按类似列名对组查询结果进行分组

Question

PostgreSQL version 8.0.4

我设法查询数据库以获取此表，我们称之为A：

从本质上讲，这告诉我获得某项成就最多的前3个等级。我将row_number()分区为achievement_id，如图所示。

我对如何查询这一点感到困惑，因此我得到了一个类似于此的表：

区别在于first_max对应row_number = 1，second_max对应row_number = 2，third_max对应row_number = 3。

我用来获取第二个表的查询：

SELECT
  achievement_id,
  level_id AS first_max,
  level_id AS second_max,
  level_id AS third_max
FROM A
WHERE row_number = 1 --Obviously this line is wrong but not sure how to approach

Answer 1

使用aggregate expressions with FILTER:

SELECT
    achievement_id,
    sum(level_id) FILTER (WHERE row_number = 1) AS first_max,
    sum(level_id) FILTER (WHERE row_number = 2) AS second_max,
    sum(level_id) FILTER (WHERE row_number = 3) AS third_max
FROM A
GROUP BY 1

SQLFiddle

Answer 2

尝试：

select 
achievement_id,
max(first_max),
max(second_max),
max(third_max)
from (
SELECT
  achievement_id,
  case when row_number = 1 then level_id end AS first_max,
  case when row_number = 2 then level_id end AS second_max,
  case when row_number = 3 then level_id end AS third_max
FROM A
) B
group by achievement_id