Sql 如何计算同一订单id的订单数量和退货数量？

我试图从列为

CustomerID

，

OrderID

，

ReturnFlag

的表中派生一个CustomerID、OrderCount和ReturnCount表

我尝试了下面的代码，它给了我两个不同的表，但我希望结果在一个输出表中

select "CustomerID", count(DISTINCT "OrderID") as OrderCount  
from "Customer" 
group by "CustomerID";

select "CustomerID", count(DISTINCT "OrderID") as ReturnCount 
from "Customer" 
where "ReturnFlag" = 1 
group by "CustomerID";

源表-

|CustomerID| |OrderID| |ReturnFlag|
|1     | |  A    | |    0 |
|2     | |   B   | |        1 |
|1     | |  C    | |    0 |
|2     | |   D   | |    0 |
|3     | |   E   | |    1 |

预期产出表-

|CustomerID| |OrderCount| |ReturnCount|
|1     | |  2      | |  0     |
|2     | |  2      | |        1  |
|3     | |  1      | |  1     |

---

使用条件聚合：

select "CustomerID",
        count(DISTINCT case when "ReturnFlag" = 1 THEN "OrderID" end) as ReturnCount 
        count(DISTINCT "OrderID") as OrderCount 
from "Customer" 
group by "CustomerID";

---

我有点困惑为什么

CustomerID

在名为

Customer

的表中不是唯一的。我希望此查询在

Orders

表上运行，而不是在customers表上运行。

使用

case

表达式进行条件聚合：

select "CustomerID",
        count(DISTINCT case when "ReturnFlag" = 1 THEN "OrderID" end) as ReturnCount 
        count(DISTINCT "OrderID") as OrderCount 
from "Customer" 
group by "CustomerID";

---

您没有指定DBMS，因此这是ANSI SQL

您可以为此使用条件聚合：

select "CustomerID", 
       count(DISTINCT "OrderID") as OrderCount, 
       count(DISTINCT "OrderID") filter (where "ReturnFlag" = 1) as return_count
from "Customer" 
group by "CustomerID";

如果DBMS不支持标准的

filter（）

子句，则可以使用大小写表达式：

select "CustomerID", 
       count(DISTINCT "OrderID") as OrderCount, 
       count(DISTINCT case when "ReturnFlag" = 1 then "OrderID" end) as return_count
from "Customer" 
group by "CustomerID";