Federation Patterns - SPARQLqueries

Why Patterns Matter

Federated queries can be slow and complex. These patterns help you structure queries efficiently, handle failures gracefully, and get the most out of distributed data.

Pattern 1: Entity Linking

Use owl:sameAs or external IDs to find the same entity across different datasets. This is the foundation of most federated queries.

Link Entities via owl:sameAs

Run ↗

PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX dbo: <http://dbpedia.org/ontology/>

SELECT ?city ?cityLabel ?population
WHERE {
  # Cities in Germany
  ?city wdt:P31 wd:Q515 ;
        wdt:P17 wd:Q183 .

  # Link to DBpedia via sameAs
  SERVICE <http://dbpedia.org/sparql> {
    ?dbCity owl:sameAs ?city ;
            dbo:populationTotal ?population .
  }

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
ORDER BY DESC(?population)
LIMIT 10

Federation Join datasets via owl:sameAs

Pattern 2: Data Enrichment

Add information from external sources that Wikidata doesn't have. This is the most common use case for federation.

Enrich with External Descriptions

Run ↗

PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX dbo: <http://dbpedia.org/ontology/>

SELECT ?museum ?museumLabel ?abstract
WHERE {
  # Art museums
  ?museum wdt:P31 wd:Q207694 .

  # Get rich description from DBpedia
  SERVICE SILENT <http://dbpedia.org/sparql> {
    ?dbMuseum owl:sameAs ?museum ;
               dbo:abstract ?abstract .
    FILTER(LANG(?abstract) = "en")
  }

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
LIMIT 5

Federation Add Wikipedia abstracts to Wikidata results

Pattern 3: Cross-Validation

Compare data between sources to find discrepancies or validate information.

Compare Population Data

Run ↗

PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX dbo: <http://dbpedia.org/ontology/>

SELECT ?country ?countryLabel ?wdPop ?dbPop (ABS(?wdPop - ?dbPop) AS ?diff)
WHERE {
  # Countries with population in Wikidata
  ?country wdt:P31 wd:Q6256 ;
           wdt:P1082 ?wdPop .

  # Get population from DBpedia
  SERVICE SILENT <http://dbpedia.org/sparql> {
    ?dbCountry owl:sameAs ?country ;
                dbo:populationTotal ?dbPop .
  }

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
ORDER BY DESC(?diff)
LIMIT 10

Federation Find largest population discrepancies

Pattern 4: Filter First

Always filter locally before federating. Sending fewer bindings to remote endpoints dramatically improves performance.

Filter Before Federation

Run ↗

PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX dbo: <http://dbpedia.org/ontology/>

SELECT ?scientist ?scientistLabel ?abstract
WHERE {
  # Filter locally: Nobel Prize winners in Chemistry
  ?scientist wdt:P31 wd:Q5 ;
             wdt:P166 wd:Q44585 .

  # Only then federate with small result set
  SERVICE SILENT <http://dbpedia.org/sparql> {
    ?dbScientist owl:sameAs ?scientist ;
                  dbo:abstract ?abstract .
    FILTER(LANG(?abstract) = "en")
  }

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

Federation Only ~200 Nobel laureates to look up

Anti-pattern: Don't query the remote endpoint first and then filter locally. This downloads massive amounts of data before filtering.

Pattern 5: OPTIONAL Federation

Use OPTIONAL with SERVICE SILENT to return results even when the remote data doesn't exist or the endpoint fails.

Optional Remote Data

Run ↗

PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>

SELECT ?uni ?uniLabel ?homepage
WHERE {
  # Universities in France
  ?uni wdt:P31 wd:Q3918 ;
       wdt:P17 wd:Q142 .

  # Optionally get homepage from DBpedia
  OPTIONAL {
    SERVICE SILENT <http://dbpedia.org/sparql> {
      ?dbUni owl:sameAs ?uni ;
              foaf:homepage ?homepage .
    }
  }

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en,fr" . }
}
LIMIT 20

Federation Returns universities even without homepage

Pattern 6: Multiple Endpoints

Combine data from several external sources in a single query. Use SERVICE SILENT for each to handle failures gracefully.

Combine Multiple Sources

Run ↗

PREFIX owl: <http://www.w3.org/2002/07/owl#>
PREFIX dbo: <http://dbpedia.org/ontology/>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>

SELECT ?artist ?artistLabel ?abstract ?homepage
WHERE {
  # Painters
  ?artist wdt:P31 wd:Q5 ;
          wdt:P106 wd:Q1028181 .

  # Source 1: DBpedia abstract
  OPTIONAL {
    SERVICE SILENT <http://dbpedia.org/sparql> {
      ?dbArtist owl:sameAs ?artist ;
                  dbo:abstract ?abstract .
      FILTER(LANG(?abstract) = "en")
    }
  }

  # Source 2: DBpedia homepage
  OPTIONAL {
    SERVICE SILENT <http://dbpedia.org/sparql> {
      ?dbArtist2 owl:sameAs ?artist ;
                   foaf:homepage ?homepage .
    }
  }

  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
LIMIT 5

Federation Multiple OPTIONAL SERVICE blocks

Best Practices Summary

Filter first — Minimize data before federating
Use SERVICE SILENT — Handle endpoint failures gracefully
Use OPTIONAL — Return results even when remote data is missing
Test separately — Debug each SERVICE block independently
Limit results — Start small, scale up carefully
Use entity links — owl:sameAs and external IDs are essential
Monitor performance — Federated queries can be slow

Key Entities

Entity	Description
`owl:sameAs`	Links identical entities across datasets
`skos:exactMatch`	Links equivalent concepts (weaker than sameAs)
`SERVICE SILENT`	Graceful failure handling for federation
`OPTIONAL`	Return results even if pattern doesn't match