I have implemented random_id generation
which solves 2 problems:
(Link to explination: http://earnestberry.com/node/13 )
The only draw back is on the DB side…if
you’re writing stored procedures, you don’t have a sequence or
counter table to look at. However, one can easily solve this by implementing
the methodology in a sp (stored-proc).
From:
development-bounces@drupal.org [mailto:development-bounces@drupal.org] On Behalf Of Amr Mostafa
Sent: Wednesday, April 25, 2007
11:21 AM
To: development@drupal.org
Subject: Re: [development] Why do
we avoid auto-incrementing db columns?
I believe one of the
reasons is that many times you want to know the ID of your new record. For
example, you inserted a new file to the 'files' table, and you would like to
store a reference to it in your xyz table.
For mysql there is mysql_insert_id(), but not for others.
On 4/25/07, Syscrusher
<syscrusher@4th.com > wrote:
On Wednesday 25 April 2007 10:29, Ezra B. Gildesgame wrote:
> I happened to discuss this with Greg Knaddison yesterday.
> MySQL 4 does not support auto-incrementing columns.
I must respectfully disagree. I happened to have a copy of the MySQL
reference book from version 3.22 sitting on my bookshelf, and AUTO_INCREMENT
was supported even then. :-) Never mind the question of why I still have such
an old book....
I also checked a PostgreSQL reference book, and they offer a SERIAL
column type that works very similarly to the AUTO_INCREMENT in MySQL. I don't
know how far back this support goes; my book is for PG version 8.something and
was copyright 2006.
I can't comment on the Microsoft SQL Server implementation of this, if any,
because I don't have any manual for that here.
Scott
--
-------------------------------------------------------------------------------
Syscrusher (Scott
Courtney) Drupal
page: http://drupal.org/user/9184
syscrusher at 4th dot
com Home
page: http://4th.com/