Embedded SPARQL Engines
The following engines can be used using rpt integrate -e engine [--loc engine-specific-location].
Embedded SPARQL engines are built into RPT and thus readily available. The following engines are currently available:
| Engine | Description |
|---|---|
| mem | The default in-memory engine based on Apache Jena. Data is discarded once the RPT process terminates. |
| tdb2 | Apache Jena's TDB2 persisent engine. Use --loc to specfify the database folder. |
| binsearch | Binary search engine that operates directly on sorted N-Triples files. Use --loc to specify the file path or HTTP(s) URL to the N-Triples file. For URLs, HTTP range requests must be supported! |
| remote | A pseudo engine that forwards all processing to the SPARQL endpoint whole URL is specified in --loc. |
| qlever | The blazing fast qlever triple store launched from its docker image via Java's TestContainers framework. Use qlever:imageName:tag to use a specific image - this image's command line interfaces for starting the server and creating the indexes must be compatible with the default image registered with RPT. |
(ARQ) Engine Configuration
The engines mem, tdb2 and binsearch build an Jena’s query engine ARQ and thus respect its configuration.
rpt integrate --set 'arq:queryTimeout=60000' --set 'arq:updateTimeout=1800000' data.ttl myUpdate.ru myQuery.rq