Question

第一篇文章。

我有一个表格，我想计算出某个值出现的行。总的来说，我想要两个count（）行，这些行依赖于另一列（Y和N）中的不同值。我认为最好在桌面上进行两次连接。但是，结果根本就没有行。 id是唯一的（主键）。

我缺少什么？它们相互取消，因为如果我只运行一个连接和相应的count（）就可以了。或者，如果有更好的方法来解决这个问题。谢谢！

 select
 a.loans,
 count(b.loans) as count_new_cust,
 count(c.loans) as count_old_cust
 FROM  applications a 
,(SELECT distinct id, loans FROM  applications where new_cust='Y') b 
,(SELECT distinct id, loans FROM  applications where new_cust='N') c
  where a.id=b.id  and b.id=c.id
  group by a.loans
  order by a.loans DESC

Answer 1

无需JOIN。您可以改为使用条件聚合：

select a.loans,
       count(case when new_cust='Y' then a.loans END) as count_new_cust,
       count(case when new_cust='N' then a.loans END) as count_old_cust
FROM  applications a 
group by a.loans
order by a.loans DESC

主表上的子查询互相取消？

1 个答案: