Re: insert to projection

Vadim Tropashko wrote:
> On Sep 6, 11:40 am, paul c <toledobythe..._at_oohay.ac> wrote:

>> If I read that worksheet right, it means we can't delete <p 1, q 'a'>
>> from a  conventional table or relvar and expect <p 1> to be negated,
>> which I don't argue with.  (whereas I don't think anybody disagrees that
>> we already can delete <p 1> and expect <p 1, q 'a'> to be negated.)

Retracting/negating the proposition the tuple stands for.

View update expresses increments/
> decrements of the base relations in terms of increments/decrements of
> the view(s). Compare it to (a relatively simple) view maintenance
> problem where increments/decrements of the view(s) in terms of
> increments/decrements of the base relations. Maintenance of the
> projection view (or any other relational expression for that matter)
> is easy: just take a projection of any incremant/decremant of the base
> relation.
>

No argument that it's conceptually easy, but in practice the base could have a dozen such projections. Received on Sun Sep 06 2009 - 21:35:19 CEST