使用 SQL 中的多个列在 GROUP BY 中返回 0 个计数

Question

我有一个 MySQL 脚本，它按城市返回每周的新用户数。 问题是它忽略了一周内没有用户的任何城市。 当那周没有用户时，我需要它返回 0。

select 
  FROM_DAYS(TO_DAYS([u.created_at:pst]) - MOD(TO_DAYS([u.created_at:pst]) -1, 7)) as week,
  a.city,
  count(u.id) as this_week_signups_to_date
from users u
left join addresses a on a.user_id = u.id
group by 
  week, city
order by 
  week DESC

任何想法我怎么能做到这一点？

Answer 1

也许这不是最流畅的解决方案，但您可以使用这样的交叉连接操作：

select week,
        case when joined.city = all_cities.city then this_week_signups_to_date else 0 end 
         as this_week_signups_to_date,
        all_cities.city
from
    (select 
           week([u.created_at:pst]) as week, --you could use week() in mysql or datepart() in t-sql
           a.city,
           count(u.id) as this_week_signups_to_date
     from users u
     left join addresses a on a.user_id = u.id --left/right join depends on your needs
     group by week, city) joined
cross join 
    (select distinct city from addresses) all_cities
order by week desc