Re: Events

From: jlfoster <jlfoster_at_sprintpcs.kawm.INVALID>
Date: Thu, 16 Nov 2006 21:31:55 -0800
Message-ID: <_bOdned_OZUH1sDYnZ2dnUVZ_sCdnZ2d_at_got.net>


"Cimode" <cimode_at_hotmail.com> wrote in message news:1163343281.254371.64670_at_k70g2000cwa.googlegroups.com...

> VJ Kumar a écrit :

> > "Dmitry A. Kazakov" <mailbox_at_dmitry-kazakov.de> wrote in
> > news:1mm81xinxq98n.1jeg2su8hzd9g.dlg_at_40tude.net:
> >
> > .......
> > > Compare it with selection of two adjacent
> > > records in a relational table. This also breaks the abstraction of
> > > independent elements on which a relation were defined.
> >
> > I think that you are not right. The relational model has no conception
> > of 'records' much less adjacent 'records', even SQL talks about rows
> > rather than records. The relational model operates with sets, not with
> > individual tuples although API might.

> > > The world is imperfect.

> Adjacency is not a relational concept.

Rows representing such things as invoices might need to capture a sequential "invoice number" and the like. There could be a need to ask for "adjacent" invoices. In short, assign some sort of sequence number and query on that. Received on Fri Nov 17 2006 - 06:31:55 CET

Original text of this message