Postgres - pair elements of one array with elements of another - postgresql

Question

The values to pair are defined in two arrays, array[1,2,3] and array['A','B','C'].
What I need to do is merge these two arrays together, where each element is paired with one at the same index in the other, so it results in array[[1,'A'],[2,'B'],[3,'C']].
How can I do that?

Answer 1

You could use UNNEST WITH ORDINALITY & join the 2 result this way:
Schema (PostgreSQL v14)
CREATE TABLE array_zip (
id INT,
a1 INT[],
a2 TEXT[]
);
INSERT INTO array_zip
VALUES (1, ARRAY [1, 2, 3], ARRAY ['A', 'B', 'C'])
, (2, ARRAY [4, 5, 6], ARRAY ['D', 'E', 'F', 'G']) -- different number of elements
;
Query #1
SELECT id, zip
FROM array_zip
CROSS JOIN LATERAL (
SELECT array_agg((aa1.v, aa2.v)) AS zip
FROM UNNEST(a1) WITH ORDINALITY AS aa1(v, i)
-- use an INNER JOIN instead to shortcut the zip of the first missing element
FULL JOIN UNNEST(a2) WITH ORDINALITY AS aa2(v, i)
USING (i)
) AS f
;
id
zip
1
{"(1,A)","(2,B)","(3,C)"}
2
{"(4,D)","(5,E)","(6,F)","(,G)"}
Query #2
You could also avoid the JOIN between the two arrays with an index-based direct access, which should be faster (but probably not by a lot, unless the arrays here are pretty big):
SELECT id, zip
FROM array_zip
CROSS JOIN LATERAL (
SELECT array_agg((a1[i], a2[i])) AS zip
-- use LEAST instead to stop the zip on the first missing element
FROM generate_series(1, GREATEST(cardinality(a1), cardinality(a2))) AS i
) AS f;
id
zip
1
{"(1,A)","(2,B)","(3,C)"}
2
{"(4,D)","(5,E)","(6,F)","(,G)"}
View on DB Fiddle
The last option I can think of is writing a function (either PL/pgSQL, or PL/v8, or...). The code would probably be easier to understand (especially if you need this feature in multiple queries) and you could handle the len(arr1) != len(arr2 -> raise case if you want/need to.