Postgres陣列行彙總

Question

我有一個遞歸查詢，其中我得到如下所示的數組行。我怎麼可能將所有行合併到一行中的一個數組中並刪除重複項？訂購併不重要。

--my_column-- 
"{431}" 
"{431,33}" 
"{431,60}" 
"{431,28}" 
"{431,1}" 
"{431,226}" 
"{431,38}" 
"{431,226,229}" 
"{431,226,227}" 
"{431,226,235}" 
"{431,226,239}" 
"{431,226,241}" 

我想下面的查詢，但我得到一個空的整數[]列

select array(select unnest(my_column) from my_table 

感謝

Answer 1

使用array_agg()與distinct和（非必要）order by從unnest()：

with my_table(my_column) as (
values 
    ('{431}'::int[]), 
    ('{431,33}'), 
    ('{431,60}'), 
    ('{431,28}'), 
    ('{431,1}'), 
    ('{431,226}'), 
    ('{431,38}'), 
    ('{431,226,229}'), 
    ('{431,226,227}'), 
    ('{431,226,235}'), 
    ('{431,226,239}'), 
    ('{431,226,241}') 
) 

select array_agg(distinct elem order by elem) 
from my_table, 
lateral unnest(my_column) elem; 

        array_agg     
--------------------------------------------- 
{1,28,33,38,60,226,227,229,235,239,241,431} 
(1 row) 

Answer 2

A nother solution without lateral subquery：

select array_agg(distinct val) from 
    (select unnest(my_column) as val from my_table) x;