Question

我正在尝试执行无法弄清楚如何写的查询。我有一个claims桌子

claims table:
 - id
 - date_received

我想获取30天内每天的新索赔计数。我想提出的幼稚解决方案就是这个。

select count(*) 
  from claims
  where date_received 
    BETWEEN CURRENT_DATE - interval '1 month' 
        AND CURRENT_DATE
  group by date_received;

这有效，但是它按确切的时间戳而不是同一天进行分组。我如何才能在同一天将其分组？

编辑

我能够找到我们更新的查询：

select date_received::date as date, count(id) as new_claims
  from claims 
  where date_received BETWEEN CURRENT_DATE - interval '1 month'
              AND CURRENT_DATE
  group by date_received::date 
  order by date;

Answer 1

大概，索赔到达的是过去的时间，而不是将来的发生时间，所以您需要：

select date_received::date, count(*)
from claims
where date_received >= CURRENT_DATE - interval '1 month' and
      date_received < CURRENT_DATE
group by date_received::date
order by date_received::date;

Answer 2

基于戈登的答案，您可以使用date_trunc(precision, timestamp)来获取所需的内容：

select date_received::date, count(*)
from claims
where date_received >= 
      date_trunc('day', CURRENT_DATE - interval '1 month') 
  and date_received < CURRENT_DATE
group by date_received::date
order by date_received::date;

有关详细信息，请参见date_trunc()。

PostgreSQL查询间隔数据

2 个答案: