Question

我有两个没有任何关系的不同表，但是我需要对两者的行数求和，并使用日期进行分组。我需要这种方式，因为我需要在元数据库中绘制图表。

我需要用它来总结用户每天观看的视频广告的数量，不幸的是，这是划分的。

我尝试了此SQL并几乎成功了：

select *, total_video_prize + total_video_stock
from (select last_video_watched, count(*) as total_video_prize from video_ads_user_historic group by DAY(last_video_watched)) as video_prize,
(select date_when_transacted, count(*) as total_video_stock from stock_historic where has_seen_video_ad = true group by DAY(date_when_transacted)) as video_stock

问题在于日期未合并。我需要last_video_watched与date_when_transacted相同，并按天分组。可以这样做吗？

Answer 1

您可以使用UNION获取日期，然后与子查询一起进行计数

select  t1.my_day , t2.total_video_prize , t3.total_video_stock
from (
    select DAY(date_when_transacted) as my_day
    from  stock_historic 
    union 
    select DAY(last_video_watched)
    from video_ads_user_historic 
) t1 
left join  (

    select DAY(last_video_watched) my_day, count(*) as total_video_prize 
    from video_ads_user_historic 
    group by DAY(last_video_watched)

) t2 on t2.myday = t1.my_day

left join  (
    selectDAY(date_when_transacted) my_day, count(*) as total_video_stock 
    from stock_historic 
    where has_seen_video_ad = true 
    group by DAY(date_when_transacted)

) t3 on t3.my_day = t1.my_day

Answer 2

无论两个表中的哪个表，每个记录都代表观看的电影。因此，将表的行与UNION ALL粘合在一起并计数：

select dt, count(*)
from
(
  select last_video_watched as dt from video_ads_user_historic
  union all
  select date_when_transacted as dt from stock_historic where has_seen_video_ad = true
) unioned
group by dt
order by dt;

如果您的日期列包含时间部分，请使用DATE(column)而不是仅列。

Answer 3

我认为最简单的解决方案就是a <- matrix(rep(1:2, 30), 2, 30) a <- paste0("p", 1:30, ":")和union all：

group by

不需要加入。

如何在两个子查询中使用group by并对行计数求和？

3 个答案: