Find position(s) in array matching a given sub-array - sql

Question

Given this table:
CREATE TABLE datasets.travel(path integer[], path_timediff double precision[]);
INSERT INTO datasets.travel
VALUES (array[50,49,49,49,49,50], array[NULL,438,12,496,17,435]);
I am looking for some kind of function or query in the PostgreSQL that for a given input array[49,50] will find the matching consecutive index values in path which is [5,6] and the corresponding element in path_timediff which is 435 in the example (array index 6).
My ultimate purpose is to find all such occurrences of [49,50] in path and all the corresponding elements in path_timediff. How can I do that?

Answer 1

Assuming you have a primary key in your table you did not show:
CREATE TABLE datasets.travel (
travel_id serial PRIMARY KEY
, path integer[]
, path_timediff float8[]
);
Here is one way with generate_subscripts() in a LATERAL join:
SELECT t.travel_id, i+1 AS position, path_timediff[i+1] AS timediff
FROM (SELECT * FROM datasets.travel WHERE path #> ARRAY[49,50]) t
, generate_subscripts(t.path, 1) i
WHERE path[i:i+1] = ARRAY[49,50];
This finds all matches, not just the first.
i+1 works for a sub-array of length 2. Generalize with i + array_length(sub_array, 1) - 1.
The subquery is not strictly necessary, but can use a GIN index on (path) for a fast pre-selection:
(SELECT * FROM datasets.travel WHERE path #> ARRAY[49,50])
Related:
How to access array internal index with postgreSQL?
Parallel unnest() and sort order in PostgreSQL
PostgreSQL unnest() with element number