Currently I’m working on the queries of interactive workload of LDBC Social Network Benchmark. It’s an opensource benchmark project with a few SPARQL queries. I’m trying to update the queries use only SPARQL features and therefore we use STARDOG to validate the queries, but I’m stucked with this query. Here you can find the used dataset and the generated log.
Can you guys help me to move forward with this query? I tried to transform the query into many forms to make it work, but I couldn’t find a working one. It seems like a STARDOG bug, so can I somehow work around this?
Thanks for the well made bug report, we’ll investigate. If there’s a workaround. we’ll post it here. In either case it’ll be fixed in the next release.
One important notice: we can execute the query with 16GB memory, but we always got NPE with 4GB maximum memory (-Xmx), so it could be correlated to memory management.
Yes, this sounds likely. Can you share the contents of
stardog.properties in your STARDOG_HOME as well as values of
$STARDOG_JAVA_ARGS, in case you set any of those?
We’ve been able to reproduce, the fix will be available in the first week of Sep. For now the best workaround is to allocate more memory, as you figured.
Thank you for your quick response!
In fact we managed to squeeze the fix into the Aug release which will be out this week. This is what good bug reports achieve, enjoy.