Re: SERIAL type feature request

From: Jan Wieck <JanWieck(at)Yahoo(dot)com>
To: Zoltan Boszormenyi <zboszor(at)dunaweb(dot)hu>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: SERIAL type feature request
Date: 2005-12-05 00:50:18
Message-ID: 43938ECA.9080301@Yahoo.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 12/4/2005 5:10 PM, Zoltan Boszormenyi wrote:
> I found this in the SQL2003 draft:
>
> "
> 4.14.7 Identity columns
>
> ... An identity column has a start value, an increment, a maximum value,
> a minimum value,
> and a cycle option. ...
> "
>
> The exact properties of a sequence. It would be a good idea to be able
> to provide
> all these the same way PostgreSQL provides CREATE SEQUENCE.

I think nobody would object to implementing support for the SQL2003
syntax. Most of that would be providing all the values that will get
forwarded into the internal sequence generation during CREATE TABLE.

The other thing needed is an extension to the default value mechanism
that overrides any given value to implement GENERATE ALLWAYS. Not too
hard either.

> Thinking about it more, 0 is a special value that a sequence created
> with defaults
> (just like the ones created for SERIAL fields) will not produce. If
> PostgreSQL

Zero is no more special than any other value and the standard you
pointed to does not talk about it either. If we implement IDENTITY as
per standard, you will either omit the value, specify DEFAULT or declare
the column GENERATE ALLWAYS.

If we had to pick any magic value I would vote for skipping 666 in all
sequence generators and use that.

Jan

--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck(at)Yahoo(dot)com #

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Michael Glaesemann 2005-12-05 00:53:21 Re: SERIAL type feature request
Previous Message Andrew Dunstan 2005-12-04 23:22:54 Re: SERIAL type feature request