| From: | Sam Mason <sam(at)samason(dot)me(dot)uk> |
|---|---|
| To: | pgsql-general(at)postgresql(dot)org |
| Subject: | Re: question on serial key |
| Date: | 2009-05-22 14:33:34 |
| Message-ID: | 20090522143334.GC5407@samason.me.uk |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Fri, May 22, 2009 at 08:41:46AM -0500, Brandon Metcalf wrote:
> I am looking for criteria on deciding whether or not to use a serial
> (auto-incrementing) key for rows in a table.
Wow, that's the second time today someone asked that!
> Intuitively, it's pretty clear to me when a serial index is called
> for. Is there a succinct set of guidelines that one could go by?
Not that I'm aware of; it's a fuzzy design choice with benefits and
costs for either option. There are lots of people who arbitrarily
pick one side which tends to make things worse, using one or the other
*exclusively* will add complication. General terms to search for are
Natural keys vs. Surrogate keys.
--
Sam http://samason.me.uk/
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Keith Hayden | 2009-05-22 14:35:39 | How to update stavaluesN columns in pg_statistics (type anyarry) |
| Previous Message | Roderick A. Anderson | 2009-05-22 14:26:35 | Re: question on serial key |