[Neo] SparQL engine?

Ryan Levering rrlevering at gmail.com
Sun Jan 10 06:45:54 CET 2010


Sorry to repeat myself everyone, but I'll answer more comprehensively  
this time so this thread doesn't trickle.  I originally wrote the  
engine for Sesame but at the time (2005) there were no other SPARQL  
interpreters for any system (besides a poorly written one for Jena),  
so I made it general purpose.  The goal of the project was to write a  
SPARQL query interpreter that could be easily plugged into any backend  
of triples.  The first version of Sesame 2 used this library and I  
wrote an adapter for Sesame 1.  The guys at OpenRDF eventually decided  
they wanted their own SPARQL interpreter, because they were working on  
a more generic query language infrastructure which converts all of  
their possible query languages (serql, sparql, etc) into a common  
query structure and then did more database specific optimizations.

Currently it stands as a library with a fairly low entry point to get  
SPARQL interpretation for any database that can hold triples.  It's a  
bit overengineered, but it does handle moderate sized data sets fairly  
well, because it is very conservative about caching.  It's not as  
efficient as if you wrote one from scratch for your own triple store,  
because it doesn't have as many hooks into the db to do certain  
predictive query optimizations and it's conservative nature.  If the  
spec has changed in the last two years it also won't have fixes, like  
you noticed.  At the very least, there is a decent jjt grammar there  
to do the language parsing if you were going to roll your own.

Someday I will go back and finish off this project by adding some more  
optimization hooks and finishing the debugger, but it won't be soon.   
If there are actual serious bugs in it, I would be happy to fix them.   
I do still get notifications from SourceForge for that project.

Ryan Levering

On Jan 9, 2010, at 11:08 PM, Marko A. Rodriguez wrote:

> Hello,
> This is sort of odd because OpenRDF Sesame already comes with a
> SPARQL-engine that works over Repository and SAIL.
> http://www.openrdf.org/doc/sesame2/2.3.0/apidocs/org/openrdf/query/parser/sparql/SPARQLParser.html
> Perhaps his is faster, more efficient? I didn't read far enough on his
> site to get why he built it. Finally, last "news" was in 2007.
> Take care,
> Marko.
> http://markorodriguez.com
> http://gremlin.tinkerpop.com
>> Hi there,
>> just found out about http://sparql.sourceforge.net/ , has anyone ever
>> tried this out? Would Neo4j benefit anything from writing an adapter
>> to it as opposed to the existing RDF components?
>> Cheers,
>> /peter neubauer
>> COO and Sales, Neo Technology
>> GTalk:      neubauer.peter
>> Skype       peter.neubauer
>> Phone       +46 704 106975
>> LinkedIn   http://www.linkedin.com/in/neubauer
>> Twitter      http://twitter.com/peterneubauer
>> http://www.neo4j.org                - Relationships count.
>> http://gremlin.tinkerpop.com    - PageRank in 2 lines of code.
>> http://www.linkedprocess.org   - Computing at LinkedData scale.
>> _______________________________________________
>> Neo mailing list
>> User at lists.neo4j.org
>> https://lists.neo4j.org/mailman/listinfo/user
> _______________________________________________
> Neo mailing list
> User at lists.neo4j.org
> https://lists.neo4j.org/mailman/listinfo/user

More information about the User mailing list