Re: [HACKERS] Fwd: Joins and links

From: Leon <leon(at)udmnet(dot)ru>
To: Bob Devine <devine(at)cs(dot)utah(dot)edu>
Cc: pgsql-hackers(at)postgreSQL(dot)org
Subject: Re: [HACKERS] Fwd: Joins and links
Date: 1999-07-08 06:37:48
Message-ID: 3784473C.317AD3A7@udmnet.ru
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Bob Devine wrote:

> Beware of adding special purpose hard-links as a way to
> skip the run-time value comparisons. A link looks attractive
> but it really only works for one-to-one relationships
> (any multi-way relationships would require a list of links
> to follow)

Not exactly. If you have a fixed set of links it a tuple, you
don't have to follow the list of them.

> and a link has all of the overhead that a
> foreign key requires.
>

We looked at the matter carefully and found no overhead like
foregn key's. Maybe you should read the thread more carefully
once again.

> As somone who has developed several commercial dbms systems,
> I would discourage doing a special "link" type. There are
> other ways to gain performance -- de-normalize your tables
> if you are doing mainly reads;

If I denormalize my tables, they will grow some five to ten
times in size.

But simply think what you are proposing: you are proposing
exactly to break RDBMS "alphabet" to gain performance! This
means that even SQL warriors see RDBMS's ideology as not
proper and as corrupt, because it hinders performance.

You are contradicting yourself!

> carefully check your storage
> layout; and, of course, buy more RAM ;-)

And what will I do with performance loss from bloated tables?

--
Leon.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Mount 1999-07-08 08:30:24 RE: [HACKERS] Updated TODO list
Previous Message Gene Sokolov 1999-07-08 06:37:02 Re: [HACKERS] Updated TODO list