Re: Can these constraint be implemented in an RDBMS ?

From: ben brugman <ben_at_niethier.nl>
Date: Tue, 2 Mar 2004 19:43:57 +0100
Message-ID: <c22kai$eur$1_at_reader10.wxs.nl>


Problems with 'serializable' can be show with less tables. But this thread was not about the 'serializable' problem, but about a constraint problem. The serializable problem can be part of that.
(If you start a new thread, I'll be happy to discuss the serializable problem of Oracle).

Reducing the example to 1-2 tables makes it less complex and that was one of the points made. A situation to complex to implement in an existing RDBMS.

ben brugman.

"Mikito Harakiri" <mikharakiri_at_iahu.com> wrote in message news:Rz31c.24$4E5.432_at_news.oracle.com...
> "ben brugman" <ben_at_niethier.nl> wrote in message
> news:4044ac1c$0$3681$4d4ebb8e_at_read.news.nl.uu.net...
> > The problem is derived from two accounts which together are not
> > allowed to be negative. (Used to illustrate Oracle's failure to
implement
> > serializability (as defined in most dictionaries and in the
> SQL-standard).)
> > Most solutions I have seen involve some application coding. And often
> > a denial that you actually want those constraints.
>
> Discussing serializability might be interesting. Could you please reduce
the
> problem to possibly no more than 1-2 tables? An example involving half
dozen
> tables is overkill.
>
>
Received on Tue Mar 02 2004 - 19:43:57 CET

Original text of this message