Federation

The Power of Linked Data

No single knowledge base contains all the information you need. Wikidata might have an entity, but DBpedia has additional details, or another specialized endpoint has domain-specific data.

SPARQL federation solves this with the SERVICE keyword. Execute parts of your query against remote endpoints and combine results seamlessly — all from a single query.

This is the vision of the Semantic Web: interconnected data that can be queried as if it were a single giant database.

How Federation Works

SELECT ?localVar ?remoteVar
WHERE {
  # Local patterns (run on current endpoint)
  ?item wdt:P31 wd:Q9143 .

  # Remote patterns (run on external endpoint)
  SERVICE <http://dbpedia.org/sparql> {
    ?dbpediaItem owl:sameAs ?item .
    ?dbpediaItem foaf:homepage ?remoteVar .
  }
}

The query engine sends the SERVICE block to the remote endpoint, receives the results, and joins them with local results based on shared variables.

Federation Examples

Wikidata + DBpedia

SELECT ?lang ?langLabel ?homepage
WHERE {
  # Get programming languages from Wikidata
  ?lang wdt:P31 wd:Q9143 .

  # Get homepage from DBpedia via sameAs link
  SERVICE <http://dbpedia.org/sparql> {
    ?dbpediaLang owl:sameAs ?lang ;
                  foaf:homepage ?homepage .
  }

  SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en" .
  }
}
LIMIT 10

SERVICE SILENT

SELECT ?lang ?langLabel ?homepage
WHERE {
  ?lang wdt:P31 wd:Q9143 .

  # SILENT: Don't fail if endpoint is unavailable
  SERVICE SILENT <http://dbpedia.org/sparql> {
    ?dbpediaLang owl:sameAs ?lang ;
                  foaf:homepage ?homepage .
  }

  SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en" .
  }
}
LIMIT 10

Popular SPARQL Endpoints

Federation Best Practices

• Minimize remote calls — Filter locally first, then federate with smaller result sets
• Use SERVICE SILENT — Prevent query failure when remote endpoints are down
• Check endpoint availability — Not all endpoints are always online
• Respect rate limits — Many public endpoints have query limits
• Use entity links — owl:sameAs, skos:exactMatch help join across datasets
• Test queries separately — Debug each SERVICE block independently first

You've Completed the Core Chapters!

Congratulations! You've learned SPARQL from basic queries to advanced federation. Continue exploring with more examples or revisit any chapter.