0

How can I load data from OpenStreetMap of a particulat area (e.g. Berlin) into the (open source) Virtuoso triple store which runs on my local computer (Ubuntu)?

I tried to download the particular OSM file and access it with sparqlify in order to convert it to RDF (or turtle etc.), which then later (at least that was my idea) could be loaded to Virtuoso using the bulk loading strategy. That did not work.

I would be happy if you could tell me if there is any other alternative how I could convert the osm files into RDF.... or, if there is a totally different approach?!

I also thought about using apache jena within Java to access the linkedgeodata access point directly, however, I guess having the data locally gives me much more performance at a later point when I run SPARQL commands.

Cheers

Mr M
  • 69
  • 8
  • [Documentation](http://vos.openlinksw.com/owiki/wiki/VOS/VirtBulkRDFLoader) is your friend. If that doesn't satisfy your need, please [reformulate your question](http://stackoverflow.com/help/how-to-ask). – TallTed Sep 20 '17 at 14:06
  • The question is, whether you have the RDF dump or just the SQL dataset? – UninformedUser Sep 20 '17 at 14:41
  • I edited my question. I hope it makes it more clear what my problem is. I do not manage to convert osm data of a particular city into RDF. The only thing I have got so far is osm data converted into a PostgreDB, which then is accessed via sparqlify... I thought I could write the information encoded in the DB as RDF and then load it into Virtuoso. That did not work out for me so far.. – Mr M Sep 20 '17 at 15:09
  • I hope you know the [LinkedGeoData](http://linkedgeodata.org) project? – UninformedUser Sep 20 '17 at 20:53
  • Yes I do. I actually followed the steps they described to (1) transform osm data into tables in Postgres and (2) use sparqlify to access it via SPARQL. Accessing it in this way is even slower than accessing LinkedGeoData via their online virtuoso web-interface. This is way I decided to establish a local virtuoso triple store as I expect it to excel the performance of sparqlify. Any help would be really appreciated... I think I am missing something out here. – Mr M Sep 21 '17 at 06:46
  • It's probably much slower because 1.) the data set is really large and 2.) you're not deploying it on a similar powerful machine. – UninformedUser Sep 21 '17 at 10:09
  • LinkedGeoData provides a SNORQL web-interface and a Virtuoso web-interface. When invoking the same query in both interfaces the Virtuoso still outperforms by far. So I would not aree. – Mr M Sep 21 '17 at 10:49
  • Maybe I should add: SNORQL uses sparqlify in the background as far as I understood it... – Mr M Sep 21 '17 at 10:57
  • The LinkedGeoData [SNORQL interface](https://github.com/kurtjx/SNORQL) passes your SPARQL queries directly to Virtuoso, which natively handles that SPARQL. From all you've said, I wonder whether we're digging into [an XY Problem](http://meta.stackexchange.com/questions/66377/what-is-the-xy-problem/66378#66378) here, rather than focusing on your real goal (which remains unknown)? – TallTed Sep 22 '17 at 15:02
  • Yes, exactly. I couldn't figure out the solution to the problem yet. – Mr M Sep 22 '17 at 15:19

0 Answers0