On Sat, Oct 29, 2016 at 6:55 AM, Kim Rose Carlsen <krc@hiper.dk> wrote:
>> try this :-D
>
>> create or replace function indf(anyelement, anyelement) returns anyelement
>> as
>> $$
>> select $1 = $2 or ($1 is null and $2 is null);
>> $$ language sql;
>>
>> CREATE VIEW view_circuit_with_status AS (
>> SELECT r.*,
>> s.circuit_status,
>> s.customer_id AS s_customer_id,
>> p.line_speed,
>> p.customer_id AS p_customer_id
>> FROM view_circuit r
>> JOIN view_circuit_product_main s
>> ON r.circuit_id = s.circuit_id
>> AND indf(r.customer_id, s.customer_id)
>> JOIN view_circuit_product p
>> ON r.circuit_id = p.circuit_id
>> AND indf(r.customer_id, s.customer_id)
>>
>> merlin
>
> This doesn't do much good. This doesn't tell the planner that the 3
> customer_ids are actually of same value, and it therefore can't filter them
> as it sees fit.
You do know you can index on a function, and the planner then keeps
stats on it when you run analyze right?