JarCastingStreaming SPARQL
Provides a robust, incremental processing of streaming results received from SPARQL servers. The StreamingResultSet iterator yields results as they are received from the server.
Javadoc
http://javadoc.io/doc/com.weblyzard.sparql/streaming-sparql/
Example code:
try (StreamingResultSet s = StreamingQueryExecutor.getResultSet("http://dbpedia.org/sparql", "SELECT ?s ?p ?o WHERE { ?s ?p ?o. } LIMIT 5")) {
while (s.hasNext()) {
System.out.println("Tupel " + s.getRowNumber() + ": " + s.next())
}
}
Command line client
Streaming SPARQL also provides a command line client for testing queries.
Usage
java -jar ./streaming-client-0.0.7-SNAPSHOT.jar
QueryEntitites [URL] [Query]
URL ... URL to the linked data repository
Query ... The query to perform on the server
Example
java -jar ./streaming-client-0.0.7-SNAPSHOT.jar http://localhost:8080/rdf4j-sesame/test "SELECT ?s ?p ?o WHERE { ?s ?p ?o. } LIMIT 5"
Background
We have been using Fuseki and RDF4j together with comprehensive result sets (> 100 Mio. tuple) which lead to instabilities with the native libraries that have been extremely difficult to debug.
Example error messages on the server site have been:
[2017-05-04 19:50:14] Fuseki WARN [1450] Runtime IO Exception (client left?) RC = 500 : org.eclipse.jetty.io.EofException
org.apache.jena.atlas.RuntimeIOException: org.eclipse.jetty.io.EofException
[2017-05-04 19:50:14] Fuseki WARN (HttpChannel.java:468) (and one from ServletHandler.java:631):
java.io.IOException: java.util.concurrent.TimeoutException: Idle timeout expired: 30001/30000 m
These problems triggered the development of Streaming SPARQL which has proven to be very robust - even for queries that take more than one hour to process and transfer multiple gigabytes of results. (Note: you will need to call getResultSet with a higher timeout to prevent TimeoutExceptions on the server).
Compatiblity
Streaming SPARQL is known to work with Jena, OpenRDF, RDF4j and Virtuoso.
Changelog
Please refer to the release page.
