On Wed, Jun 21, 2006 at 04:12:08PM +0200, Alban Hertroys wrote:
> Martijn van Oosterhout wrote:
> >ISTM that what would really work well is some kind of "Merge Sort" node
> >that would work by having multiple subnodes which are already sorted
> >and merging them into one sorted list.
>
> Would...
Martijn van Oosterhout wrote:
ISTM that what would really work well is some kind of "Merge Sort" node
that would work by having multiple subnodes which are already sorted
and merging them into one sorted list.
Would... So this isn't available yet?
The planner would use this whenever it saw a
Alban Hertroys wrote:
Jim Nasby wrote:
Probably a better bet would be going to 8.1 and using constraint
elimination.
Maybe you mean constraint exclusion?
If so, is that going to help excluding partitions (basically the same
thing, it seems) from a query based on an ORDER BY and a LIMIT?
On Wed, Jun 21, 2006 at 11:56:27AM +0200, Alban Hertroys wrote:
> Say we take the query I posted:
> "SELECT * FROM mm_posrel ORDER BY number DESC LIMIT 25;"
> and the knowledge that this table is inherited by two other tables, with
> number being unique across them (though PostgreSQL probabl
Jim Nasby wrote:
On Jun 19, 2006, at 7:00 AM, Alban Hertroys wrote:
Now all we need to do is getting MMBase to do its queries like this :P
Probably a better bet would be going to 8.1 and using constraint
elimination.
I searched the documentation, google and wikipedia for "constraint
eli
On 6/19/06, Alban Hertroys <[EMAIL PROTECTED]> wrote:
I found a way that works, and is indeed quite a bit faster. It is evenuglier than what you proposed. The problem wasn't the "order by" in thesubquery, but the "order by" combined with the "union":
sorry, i always forget about the fact that unio
On Jun 19, 2006, at 7:00 AM, Alban Hertroys wrote:
Now all we need to do is getting MMBase to do its queries like this :P
Probably a better bet would be going to 8.1 and using constraint
elimination.
--
Jim C. Nasby, Sr. Engineering Consultant [EMAIL PROTECTED]
Pervasive Software h
hubert depesz lubaczewski wrote:
in case you can't, do something similar to this:
select * from
(
select * from only table_a order by number desc limit 25
union
select * from only table_b order by number desc limit 25
union
select * from only table_c order by number desc limit 25
) x
order by
Hubert said 'do "something similar" to this'... The syntax was
incorrect for sure.
If you wish to be able to do ORDER BY and LIMIT in UNION queries, you
need to make them sub-queries "something similar" to this:
select * from
(
select * from (select * from only table_a order by number desc limi
hubert depesz lubaczewski wrote:
ditch the inheritance. it is no good, and makes everything too
complicated to work with.
Would love to, but that's what the system we use generates (MMBase, for
the record). We can probably rework the generated tables, but it isn't
entirely certain that won't
On 6/16/06, Alban Hertroys <[EMAIL PROTECTED]> wrote:
We really need this solved. Isn't anybody able to shed some light onthis? Is it possible to make this query use an index scan, preferablyw/o disabling sequential scanning?ditch the inheritance. it is no good, and makes everything too complicated
Alban Hertroys wrote:
Hi all,
We're using some 3rd party product that uses inheritence, and the
following query is rather slow on PostgreSQL 7.4.7 (debian stable). Any
suggestions how to speed it up?
We really need this solved. Isn't anybody able to shed some light on
this? Is it possible t
Alban Hertroys wrote:
Hi all,
We're using some 3rd party product that uses inheritence, and the
following query is rather slow on PostgreSQL 7.4.7 (debian stable). Any
suggestions how to speed it up?
A few more datapoints:
- Database was vacuum full analyzed just before the query.
- The same
13 matches
Mail list logo
