From: | Csaba Nagy <nagy(at)ecircle-ag(dot)com> |
---|---|
To: | PFC <lists(at)peufeu(dot)com> |
Cc: | NikhilS <nikkhils(at)gmail(dot)com>, Dimitri Fontaine <dfontaine(at)hi-media(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: COPY Transform support |
Date: | 2008-04-03 14:57:53 |
Message-ID: | 1207234673.8259.45.camel@PCD12478 |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thu, 2008-04-03 at 16:44 +0200, PFC wrote:
> CREATE FLATFILE READER mydump (
> id INTEGER,
> date TEXT,
> ...
> ) FROM file 'dump.txt'
> (followed by delimiter specification syntax identical to COPY, etc)
> ;
Very cool idea, but why would you need to create a reader object
first ? You should be able to use COPY directly with the target table
being omitted, meaning the copy will not pump it's result in the target
but be equivalent to a select... and use it in any place where a select
can be used. This would have absolutely no new syntax, just the rules
changed...
Now that I had a second look you actually need the field definitions to
meaningfully interpret the file, but then why not use a record
specification instead of the table in the normal COPY command ? I'm not
sure if there's any existing syntax for that but I would guess yes...
In any case, such a feature would help a lot in processing input files
based also on other existing data in the DB.
Cheers,
Csaba.
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2008-04-03 15:00:26 | Re: COPY Transform support |
Previous Message | Andrew Dunstan | 2008-04-03 14:55:39 | Re: modules |