From: Rhaoni Chiu Pereira <rhaoni@sistemica.info>
To: josh@agliodbs.com
Cc: PostgreSQL ADMIN <pgsql-admin@postgresql.org>
Cc: PostgreSQL Performance <pgsql-performance@postgresql.org>
Subject: Re: [PERFORM] How to force an Index ?
Date: Thu, 18 Sep 2003 11:45:09 -0300
Message-ID: <1063896309.3f69c4f5cd16a@sistemica.info> (raw)
In-Reply-To: <200309171538.33515.josh@agliodbs.com>
References: <1063837067.3f68dd8b5a62b@sistemica.info>
<200309171538.33515.josh@agliodbs.com>
I solve this problem doing this:
create function date_to_yyyymm( timestamp ) returns gsames00.ano_mes%type as
'select to_char($1, ''YYYYMM'');
' language sql immutable strict;
And changing the SQL where clause:
... gsames00.ano_mes = to_char(ftnfco00.data_emissao,'YYYYMM') AND ...
to:
... gsames00.ano_mes = date_to_yyyymm(ftnfco00.data_emissao) AND ...
Then it uses the gsames00 index instead of a SeqScan 'cuz it is camparing
same data type, but .. I don't want to create this function 'cuz this aplication
is used with Oracle too.
I need to know if there is a way to set the to_char output to varchar instead of
text !
Any Idea ? So, this way I wont have to change my aplication source.
Atenciosamente,
Rhaoni Chiu Pereira
Sistêmica Computadores
Visite-nos na Web: http://sistemica.info
Fone/Fax : +55 51 3328 1122
Citando Josh Berkus <josh@agliodbs.com>:
<> Rhaoni,
<>
<> First off, thanks for posting such complete info.
<>
<> > ... gsames00.ano_mes = to_char(ftnfco00.data_emissao,'YYYYMM') AND ...
<> >
<> > ftnfco00.data_emissao is a timestamp. When I run the explain analyze it
<>
<> says:
<> >
<> > ...
<> > -> Seq Scan on gsames00 (cost=100000000.00..100000006.72 rows=372
<> width=10)
<> > (actual time=0.01..0.96 rows=372 loops=19923)
<> > ...
<>
<> Your problem is that you're comparing against a calculated expression based
<> on
<> ftnfco00, which is being filtered in about 18 other ways. As a result, the
<>
<> planner doesn't know what to estimate (see the cost estimate of 100000000,
<> which is a "blind guess" values) and goes for a seq scan.
<>
<> Can I ask you to try this workaround, to create an expressional index on
<> ftnfco00 (assuming that data_emmisao is of type DATE)
<>
<> create function date_to_yyyymm( date ) returns text as
<> 'select to_char($1, ''YYYYMM'');
<> ' language sql immutable strict;
<>
<> create index idx_data_yyyymm on ftnfco00(date_to_yyyymm(data_emmisao));
<>
<> --
<> -Josh Berkus
<> Aglio Database Solutions
<> San Francisco
<>
<>
<> ---------------------------(end of broadcast)---------------------------
<> TIP 4: Don't 'kill -9' the postmaster
<>
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 from your mail client; the link fills in
the headers below: reply via email
Some clients, webmail especially, drop In-Reply-To and References from
a link, and the reply then starts a new thread: use the mbox below.
To: pgsql-performance@postgresql.org
Cc: rhaoni@sistemica.info, josh@agliodbs.com, pgsql-admin@postgresql.org
Subject: Re: [PERFORM] How to force an Index ?
In-Reply-To: <1063896309.3f69c4f5cd16a@sistemica.info>
References: <1063837067.3f68dd8b5a62b@sistemica.info> <200309171538.33515.josh@agliodbs.com> <1063896309.3f69c4f5cd16a@sistemica.info>
* 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