Question

有一个表格“付款”

user_id     payment_time    amount    sale_type
1            2018-04-01      10       cash
1            2018-04-01      10       cash
1            2018-04-01      10       cash
1            2018-04-01      20       bank
2            2018-04-01      10       cash
2            2018-04-01      10       cash

需要现金总额。

我不明白为什么这个查询会给出错误的结果：

select SUM(CASE WHEN p1.sale_type='cash' THEN p1.amount ELSE 0 END) 
as cash
FROM 
(SELECT distinct user_id, SUM(amount) AS amount, sale_type FROM payments where 
payment_time = '2018-04-01' group by user_id) p1

Answer 1

对于内部查询，您需要在sale_type语句中添加GROUP BY列，对于您的查询样式，正确的结果应为group by user_id, sale_type。

P.S。实际上，我认为您不需要子查询。

上面的查询给出的结果为60，而

select SUM(CASE WHEN p1.sale_type='cash' THEN p1.amount ELSE 0 END) as cash
  from 
 (select distinct user_id, SUM(amount) AS amount, sale_type 
    from payments 
   where payment_time = date'2018-04-01' 
   group by user_id, sale_type) p1;

或

select SUM(CASE WHEN sale_type='cash' THEN amount ELSE 0 END) as cash 
  from payments 
 where payment_time = date'2018-04-01'

为40列提供cash

SQL Fiddle Demo

Answer 2

为什么不使用为此目的而设计的“具有”子句。

SELECT SUM(amount) AS cash FROM payments
WHERE payment_time = '2018-04-01'
GROUP BY sale_type
HAVING sale_type= 'cash'

Answer 3

我认为您可能根本不需要在子查询或整个子查询中使用distinct。

select p.user_id as id, sum(case when p.sale_type = 'cash' then p.amount else 0 end) as amount
from payments p 
where p.payment_time = '2018-04-01' 
group by p.user_id

或不包含case

select p.user_id, sum(p.amount)
from payments p 
where p.sale_type = 'cash' and p.payment_time = '2018-04-01' 
group by p.user_id

关于案件陈述中和的正确使用

3 个答案: