MarkLogic memory limit for SPARQL?

1
votes

I am in the situation to test several "select *" kind of SPARQL queries against few TB of triple data, of course not in production. However, we only have limited machine resources (4GB memory) to test the queries.

I understand that it requires more memory, but are there any alternatives to run queries and get results? (Time consuming is welcome)

1
sparqlmarklogic
Please paste example query... - Stanislav Kralin
"few TB" ... to be honest, 4GB is nothing today and something has to be loaded into main memory to work on it, at least some database index (for RDF usually multiple, e.g. spo, pos, etc.) - UninformedUser

1 Answers

0
votes

My laptop has 32GB of RAM, so that sounds under-resourced even for a dev server. Having said that, for any particular query I would look for ways to reduce the number of triples you're running against. Are your triples segmented into graphs, and if so, can a query be directed against one graph? Another reducing strategy is to use the $query parameter to sem:sparql to identify documents that hold the triples you care about.