PostgreSQL merge recursive query and JOIN - sql

Question

I have the following schema:
CREATE TABLE tbl_employee_team
(
employee_id int,
teams_id int
);
INSERT INTO tbl_employee_team
VALUES
(1, 2),
(1, 3),
(1, 4);
CREATE TABLE tbl_team_list_serv
(
service_id int,
team_id int
);
INSERT INTO tbl_team_list_serv
VALUES
(7, 2),
(9, 3),
(10, 4);
CREATE TABLE tbl_service
(
id int,
parent int
);
INSERT INTO tbl_service
VALUES
(5, null),
(6, 5),
(7, 6),
(8, null),
(9, 8),
(10, null);
For the sake of simplicity I declared:
1 as employee_id
2, 3, 4 as team_id
5 -> 6 -> 7 as service (5 is the main service)
8 -> 9 (8 is the main service)
10 (10 is the main service)
To retrieve the services the employee belongs to I query
SELECT ls.service_id FROM tbl_team_list_serv ls
JOIN tbl_employee_team t ON ls.team_id=t.teams_id WHERE t.employee_id = 1
To get the main service from the services I use
WITH RECURSIVE r AS
(
SELECT id, parent, 1 AS level
FROM tbl_service
WHERE id = 7 /*(here's I need to assign to every id from the JOIN)*/
UNION
SELECT tbl_service.id, tbl_service.parent, r.level + 1 AS level
FROM tbl_service
JOIN r
ON r.parent = tbl_service.id
)
SELECT id FROM r WHERE r.level = (SELECT max(level) FROM r)
My question is how do I merge the two queries?
Based on the data above I want to finally get a list of ids which is in this case:
5, 8, 10
Also, I want my recursive query to return the last row (I don't think that the solution with level is elegant)
SQLFiddle can be found here
Thanks in advance

Answer 1

I feel like you already did most of the work for this question. This is just a matter of the following tweaks:
Putting the logic for the first query in the anchor part of the CTE.
Adding the original service id as a column to remember the hierarchy.
Tweaking the final logic to get one row per original service.
As a query:
WITH RECURSIVE r AS (
SELECT ls.service_id as id, s.parent, 1 as level, ls.service_id as orig_service_id
FROM tbl_team_list_serv ls JOIN
tbl_employee_team t
ON ls.team_id = t.teams_id JOIN
tbl_service s
ON ls.service_id = s.id
WHERE t.employee_id = 1
UNION ALL
SELECT s.id, s.parent, r.level + 1 AS level, r.orig_service_id
FROM tbl_service s JOIN
r
ON r.parent = s.id
)
SELECT r.id
FROM (SELECT r.*,
MAX(level) OVER (PARTITION BY orig_service_id) as max_level
FROM r
) r
WHERE r.level = max_level;
Here is a db<>fiddle.