Postgresql按列获取最大行组

Question

我正在尝试从 table 中的每日计数总和中获取最大行。我看过几个看起来相似的帖子，但它似乎不起作用。我试图跟随 Get MAX row for GROUP in MySQL 但它在 Postgres 中不起作用。这是我的

select source, SUM(steps) as daily_steps, to_char("endTime"::date, 'MM/DD/YYYY') as step_date
 from activities
 where user_id = 1
 and "endTime" <= CURRENT_TIMESTAMP + INTERVAL '1 day'
 and "endTime" >= CURRENT_TIMESTAMP - INTERVAL '7 days' 
 group by source, to_char("endTime"::date, 'MM/DD/YYYY')

这个return下面

source, daily_steps, step_date
"walking";750;"11/17/2015"
"walking";821;"11/22/2015"
"walking";106;"11/20/2015"
"running";234;"11/21/2015"
"running";600;"11/24/2015"

我希望结果 return 仅具有 daily_steps 最大值的行（按来源）。结果应该看起来像

source, daily_steps, step_date
"walking";821;"11/22/2015"
"running";600;"11/24/2015"

Answer 1

Postgres 提供了方便的 distinct on 语法：

select distinct on (a.source) a.*
from (select source, SUM(steps) as daily_steps, to_char("endTime"::date, 'MM/DD/YYYY') as step_date
      from activities a
      where user_id = 1 and
            "endTime" <= CURRENT_TIMESTAMP + INTERVAL '1 day' and
            "endTime" >= CURRENT_TIMESTAMP - INTERVAL '7 days' 
      group by source, to_char("endTime"::date, 'MM/DD/YYYY')
     ) a
order by a.source, daily_steps desc;

Postgresql按列获取最大行组

Postgresql get max row group by column

sql

postgresql

greatest-n-per-group