2017-05-17 00:00:04', '2017-05-17 00:00:04'),
(6, 1, '評論3', '2017-05-17 00:00:05', '2017-05-17 00:00:05'),
(7, 4, '評論1', '2017-05-17 00:00:06', '2017-05-17 00:00:06'),
(8, 4, '評論2', '2017-05-17 00:00:07', '2017-05-17 00:00:07'),
(9, 4, '評論3', '2017-05-17 00:00:08', '2017-05-17 00:00:08'),
(10, 4, '評論4', '2017-05-17 00:00:09', '2017-05-17 00:00:09'),
(11, 3, '評論2', '2017-05-17 00:00:10', '2017-05-17 00:00:10');
select * from comment;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime | lastmodify |
+----+---------+---------+---------------------+---------------------+
| 1 | 1 | 評論1 | 2017-05-17 00:00:00 | 2017-05-17 00:00:00 |
| 2 | 1 | 評論2 | 2017-05-17 00:00:01 | 2017-05-17 00:00:01 |
| 3 | 2 | 評論1 | 2017-05-17 00:00:02 | 2017-05-17 00:00:02 |
| 4 | 2 | 評論2 | 2017-05-17 00:00:03 | 2017-05-17 00:00:03 |
| 5 | 3 | 評論1 | 2017-05-17 00:00:04 | 2017-05-17 00:00:04 |
| 6 | 1 | 評論3 | 2017-05-17 00:00:05 | 2017-05-17 00:00:05 |
| 7 | 4 | 評論1 | 2017-05-17 00:00:06 | 2017-05-17 00:00:06 |
| 8 | 4 | 評論2 | 2017-05-17 00:00:07 | 2017-05-17 00:00:07 |
| 9 | 4 | 評論3 | 2017-05-17 00:00:08 | 2017-05-17 00:00:08 |
| 10 | 4 | 評論4 | 2017-05-17 00:00:09 | 2017-05-17 00:00:09 |
| 11 | 3 | 評論2 | 2017-05-17 00:00:10 | 2017-05-17 00:00:10 |
+----+---------+---------+---------------------+---------------------+
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
在comment表中，每個用戶最后評論的內容就是id為6,4,11,10的記錄。
使用group by查詢
select * from comment group by user_id;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime | lastmodify |
+----+---------+---------+---------------------+---------------------+
| 1 | 1 | 評論1 | 2017-05-17 00:00:00 | 2017-05-17 00:00:00 |
| 3 | 2 | 評論1 | 2017-05-17 00:00:02 | 2017-05-17 00:00:02 |
| 5 | 3 | 評論1 | 2017-05-17 00:00:04 | 2017-05-17 00:00:04 |
| 7 | 4 | 評論1 | 2017-05-17 00:00:06 | 2017-05-17 00:00:06 |
+----+---------+---------+---------------------+---------------------+
1
2
3
4
5
6
7
8
9
可以看到結果，分組后只會返回分組內的第一條數據。因為group by語法沒有進行組內排序的功能，只會按mysql默認的排序顯示。
如何才能對group by分組內的數據進行排序了，這個需要根據不同的需求處理。
1.id最大的，評論時間肯定最新
這種情況我們可以使用id代替時間去搜尋並組內排序，使用max(id)就可以獲取到每個分組中最大的評論id(即最新的評論)
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 | 評論3 | 2017-05-17 00:00:05 | 2017-05-17 00:00:05 |
| 4 | 2 | 評論2 | 2017-05-17 00:00:03 | 2017-05-17 00:00:03 |
| 11 | 3 | 評論2 | 2017-05-17 00:00:10 | 2017-05-17 00:00:10 |
| 10 | 4 | 評論4 | 2017-05-17 00:00:09 | 2017-05-17 00:00:09 |
+----+---------+---------+---------------------+---------------------+
1
2
3
4
5
6
7
8
9
2.id與評論時間沒有關系，id大的評論時間可能不是最新
這種情況我們就需要使用max(addtime)來獲取最新的評論，但因為不同用戶的評論時間有可能相同，因此還需要加多user_id這個條件去查詢。
重新創建測試數據
truncate table comment;
INSERT INTO `comment` (`id`, `user_id`, `content`, `addtime`, `lastmodify`) VALUES
(1, 1, '評論1', '2017-05-17 00:00:00', '2017-05-17 00:00:00'),
(2, 1, '評論2', '2017-05-17 00:10:01', '2017-05-17 00:10:01'),
(3, 2, '評論1', '2017-05-17 00:10:02', '2017-05-17 00:10:02'),
(4, 2, '評論2', '2017-05-17 00:00:03', '2017-05-17 00:00:03'),
(5, 3, '評論1', '2017-05-17 00:10:04', '2017-05-17 00:10:04'),
(6, 1, '評論3', '2017-05-17 00:00:05', '2017-05-17 00:00:05'),
(7, 4, '評論1', '2017-05-17 00:00:06', '2017-05-17 00:00:06'),
(8, 4, '評論2', '2017-05-17 00:10:07', '2017-05-17 00:10:07'),
(9, 4, '評論3', '2017-05-17 00:00:08', '2017-05-17 00:00:08'),
(10, 4, '評論4', '2017-05-17 00:00:09', '2017-05-17 00:00:09'),
(11, 3, '評論2', '2017-05-17 00:00:10', '2017-05-17 00:00:10');
select * from comment;
+----+---------+---------+---------------------+---------------------+
| id | user_id | content | addtime | lastmodify |
+----+---------+---------+---------------------+---------------------+
| 1 | 1 | 評論1 | 2017-05-17 00:00:00 | 2017-05-17 00:00:00 |
| 2 | 1 | 評論2 | 2017-05-17 00:10:01 | 2017-05-17 00:10:01 |
| 3 | 2 | 評論1 | 2017-05-17 00:10:02 | 2017-05-17 00:10:02 |
| 4 | 2 | 評論2 | 2017-05-17 00:00:03 | 2017-05-17 00:00:03 |
| 5 | 3 | 評論1 | 2017-05-17 00:10:04 | 2017-05-17 00:10:04 |
| 6 | 1 | 評論3 | 2017-05-17 00:00:05 | 2017-05-17 00:00:05 |
| 7 | 4 | 評論1 | 2017-05-17 00:00:06 | 2017-05-17 00:00:06 |
| 8 | 4 | 評論2 | 2017-05-17 00:10:07 | 2017-05-17 00:10:07 |
| 9 | 4 | 評論3 | 2017-05-17 00:00:08 | 2017-05-17 00:00:08 |
| 10 | 4 | 評論4 | 2017-05-17 00:00:09 | 2017-05-17 00:00:09 |
| 11 | 3 | 評論2 | 2017-05-17 00:00:10 | 2017-05-17 00:00:10 |
+----+---------+---------+---------------------+---------------------+
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
符合條件的應該是id為2,3,5,8的記錄
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 | 評論2 | 2017-05-17 00:10:01 | 2017-05-17 00:10:01 |
| 3 | 2 | 評論1 | 2017-05-17 00:10:02 | 2017-05-17 00:10:02 |
| 5 | 3 | 評論1 | 2017-05-17 00:10:04 | 2017-05-17 00:10:04 |
| 8 | 4 | 評論2 | 2017-05-17 00:10:07 | 2017-05-17 00:10:07 |
+------+---------+---------+---------------------+---------------------+
1
2
3
4
5
6
7
8
9
10
11
12
使用right join可以減少外層的數據集。
where user_id is not null 可以使group by user_id時使用索引。
————————————————
版權聲明：本文為CSDN博主「傲雪星楓」的原創文章，遵循CC 4.0 BY-SA版權協議，轉載請附上原文出處鏈接及本聲明。
原文鏈接：https://blog.csdn.net/fdipzone/article/details/72453553
轉 https://blog.csdn.net/fdipzone/article/details/72453553