Re: Creating a sequence

From: Craig Ringer <craig(at)postnewspapers(dot)com(dot)au>
To: Chris <cmattmiller(at)gmail(dot)com>
Cc: pgsql-jdbc(at)postgresql(dot)org
Subject: Re: Creating a sequence
Date: 2008-04-23 17:00:51
Message-ID: 480F6B43.8080407@postnewspapers.com.au
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-jdbc

Chris wrote:
> That was my problem, I didn't know where or how to create the sequence.
> Got it to work now. :D
> So, should the standard practice be to create a sequence for every table?

That leads in to the synthetic vs natural primary key debate, which
isn't worth getting into here.

If you have a synthetic key on a table (a "meaningless" integer
identifier generated by the application or database) then a sequence is
an extremely good way of generating such identifiers. The use of a
sequence permits multiple concurrent inserts without risk of conflict
and without the need for locking.

You *WILL* get gaps in your primary key values, though, where
transactions have rolled back. Your application should be designed not
to care about this.

By the way, I pointed out the stupid way of setting a sequence start
value. What I should've written was:

CREATE SEQUENCE employee_emp_uid_seq;
SELECT setval('employee_emp_uid_seq',
(SELECT max(emp_uid) FROM employee));

--
Craig Ringer

In response to

Browse pgsql-jdbc by date

  From Date Subject
Next Message Olaf Tomczak 2008-04-23 19:28:57 Re: Query parameters limit in postgres jdbc driver?
Previous Message Chris 2008-04-23 16:01:42 Login Roles