Question

尽管脚本仅包含每周数据，但我正在尝试按日期对数据进行排序。

下面是一个示例SQL：

SELECT
  CONCAT(CONCAT('Week', ''), 
         EXTRACT(WEEK FROM sale_date ::date + '1 day'::interval)
        ),
  COUNT(*)
FROM sales
WHERE sale_date between '2018-12-29' AND '2019-01-04'
GROUP BY CONCAT(CONCAT('Week', ''), 
                EXTRACT(WEEK FROM sale_date ::date + '1 day'::interval)
               )
ORDER BY 1 ASC

上面的脚本运行得很好，但是当对最终输出进行排序时，它首先将Week1排序，最后将Week 52排序，而我试图按实际日期排序，使得W52首先出现（自2018-12起），而W1出现。最后。

当前输出：

Week1,101
Week52,2350

预期输出：

Week52,2350
Week1,101

我正在使用Amazon Redshift DB。

Answer 1

这可以是你行，它需要添加一个额外的列到查询

select date_trunc('WEEK',sale_date ::date + interval '1 day'),
  CONCAT(CONCAT('Week', ''),
              EXTRACT(WEEK FROM sale_date  ::date + '1 day'::interval)),
       count(*)
from sales
where sale_date  between '2018-12-29' and '2019-01-04'
group by date_trunc('WEEK',sale_date ::date+ interval '1 day'),
         CONCAT(CONCAT('Week', ''),
                EXTRACT(WEEK FROM sale_date  ::date + '1 day'::interval))
order by date_trunc('WEEK',sale_date ::date+ interval '1 day') asc;

Answer 2

这应该有效：

SELECT week, count FROM (
    SELECT week, MAX(sale_date) max_date, COUNT(*) count FROM (
        SELECT sale_date,
            CONCAT('Week', EXTRACT(WEEK FROM sale_date::date + '1 day'::interval)) AS week
        FROM sales
        WHERE sale_date between '2018-12-29' AND '2019-01-04'
    )
    GROUP BY week
) ORDER BY max_date ASC

如果每个空间超过一年，将会得到奇怪的结果。

在Amazon Redshift中按日期对每周明智数据进行排序

2 个答案: