Help:SPARQLStore

From semantic-mediawiki.org
(Redirected from SPARQL store)
Jump to: navigation, search
Has function class name::SPARQLStore
Has description::'''SPARQLStore''' the component to establish a connection between SMW and a triple store.
Further Information
Short name: Has short name::SPARQLStore
Provided by: Has component::Semantic MediaWiki
Added: Has minimum version::1.6
Removed: Has maximum version::still in use
Keyword : {{#ask: Has keyword::RDF|?has description|limit=0|searchlabel=RDF }} · {{#ask: Has keyword::SPARQL|?has description|limit=0|searchlabel=SPARQL }} · {{#ask: Has keyword::Triple store|?has description|limit=0|searchlabel=Triple store }}
Table of Contents

The SPARQLStore is the name of the component that handles the communication (including data management) between Semantic MediaWiki and a triple-store and the transformation of #ask into a SPARQL query representation.

Features and limitations

Scalability and stability

Starting with SMW 2.3, SPARQLStore supports all native features of the SQLStore including redirects for queries and property/category hierarchies (given that the back-end supports SPARQL 1.1)

  • W3C RDF Store Benchmarking lists RDF benchmarks, benchmarking results and papers
  • Comparison on Blazegraph vs. Sesame vs. Virtuoso
  • Given the fact that DBpedia is facilitating Virtuoso 7 successfully (probably the commercial version) [0] it should be regarded as reliable and stable [2, 3, 4]
  • Quoting a Sesame developer "...on more recent hardware Sesame's native store scales to about 150-200 million triples. Sesame development focuses on API usability, standards compliance, parsing performance ..." but supports graph database like Systap Blazegraph or Ontotext GraphDB. [1]
  • If SPARQL compliance is required Fuseki and Sesame are good candidates. If authorization is required Virtuoso and Fuseki (Version 2 with Shiro) is the choice to make. For a comparison on large datasets, see [5].
  • 4store showed a rather degrading performance when run our test suite in comparison to any other supported repository (also it is missing SPARQL 1.1 support)
  • ARC2 (as a native PHP/MySQL) seems not a real alternative due to "ARC2 is now in a stable state with no further feature additions planned". [6]
  • Neo4J has not been tested (see comments made about Virtuoso/Neo4J [7])
  • Blazegraph has been selected by the Wikidata project to serve the "Wikidata Query Service" [8]

Available repository connectors

{{#ask:

 Has keyword::Database connector OR Has keyword::Repository connector
Has keyword::SPARQLStore 
|?Has short name=Connector
|?Has description=Description
|?Has minimum version=SMW
|format=broadtable
|mainlabel=Help page

}}

Available configuration parameters

SPARQL requests, whether queries or updates, are exchanged through web services. This means that requests are sent to and data is received from URLs that specify the location of the according service. This location is determined by the SPARQL database and by its configuration. The respective configuration parameters are shown in the following table:

{{#ask:

Has keyword::SPARQLStore
|?Has description=Description
|?Has minimum version=SMW
|format=broadtable
|sort=Has minimum version
|order=ascending
|mainlabel=Parameter

}}

Notes

[0] http://www.openlinksw.com/dataspace/doc/dav/weblog/Comunity%20blog/1833

[1] https://groups.google.com/forum/#!topic/sesame-users/XDrMDka82II

[2] http://db-engines.com/en/system/Jena%3BVirtuoso [3] http://db-engines.com/en/system/Jena%3BSesame%3BVirtuoso [4] http://stackoverflow.com/questions/304920/which-triplestore-for-rapid-semantic-web-development/350726#350726

[5] http://www.w3.org/wiki/LargeTripleStores

[6] https://github.com/semsol/arc2

[7] https://github.com/SemanticMediaWiki/SemanticMediaWiki/issues/586#issuecomment-60571556

[8] http://blog.blazegraph.com/?p=826

See also