Re: Query/database optimization

Eugeny N Dzhurinsky <bofh@redwerk.com> writes:
> [slow query]

The bulk of your time seems to be going into this indexscan:

>                                  ->  Index Scan using task_scheduler_icustomer_id on task_scheduler ts
(cost=2.03..11.51rows=1 width=51) (actual time=2.785..2.785 rows=0 loops=4161) 
>                                        Index Cond: ("outer".customer_id = ts.customer_id)
>                                        Filter: ((get_available_pages(task_id, customer_id) > 0) AND
((get_check_period(task_id,next_check) IS NULL) OR ((date_part('epoch'::text, get_check_period(task_id, next_check)) -
date_part('epoch'::text,(timenow())::timestamp without time zone)) < 3600::double precision)) AND (status <> 1) AND
((((start_time)::timewith time zone < ('now'::text)::time(6) with time zone) AND ((stop_time)::time with time zone >
('now'::text)::time(6)with time zone)) OR ((start_time IS NULL) AND (stop_time IS NULL))) AND (NOT (hashed subplan))) 
>                                        SubPlan
>                                          ->  Unique  (cost=2.02..2.03 rows=1 width=4) (actual time=0.617..0.631
rows=3loops=1) 
> ...

I kinda doubt that the index search itself is that slow --- doubtless
the problem comes from having to evaluate that filter condition on a lot
of rows.  How fast are those functions you're calling?

            regards, tom lane