Question

我有以下表结构：

CREATE TABLE `client_notes` (
`id` bigint(20) NOT NULL,
`client_id` int(11) NOT NULL,
`user_id` int(11) NOT NULL,
`note` text CHARACTER SET utf8 COLLATE utf8_unicode_ci NOT NULL,
`datepost` datetime NOT NULL
) ENGINE=MyISAM

这里我存储了用户为客户插入的注释。每个client_id都有一些注释，由user_id添加。目标是计算经纪人添加第一个注释的次数（额外选项可以是本周的期间，例如在where子句中）

+----+----------+-----------------------------+
| id | broker_id| client | datepost           |
+----+---------+------------------------------+
|  1 |    1     |   10   | 29-11-2017 10:00:00| <- this is first note for this client
|  2 |    2     |   10   | 29-11-2017 10:20:00|
|  3 |    2     |   15   | 28-11-2017 10:10:00| <- this is first note for this client
|  4 |    1     |   15   | 28-11-2017 10:20:00|
|  5 |    1     |   15   | 28-11-2017 10:30:00|
+----+----------+--------+--------------------+

输出应为：

+-----------+-------+
| broker_id | count |
+-----------+-------+
|     1     |   1   |
|     2     |   1   |
+-----------+-------+

Answer 1

首先，不要在简单的SQL标识符上使用不必要的和非标准的引号：

-- Note: example data column names don't match above table definition
select min(datepost) as datepost, client_id from ENGINE group by client_id

首先获得＆＃34;＆＃34;任何意味着订单和最低限度。您希望每个客户的日期最短：

select broker_id, count(*) as 'count'
from ENGINE as E
join (
    select min(datepost) as datepost, client_id from ENGINE group by client_id
) as A
on E.client_id = A.client_id
and E.datepost = A.datepost
group by broker_id

任何SELECT的结果都是一个表，可以操作，

raw_input()

如何仅按用户计算第一个出现的组

1 个答案: