Message-ID: <40FFC654.9000903@cc.com.pl>
Date: Thu, 22 Jul 2004 15:51:16 +0200
From: Michal Dobaczewski <mdobaczewski@cc.com.pl>
User-Agent: Mozilla Thunderbird 0.7.2 (Windows/20040707)
MIME-Version: 1.0
To: Peter Eisentraut <peter_e@gmx.net>
Cc: pgsql-interfaces@postgresql.org
Subject: Re: libpq: usage of PQoidValue to obtain serial primary
References: <40FFA197.9080106@cc.com.pl> <200407221444.56693.peter_e@gmx.net>
In-Reply-To: <200407221444.56693.peter_e@gmx.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Peter Eisentraut wrote:

> Am Donnerstag, 22. Juli 2004 13:14 schrieb Michal Dobaczewski:
> 
>>Browsing the general list I've just read a post claiming that oid are
>>not guaranteed to be unique in the table. This is also stated in the
>>documentation - which seems to have escaped me this far. So I assume the
>>approach presented above is flawed.
>>
>>I would like to ask what is a most elegant way to deal with such issues?
> 
> 
> Create a unique constraint on the oid column.

Thanks for the response.

I wonder: how will postgres behave if it happens to generate a 
repetitive oid for a table with such constraint? I understand it will 
work it out somehow internally and get a different oid, but I would like 
to be sure. It doesn't create a risk of inserts failing at random, does it?

I also understand this theoretically limits the number of rows in a 
table to 2^32 whereas without such constraint there is no set limit - is 
that true? It's not a problem, we don't have such big tables so far but 
it would be interesting to know.

Regards,

Michal Dobaczewski.