在 Postgres 中的数组字段上应用聚合函数？

Question

是否可以对 integer[] 字段（或其他数字数组）中的所有值应用聚合（如 avg()、stddev()）？

CREATE TABLE widget
(
  measurement integer[]
);

insert into widget (measurement) values ( '{1, 2, 3}');

select avg(measurement::integer[]) from widget;

ERROR:  function avg(integer[]) does not exist
LINE 4: select avg(measurement::integer[]) from widget;
               ^
HINT:  No function matches the given name and argument types. You might need to add explicit type casts.

********** Error **********

ERROR: function avg(integer[]) does not exist
SQL state: 42883
Hint: No function matches the given name and argument types. You might need to add explicit type casts.
Character: 71

我可以通过将数组分成多行来解决

select avg(m)::float from (select unnest(measurement) m from widget) q;

但它不那么优雅。

谢谢你。

Answer 1

您可以创建这样的简单函数：

create function array_avg(_data anyarray)
returns numeric
as
$$
    select avg(a)
    from unnest(_data) as a
$$ language sql;

并像这样查询

select avg(array_avg(measurement))
from widget;

或者你可以简单地做

select avg((select avg(a) from unnest(measurement) as a))
from widget;

sql小提琴演示

Answer 2

如果有人想知道如何在保留其他属性的同时做到这一点，无需分组，横向连接提供了一个简单的解决方案：

select Point, High, Low, Average
    from GridPoint G
    join lateral (
        select Max(V) High, Min(V) Low, Avg(V) Average
            from Unnest(G.Values) V
    ) _ on true