View on GitHub



The code for our leapfrog implementation for Apache Jena is available here.


The dataset used was a reduced version of the Wikidata truthy dump from November 15, 2018. The original dump and its reduced version are available at zenodo.

Repeating the experiments


Getting the repo and the dataset

Create the database for Jena and leapfrog

Create the database for Blazegraph

Create the database for Virtuoso Opensource

Run the benchmark

Building the dataset

Getting random queries for the benchmark

For each query pattern we created a java program that will find 50 random sets of properties with at least 1 result. The jars are in the find-queries folder. To find a query, you need to execute java -jar find_XYZ.jar [jena-database-location] properties_wikidata.txt, where properties_wikidata.txt is a file with the properties that can be chosen.


You can find our results in our repository