我有下一个数据表:
user group dt_action
----- ----- ---------
John salvage 2016-3-2
Dennis fire 2016-1-1
Martha fire 2016-12-23
John rescue 2016-1-20
John salvage 2017-1-26
Developer NULL 2016-5-6
Dennis several 2016-4-29
Martha fire 2003-1-1
我的想法是按类别" group"进行分组,计算每个用户的gorup数量,但是在datetime字段中,我想保持每个组的最小日期为sumarized。结果应该是这样的:
user group count_group dt_action_min
---- ----- ----------- -------------
John salvage 2 2016-3-2
John rescue 1 2016-1-20
Dennis fire 1 2016-1-1
Dennis several 1 2016-4-29
Martha fire 2 2003-1-1
Developer NULL 1 2016-5-6
这是一个很久以前偷了我梦想的问题,但我能搞清楚! 非常感谢,
达尼
答案 0 :(得分:1)
创建表/ INSERT数据
CREATE TABLE DATA
(`user` VARCHAR(9), `group` VARCHAR(7), `dt_action` VARCHAR(10))
;
INSERT INTO DATA
(`user`, `group`, `dt_action`)
VALUES
('John', 'salvage', '2016-3-2'),
('Dennis', 'fire', '2016-1-1'),
('Martha', 'fire', '2016-12-23'),
('John', 'rescue', '2016-1-20'),
('John', 'salvage', '2017-1-26'),
('Developer', NULL, '2016-5-6'),
('Dennis', 'several', '2016-4-29'),
('Martha', 'fire', '2003-1-1')
;
<强>查询强>
用您自己的表名替换数据
SELECT
`data`.`user`
, `data`.`group`
, `data_group`.`count_group`
, `data_group`.`min_dt_action`
FROM
`data`
INNER JOIN (
SELECT
`data`.user
, `data`.`group`
, COUNT(*) count_group
, MIN(`data`.`dt_action`) min_dt_action
FROM
`data`
GROUP BY
`data`.user
, `data`.`group`
) data_group
ON
`data`.`user` = data_group.`user`
AND
`data`.dt_action = data_group.min_dt_action
ORDER BY
`data_group`.`count_group` DESC
<强>结果强>
user group count_group min_dt_action
--------- ------- ----------- ---------------
Martha fire 2 2003-1-1
John salvage 2 2016-3-2
John rescue 1 2016-1-20
Developer (NULL) 1 2016-5-6
Dennis fire 1 2016-1-1
Dennis several 1 2016-4-29