Re: AW: Re: postgres TODO - Mailing list pgsql-hackers

From The Hermit Hacker
Subject Re: AW: Re: postgres TODO
Date
Msg-id Pine.BSF.4.21.0007101140130.3314-100000@thelab.hub.org
Whole thread Raw
In response to AW: Re: postgres TODO  (Zeugswetter Andreas SB <ZeugswetterA@wien.spardat.at>)
Responses Re: AW: Re: postgres TODO  (Bruce Momjian <pgman@candle.pha.pa.us>)
List pgsql-hackers
On Mon, 10 Jul 2000, Zeugswetter Andreas SB wrote:

> 
> 
> > > Peter Eisentraut wrote:
> > > 
> > > > Bruce Momjian writes:
> > > > 
> > > > >       * Add function to return primary key value on INSERT
> > > > 
> > > > I don't get the point of this. Don't you know what you 
> > inserted? For
> > > > sequences there's curval()
> > > 
> > > Mmmhhh... it means that we can assume no update to the 
> > sequence value
> > > between the insert and the curval selection?
> > 
> > No curval() is per-backend value that is not affected by other users. 
> > My book has a mention of that issue, and so does the FAQ.
> 
> Not all default values need to be a sequence, thus imho
> this function is a useful extension. ODBC has it too.

actually, had thought about this too over the weekend ... if I define a
'serial' type, it right now creates a sequence for that ... if I recall
correctly, that was purely a kludge until someone built a better 'serial'
...

having an INSERT return the value of 'serial' that was used would save a
second SELECT call *and* eliminate the requirement for an app programmer
to have to know to do a 'SELECT curval('table_field_seq');' ...




pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Progress report: intraquery memory recovery in executor
Next
From: Thomas Lockhart
Date:
Subject: Re: Templates