Re: is pivoted phones view updateable?

From: NENASHI, Tegiri <tnmail42_at_gmail.com>
Date: 10 Nov 2006 08:25:26 -0800
Message-ID: <1163175926.785766.32860_at_h54g2000cwb.googlegroups.com>


Cimode wrote:
> Aloha Kakuikanu a écrit :
>
> > Tchuzhie Giri wrote:
> > > It is not "OK" so long as one does not have your rules of updatability.
> > > You put a pile of symbols on the pile of symbols: it does not have
> > > sense without the rules. Please provide the rules, simple examples
> > > and then one can go to more complex examples.
> >
> > Given two relation variables
> >
> > Y(n, #) -- aka WorkPhones(name, number)
> > and
> > Z(n, #) -- aka CellPhones(name, number)
> >
> > and two relation constants
> >
> > A = {(t='work')}
> >
> > and
> >
> > B = {(t='cell')}
> >
> > is the view
> >
> > X = (Y /\ A) \/ (Z /\ B) -- aka ConsolidatedPhones
> >
> > updateable? There is no minus operator involved. The only two RA
> > operators are join /\ and union \/.
>
> How exactly do you intersect Y and A AND Z and B with no FK? What unary
> relation is t?

It is the private notation of the person that posted:

/\ -- join
\/ -- union

Y join A is Cartesian product, also Z join B:

Let Y be {('y1', 1), ('y2', 2)} then Y join A = {('y1', 1, 'work'), ('y2', 2, 'work)} et cetera.

The question is to arbitrate the insertion of the union with 'work' and 'cell'. C.J. Date has offered the thought and also critiqued it.

--
Tegi
Received on Fri Nov 10 2006 - 17:25:26 CET

Original text of this message