在 SQl 中按 GROUP 限制 (postgresql)

Question

下面是我的 table 学生（原始有 100K+ 行，只显示一组）：

id	name	class	marks
1	Ryan	5	8
2	Max	5	7
3	Max1	5	10
4	Ryan1	6	8
5	Max2	6	10
6	Ryan2	6	7
7	Ryan3	7	8
8	Max3	7	10
9	Ryan4	7	7

我想为每个 class（5 和 7）获取两行标记 <= 10，也按 class 排序，标记 ASC

因此，预期结果将是：-

id	name	class	marks
1	Ryan	5	8
3	Max1	5	10
7	Ryan3	7	8
8	Max3	7	10

为了执行下面我试过：-

SELECT DISTINCT t_out.class, t_top.marks, t_top.name 
FROM students t_out 
  JOIN LATERAL (
    select * 
    from students t_in 
    where t_in.class = t_out.class 
    ORDER BY t_in.id ASC
  ) t_top ON TRUE 
WHERE t_top.marks <= 10 
 AND (t_out.class = 5 OR t_out.class = 7) 
ORDER BY t_top.marks DESC LIMIT 2;

Result on original database:- it's loading since long time

Result on sample :- Error: near line 20: near "(": syntax error

Answer 1

你会使用 row_number():

select s.*
from (select s.*,
             row_number() over (partition by class order by marks desc) as seqnum
      from students s
      where marks < 10 and class in (5, 7)
     ) s
where seqnum <= 2
order by class, marks;

注意：您的问题有点令人费解。您似乎想要两行，每个 class 的最高标记按标记降序排列。

编辑：

根据您的评论：

select s.*
from (select s.*,
             row_number() over (partition by class order by marks desc) as seqnum,
             count(*) over (partition by class) as cnt
      from students s
      where marks < 10 
     ) s
where seqnum <= 2 and cnt >= 2
order by class, marks;

在 SQl 中按 GROUP 限制 (postgresql)

LIMIT by GROUP in SQl (postgresql)

sql

postgresql

group-by

sql-order-by