Re: View Updates Problem Statement

"Mikito Harakiri" <mikharakiri_at_ywho.com> wrote in message news:BBa7b.29$741.241_at_news.oracle.com...
> "Bob Badour" <bbadour_at_golden.net> wrote in message
> news:O3b7b.733$QQ.69871762_at_mantis.golden.net...
> > "Mikito Harakiri" <mikharakiri_at_ywho.com> wrote in message
> > news:Yw97b.27$741.210_at_news.oracle.com...
> > > 1. We need to consider more than one view in order to be able solve
> > > ambiguous cases.
> >
> > That doesn't make any sense because the user did not mention any other
> view.
>
> That entirely depends upon view update context. Suppose we have "data
> integration" problem.
>
> Database A has relations RA1, RA2, RA3, ...
> Database B has relations RB1, RB2, ...
>
> Views V1, V2, V3, ... establish the mapping between the 2 databases:
>
> V1*RA1 = RB1
> V2*RA2 = RB2
> ...
>
> Problem. Translate each update in database B into database A.
>
> We, therefore, have at least one view update problem with many views.

Insofar as the user names multiple relvars while updating B, the update involves multiple views. If the user names only a single relvar in B, the update involves only a single view.

> Now let's change "relations in the database A" to "base relations", and
> "relations in the database B" into "views" (in the same database). That
> sounds like original view updates formulation.
>
> What is wrong with user knowing the effect of update on a set of views V1,
> V2, ... ?

I never claimed there is anything wrong with a user having such knowledge; although, I find it unreasonable for a dbms to require more knowledge of a user than is necessary.

> > > 3. Constraints are views that evaluate to constants. For example, if
we
> > > define
> > >
> > > view V3
> > > select max(sal) from (
> > > select sal from emp
> > > union
> > > select 1000 from DEE
> > > )
> > >
> > > then constraint "each emp has salary no more than 1000" can
> algebraically
> > be
> > > written as
> > >
> > > V3*x = 1000
> > >
> > > where "V3*x" means view V3 applied to a set of base relations x, and
> > "1000"
> > > means an output relation that has one tuple with a single value 1000.
> >
> > What set of relations are you talking about? V3 references emp and DEE.
Or
> > are you talking about the possible values of emp and DEE?
>
> Date's DEE is a constant relation containing one tuple. I introduced it in
> order to transform inequality into equality. Emp is the only relation
> variable that is used in the V3. But it can be many. We can assume that a
> view input is the set of all base relations, while view output doesn't
> necessarily depend upon all of them.

So, you were talking about a set of relvars. Is that a correct interpretation? Received on Tue Sep 09 2003 - 17:21:46 CEST