关于分组 2 列的 SQL 选择查询

Question

我正在努力解决这个问题 - 看起来很简单

下表记录了哪个用户观看了哪个视频

topic_user_id |主题ID |用户身份 1 | 10 | 3 2 | 10 | 4 3 | 11 | 5 4 | 11 | 3 5 | 12 | 6 6 | 13 | 6 7 | 14 | 7 8 | 11 | 8 9 | 11 | 9 10 | 14 | 10 11 | 15 | 11 12 | 16 | 11 13 | 17 | 11

现在要了解有多少用户观看了特定视频 - 我有以下查询。

有多少用户观看了特定视频

从 topic_user 中选择 count(distinct(user_id)) 作为“用户数”，topic_id 其中 user_id 不为空 按 topic_id 分组

输出

用户数量 | topic_id 2 | 10 4 | 11 1 | 12 1 | 13 2 | 14 1 | 15 1 | 16 1 | 17 读作：2 位用户观看了主题 10 ，4 位用户观看了主题 11 等等

这很好用 - 但我正在寻找的是找到：

有多少用户观看了 1 个视频
有多少用户观看了 2 个视频
有多少用户观看了 3 个视频

输出应该是这样的

用户数量 |观看的视频数 6 | 1 2 | 2 1 | 3 阅读为 - 6 人仅观看了 1 个视频，2 人观看了 2 个视频，依此类推。

需要一些帮助。

提前致谢

Answer 1

可能有更简单的方法，但子查询可以工作

select
    videos as 'Number of videos',
    count(user_id) as 'Num of Users'
from (
   select
        count(distinct(topic_id)) as videos,
        user_id from topic_user
   group by
        user_id
 ) sub
 group by
    videos

Answer 2

select u.n as `Number of Video watched`, count(t.user_id)
from
    (select user_id, count(*) as `cnt`
    from topic_user 
    group by user_id) t INNER JOIN(
        SELECT 1 as `n` FROM DUAL 
        UNION 
        SELECT 2 as `n` FROM DUAL 
        UNION
        SELECT 3 as `n` FROM DUAL) u ON u.n = t.cnt
group by u.n

小提琴：http://sqlfiddle.com/#!2/844ee/15

Answer 3

试试这个::

Select first_table.count(distinct(user_id))  as 'Number of Users', SUM(Count(topic_id))
from
(select count(distinct(user_id)) as 'Number of Users',topic_id from topic_user
where user_id is not null
group by topic_id) as first_table group by Count(topic_id)

Answer 4

试试这个

SELECT videos AS `Number of videos`,COUNT(user_id) AS `Num of Users`
FROM (SELECT COUNT(DISTINCT(topic_id)) AS videos,user_id FROM topic_user
GROUP BY user_id) sub_query
GROUP BY
videos

Answer 5

“有多少用户观看了 n 个视频”可以解释为两种方式：

1) 无论标题如何，有多少用户观看了 n 个视频？同一个视频标题可能被观看了两次。

select videoCount, count(*) as userCount
  from (
    select count(*) as videoCount
      from topic_user
    group by user_id
  ) t
 group by videoCount

2) 有多少用户观看了 n 个不同的视频标题。

select videoCount, count(*) as userCount
  from (
    select count(distinct topic_id) as videoCount
      from topic_user
    group by user_id
  ) t
 group by videoCount