Re: Can these constraint be implemented in an RDBMS ?

"ben brugman" <ben_at_niethier.nl> wrote in message news:c22kqb$h0e$1_at_reader10.wxs.nl...
>
> "Alfredo Novoa" <alfredo_at_ncs.es> wrote in message
> news:4044a2a3.20400314_at_news.wanadoo.es...
> > On Tue, 2 Mar 2004 15:19:07 +0100, "ben brugman" <ben_at_niethier.nl>
> > wrote:
> >
> >
> > I don't think they are very complex views.
>
> The number of tables in the view became large.
> The optimiser did not always choose a path as wel as in the 'simpler'
views.
>
>
> >
> > > This makes the views far more complex and probably
> > >very slow because the access has to be evaluated for each row.
> > >(Over and over again).
> >
> > They would not be slow with a decent RDBMS.
>
> First we implemented a view which did check if the user was allowed
> to access certain parts. Because of the response times this gave,
> it was decided that the check would be done once and be valid for
> 24 hours. (Our 'RDBMS'ses where Oracle and SQL-server, I do not
> know if they are 'decent' or 'RDBMS' or a 'decent RDBMS' or not).

Um, would that be like a snapshot? If not, how would it differ? (No, they are not particularly decent. Why do you lack the ability to judge for yourself?)

> thanks,
> ben
>
> >
> > >Do you have a solutions for this type of problem within your 'RDBMS' ?
> >
> > I still have not implemented that :( I don't have a lot of time.
> > Although it is rather easy to implement.
> >
> > >Is access control part of a 'RDBMS' ?
> >
> > It should be.
> >
> > Regards
> > Alfredo
> >
>
>
Received on Thu Mar 04 2004 - 00:11:16 CET