最后X个订单的平均值

Question

我需要查找每个客户最近X个订单的平均金额。我的数据的结构为这样：

• 客户ID
• 总额
• 日期

我试图通过partitionning客户ID，然后按日期排序，但我不能求其平均值顶十的

Answer 1

另一种选择是一个CTE和ROW_NUMBER（）。以下将给你客户的平均订单。

Declare @YourTable table (CustID int,OrderDate Date,OrderAmount int) 
Insert Into @YourTable values 
(1,'2016-01-18',2500), 
(1,'2016-02-13',5000), 
(1,'2016-03-31',3000), 
(2,'2016-03-18',1800), 
(2,'2016-04-13',2200), 
(2,'2016-05-31',2500) 

;with cteBase as (
    Select * 
      ,RowNr=Row_Number() over (Partition By CustID Order By OrderDate Desc) 
    From @YourTable 
) 
Select CustID 
     ,AvgOrder = avg(OrderAmount) 
From cteBase 
Where RowNr<=2 
Group By CustID 

返回

CustID AvgOrder 
1  4000 
2  2350 

Answer 2

使用ROW_NUMBER窗函数

select [Customer ID],Avg([Total Amount]) as avg_amt 
(
select row_number()over(partition by [Customer ID] order by [Date] desc) rn, * 
from yourtable 
) A 
Where Rn <= 5 --change it based on requirement 
Group by [Customer ID] 

貌似当前表是事务表，所以你可能有一个单独的表称为Customer其中[Customer ID]是独一无二的，那么你可以用这个方法也

SELECT c.[Customer ID], 
     Avg([Total Amount]) 
FROM customers c 
     CROSS apply (SELECT TOP 5 [Total Amount] 
        FROM yourtable y 
        WHERE c.[Customer ID] = y.[Customer ID] 
        ORDER BY [Date] DESC) cs 
GROUP BY [Customer ID] 

上面的查询可以用OUTER APPLY被改变，以获得列表所有customers即使他没有与平均Total Amount零交易。

Answer 3

一种方法使用row_number()：

select customerid, avg(totalamount) 
from (select t.*, 
      row_number() over (partition by customerid order by date desc) as seqnum 
     from t 
    ) t 
where seqnum <= x;