Re: Slow query: Select all buildings that have >1 pharmacies and >1 schools within 1000m - Mailing list pgsql-performance

From Stefan Keller
Subject Re: Slow query: Select all buildings that have >1 pharmacies and >1 schools within 1000m
Date
Msg-id CAFcOn2-fYtZZn0bCiB5F9oF9H43fFvUqYLbZ4TyAPvr_X4nF5A@mail.gmail.com
Whole thread Raw
In response to Re: Slow query: Select all buildings that have >1 pharmacies and >1 schools within 1000m  (Jeff Janes <jeff.janes@gmail.com>)
Responses Re: Slow query: Select all buildings that have >1 pharmacies and >1 schools within 1000m
List pgsql-performance
Hi

2012/8/8 Jeff Janes <jeff.janes@gmail.com>:
> On Tue, Aug 7, 2012 at 5:07 PM, Stefan Keller <sfkeller@gmail.com> wrote:
>> Hi Craig
>>
>> Clever proposal!
>> I slightly tried to adapt it to the hstore involved.
>> Now I'm having a weird problem that PG says that "relation 'p' does not exist".
>> Why does PG recognize table b in the subquery but not table p?
>> Any ideas?
>
> I don't think it does recognize b, either.  It just fell over on p
> before it had a chance to fall over on b.

No, the b get's recognized. See my original query.
That's a strange behaviour of the SQL parser which I can't understand.

> I think you have to use WITH if you want to reference the same
> subquery in multiple FROMs.

I'll try that with CTE too.

> Another approach would be to add explicit conditions for there being
> at least 1 school and 1 pharmacy within distance.  There can't be >1
> unless there is >=1, but the join possibilities for >=1 (i.e. "where
> exists" rather than "where (select count(*)...)>1" )  are much more
> attractive than the ones for >1.
>
> Cheers,
>
> Jeff

You mean, first doing a select on existence and then apply the count
condition later?

Stefan

pgsql-performance by date:

Previous
From: Ondrej Ivanič
Date:
Subject: Re: Postgresql - performance of using array in big database
Next
From: Jeff Janes
Date:
Subject: Re: Slow query: Select all buildings that have >1 pharmacies and >1 schools within 1000m