Triggering an update on a second table after insert - postgresql

Question

I have two tables, A and P.
A
________________
id | num_cars
----------
1 | 2
2 | 0
3 0
P
__________________________
id_driver | id_car
--------------------------
1 | Porsche
1 | BMW
A.id and P.id_driver referes to the same person. I created the below trigger. The idea is, every time I add a new row in P for an existing driver its correspondent row in A must be updated with the number of total cars owned by the person with that id.
CREATE OR REPLACE FUNCTION update_a() RETURNS trigger AS $$
BEGIN
IF TG_OP = 'INSERT' THEN
UPDATE A a
SET num_cars = (SELECT COUNT(NEW.id_driver)
FROM P p
WHERE (a.id = p.id_driver AND a.id=NEW.id_driver));
ELSIF TG_OP = 'DELETE' THEN
UPDATE A a
SET num_cars = num_cars - 1
WHERE a.id = OLD.id_driver AND a.num_cars<>0;
END IF;
RETURN NULL;
END;
$$ LANGUAGE plpgsql;
CREATE TRIGGER add_car
AFTER INSERT OR DELETE ON PARTICIPATION
FOR EACH ROW EXECUTE PROCEDURE update_a();
The trigger works fine when I add a row in B for a driver. However, if I then add a row for a different driver in B the rest of the rows in A are set back to 0. I would like the procedure to run only when A.id = P.id_driver. How can I do this?

Answer 1

The update query makes a cross product between A and P, and therefore updates the entire table, counting 0 cars most of the time.
You would need to restrict the update to the proper driver only, and also to compute the number of cars only for this driver:
UPDATE A a
SET num_cars = (SELECT COUNT(*)
FROM P p
WHERE p.id_driver = NEW.id_driver)
WHERE a.id = NEW.id_driver;