如果间隔不是 "full"，如何使 OVER() 中的结果无效

Question

假设我有以下数据集和查询：

WITH results as (
    SELECT 'DAL' as team, 2010 as season, 6 as wins union
    SELECT 'DET' as team, 2010 as season, 6 as wins union
    SELECT 'DET' as team, 2011 as season, 10 as wins union
    SELECT 'DET' as team, 2012 as season, 4 as wins union
    SELECT 'DET' as team, 2013 as season, 7 as wins union
    SELECT 'DET' as team, 2014 as season, 11 as wins union
    SELECT 'DET' as team, 2015 as season, 7 as wins union
    SELECT 'DET' as team, 2016 as season, 9 as wins union
    SELECT 'DET' as team, 2017 as season, 9 as wins union
    SELECT 'DET' as team, 2018 as season, 6 as wins union
    SELECT 'DET' as team, 2019 as season, 3 as wins
) SELECT team, season, wins,
    AVG(wins) OVER (partition by team order by season rows between 2 preceding and current row) AS avg_wins_last_two_seasons
from results;

这将return以下结果：

# team, season, wins, avg_wins_last_two_seasons
DAL, 2010, 6, 6.0000
DET, 2010, 6, 6.0000
DET, 2011, 10, 8.0000

但是，如果 window 的大小不是 'complete' - 即前面没有两行 - 那么我希望结果为 NULL而不是对减小的 window 大小取平均值——例如，对于 2010，它将 avg 计算为 [6]/1 = 6，但我希望它计算为：[NULL, NULL, 6] / 3 = NULL .我该怎么做？

Answer 1

您可以使用 row_number() 和 case 表达式：

select team, season, wins,
    case when row_number() over(partition by team order by season) > 2
        then avg(wins) over (
            partition by team 
            order by season 
            rows between 2 preceding and current row
        ) 
    end as avg_wins_last_two_seasons
from results;

这确保在计算 window 平均值之前至少有两行；如果没有，那么这个 returns null 代替。

Answer 2

@GMB 在这里有公认的答案，但这里有一个细微的变化，使用命名的 window 以提高可读性并显示两个选项（null 和允许不完整的 window 大小）：

WITH results as (
    SELECT 'DAL' as team, 2010 as season, 6 as wins union
    SELECT 'DET' as team, 2010 as season, 6 as wins union
    SELECT 'DET' as team, 2011 as season, 10 as wins union
    SELECT 'DET' as team, 2012 as season, 4 as wins union
    SELECT 'DET' as team, 2013 as season, 7 as wins union
    SELECT 'DET' as team, 2014 as season, 11 as wins union
    SELECT 'DET' as team, 2015 as season, 7 as wins union
    SELECT 'DET' as team, 2016 as season, 9 as wins union
    SELECT 'DET' as team, 2017 as season, 9 as wins union
    SELECT 'DET' as team, 2018 as season, 6 as wins union
    SELECT 'DET' as team, 2019 as season, 3 as wins
) SELECT team, season, wins
    ,ROUND(AVG(wins) OVER trailing_2, 1) AS avg_wins_trailing_2
    ,IF(COUNT(1) OVER trailing_2 < 3, NULL, ROUND(AVG(wins) OVER trailing_2,1))AS avg_wins_trailing_2_if_full_window
FROM results
WINDOW trailing_2 AS (
    PARTITION BY team
    ORDER BY season
    ROWS BETWEEN 2 PRECEDING AND CURRENT ROW
);

# team, season, wins, avg_wins_trailing_2, avg_wins_trailing_2_if_full_window
DAL, 2010, 6, 6.0, null
DET, 2010, 6, 6.0, null
DET, 2011, 10, 8.0, null
DET, 2012, 4, 6.7, 6.7
DET, 2013, 7, 7.0, 7.0

如果间隔不是 "full"，如何使 OVER() 中的结果无效

How to nullify result in OVER() if the interval isn't "full"

mysql

sql

average

aggregate-functions

window-functions