如何计算 ID 的组/字段值的数量？答案

【问题标题】：How can I count the number of groups/field values for an ID?如何计算 ID 的组/字段值的数量？
【发布时间】：2018-01-29 23:40:19
【问题描述】：

我有一张详细信息表，我想计算具有 2 个或更多不同 action_taken 值的 user_id。

详情：

| user_id | action_taken          | action_date |
|---------|-----------------------|-------------|
| 1234    | clicked on a link     | 1/1/2017    |
| 1234    | went to the home page | 1/5/2017    |
| 1234    | clicked on a link     | 1/7/2017    |
| 1234    | clicked on a link     | 1/9/2017    |
| 1234    | changed password      | 1/11/2017   |
| 1234    | clicked on a link     | 1/13/2017   |
| 9876    | went to the home page | 2/1/2017    |
| 9876    | went to the home page | 2/5/2017    |
| 9876    | went to the home page | 2/7/2017    |
| 9876    | went to the home page | 2/9/2017    |
| 5566    | clicked on a link     | 1/1/2017    |
| 5566    | clicked on a link     | 1/5/2017    |
| 5566    | changed password      | 1/7/2017    |
| 5566    | clicked on a link     | 1/9/2017    |
| 4433    | went to the home page | 1/5/2017    |

期望的输出：

输出1：返回actions_taken不同值的个数。

| user_id | number_dift_action_taken_values |
|---------|---------------------------------|
| 1234    | 3                               |
| 4433    | 1                               |
| 5566    | 2                               |
| 9876    | 1                               |

输出 2：仅返回具有 >=2 个不同 action_taken 值的 user_id。

| user_id |
|---------|
| 1234    |
| 5566    |

这是我目前无法使用的：http://rextester.com/TUL87833。 HAVING 子句在详细信息中计算属于每个组的行数，不是 GROUP BY user_id, action_taken 子句指定的组数...

select
    user_id
    ,action_taken
    ,count(*)
from
    tbl
group by
    user_id
    ,action_taken
having count(*) >=2;

| user_id | action_taken          | count |
|---------|-----------------------|-------|
| 1234    | clicked on a link     | 4     |
| 5566    | clicked on a link     | 3     |
| 9876    | went to the home page | 4     |

【问题讨论】：

标签： sql postgresql

【解决方案1】：

输出 1

SELECT user_id,
       COUNT(DISTINCT action_taken) number_dift_action_taken_values
  FROM t_tab
 GROUP BY user_id

结果

user_id number_dift_action_taken_values
1234    3
4433    1
5566    2
9876    1

输出 2

SELECT user_id
  FROM t_tab
 GROUP BY user_id
HAVING COUNT(DISTINCT action_taken) >= 2

结果

user_id
1234
5566

【讨论】：

【解决方案2】：

对于每个输出，我将从以下内容开始 #:

输出 1：

select 
user_id
,count(distinct action_taken) as number_dift_action_taken_values
into #Output1_tbl
from details_tbl 
group by user_id

输出 2：

select *
from #Output1_tbl
where number_dift_action_taken_values >= 2

【讨论】：