pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Edoardo Ceccarelli <eddy@axa.it>
Cc: Christopher Kings-Lynne <chriskl@familyhealth.com.au>
Cc: Nick Barr <nicky@chuckie.co.uk>
Cc: pgsql-performance@postgresql.org
Subject: Re: slow seqscan
Date: Wed, 21 Apr 2004 12:10:02 +0200
Message-ID: <4086487A.4090703@axa.it> (raw)
In-Reply-To: <408641B7.5090808@axa.it>
References: <40862003.2090201@axa.it>
	<4086272E.2080102@chuckie.co.uk>
	<40863228.907@axa.it>
	<408636A5.2050305@familyhealth.com.au>
	<408641B7.5090808@axa.it>

tried the

enable_seqscan = false

and I'm having all index scans, timing has improved from 600ms to 18ms

wondering what other implications I might expect.




Edoardo Ceccarelli ha scritto:

>
>> What happens if you go:
>>
>> CREATE INDEX annuncio400_rubric_testo_idx ON annuncio400(rubric, 
>> LOWER(testo));
>>
>> or even just:
>>
>> CREATE INDEX annuncio400_rubric_testo_idx ON annuncio400(LOWER(testo));
>>
> I wasn't able to make this 2 field index with lower:
>
> dba400=# CREATE INDEX annuncio400_rubric_testo_idx ON 
> annuncio400(rubric, LOWER(testo));
> ERROR:  parser: parse error at or near "(" at character 71
>
> seems impossible to creat 2 field indexes with lower function.
>
> The other one does not make it use the index.
>
>
>>> But the strangest thing ever is that if I change the filter with 
>>> another one that represent a smaller amount of data  it uses the 
>>> index scan!!!
>>
>>
>>
>> What's strange about that?  The less data is going to be retrieved, 
>> the more likely postgres is to use the index.
>>
> can't understand this policy:
>
> dba400=# SELECT count(*) from annuncio400 where rubric='DD';
> count
> -------
>  6753
> (1 row)
>
> dba400=# SELECT count(*) from annuncio400 where rubric='MA';
> count
> -------
>  2165
> (1 row)
>
> so it's using the index on 2000 rows and not for 6000?  it's not that 
> big difference, isn't it?
>
>
>> I suggest maybe increasing the amount of stats recorded for your 
>> rubrik column:
>>
>> ALTER TABLE annuncio400 ALTER rubrik SET STATISTICS 100;
>> ANALYZE annuncio400;
>>
> done, almost the same, still not using index
>
>> You could also try reducing the random_page_cost value in your 
>> postgresql.conf a little, say to 3 (if it's currently 4).  That will 
>> make postgres more likely to use index scans over seq scans.
>>
>
> changed the setting on postgresql.conf, restarted the server,
> nothing has changed.
>
> what about setting this to false?
> #enable_seqscan = true
>
> thanks again
> Edoardo
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: you can get off all lists at once with the unregister command
>    (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)
>
>
>



view thread (12+ messages)  latest in thread

Message-ID: <4086487A.4090703@axa.it>
Permalink:  ../4086487A.4090703@axa.it/
Also on:    postgresql.org/message-id/4086487A.4090703@axa.it

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: pgsql-performance@postgresql.org
  Cc: eddy@axa.it, chriskl@familyhealth.com.au, nicky@chuckie.co.uk
  Subject: Re: slow seqscan
  In-Reply-To: <4086487A.4090703@axa.it>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox