User:Daniel/Queries/Graph profiling

About

This page assists in the collection and curation of queries that

• can be used with no or minimal modification on multiple graphs
• provide insights into the graph structure

Generic queries

Number of triples in the graph

SELECT (COUNT(*) AS ?count) WHERE {
  ?s ?p ?o
}

Try it!

Examples of distinct relationships in the graph

SELECT DISTINCT ?p WHERE {
 ?s ?p ?o .
}
LIMIT 100

Try it!

• This currently fails (due to memory issues) for the DBLP SPARQL endpoint but
  • they provide detailed documentation of their RDF schema
  • the same query works on the Freiburg QLever instance for Wikidata

Number of distinct relationships in the graph

SELECT (COUNT(DISTINCT ?p) AS ?count) WHERE {
  ?s ?p ?o
}

Try it!

Number of types of entities in the graph

This requires knowing what the graph's local equivalent of "instance of" is. For instance, in the EU knowledge graph, this would be P35, leading to the following query.'

SELECT (count(DISTINCT ?type ) as ?count) WHERE {
  ?s <https://linkedopendata.eu/entity/P35> ?type . 
}

Types of entities in the graph

SELECT DISTINCT ?type WHERE {
  ?s wdt:P35 ?type . 
}

Queries for Wikibase instances

# Properties by frequency in a random sample 
SELECT DISTINCT ?property ?propertyLabel (COUNT(*) AS ?count) WHERE {
  SERVICE bd:sample { ?item ?p [ ] . bd:serviceParam bd:sample.limit 1000 }
  
  ?property a wikibase:Property;
              wikibase:claim ?p.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
GROUP BY ?property ?propertyLabel
ORDER BY DESC(?count)

Try it!

Queries for a specific set of KGs

• Relative coverage of zbMATH authors in MaRDI vs. Wikidata

See also

• NFDI Core Ontology
• SPARQL Endpoint Metadata Check
  • all fine with https://sparql.uniprot.org/
  • warnings for https://query.portal.mardi4nfdi.de/proxy/wdqs/bigdata/namespace/wdq/sparql
  • warnings for https://database.factgrid.de/sparql
• Vocabulary of Interlinked Datasets (VoID)
  • VoID vocabulary definition
  • Describing Linked Datasets with the VoID Vocabulary
  • VoID generator
• LLM-based SPARQL Query Generation from Natural Language over Federated Knowledge Graphs