Question

我有两个表，coaches和awardscoaches。 coaches给了我教练和球队之间的关系。 awardscoaches包括教练奖励通知。以下是他们的结构：

教练

coachid: string   (the id of each coach, it is the primary key in this table)
tmid:    string   (the team id)

awardscoaches

coachid: string   (the id of each coach)
award:   string   (the award the coach got, each coach may have more than one award, so the primary key in this table is the combine of coachid and award)

现在我要写一个查询，找出哪个教练获得最多奖项的每个团队。

下面是我目前拥有的sql：

select c.tmid tmid, max(a.count) count from coaches c 
inner join (select coachid, count(award) count 
from awardscoaches group by coachid) a 
on a.coachid = c.coachid group by c.tmid;

此查询将返回每个团队的最大奖励数。但我不知道如何选择结果集中的coachid，因为我只能从组中选择字段。我正在寻找一个通用的sql语句来实现这个要求。

我在下面尝试了命令：

select coachid,tmid,award_count 
from (select coachid,tmid,award_count
      ,rank() over(partition by tmid order by award_count desc) as rnk
      from (select a.coachid, count(*) over(partition by a.coachid) as award_count,c.tmid 
            from awardscoaches a
            join coaches c on c.coachid=a.coachid
           ) t
      ) t
where rnk = 1

但得到重复的行如下：

murrabr01c  WAS 17
murrabr01c  WAS 17
murrabr01c  WAS 17
murrabr01c  WAS 17
krommbo01c  WIJ 10
krommbo01c  WIJ 10
krommbo01c  WIJ 10
krommbo01c  WIJ 10
wattto01c   WIN 7
wattto01c   WIN 7
wattto01c   WIN 7

Answer 1

max(struct(a.cnt,a.coachid)).col2   as coachid

select      c.tmid                              as tmid
           ,max(struct(a.cnt,a.coachid)).col2   as coachid 

from                    coaches c 

            join       (select      coachid
                                   ,count(*)    as cnt

                        from        awardscoaches 

                        group by    coachid
                        ) a 

            on          a.coachid = 
                        c.coachid 

group by    c.tmid
;

如何在两个表连接中进行sql查询包括group by和max value

1 个答案: