From: Paul Gearon <gearon@ieee.org>

Date: Wed, 8 Jul 2009 14:20:06 -0500

Message-ID: <a25ac1f0907081220g6021be7er440d6d9f5dfd187b@mail.gmail.com>

To: Simon Schenk <sschenk@uni-koblenz.de>

Cc: "public-rdf-dawg@w3.org" <public-rdf-dawg@w3.org>

Date: Wed, 8 Jul 2009 14:20:06 -0500

Message-ID: <a25ac1f0907081220g6021be7er440d6d9f5dfd187b@mail.gmail.com>

To: Simon Schenk <sschenk@uni-koblenz.de>

Cc: "public-rdf-dawg@w3.org" <public-rdf-dawg@w3.org>

On Wed, Jul 8, 2009 at 11:31 AM, Simon Schenk<sschenk@uni-koblenz.de> wrote: > Hi, > > I've been fiddling with algebra expressions for MINUS and I am wondering > whether the following capture the semantics of antijoin with restriction > to shared variables: > > A \ (A |X| [Projection to vars in A] B) > > If A and B do not share a variable, the projection results in a relation > with zero colums. I am not sure, whether this is even defined. > > Does anyone of the relational algebra gurus out there have a clue? Not a guru, but this is similar to how I described it in that link no one ever comments on. :-) In my case, I describe it as: A intersection ([project to vars in A] complement (A |X| B)) Where "complement" is defined in terms of the known data (ie. the assertions in the database). By this definition, "intersection-complement" should be equivalent to your \ operator. Also, the projection can be done before or after the inner join and complement with the same result. So yes, I agree with you. (BTW, the actual notation I used was: A - B = A ∩ π(¬'(A ⋈ B)) but I don't know how well that will show up in email) Regards, PaulReceived on Wednesday, 8 July 2009 19:20:48 UTC

*
This archive was generated by hypermail 2.3.1
: Wednesday, 7 January 2015 15:00:57 UTC
*