Postgres GROUP BY数组列

Question

I use postgres & have a table like this :

id   | arr
-------------------
 1   | [A,B,C]
 2   | [C,B,A]
 3   | [A,A,B]
 4   | [B,A,B]

I created a GROUP BY 'arr' query.

SELECT COUNT(*) AS total, "arr" FROM "table" GROUP BY "arr"

... and the result :

total | arr
-------------------
 1    | [A,B,C]
 1    | [C,B,A]
 1    | [A,A,B]
 1    | [B,A,B]

BUT, since [A,B,C] and [C,B,A] have the same elements, so i expected the result should be like this :

total | arr
-------------------
  2   |   [A,B,C]
  2   |   [A,A,B]

Did i miss something (in query) or else? Please help me..

Answer 1

You do not need to create a separate function to do this. It can all be done in a single statement:

select array(select unnest(arr) order by 1) as sorted_arr, count(*)
from t
group by sorted_arr;

Here is a rextester .

Answer 2

[A,B,C] and [C,B,A] are different arrays even if they have the same elements they are not in the same position, they will never be grouped by a group by clause, in case you want to make them equivalent you'd need to sort them first.

On this thread you have info abour sorting arrays.

You should do something like:

SELECT COUNT(*) AS total, array_sort("arr") FROM "table" GROUP BY array_sort("arr")

After creating a sort function like the one proposed in there:

CREATE OR REPLACE FUNCTION array_sort (ANYARRAY)
RETURNS ANYARRAY LANGUAGE SQL
AS $$
SELECT ARRAY(SELECT unnest($1) ORDER BY 1)
$$;