SPARQL vs GQL¶

This guide compares SPARQL (W3C standard for RDF) with GQL (ISO standard for property graphs). Both are powerful graph query languages, but they're designed for different data models.

Data Model Differences¶

Query Syntax Comparison¶

Finding Entities¶

PREFIX foaf: <http://xmlns.com/foaf/0.1/>

SELECT ?name
WHERE {
    ?person a foaf:Person .
    ?person foaf:name ?name
}

MATCH (person:Person)
RETURN person.name

Finding Relationships¶

PREFIX foaf: <http://xmlns.com/foaf/0.1/>

SELECT ?name ?friendName
WHERE {
    ?person foaf:name ?name .
    ?person foaf:knows ?friend .
    ?friend foaf:name ?friendName
}

MATCH (person:Person)-[:KNOWS]->(friend:Person)
RETURN person.name, friend.name

Filtering¶

SELECT ?name ?age
WHERE {
    ?person foaf:name ?name .
    ?person foaf:age ?age
    FILTER(?age > 30)
}

MATCH (person:Person)
WHERE person.age > 30
RETURN person.name, person.age

Optional Matches¶

SELECT ?name ?email
WHERE {
    ?person foaf:name ?name
    OPTIONAL { ?person foaf:mbox ?email }
}

MATCH (person:Person)
OPTIONAL MATCH (person)-[:HAS_EMAIL]->(email:Email)
RETURN person.name, email.address

Aggregations¶

SELECT ?country (COUNT(?person) AS ?count)
WHERE {
    ?person foaf:based_near ?country
}
GROUP BY ?country
HAVING (COUNT(?person) > 10)
ORDER BY DESC(?count)

MATCH (person:Person)-[:LIVES_IN]->(country:Country)
WITH country, COUNT(person) AS count
WHERE count > 10
RETURN country.name, count
ORDER BY count DESC

Path Traversal¶

# All ancestors (transitive closure)
SELECT ?ancestor
WHERE {
    ?person ex:parent+ ?ancestor
}

MATCH (person:Person)-[:PARENT*1..]->(ancestor:Person)
RETURN ancestor

Union Queries¶

SELECT ?contact
WHERE {
    { ?person foaf:mbox ?contact }
    UNION
    { ?person foaf:phone ?contact }
}

MATCH (person:Person)-[:HAS_EMAIL]->(e:Email)
RETURN e.address AS contact
UNION
MATCH (person:Person)-[:HAS_PHONE]->(p:Phone)
RETURN p.number AS contact

Feature Comparison¶

When to Use Each¶

Use SPARQL When:¶

• Working with RDF/Linked Data
• Need semantic reasoning (RDFS/OWL inference)
• Querying knowledge graphs with ontologies
• Federated queries across multiple endpoints
• Data follows W3C standards (Dublin Core, FOAF, Schema.org)

Use GQL When:¶

• Working with property graph data
• Need intuitive pattern matching syntax
• Building application databases
• Relationships have properties (weights, timestamps)
• Prefer ASCII-art style query patterns

Grafeo Support¶

Grafeo supports both query languages:

# Enable both features
[dependencies]
grafeo = { version = "0.5", features = ["gql", "sparql"] }

import grafeo

db = grafeo.GrafeoDB()

# SPARQL query
result = db.execute_sparql("""
    PREFIX foaf: <http://xmlns.com/foaf/0.1/>
    SELECT ?name
    WHERE { ?person foaf:name ?name }
""")

import grafeo

db = grafeo.GrafeoDB()

# GQL query
result = db.execute("""
    MATCH (person:Person)
    RETURN person.name
""")

Performance Considerations¶

Migration Tips¶

SPARQL to GQL¶

1. Replace triple patterns with node/edge patterns
2. Convert FILTER to WHERE clauses
3. Replace property paths with variable-length patterns
4. Map IRIs to node labels and properties

GQL to SPARQL¶

1. Define appropriate prefixes for the domain
2. Model node properties as separate triples
3. Convert edge patterns to predicate URIs
4. Use OPTIONAL for optional relationships

Further Reading¶

• SPARQL 1.1 Specification
• GQL Standard (ISO/IEC 39075)
• LPG vs RDF Data Models
• GQL vs SPARQL Concepts