Question

我想要实现的目标：

数据库是postgres。我们为用户订购了购买日期和user_id。我想得到订购此期间和上一期的用户数（比如说它是一周，然后是本周和上周订购的人）。

我是SQL的新手，能够获得本周的用户数，但这就是它停止的地方。

SELECT
    COUNT(distinct(user_id))
FROM
    orders
WHERE
    shipment_state = 'shipped' AND
    delivery_date >= '2017-11-01' AND
    delivery_date <= '2017-11-10'

我现在该如何制作，以便获得那段时间内订购的人数，以及11月1日前一周的f.ex？

Answer 1

这是一个有点蛮力，但假设“这个时期”意味着十一月，“最后一个时期”意味着十月，这将是一种方式：

with cte as (
  SELECT
    user_id,
    sum (case when delivery_date 
        between '2017-10-01' and '2017-10-31' then 1 else 0 end) as last,
    sum (case when delivery_date 
        between '2017-11-01' and '2017-11-30' then 1 else 0 end) as this
  FROM
    orders
  where
    shipment_state = 'shipped' and
    delivery_date >= '2017-10-01'
  group by user_id
)
select count (*)
from cte
where last > 0 and this > 0

如果没有对值进行硬编码，你可以做几周的一种方法是做这样的事情：

delivery_date between date_trunc ('week', current_date) and
   date_trunc ('week', current_date) + interval '6 days'

对于本周以及之前一周的回升一周：

delivery_date between date_trunc ('week', current_date) - interval '1 week' and
   date_trunc ('week', current_date) - interval '1 day'

Answer 2

我会这样做：

SELECT COUNT(DISTINCT CASE WHEN delivery_date >= '2017-11-01' AND delivery_date <= '2017-11-10'
                           THEN user_id
                      END),
       COUNT(DISTINCT CASE WHEN delivery_date >= '2017-11-01' - INTERVAL '1 week' AND delivery_date < '2017-11-01'
                           THEN user_id
                      END)      
FROM orders o
WHERE shipment_state = 'shipped' AND
      delivery_date >= '2017-11-01' - INTERVAL '1 week' AND
      delivery_date <= '2017-11-10';

Answer 3

       SELECT COUNT(DISTINCT CASE WHEN delivery_date >= '2017-11-01' 
    AND delivery_date <= '2017-11-10'
                                   THEN user_id
                              END)
        FROM orders o
        WHERE user_id = (SELECT DISTINCT user_id 
            FROM orders o    
            WHERE
                shipment_state = 'shipped' AND
                      delivery_date >= '2017-11-30' - INTERVAL '1 week' AND
                      delivery_date <= '2017-11-01')
  AND
   shipment_state = 'shipped' AND
   delivery_date >= '2017-11-01' - INTERVAL '1 week' AND
   delivery_date <= '2017-11-10';

从两个部分获得不同的数量

3 个答案: