Re: Are there performance advantages in storing bulky field in separate table?

From: Chris Browne <cbbrowne(at)acm(dot)org>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: Are there performance advantages in storing bulky field in separate table?
Date: 2009-04-08 20:31:33
Message-ID: 87fxgikhui.fsf@dba2.int.libertyrms.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

ianmayo(at)tesco(dot)net (Ian Mayo) writes:
> On Wed, Apr 8, 2009 at 8:13 PM, Robert Treat
> <xzilla(at)users(dot)sourceforge(dot)net> wrote:
>> Maybe I've been reading too much Pascal again lately, but if only 1% of your
>> rows are going to have data in this column, personally, I'd put it in a
>> separate table.
>
> thanks for that Robert - it does match my (completely groundless)
> first impression.
>
> In the nature of debate, would you mind passing on the pascal-related
> reasons why you'd put the data in another table?

Fabian Pascal's thesis is that you shouldn't have NULLs altogether, as
this leads to having to support the 3-or-more-valued logic of NULLs.
The "Third Manifesto" declines to support having NULLs in relations.

Hugh Darwen wrote the relevant paper explaining how to avoid them:
http://web.onetel.com/~hughdarwen/TheThirdManifesto/Missing-info-without-nulls.pdf

I tend to agree that it is reasonable to go to *some* effort to avoid
having NULL values.

Unfortunately, it seems to me that Darwen's papers elaboration on the
issue doesn't present a solution that is without points to criticize.
He does nicely describe how you may indicate various reasons why you
might have missing information. This is both good and bad...

- It is good because it provides unambiguous ways to determine why
the data was missing. NULL leaves that ambiguous.

- It is Not So Good because it replaces the 3-value-logic of NULLs
with an "as many values for logic as we have kinds of unknown
values," which is more like a 5- or 6-value logic.

http://en.wikipedia.org/wiki/Fabian_Pascal

"Pascal is known for his sharp criticisms of the data management
industry, trade press, current state of higher education, Western
culture and alleged media bias. Pascal advocates strict adherence to
the principles of the relational model, and argues that departing from
the model in the name of pragmatism is responsible for serious data
management troubles. Criticism of Pascal's advocacy often centers
around his polemical style, which some perceive as overly
confrontational and unprofessional.

He has retired from the technological industry and now does political
commentary, specially on Middle East issues."
--
let name="cbbrowne" and tld="cbbrowne.com" in name ^ "@" ^ tld;;
http://linuxdatabases.info/info/linuxxian.html
"The only thing better than TV with the sound off is Radio with the
sound off." -- Dave Moon

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Tom Lane 2009-04-08 20:42:06 Re: some external sql not working in psql
Previous Message James B. Byrne 2009-04-08 20:19:18 Re: No return from trigger function