Skip to main content

Introduction to Reasoning

Reasoning allows Fluree to infer new facts from your data based on logical rules. This guide introduces the concepts—see the Reasoning Reference for complete details.

What is Reasoning?​

Reasoning (or inference) lets you define relationships between classes and properties, and Fluree will automatically derive implied facts.

For example, if you define:

  • "Manager" is a subclass of "Employee"
  • Alice is a Manager

Fluree can infer that Alice is also an Employee, even if you never explicitly stated it.

Class Equivalences​

You can declare that classes from different vocabularies mean the same thing:


{
"@id": "ex:Person",
"owl:equivalentClass": {"@id": "schema:Person"}
}

Now data typed as ex:Person will also match queries for schema:Person.

OWL and RDFS Classes​

A common use case: declaring that OWL classes are equivalent to RDFS classes:


{
"@id": "owl:Class",
"owl:equivalentClass": {"@id": "rdfs:Class"}
}

This is useful when working with ontologies that use OWL vocabulary.

Property Equivalences​

Properties can also be declared equivalent:


{
"@id": "ex:email",
"owl:equivalentProperty": {"@id": "schema:email"}
}

Data stored with ex:email will match queries using schema:email.

Subclass Relationships​

Define class hierarchies:


{
"@id": "ex:Manager",
"rdfs:subClassOf": {"@id": "ex:Employee"}
}

All Managers are automatically considered Employees.

Subproperty Relationships​

Properties can have hierarchies too:


{
"@id": "ex:workEmail",
"rdfs:subPropertyOf": {"@id": "ex:email"}
}

A query for ex:email will also return ex:workEmail values.

Inverse Properties​

Define bidirectional relationships:


{
"@id": "ex:manages",
"owl:inverseOf": {"@id": "ex:reportsTo"}
}

If Alice manages Bob, Fluree can infer that Bob reportsTo Alice.

When to Use Reasoning​

  • Integrating data from multiple sources with different vocabularies
  • Building flexible queries that match related concepts
  • Modeling real-world hierarchies (categories, roles, types)
  • Avoiding redundant data entry by inferring relationships

Performance Considerations​

Reasoning adds computation. Use it when:

  • You need vocabulary interoperability
  • The logical relationships are stable
  • Query flexibility outweighs the overhead

Next Steps​