Uploaded image for project: 'Blazegraph (by SYSTAP)'
  1. Blazegraph (by SYSTAP)
  2. BLZG-969

large sparql insert optimization slow?

    XMLWordPrintable

    Details

      Description

      recording this as an issue but it may be a no-fix because of unreasonable expectations.

      The attached sparql update can be executed as:
      time curl -H 'Accept: text/csv' -s --data-urlencode update@query-3.sparql $ENDPOINT

      If before doing this, one does a DROP ALL, then the insert is guaranteed to fail.
      On my system it takes over 30 seconds to do this, and the time is taken as follows:

      Time (ms)	Name
      15944	com.bigdata.rdf.sparql.ast.StaticAnalysis.getDefinitelyProducedBindings(IBindingProducerNode, Set, boolean)
      15879	com.bigdata.rdf.sparql.ast.StaticAnalysis.getDefinitelyIncomingBindings(IGroupMemberNode, Set)
      10511	com.bigdata.rdf.sparql.ast.optimizers.StaticOptimizer.hasSharedVars(StaticOptimizer$IJoinDimension, StaticOptimizer$IJoinDimension)
      8239	java.util.HashSet.add(Object)
      7531	com.bigdata.rdf.sparql.ast.StatementPatternNode.getProducedBindings()
      5373	java.util.LinkedHashSet.<init>()
      5186	java.util.AbstractCollection.addAll(Collection)
      3715	com.bigdata.bop.CoreBaseBOp.getProperty(String, Object)
      2184	com.bigdata.rdf.sparql.ast.StaticAnalysis.getMaybeProducedBindings(JoinGroupNode, Set, boolean)
      2023	java.util.LinkedHashMap.get(Object)
      1133	com.bigdata.rdf.sparql.ast.optimizers.StaticOptimizer.getFirstJoin()
      

      The issue is that, while this is a big update, the apparent time wastage is annoying

        Attachments

          Activity

            People

            Assignee:
            mrpersonick mrpersonick
            Reporter:
            jeremycarroll jeremycarroll
            Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: