From: | Ron Peterson <rpeterson(at)yellowbank(dot)com> |
---|---|
To: | Fred_Zellinger(at)seagate(dot)com |
Cc: | pgsql-general(at)hub(dot)org |
Subject: | Re: Large Tables(>1 Gb) |
Date: | 2000-06-30 19:25:13 |
Message-ID: | 395CF419.EDFDEB12@yellowbank.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Fred_Zellinger(at)seagate(dot)com wrote:
> However, there is still something bugging me. Even though many people
> related stories of 7.5 Gb+ Dbs, I still can't make that little voice in me
> quit saying "breaking things into smaller chunks means faster work"
> theories.
>
> There must exist a relationship between file sizes and DB performance.
If your data doesn't completely fit into main memory, at least some of
it will have to be saved off-line somewhere. Your question is: should
the off-line portion be split into more than one file to speed
performance?
I won't try to be precise here. There are good textbooks on the subject
if your interested. I've just been reading one, actually, but it's at
home and I don't remember the name :( Knuth would of course be good
reading on the subject.
Maybe think of it this way: what's the difference between one file and
two, really? You've basically just got a bunch of bits on a block
device, either way. By saving your data to a single file, you have more
control of the data layout, so you can organize it in the manner most
appropriate to your needs.
________________________
Ron Peterson
rpeterson(at)yellowbank(dot)com
From | Date | Subject | |
---|---|---|---|
Next Message | Mitch Vincent | 2000-06-30 20:45:16 | Trigger programming.. |
Previous Message | Gilles DAROLD | 2000-06-30 19:17:10 | CURSOR problem |