pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
Yet another question about not use on indexes
3+ messages / 3 participants
[nested] [flat]

* Yet another question about not use on indexes
@ 2003-05-11 13:16  Xevi Serrats <tomcatbsd@yahoo.es>
  0 siblings, 2 replies; 3+ messages in thread

From: Xevi Serrats @ 2003-05-11 13:16 UTC (permalink / raw)
  To: pgsql-performance

Hi,

I have created a table whit some indexes. I analize
the query of this table and never use index.

After this, I create a more simplistic table with two
columns and one index and the query uses the index.

Look at this:

pfc=# \d document
                                  Table
"public.document"
   Column   |           Type           |              
      Modifiers
------------+--------------------------+----------------------------------------------------
 codi       | integer                  | not null
default nextval('seq_document'::text)
 nom        | character varying(32)    | not null
 descripcio | text                     |
 formulari  | integer                  |
 fitxer     | character varying(32)    |
 tamany     | integer                  | default -1
 data       | timestamp with time zone | default
('now'::text)::timestamp(6) with time zone
Indexes: document_pkey primary key btree (codi),
         ind_doc1 btree (codi),
         ind_document btree (formulari)
Triggers: RI_ConstraintTrigger_19414,
          RI_ConstraintTrigger_19418,
          RI_ConstraintTrigger_19419,
          actualitzaritemcercadocument,
          altaitemcercadocument,
          baixaitemcercadocument,
          eliminaracldocument,
          eliminaravaluaciodocument

pfc=# explain select * from document where codi=2;
                        QUERY PLAN
----------------------------------------------------------
 Seq Scan on document  (cost=0.00..1.19 rows=1
width=120)
   Filter: (codi = 2)
(2 rows)

This query must use index document_pkey but explain
tells us that the query does a Sequencial scan on
table document.

Look at this simplistic case:

pfc=# \d prova
            Table "public.prova"
 Column |         Type          | Modifiers
--------+-----------------------+-----------
 codi   | integer               | not null
 nom    | character varying(30) |
Indexes: prova_pkey primary key btree (codi)

pfc=# explain select * from prova where codi=1234;
                               QUERY PLAN
-------------------------------------------------------------------------
 Index Scan using prova_pkey on prova 
(cost=0.00..5.99 rows=1 width=37)
   Index Cond: (codi = 1234)
(2 rows)

Now the query uses index, explain tell something about
index scan using index prova_pkey.

What is the diference with two cases? What must I do?
It is a bug? I need do something else? 

Thanks a lot for helping me.

Regards,

Xevi.

_______________________________________________________________
Yahoo! Messenger
Nueva versión: Webcam, voz, y mucho más ¡Gratis! 
Descárgalo ya desde http://messenger.yahoo.es




^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: Yet another question about not use on indexes
@ 2003-05-11 14:08  Richard Huxton <dev@archonet.com>
  parent: Xevi Serrats <tomcatbsd@yahoo.es>
  1 sibling, 0 replies; 3+ messages in thread

From: Richard Huxton @ 2003-05-11 14:08 UTC (permalink / raw)
  To: Xevi Serrats <tomcatbsd@yahoo.es>; pgsql-performance

On Sunday 11 May 2003 2:16 pm, Xevi Serrats wrote:
> Hi,
>
> I have created a table whit some indexes. I analize
> the query of this table and never use index.

> pfc=# \d document
>                                   Table
> "public.document"
>    Column   |           Type           |
>       Modifiers
> ------------+--------------------------+-----------------------------------
>----------------- codi       | integer                  | not null
> default nextval('seq_document'::text)
>  nom        | character varying(32)    | not null
etc...

> pfc=# explain select * from document where codi=2;
>                         QUERY PLAN
> ----------------------------------------------------------
>  Seq Scan on document  (cost=0.00..1.19 rows=1
> width=120)
>    Filter: (codi = 2)
> (2 rows)

1. Have you done a VACUUM ANALYSE?
2. How many rows are in this table?
3. Can you post the output of EXPLAIN ANALYSE SELECT... - that actually runs 
the query.

-- 
  Richard Huxton




^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: Yet another question about not use on indexes
@ 2003-05-11 14:15  Tom Lane <tgl@sss.pgh.pa.us>
  parent: Xevi Serrats <tomcatbsd@yahoo.es>
  1 sibling, 0 replies; 3+ messages in thread

From: Tom Lane @ 2003-05-11 14:15 UTC (permalink / raw)
  To: Xevi Serrats <tomcatbsd@yahoo.es>; +Cc: pgsql-performance

=?iso-8859-1?q?Xevi=20Serrats?= <tomcatbsd@yahoo.es> writes:
> pfc=# explain select * from document where codi=2;
>                         QUERY PLAN
> ----------------------------------------------------------
>  Seq Scan on document  (cost=0.00..1.19 rows=1 width=120)
>    Filter: (codi = 2)
> (2 rows)

Judging from the cost estimate, this table is too small to bother with
an indexscan.

			regards, tom lane




^ permalink  raw  reply  [nested|flat] 3+ messages in thread


end of thread, other threads:[~2003-05-11 14:15 UTC | newest]

Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2003-05-11 13:16 Yet another question about not use on indexes Xevi Serrats <tomcatbsd@yahoo.es>
2003-05-11 14:08 ` Richard Huxton <dev@archonet.com>
2003-05-11 14:15 ` Tom Lane <tgl@sss.pgh.pa.us>

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