mysql的group by語法可以依照指定的規則將資料分組，分組就是將一個資料集分割成若干個小區域，然後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，實作群組內排序的方法。
相關mysql影片教學推薦：《mysql教學》
mysql的group by語法可以將資料分組，但是分組後的資料並不能進行組內排序。
例如一個評論表有多個使用者評論，需要取得每個使用者最後評論的內容。
建立測試資料表及資料
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，每位使用者最後評論的內容就是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 | 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 | +----+---------+---------+---------------------+---------------------+
可以看到結果，分組後只會傳回分組內的第一個資料。因為group by語法沒有進行群組內排序的功能，只會依mysql預設的排序顯示。
如何才能對group by分組內的資料進行排序了，這個需要依照不同的需求處理。
1.id最大的，評論時間肯定最新
這種情況我們可以用id代替時間去搜尋並組內排序，使用max(id)就可以取得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 | 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與評論時間沒有關係，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，但因為不同使用者的評論時間有可能相同，因此也需要加多user_id這個條件去查詢。
重新建立測試資料
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 | 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可以減少外層的資料集。
where user_id is not null 可以讓group by user_id時使用索引。
這篇文章解釋了mysql group by 群組內排序的方法 ，更多相關內容請關注php中文網。
相關推薦：
#詳解在mysql查詢時，offset過大影響效能的原因與最佳化方法
##
以上是講解mysql group by 群組內排序的方法的詳細內容。更多資訊請關注PHP中文網其他相關文章！
