Ontology NL Search

SPARQL 1.1 Overview ​

SPARQL 1.1 is the W3C query language for RDF graphs. Queries match graph patterns, bind variables, and then shape the result set with modifiers such as ORDER BY, LIMIT, and DISTINCT.

Core query shape ​

sparql
PREFIX ex: <http://example.org/>

SELECT ?subject ?label
WHERE {
  ?subject ex:label ?label .
}
LIMIT 10

Triple patterns and variables ​

  • RDF data is queried as subject-predicate-object triples.
  • Variables start with ? or $.
  • A solution is produced when all required patterns match.
sparql
SELECT ?person ?name WHERE {
  ?person <http://xmlns.com/foaf/0.1/name> ?name .
}

FILTER ​

FILTER narrows the solutions produced by graph matching.

sparql
PREFIX ex: <http://example.org/>

SELECT ?item WHERE {
  ?item ex:score ?score .
  FILTER(?score >= 10)
}

OPTIONAL ​

OPTIONAL keeps the main match even when extra data is missing.

sparql
PREFIX ex: <http://example.org/>

SELECT ?item ?comment WHERE {
  ?item ex:title ?title .
  OPTIONAL { ?item ex:comment ?comment }
}

If the optional pattern does not match, the variables inside it remain unbound.

UNION ​

UNION matches either branch.

sparql
PREFIX ex: <http://example.org/>

SELECT ?item WHERE {
  { ?item ex:status "draft" }
  UNION
  { ?item ex:status "published" }
}

Property paths ​

SPARQL 1.1 property paths let one pattern traverse multiple edges.

| Syntax | Meaning | | ------------------- | ------------ | ----------- | | ex:parent/ex:name | sequence | | ex:parent | ex:guardian | alternative | | ^ex:parent | inverse path | | ex:child* | zero or more | | ex:child+ | one or more | | ex:child? | zero or one |

sparql
PREFIX ex: <http://example.org/>

SELECT ?ancestor WHERE {
  ?start ex:parent+ ?ancestor .
}

Solution modifiers ​

Common modifiers:

  • DISTINCT
  • ORDER BY
  • LIMIT
  • OFFSET
  • GROUP BY
  • HAVING
sparql
PREFIX ex: <http://example.org/>

SELECT ?category (COUNT(?item) AS ?count)
WHERE {
  ?item ex:category ?category .
}
GROUP BY ?category
ORDER BY DESC(?count)

Good defaults ​

  • Use PREFIX declarations for readability.
  • Prefer OPTIONAL only when missing data is acceptable.
  • Use STR() before string functions when a term may be an IRI.
  • Keep FILTER expressions close to the variables they constrain.

References ​