Question

我有两个表费用和员工。我想获得每位员工的总费用，以及已退款的总费用。退款定义为1，未退款为0.

目前，我的查询返回以下内容，但应该为每位员工添加值并仅返回一行：

select  distinct(emp.employeeid) as EmpID, sum(ex.amount) as TotalExpenses,
(select SUM( ex.amount ) where ex.Refunded =1) as refunded,
(select SUM( ex.amount ) where ex.Refunded =0) as notrefunded
from Expenses ex 
left outer join  Employee emp
on ex.EmployeeID = emp.EmployeeID 
group by  emp.employeeid,ex.amount, ex.Refunded  

48  10.00   10.00   NULL
48  16.00   NULL    16.00
49  2.00    2.00    NULL
49  10.00   NULL    10.00

Answer 1

这是子查询的一种奇怪用法，使用distinct时不需要group by。我想这是你想要的查询：

select emp.employeeid as EmpID, sum(ex.amount) as TotalExpenses,
       sum(case when e.Refunded = 1 then ex.amount else 0 end) as refunded,
       sum(case when e.Refunded = 0 then ex.amount else 0 end) as notrefunded
from Employee emp left outer join
     Expenses ex 
     on ex.EmployeeID = emp.EmployeeID 
group by emp.employeeid;

注意：我改变了left outer join的顺序。您按员工ID进行分组。似乎不太可能为不存在的员工支付费用。无论如何，汇总是emp.employeeid，表明你真的希望匹配。

Answer 2

@Gordon Linoff的答案是理想的解决方案。但是，它并没有解决您在初始查询中犯的一些错误。

Distinct是select语句的子句，而不是函数。它始终适用于所有返回的字段。
使用相关子查询（即列列表中出现的子查询）时，需要将子查询连接到主查询并分别查询表。如上所述，您只能获取该行中找到的值sum。要正确地做到这一点，它需要是这样的：
```
(select SUM( ex2.amount ) 
   from Expenses ex2 
  where ex2.Refunded =1 
        and ex2.employeeid = e.ex2.employeeid) as refunded
```

带有SUM和分组的SQL查询CASE

2 个答案: