您可以在单个Redshift查询中一起使用AVG（）和PERCENTILE_DISC（）函数吗？

Question

我很难将这两个功能组合到一个查询中，例如：

WITH time_values AS (
    SELECT
        (end_time - start_time) * 1.0 / 3600000000 AS num_hours
    FROM table
    WHERE
        end_time >= 1493596800000000
        AND start_time < 1493683200000000
)

SELECT
    PERCENTILE_DISC(0.25) WITHIN GROUP (ORDER BY num_hours) OVER() AS p25,
    PERCENTILE_DISC(0.50) WITHIN GROUP (ORDER BY num_hours) OVER() AS p50,
    PERCENTILE_DISC(0.80) WITHIN GROUP (ORDER BY num_hours) OVER() AS p80,
    PERCENTILE_DISC(0.99) WITHIN GROUP (ORDER BY num_hours) OVER() AS p99,
    AVG(num_hours)
FROM time_values;

这返回 ERROR: column "time_values.num_hours" must appear in the GROUP BY clause or be used in an aggregate function

Answer 1

AVG()既可以聚合（需要按某列分组）又可以是窗口函数（需要frame子句）。出现错误是因为您PERCENTILE已经使用了窗口函数，并且AVG函数不明显，但没有框架子句。要AVG在同一查询中使用，您需要模拟如下子句

AVG(num_hours) OVER ()