Sorry Guy,
was just a little tired yesterday.
> Err... you're right... one of us say the same thing when I show the
> Volker mail...
Try to make a group by in the inline-view, so you will get something
like this:
>
> On Mon, 4 Aug 2003 12:02:46 -0300, "Fernando Papa" <[EMAIL PROTECTED]>
> wrot
On Mon, 04 Aug 2003 16:10:18 +0200, I wrote:
>SELECT DISTINCT ON (
>cp.id_instalacion,
>cp.id_contenido,
>cp.generar_vainilla,
>cp.fecha_publicacion
> )
Cut'n'paste error! fecha_publicacion should no
"Fernando Papa" <[EMAIL PROTECTED]> writes:
> Thanks Tom. I vaccumed full every night. Now I drop function index and
> change the upper. Nothing change (I know, total time rise because we are
> doing other things on database now).
> -> Seq Scan on cont_publicacion
> (cost=0.0
--Mensaje original-
De: Tom Lane [mailto:[EMAIL PROTECTED]
Enviado el: lunes, 04 de agosto de 2003 18:28
Para: Fernando Papa
CC: [EMAIL PROTECTED]
Asunto: Re: [PERFORM] I can't wait too much: Total runtime 432478.44
msec
"Fernando Papa" <[EMAIL PROTECTED]> writes:
>
"Fernando Papa" <[EMAIL PROTECTED]> writes:
>-> Nested Loop (cost=0.00..1828.46 rows=1 width=367)
> (actual time=7525.51..436843.27 rows=40 loops=1)
> Join Filter: (("inner".id_contenido =
> "outer".id_contenido) AND ("inner".id_instalacion =
> "outer".id_inst
I was play with nested loops, and I found this:
Original explain:
Limit (cost=9.75..9.76 rows=1 width=479) (actual
time=436858.90..436858.93 rows=8 loops=1)
-> Sort (cost=9.75..9.76 rows=1 width=479) (actual
time=436858.88..436858.89 rows=8 loops=1)
Sort Key: cont_publicacion.fe
Err... you're right... one of us say the same thing when I show the
Volker mail...
-Mensaje original-
De: Manfred Koizar [mailto:[EMAIL PROTECTED]
Enviado el: lunes, 04 de agosto de 2003 12:17
Para: Fernando Papa
CC: Volker Helm; [EMAIL PROTECTED]
Asunto: Re: [PERFORM] I can'
On Mon, 4 Aug 2003 12:02:46 -0300, "Fernando Papa" <[EMAIL PROTECTED]>
wrote:
>FROM
>cont_contenido
>,juegos_config
>,cont_publicacion
>,(SELECT max(cp1.fecha_publicacion) as max_pub
> --change here
>
ion cp1 (cost=0.00..12.48 rows=448
width=8) (actual time=0.05..11.62 rows=448 loops=16) Total runtime:
2250.92 msec(20 rows)
The
problem was the subquery, no doubt.
-----Mensaje original-----De: Volker Helm
[mailto:[EMAIL PROTECTED] Enviado el: lunes, 04 de agosto de 2003
11:45Para: Fe
I create the index, but doesn't help too much:
QUERY
PLAN
---
Filter: ((id_instalacion = 2::numeric) AND
(id_sbc = 619::numeric) AND (id_tipo = 2::numeric))
Total runtime: 449765.69 msec
(20 rows)
-Mensaje original-
De: Josh Berkus [mailto:[EMAIL PROTECTED]
Enviado el: viernes, 01 de agosto de 2003 18:32
Para: Christopher Browne; Fe
=1)
Filter: ((id_instalacion = 2::numeric) AND (id_sbc =
619::numeric) AND (id_tipo = 2::numeric))
Total runtime: 465088.66 msec
(21 rows)
-Mensaje original-
De: Christopher Browne [mailto:[EMAIL PROTECTED]
Enviado el: viernes, 01 de agosto de 2003 18:27
Pa
On Fri, 1 Aug 2003 18:17:17 -0300, "Fernando Papa" <[EMAIL PROTECTED]>
wrote:
> AND cont_publicacion.fecha_publicacion = (SELECT
>max(cp1.fecha_publicacion)
> FROM cont_publicacion cp1
> WHERE cp1.id_instalacion =
>cont_publicacion.id_instalacion
>
From: ""Fernando Papa"" <[EMAIL PROTECTED]>
> AND upper(cont_publicacion.generar_Vainilla) = 'S'
> Filter: ((upper((generar_vainilla)::text) = 'S'::text) AND
(fecha_publicacion = (subplan)))
using a functional index on this field should help
create index idx_generar_vainilla_ci on cont_public
Fernando,
> AND cont_publicacion.fecha_publicacion = (SELECT max(cp1.fecha_publicacion)
> FROM cont_publicacion cp1
> WHERE cp1.id_instalacion = cont_publicacion.id_instalacion
> AND cp1.id_contenido = cont_publicacion.id_contenido
>
I'd point at the following as being a sterling candidate for being a
cause of this being slow...
AND cont_publicacion.fecha_publicacion = (SELECT max(cp1.fecha_publicacion)
FROM cont_publicacion cp1
WHERE cp1.id_instalacion = cont_publicacion.id_instalacion
Title: Mensaje
Hi
all!
Really I don't know
what happened with this query. I'm running PG 7.3.1 on solaris, vaccumed (full)
every nigth.
The cardinality of
each table was:
cont_contenido: 97 rows
juegos_config: 40 rows
cont_publicacion: 446 rows
not huge
tables...
however, this query
17 matches
Mail list logo
