Skip site navigation (1) Skip section navigation (2)

Re: Last inserted id

From: Dave Page <dpage(at)vale-housing(dot)co(dot)uk>
To: pgsql-odbc(at)postgresql(dot)org
Subject: Re: Last inserted id
Date: 2001-11-12 10:36:19
Message-ID: (view raw, whole thread or download thread mbox)
Lists: pgsql-odbc

> -----Original Message-----
> From: Jean-Michel POURE [mailto:jm(dot)poure(at)freesurf(dot)fr] 
> Sent: 12 November 2001 10:06
> To: pgsql-odbc(at)postgresql(dot)org
> Subject: Re: [ODBC] Last inserted id
> At 09:55 12/11/01 +0000, you wrote:
> >'Cos that's not safe either - you might get multiple rows 
> inserted at 
> >the same time - unlikely, but certainly possible.
> In most cases a simple timestamp is enough because you add some other 
> parameter to you query (ex:company name, or whatever field 
> you inserted). 
> For example, if you insert values with First_name=Dave, 
> Second_name=Page, 
> Timestamp=current time and query the resulting row, there is 
> *very little* 
> chance to get rows from other users.
> I guess this is the recommended way (see Access howto).

There is *very little* chance, but very little is not *no chance* (bear in
mind that at this point in time in the example we're discussing no primary
key value is yet known).

I would always suggest that people do a select nextval('seq') followed by a
suitable insert. It requires no locking, is definitely multi user safe and
will *always* correctly identify the row. The only downside is that it
requires a quick select before the insert, but if this minor speed sacrifice
was that much of an issue then VB+ADO+ODBC is possibly not the technology to
use in such a project anyway IMHO.

Regards, Dave.

pgsql-odbc by date

Next:From: Lucas GrijanderDate: 2001-11-12 14:12:42
Subject: Lotus Wordpro
Previous:From: Jean-Michel POUREDate: 2001-11-12 10:06:09
Subject: Re: Last inserted id

Privacy Policy | About PostgreSQL
Copyright © 1996-2017 The PostgreSQL Global Development Group