雪花 - Dense_rank 从 2 而不是 1 开始

Snowflake - Dense_rank starting at 2 rather than 1

我在查询中使用了以下语句:

 CASE WHEN apercent IS NULL THEN NULL ELSE dense_rank() over (partition by adate order by apercent desc) END as arank

出于某种原因,排名结果从 2 而不是 1 开始。没有绑定值,但有一些 NULL 值。我假设这是因为 NULL。有什么方法可以更新它以强制排名从 1 开始?

当前结果:

arank apercent
2 23.57
3 13.61
4 10.67
5 6.14
6 6.00
7 5.79
8 5.57
9 5.27
10 4.80
NULL NULL

期望的结果:

arank apercent
1 23.57
2 13.61
3 10.67
4 6.14
5 6.00
6 5.79
7 5.57
8 5.27
9 4.80
NULL NULL

一个简单的解决方案是使用 NULLS LAST:

(CASE WHEN apercent IS NOT NULL
      THEN dense_rank() over (partition by adate order by apercent desc nulls last)
 END) as arank

或者,您可以按 NULL 值进行分区,这样它们就不会包含在非 NULL 值的计数中:

(CASE WHEN apercent IS NOT NULL
      THEN dense_rank() over (partition by adate, apercent is null order by apercent desc)
 END) as arank

如果您的百分比不能为负数,您可以使用 coalesce

模仿 nulls last
case when apercent is null then null 
     else dense_rank() over (partition by adate order by coalesce(apercent,0.00) desc) 
end as arank