[redland-dev] SPARQL, Redland and circular queries

[Kieron Taylor]

Here's a teaser:

I've got some data that forms a fairly complicated network. The 
predicate <knows> needs to be symmetric. Can I get Redland to infer 
using and OWL file that

<Bob> <knows> <Harry> .

from the fact that

<Harry> <knows> <Bob> . ?

Now if

<Harry> <knows> <Alice> .
<Alice> <knows> <Charles> .

Let's say I wish to know all the possible connections for people who are 
once removed from each other. Using SPARQL I might ask:

SELECT ?person,?thirdperson WHERE
{	?person <knows> ?secondperson .
	?secondperson <knows> ?thirdperson .
}";

This will return:
?person	?thirdperson
Bob	Alice
Bob	Bob
Harry	Charles
Harry	Harry

The problem is that I don't know how to say that ?person may not be the 
same as ?thirdperson. It is easy to add a ?person != ?thirdperson 
requirement, but this becomes exceedingly awkward if I am looking for 
links to a fourth, fifth, sixth person etc. Soon my query has as many != 
permutations as I am getting answers.

Using the DISTINCT keyword does not seem to help. Is there some function 
of SPARQL that I am missing, or do I need to write my own algorithm to 
use Redland's functions and exclude routes manually?

Thanks

Kieron