amazon-redshift - Redshift 中的 ROW_NUMBER() 从每组中选择最大的行？

Question

我需要根据 COUNT(1) 字段从每个组中选择一行。

在其他数据库中，我会使用 ROW_NUMBER() 函数，在 redshift 中它是 unsupported yet .

Answer 1

答案是使用 SUM(1) OVER(PARTITION BY group_field ORDER BY order field ROWS UNBOUNDED PRECEDING) 结构如下:

SELECT id,
       name,
       cnt
FROM
  (SELECT id,
          name,
          count(*) cnt,
          sum(1) over (partition BY id ORDER BY cnt DESC ROWS UNBOUNDED PRECEDING) AS row_number
   FROM table
   GROUP BY id,
            name)
WHERE row_number = 1
ORDER BY name