如何在 AWS redshift 中计算不同时间 window 的不同

Question

SELECT count(DISTINCT(c.visitid)) as count1,
       t.prod
FROM x.t1 c
JOIN y.t1 t
    ON c.headingid = t.prod_heading_id
WHERE
        c.eventtimestamp  BETWEEN '2021-01-01' AND '2021-04-02'
    AND c.evaluation > 0    
GROUP BY t.prod
ORDER BY count1 DESC
LIMIT 100

我还有时间 window 从“2020-01-01”到“2020-04-02”，我想按组进行与 count2 相同的计数。

Answer 1

您可以使用条件聚合：

SELECT count(DISTINCT c.visitid) filter (where c.eventtimestamp  BETWEEN '2021-01-01' AND '2021-04-02') as cnt1,
       count(DISTINCT c.visitid) filter (where c.eventtimestamp  BETWEEN '2020-01-01' AND '2020-04-02') as cnt2,
       t.prod
FROM x.t1 c JOIN
     y.t1 t
     ON c.headingid = t.prod_heading_id
WHERE c.evaluation > 0 AND
      c.eventtimestamp  BETWEEN '2020-01-01' AND '2021-04-02'
GROUP BY t.prod
ORDER BY cnt1 DESC;

在 Redshift（或许多其他数据库）中，语法为：

SELECT count(DISTINCT case when c.eventtimestamp  BETWEEN '2021-01-01' AND '2021-04-02' then c.visitid end) as cnt1,
       count(DISTINCT case when c.eventtimestamp  BETWEEN '2020-01-01' AND '2020-04-02' then c.visitid end) as cnt2,
       t.prod
FROM x.t1 c JOIN
     y.t1 t
     ON c.headingid = t.prod_heading_id
WHERE c.evaluation > 0 AND
      c.eventtimestamp  BETWEEN '2020-01-01' AND '2021-04-02'
GROUP BY t.prod
ORDER BY cnt1 DESC;

如何在 AWS redshift 中计算不同时间 window 的不同

How to count distinct in different time window in AWS redshift

sql

amazon-redshift