Question

我有一张与此类似的表格：

id   sml  desc amount
1     23   bla   100
2     23   bla   50
3     23   bla   30
4     23   bla   20
1     40   bla1  100
2     40   bla2  23
3     40   bla3  34
4     40   bla4  34

这是其他表的几个连接的结果。我想为每个amount值选择sml列中值最高的五名员工。

下面的示例输出:(我的示例中没有五个sml所以我只使用显示两个。另外，如果第5和第6个值相等，我会喜欢返回两者）

id   sml  desc amount
1     23   bla   100
2     23   bla   50
1     40   bla1  100
2     40   bla2  100
3     40   bla3  34

我尝试了以下查询：

select [MS_EmployeeNumber],[SA_Semel],[Description],sum([MS_Amount]) as Amount,
rank() over (Partition BY [SA_Semel], [Description]

select [MS_EmployeeNumber],[SA_Semel],[Description],sum([MS_Amount]) as Amount,
rank() over (Partition BY [SA_Semel], [Description]
                ORDER BY sum([MS_Amount]) Desc) AS Rank
from [dbo].[MonthlySalary] as sml
inner join [dbo].[SmlAnalyse] as SA on SA.SA_Semel = sml.MS_Semel
inner join [dbo].[SalarySMLlist] as list on list.Semel = SA.SA_Semel
group by [MS_EmployeeNumber],[SA_Semel],[Description]
order by SA_Semel, Amount desc

当我运行这个排名在所有表上完成他的工作并考虑我的要求在MS_Semel上排名。

Answer 1

我认为您的查询过于复杂。 rank（或者更确切地说dense_rank，根据您似乎处理重复的方式）已经为您提供了sml的排名，现在您只需要相应地进行过滤：

SELECT [id], [sml], [desc], [amount]
FROM   (SELECT [id], [sml], [desc], [amount], 
               DENSE_RANK() OVER (PARTITION BY [sml] ORDER BY [amount] DESC) rk
        FROM   [dbo].[MonthlySalary]) t
WHERE  rk <= 5

为每个类别选择前5名薪水

1 个答案: