With my limited knowledge of Postgresql, I created the indices myself.
However, it did not improve the performance much.
I am trying to get a few benchmark numbers and would post them asap.
./harshal
On Wed, 24 Oct 2007 Richard Dale wrote :
>On Monday 22 October 2007 07:36:01 Harshal Ganpatrao Hayatnagarkar wrote:
> > Hi,
> >
> > I was working on this problem and finally could arrive with solution.
> >
> > First thing to mention is - '<myns:Person rdf:ID="harshal">' does not work.
> > It has to be '<rdf escription rdf:about="harshal"><rdf:type
> > rdf:resource="Person">' and so on.
> >
> > Secondly, I could not extract better performance from 'Redland +
> > Postgresql', which I got from RDFLite. So finally I decided to switch to
> > 'Redland + MySQL'. It did not work outrightly, but after making some minor
> > code changes in 'redland.rb' (i.e. by adding another method
> > 'initialize_mysql' immitating 'initialize_postgresql'. It could have been
> > better, like combining them into one method, but I avoided that).
>The Redland Postgres adaptor doesn't create any indexes which is one reason it
>is slow. I did send a patch to the Redland list to create indexes, but nobody
>commented on it. I noticed afterwards that the indexes for the MySQL adaptor
>are described in a turtle file, and so perhaps there would be more chance of
>getting a patch in for Postgres if it worked the same way.
>
>-- Richard
>_______________________________________________
>ActiveRDF mailing list
>ActiveRDF lists.deri.org
>http://lists.deri.org/mailman/listinfo/activerdf
|