sql 不是聚合函数的字段不gounby可以吗

[qq_27561573]

能写个吗


select
u.puserid as puserid,
u.username as username,
m.reserved3 as reserved3,
u.extend as extend,
to_char(m.deal_time, 'YYYY-MM-DD') datetime,
count(distinct(m.id)) as receive
from
a r, b m, c u
group by u.puserid, m.deal_time, u.username, m.reserved3,u.extend

[gw6328]

说明你没有理解到group by .理解了就明白不能，就算一样的内容，也是用group by 或者 max() min()这些方法来取。

[中国风]

不行，只有mysql可这样用，但多表联接时会存在不可预知的结果，你的语句可改为

select
    u.puserid as puserid,
      u.username as username,   
      m.reserved3 as reserved3,
      u.extend as extend,
      to_char(m.deal_time, 'YYYY-MM-DD') datetime,
      count(distinct(m.id)) as receive
from
  a  r, b  m, c  u
 group by u.puserid,to_char(m.deal_time, 'YYYY-MM-DD'), u.username, m.reserved3,u.extend

[中国风]

to_char(m.deal_time, 'YYYY-MM-DD')--放在group by 里

[qq_27561573]

我想把不是聚合的不放到group by里

[kingtiy]

不想group by 那么多字段,可以先group by key字段. 出了结果,再去关联属性字段.