我正在尝试编写一个查询(如果可能)以根据多个条件对ID进行排名。
我的桌子是这样的:
id group subgroup value
1 A Q 12
2 A Z 10
3 B Z 14
4 A Z 20
5 B W 20
我尝试了以下查询:
SELECT id,
CASE WHEN group = 'A' THEN ROW_NUMBER() OVER (PARTITION BY group ORDER BY SUM(value) DESC) AS rank_group
CASE WHEN group = 'A' AND subgroup = 'Z' THEN ROW_NUMBER() OVER (PARTITION BY group, subgroup ORDER BY SUM(value) DESC) AS rank_subgroup
FROM table
GROUP BY group, subgroup
但最终却是这样的:
id rank_group rank_subgroup
1 1 1
1 2 2
我想获取每个不同的id并根据case语句的条件返回排名,但是看起来像添加所需的分区会导致乘法,因为group by是必要的。我可以为每一列编写单独的查询,但如果可能的话,我想避免。
答案 0 :(得分:0)
您想要这样的东西吗?
select t.*,
dense_rank() over (order by sumg, group),
dense_rank() over (partition by group order by sumsg, subg),
from (select t.*,
sum(value) over (partition by group) as sumg,
sum(value) over (partition by group, subgroup) as sumsg
from t
) t;
这是我解释您可能想要的最好的猜测。