I'm still early on in my project, so this is not in response to a problem. I'm concerned about latency of response - when I add an entry to the triplestore, I'd like the LARQ index to process it right away. However, I'd like several indexers, each running on different machines, and so I need to notify them of what changed. It would be convenient if they could just be pinged and they would retrieve the new triples to index from the datastore, but if I can leverage any existing infrastructure to do this then I don't need to write such functionality myself.
- Mark -----Original Message----- From: Paolo Castagna [mailto:[email protected]] Sent: Monday, March 12, 2012 4:10 PM To: [email protected] Subject: Re: Incremental LARQ indexing Mark Roth wrote: > What is the best way to incrementally update a LARQ index if my SDB SQL > Database is being updated from a different machine? Hi Mark, how do you currently update your RDF data in SDB? Do you use the Jena APIs? Or, SPARQL Update via Joseki/Fuseki? How frequent and big|small are your updates? Does an update need to be immediately visible to the users or your app could live with a time gap? LARQ documentation is here: http://incubator.apache.org/jena/documentation/larq/ larq.larqbuilder sources are here: https://svn.apache.org/repos/asf/incubator/jena/Jena2/LARQ/trunk/src/main/java/larq/larqbuilder.java It uses the ModDatasetAssembler therefore it should work with SDB's datasets as well. If your updates aren't frequent and your RDF dataset not massive, you could rebuild your Lucene index nightly scheduling a cron job or equivalent. I tend to use LARQ with TDB and Fuseki, but if you or someone else report a problem with LARQ and SDB, I'll look into it. Has your question came up because you found a problem? Cheers, Paolo > > Thanks, > > - Mark > >
