Hello Kjetil, I'll test the case with slightly updated Virtuoso in 2 hrs. It is really strange to see a report about DESCRIBE that returns redundant data because typically reported cases demonstrate lack of triples. (When DESCRIBE query does not specify the graph, graphs are ignored if they do not declare rdf:type of a subject, that's confusing).
Best Regards, Ivan. On Tue, 2008-12-16 at 15:44 +0100, Kjetil Kjernsmo wrote: > On Monday 15 December 2008 18:15:38 Hugh Williams wrote: > > Can you provide a simple dataset and test program to recreate this > > problem, as it we can see the problem first hand then it can be fixed ? > > Now I have tried to provide a minimal example all day to see if I can > reproduce the problem with a minimal amount of data, and a minimal servlet, > but I have failed. > > The obvious way to produce the minimal dataset was to run the query against > the old endpoint, which would give me RDF/XML-ABBREV that represented the > expected graph. So that's what I did, and I have attached that expected file. > And guess what, when I stuff this into Virtuoso, it works just as expected > (except for the namespace prefixes, but they doesn't matter)! > > When I reload my entire dataset, then it falls back to the old behaviour. In > this case, the serialisation which in RDF/XML-ABBREV is 6.4 KB becomes more > than 2 MB, so this is a big deal, and that's just surplus XML markup, it is > actually less data. > > So, I have tried to load just smaller parts of the dataset, but that's been > unsuccessful too... > > I'm not sure the problem is in any Virtuoso code at all. As previously > stated, > I have verified that the model resulting from the query execution contains > the triples that are missing in the final serialised form. I attach a > stripped down version of the servlet that does the serialisation. As you can > see, there is some code there that's commented out that will list the > rdf:types. These are missing from the serialisation. > > Even though I have failed to provide a stripped down example, this is all > open > source and code and data are on the web. To load the whole dataset, you would > need: > https://submarine.computas.com/sublima/trunk/blocks/sublima-app/src/main/resources/rdf-data/sublima-ns.ttl > https://submarine.computas.com/sublima/trunk/blocks/sublima-app/src/main/resources/rdf-data/smil-test-data.ttl > https://submarine.computas.com/sublima/trunk/blocks/sublima-app/src/main/resources/rdf-data/lingvoj-excerpt.rdf > https://submarine.computas.com/sublima/trunk/blocks/sublima-app/src/main/resources/rdf-data/information-model.ttl > > I used the following statements to import them: > SQL> ttlp (file_to_string_output > ('/usr/local/src/sublima/blocks/sublima-app/src/main/resources/rdf-data/information-model.ttl'), > 'http://rabbit.computas.int:8180/sublima-webapp-1.0-SNAPSHOT/', 'smiltest', > 0); > Done. -- 283 msec. > SQL> ttlp (file_to_string_output > ('/usr/local/src/sublima/blocks/sublima-app/src/main/resources/rdf-data/smil-test-data.ttl'), > 'http://rabbit.computas.int:8180/sublima-webapp-1.0-SNAPSHOT/', 'smiltest', > 0); > Done. -- 11990 msec. > SQL> ttlp (file_to_string_output > ('/usr/local/src/sublima/blocks/sublima-app/src/main/resources/rdf-data/sublima-ns.ttl'), > 'http://rabbit.computas.int:8180/sublima-webapp-1.0-SNAPSHOT/', 'smiltest', > 0); > Done. -- 33 msec. > SQL> DB.DBA.RDF_LOAD_RDFXML (file_to_string > ('/usr/local/src/sublima/blocks/sublima-app/src/main/resources/rdf-data/lingvoj-excerpt.rdf'), > 'http://rabbit.computas.int:8180/sublima-webapp-1.0-SNAPSHOT/', 'smiltest'); > > The attached servlet can also be found in (a work in progress) > https://submarine.computas.com/mediasone-maven/trunk/backend/ > > I'm really sorry I couldn't find something that's easier to debug, but I > wanted to share the status with you nevertheless. I'll continue to work on > this problem. > > Kind regards > > Kjetil Kjernsmo > ------------------------------------------------------------------------------ > SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. > The future of the web can't happen without you. Join us at MIX09 to help > pave the way to the Next Web now. Learn more and register at > http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ > _______________________________________________ Virtuoso-users mailing list > Virtuoso-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/virtuoso-users
