Re: Suboptimal query plan when using expensive BCRYPT functions

On Mon, Mar 24, 2014 at 12:08 AM, Heikki Linnakangas
<hlinnakangas(at)vmware(dot)com> wrote:
> On 03/22/2014 02:59 AM, Erik van Zijst wrote:
>>
>> Is there any way I can get postgres to perform the hash calculations
>> on the *result* of the other parts of the where clause, instead of the
>> other way around? Or else rewrite the query?
>
>
> The planner doesn't know that the crypt function is expensive. That can be
> fixed with "ALTER FUNCTION crypt(text, text) COST <high value>". Even with
> that, I'm not sure if the planner is smart enough to optimize the query the
> way you'd want, but it's worth a try.

Thanks. That's the kind of hint I was looking for.

I just gave it a go, but unfortunately it's not enough to get the
optimizer to change the plan, regardless of the cost value.

Cheers,
Erik