Re: Query results caching? - Mailing list pgsql-general

From Ben-Nes Yonatan
Subject Re: Query results caching?
Date
Msg-id 430D82DF.9040405@canaan.co.il
Whole thread Raw
In response to Re: Query results caching?  ("Jim C. Nasby" <jnasby@pervasive.com>)
List pgsql-general
Jim C. Nasby wrote:

>On Tue, Aug 23, 2005 at 12:27:39AM +0200, Ben-Nes Yonatan wrote:
>
>
>>Jim C. Nasby wrote:
>>
>>
>>>Emptying the cache will not show real-life results. You are always going
>>>to have some stuff cached, even if you get a query for something new. In
>>>this case (since you'll obviously want those indexes there), after some
>>>amount of time you will have most (if not all) of the non-leaf index
>>>pages cached, since they take a fairly small amount of memory and are
>>>frequently accessed. This makes index traversal *much* faster than your
>>>initial case shows, even if you query on something different each time.
>>>Testing with a completely empty cache just isn't that realistic.
>>>
>>>
>>As far as I understand it at my situation where all of the data is
>>deleted and inserted each day from the start (INDEX will get lost with
>>it..) & the endless variety of possible keywords search's & the immense
>>size of the tables, the following reason wont last.. or am I wrong here?
>>
>>
>
>You're wrong - to an extent. Remember that while you're loading all that
>data it's also being cached. Now, some of it will probably end up
>falling out of the cache as all the data is read in, but you certainly
>won't be starting from the clean slate that you're looking for.
>
>
Ok I guess that if all of you are telling me this over and over then it
probably got some point in it :), I guess that I'll just see it work by
time.

Thanks alot again (I really appreciate it),
Ben-Nes Yonatan
Canaan Surfing ltd.
http://www.canaan.net.il


pgsql-general by date:

Previous
From: David Goodenough
Date:
Subject: Re: Postgresql replication
Next
From: Ben-Nes Yonatan
Date:
Subject: Re: Query results caching?