| From: | Richard Huxton <dev(at)archonet(dot)com> |
|---|---|
| To: | Phoenix Kiula <phoenix(dot)kiula(at)gmail(dot)com> |
| Cc: | Richard Broersma Jr <rabroersma(at)yahoo(dot)com>, Postgres General <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: Partial index with regexp not working |
| Date: | 2007-09-12 12:34:55 |
| Message-ID: | 46E7DCEF.6080806@archonet.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Phoenix Kiula wrote:
>
> Ok, I've hit a snag about this index. I think it's to do with how my
> regex is structured. Basically this column can have either IP
> addresses, or alphanumeric user IDs. If it is not an IP address, it is
> a registered user ID. What is the best way of ascertaining that a
> column value is *not* an IP address?
>
> I tried this:
>
> select * from trader where trader_id !~ '[0-9]+\.[0-9]+\.[0-9]+\.';
>
> And this works, but I wonder if a partial index on a negative
> condition ("!~") will be slower than a positive condition?
To be honest, I'd probably just have a separate column "uid_type", set
it when creating the user and then just have a partial index WHERE
uid_type='IP'
--
Richard Huxton
Archonet Ltd
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Asko Oja | 2007-09-12 12:46:25 | Re: Scalability Design Questions |
| Previous Message | Richard Huxton | 2007-09-12 12:33:24 | Re: Database/Table Design for Global Country Statistics |