SPARQL on SQL Experiences

Legacy questions from answers.semanticweb.com
#1

Hi, i want to set up a SPARQL Endpoint with different RDBMS as a background. Since the data in the RDBMS is updated regulary (and its quite a lot) I wouldn't want to replicate all as RDF. Instead I would like to set up a SPARQL-SQL-Mapper. After reading publications from the W3C on the topic (W3C RDB2RDF) and some research of my own I tried the D2R Server and Virtuoso.
D2R was very easy to configure and is a very nice tool. However it seems to be a little slow (I still want to run some tests on some other DBs that I am currently using)
Virtuoso on the other hand was very uncomfortable to configure i felt. I was so far only able to get a basic example running with the Virtuoso generated mappings. So far that is not useful at all but I was not able to write a custom mapping including all the necessary steps.

So my question (finally) is: What are your experiences with setting up an SPARQL Endpoint on top of (one or many) RDBMS? Is there a best practice? Whats the best software for the task? Any thoughts...

Thanks
Jan

#2

Many approaches use the D2R's SPARQL endpoint (http://www4.wiwiss.fu-berlin.de/bizer/d2r-server/) to expose relational data as SPARQL results. DBLP (http://dblp.l3s.de/d2r/), MusicBrainz (http://dbtune.org/musicbrainz/) etc., etc. use this approach effectively.

#3

Support for R2RML is planned for Virtuoso to provide a higher level mapping layer between relational and RDF / Linked Data concepts, building on the core Meta Schema mapping language used for creating the current RDF Views supported and exposing is true power.

#4

Spyder from Revelytix implements R2RML standard. It is in Beta. I only know of the product but have not tried it. Hope it helps.