Date: Sun, 1 Jun 2008 12:37:50 -0400
"paul c" <toledobysea_at_ac.ooyah> wrote in message news:Ulz0k.25$i01.23_at_pd7urf2no...
> Brian Selzer wrote:
>> "paul c" <toledobysea_at_ac.ooyah> wrote in message
> ...>>> predicate must be disjunctive, but I'd say that the machinery of the RM
>>> When the relations in a union view expression are union-compatible
>>> (so-called), I wish somebody could show me how any disjunction is
>>> involed in the tuples of the union view. Many people say that the
>>> has completely discounted that.
>> I don't see how that could be. If you start out with two base relations
>> with the same heading, then each of those relations has its own distinct
>> predicate. The predicate of a view that is the result of the union of
>> the two base relations is obviously just a disjunction of the predicates
>> of the base relations, and each tuple in the view originated from either
>> or both of the base relations, and thus satisfies at least one but
>> possibly both of the predicates of the base relations.
> > > I'd say you are not alone. However, take the union of two base relations > and assign it to another base relation (as opposed to a view). Would you > say the predicate of the result is also disjunctive? >
Not necessarily. The predicate of the base relation on the left-hand side of the assignment would supercede the predicate of the derived relation on the right-hand side. But how what is assigned came about is not recorded in the database, so the fact that an assignment is the result of a union has no bearing on the predicate of the target of the assignment. The same resulting database value could have just as easily been the result of an insert that just coincidentally happened to match the union of the original two base relations. Received on Sun Jun 01 2008 - 18:37:50 CEST