从一列中选择不同的值并对另一列求和

Question

我在这里显示了这个存储过程，它计算每个部门的订单数量、拒绝和拒绝百分比，但我有一个问题：为了对订单数量求和，我只需要考虑不同的订单号来对订单数量求和，但对于拒绝数量，应考虑所有行。如何做到这一点？

存储过程：

CREATE PROCEDURE orderqtyper
    @fromDate Date,
    @toDate Date
AS
    SELECT 
        Department, 
        SUM(Order_Qty) AS [Order Qty],
        SUM(Rejection_Qty) AS [Rejection Qty],
        FORMAT((SUM(Rejection_Qty) * 100.0 / NULLIF(SUM(Order_Qty), 0) / 100), 'P') AS Percentage   
    FROM 
        Semicon_NPD
    WHERE 
        (Date BETWEEN @fromDate AND @toDate)
    GROUP BY 
        Department
    ORDER BY 
        Percentage DESC

示例表：

当前结果：

预期结果（如果订单号相同，则订单数量应仅取唯一值之和，而拒绝数量应取所有值之和）：

Answer 1

您可以将DISTINCT 用作 SUM(DISTINCT )

create proc orderqtyper
@fromDate Date,
@toDate Date
as
SELECT Department, 
    SUM(DISTINCT Order_Qty)as [Order Qty],
    SUM(Rejection_Qty) as [Rejection Qty],
   Format((SUM(Rejection_Qty) * 100.0 / NULLIF(SUM(Order_Qty), 0)/100),'P')  AS Percentage   
FROM Semicon_NPD
Where (Date between  @fromDate  and  @toDate)
GROUP BY Department
order by Percentage desc
go