The blank node (1) just means I don't care what the subject is (it will
match a blank node or a URI-identified resource - it cannot be a literal
as it's in subject position), just that it's the centre of the molecule
and has a name attached to it.
I'm not sure you can set the parameters and build the index (2) all
within one update/transaction.
Barry
On 06/02/13 01:49, Siow Boon Lin Eugene wrote:
Hi,
Thanks for the reply Barry!
I could not get this to work.
1.
SELECT ?name
{[] luc:myIndex "eugene male"; :hasName ?name}
Why the bnode?
The OWLIM documentation states using ?o luc:myIndex "eugene" and ?o is a
literal. Here the bnode is E1?
2. So this will only work with molecule size of 2? I suspected it could be a
problem with my index, so I included my update query that created the index.
PREFIX luc:<http://www.ontotext.com/owlim/lucene#>
INSERT DATA {
luc:moleculeSize luc:setParam "2".
luc:index luc:setParam "literals, bnodes".
luc:myIndex luc:createIndex "true".
}
I am not using any bnodes, however, I included it in case this was necessary to
do the query you provided.
So I tried this with a variety of options to see if the size of my index will
change in 'AppData/OpenRDFSesame/repositories/repo/storage/lucene'
If I change the molecule size to 3 or 4, the size of the index does not change,
which is strange.
If I choose bnodes alone it reduces from 80mb to 60mb.
If I choose literals and bnodes, literals alone, with moleculesize 2, 3,4 - the
index remains at 80mb.
Am I doing anything wrong during the index creation?
Regards,
Eugene
--------------------------------------------------
The beauty of the Lucene search is that you can combine it with graph patterns,
e.g. (presuming you created myIndex):
SELECT ?name
{[] luc:myIndex "eugene male"; :hasName ?name}
Barry
On 04/02/13 09:00, Siow Boon Lin Eugene wrote:
Hi,
As I understand, when I create the Lucene index, I create RDF
molecules and if I have specified a molecule size of 2, this means it
will index the surrounding nodes till 2 hops away from the literal.
Is this right?
As such, is it possible to use proximity in Lucene search to do
something like this:
Suppose I have:
?s | ?p | ?o
E1 | hasName | "Eugene"
E1 | hasGender | "Male"
Can I use this Lucene query "eugene male"~2 and retrieve the ?o "Eugene"?
If not, what would be a good way to do this to exploit the properties
of the RDF molecule.
Thanks!
Eugene
_______________________________________________
Owlim-discussion mailing list
Owlim-discussion@ontotext.com
http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://ontomail.semdata.org/pipermail/owlim-discussion/attachments/20130204/d80eb106/attachment-0001.html>
------------------------------
_______________________________________________
Owlim-discussion mailing list
Owlim-discussion@ontotext.com
http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
End of Owlim-discussion Digest, Vol 13, Issue 3
***********************************************
_______________________________________________
Owlim-discussion mailing list
Owlim-discussion@ontotext.com
http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
_______________________________________________
Owlim-discussion mailing list
Owlim-discussion@ontotext.com
http://ontomail.semdata.org/cgi-bin/mailman/listinfo/owlim-discussion
