X-Original-To: pgsql-sql-postgresql.org@postgresql.org Received: from localhost (mx1.hub.org [200.46.208.251]) by postgresql.org (Postfix) with ESMTP id 78E569FB24D for ; Tue, 29 Aug 2006 11:28:52 -0300 (ADT) Received: from postgresql.org ([200.46.204.71]) by localhost (mx1.hub.org [200.46.208.251]) (amavisd-new, port 10024) with ESMTP id 59145-07 for ; Tue, 29 Aug 2006 11:28:46 -0300 (ADT) X-Greylist: domain auto-whitelisted by SQLgrey- Received: from py-out-1112.google.com (py-out-1112.google.com [64.233.166.179]) by postgresql.org (Postfix) with ESMTP id 707F09FB227 for ; Tue, 29 Aug 2006 11:28:46 -0300 (ADT) Received: by py-out-1112.google.com with SMTP id d80so2861356pyd for ; Tue, 29 Aug 2006 07:28:45 -0700 (PDT) DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:sender:to:subject:cc:in-reply-to:mime-version:content-type:references:x-google-sender-auth; b=iW+llCvmmGkKKNTLHfEucqu8PkAPaxDVrj13R0RAD/fV5oLuOPIpkiEHGCfHx7gH4hI+T2ajDAbfrQcbny/rAR6wzlSPm4p3PzdoqyB8o5sUeJQ79Z9ObflJXTIkVPBcv3BuH2gHHrKt6foRV4AdwlFPOvuP67Vz1R9GG78/PUI= Received: by 10.35.61.2 with SMTP id o2mr15032621pyk; Tue, 29 Aug 2006 07:28:42 -0700 (PDT) Received: by 10.35.80.9 with HTTP; Tue, 29 Aug 2006 07:28:42 -0700 (PDT) Message-ID: Date: Tue, 29 Aug 2006 09:28:42 -0500 From: "Aaron Bono" To: "Travis Whitton" Subject: Re: Performance Problem with sub-select using array Cc: pgsql-sql@postgresql.org In-Reply-To: MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_Part_9814_23165280.1156861722299" References: X-Google-Sender-Auth: a55f814424d11dfa X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.737 tagged_above=0 required=5 tests=AWL, DNS_FROM_RFC_ABUSE, HTML_40_50, HTML_MESSAGE, SPF_PASS X-Spam-Level: X-Archive-Number: 200608/229 X-Sequence-Number: 26187 ------=_Part_9814_23165280.1156861722299 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Make sure you do a reply to all so you include the list.... On 8/28/06, Travis Whitton wrote: > > Ok, I actually got down to business with EXPLAIN ANALYZE. My performance > was actually suffering from the DISTINCT in my SQL query and not the > subquery, which I guess isn't run repeatedly since it's not constrained and > can be cached by the optimizer? Bottom line is, by replacing DISTINCT with > DISTINCT ON all my index conditions show up in the EXPLAIN output. Best of > all: Total runtime: 353.588 ms. Thanks for the help. > > Travis > > > On 8/28/06, Travis Whitton wrote: > > > > I'm pretty sure you're right, which leads me to my next question. Is it > > possible to pass a column from an outer query to a subquery? For example, is > > there a way to do something like. > > > > SELECT owners.id AS owner_id, array(SELECT dogs.name WHERE owners.id = > > owner_id) ... > > > > I would just do a normal inner-join, but then I get a row for each item > > that would otherwise come back nicely packaged in the array. The overhead of > > rearranging the data takes even more time than the subquery approach. > > > > > > On 8/28/06, Aaron Bono < postgresql@aranya.com> wrote: > > > > > > On 8/24/06, Travis Whitton wrote: > > > > > > > Hello all, I'm running the following query on about 6,000 records > > > > worth of data, and it takes about 8 seconds to complete. Can anyone provide > > > > any suggestions to improve performance? I have an index on two columns in > > > > the transacts table (program_id, customer_id). If I specify a number for > > > > customer.id in the sub-select, query time is reduced to about 2 > > > > seconds, which still seems like a million years for only 6,000 records, but > > > > I'm guessing that the sub-select can't resolve the id since it's done before > > > > the outer query, so it scans the entre recordset for every row? Transacts is > > > > a many to many table for customers and programs. I know this query doesn't > > > > even reference any columns from programs; however, I dynamically insert > > > > where clauses to constrain the result set. > > > > > > > > SELECT distinct customers.id, first_name, last_name, address1, > > > > contact_city, contact_state, primary_phone, email, array(select > > > > programs.program_name from transacts, programs where customer_id = > > > > customers.id and programs.id = transacts.program_id and > > > > submit_status = 'success') AS partners from customers, transacts, programs > > > > where transacts.customer_id = customers.id and transacts.program_id= > > > > programs.id > > > > > > > > > > > > > My guess is that your problem is that you may be getting 6000 rows, > > > but the array(select ....) is having to run once for each of record returned > > > (so it is running 6000 times). > > > > > > Try an explain analyze: http://www.postgresql.org/docs/7.4/interactive/sql-explain.html > > > - that will reveal more of where the performance problem is. > > > > > ================================================================== Aaron Bono Aranya Software Technologies, Inc. http://www.aranya.com http://codeelixir.com ================================================================== ------=_Part_9814_23165280.1156861722299 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline Make sure you do a reply to all so you include the list....

On 8/28/06, Travis Whitton <tinymountain@gmail.com > wrote:
Ok, I actually got down to business with EXPLAIN ANALYZE. My performance was actually suffering from the DISTINCT in my SQL query and not the subquery, which I guess isn't run repeatedly since it's not constrained and can be cached by the optimizer? Bottom line is, by replacing DISTINCT with DISTINCT ON all my index conditions show up in the EXPLAIN output. Best of all: Total runtime: 353.588 ms. Thanks for the help.

Travis


On 8/28/06, Travis Whitton <tinymountain@gmail.com> wrote:
I'm pretty sure you're right, which leads me to my next question. Is it possible to pass a column from an outer query to a subquery? For example, is there a way to do something like.

SELECT owners.id AS owner_id, array(SELECT dogs.name WHERE owners.id = owner_id)  ...

I would just do a normal inner-join, but then I get a row for each item that would otherwise come back nicely packaged in the array. The overhead of rearranging the data takes even more time than the subquery approach.


On 8/28/06, Aaron Bono < postgresql@aranya.com> wrote:
On 8/24/06, Travis Whitton <tinymountain@gmail.com > wrote:
Hello all, I'm running the following query on about 6,000 records worth of data, and it takes about 8 seconds to complete. Can anyone provide any suggestions to improve performance? I have an index on two columns in the transacts table (program_id, customer_id). If I specify a number for customer.id in the sub-select, query time is reduced to about 2 seconds, which still seems like a million years for only 6,000 records, but I'm guessing that the sub-select can't resolve the id since it's done before the outer query, so it scans the entre recordset for every row? Transacts is a many to many table for customers and programs. I know this query doesn't even reference any columns from programs; however, I dynamically insert where clauses to constrain the result set.

SELECT distinct customers.id, first_name, last_name, address1, contact_city, contact_state, primary_phone, email, array(select programs.program_name from transacts, programs where customer_id = customers.id and programs.id = transacts.program_id and submit_status = 'success') AS partners from customers, transacts, programs where transacts.customer_id = customers.id and transacts.program_id = programs.id

 
My guess is that your problem is that you may be getting 6000 rows, but the array(select ....) is having to run once for each of record returned (so it is running 6000 times).

Try an explain analyze: http://www.postgresql.org/docs/7.4/interactive/sql-explain.html - that will reveal more of where the performance problem is.

==================================================================
   Aaron Bono
   Aranya Software Technologies, Inc.
   http://www.aranya.com
   http://codeelixir.com
================================================================== ------=_Part_9814_23165280.1156861722299--