Question

ORACLE,为什么子查询中SELECT数可以大于GROUP BY

举例如下：
select t.purchasecode,
t.a
from (
SELECT coun.purchasecode,
coun.scandate,
count(*) a
FROM APP_COUNTERFOIL COUN
group by COUN.PURCHASECODE
)
这种情况是对的，但是
coun.scandate,
count(*) a
FROM APP_COUNTERFOIL COUN
group by COUN.PURCHASECODE
这种情况是错的，为什么？
修正一下，第二个SQL应该是
coun.scandate,
count(*) a
FROM APP_COUNTERFOIL COUN
group by COUN.PURCHASECODE

Answer 1

你举的这两个例子都是错的，group by后的分组列必需小于等于select 后的分组列

追问

实际上是可以运行的，第二个写错了，应该是
select coun.purchasecode,
count(*) a
FROM APP_COUNTERFOIL COUN
group by COUN.PURCHASECODE

Answer 2

coun.scandate,
count(*) a
FROM APP_COUNTERFOIL COUN
group by COUN.PURCHASECODE

这段代码中你前面查询的没有要分组的字段

追问

第二个错了，是
select coun.purchasecode,
count(*) a
FROM APP_COUNTERFOIL COUN
group by COUN.PURCHASECODE

Answer 3

第一种是错误的写法

追问

但是可以运行，怎么回事？