sql 根据分组删除

Question

我有以下查询：

select user_id, trajectory_id, count(trajectory_id) from point group by user_id, trajectory_id

它returns结果如下：

而且我想删除 point table 中的所有记录，其中 count 在生成的 table 中小于 5。

现在我有以下查询，但它没有编译。它有什么问题，我该如何解决？

delete 
from point
where count(trajectory_id) < 5
GROUP BY user_id, trajectory_id

我正在使用 Postgres。

Answer 1

delete p1
from point p1
join
(
  select user_id, trajectory_id 
  from point 
  group by user_id, trajectory_id
  having count(trajectory_id) < 5
) p2 on p1.user_id = p2.user_id
    and p1.trajectory_id = p2.trajectory_id

Answer 2

我会用一个简单的相关子查询来做到这一点：

delete from point
    where (user_id, trajectory_id) in (select user_id, trajectory_id
                                       from point
                                       group by user_id, trajectory_id
                                       having count(*) < 5
                                      );

Answer 3

上面接受的答案有一些语法错误，下面应该可以解决问题，

delete 
from point p1
using
(
  select user_id, trajectory_id 
  from point 
  group by user_id, trajectory_id
  having count(trajectory_id) < 5
) p2 
where p1.user_id = p2.user_id
    and p1.trajectory_id = p2.trajectory_id

sql 根据分组删除

sql delete based on group by

sql

postgresql

sql-delete