Overview of DFRNT data modelling

Overview of DFRNT data modelling

The data model in DFRNT data products follows the TerminusDB graph metamodel 1:1, and enables easy data modelling and advanced visualisation

The data modelling capabilities of DFRNT.com involves five categories of data structures: Records, Properties, Document Types, Traits (subdocuments), Foreign entities, TaggedUnions and Enums. They are all representations of the underlying TerminusDB meta model concepts.

The data modeller provides 1:1 mapping between the graphical field definitions and the JSON-LD schema definitions for Types, Traits and Enums from TerminusDB, which makes it a great playground to exchange JSON-LD data definitions.


Records is where instance data lives, your data. They are called documents in TerminusDB lingo and the RDF triples they involve are represented using the JSON-LD format, and the structure is controlled through layered JSON-LD class constructs. Records may include the following kinds of properties (data fields):

  • Base types (strings, numeric, geographical, time, etc.)
  • Links to other records of a type (documents in TerminusDB)
  • Traits (deep and controlled hierarchical data structures, subdocuments in TerminusDB, with a lifecycle connected to the base document the triples are attached to)
  • Enum constants (enum in TerminusDB)
  • Foreign (identifiers for documents in other data products with different prefixes)
  • TaggedUnion (ability to build disjoint property sets, mutually exclusive property definitions)


Properties hold the data and references between records in the system. Beneath the data model, it’s RDF triples all the way down. The RDF triples are constrained by a schema checker that checks that all data is consistent. Instead of OWL and SHACL, TerminusDB has an enterprise-orientated (vs academic) data model with both shapes and constraints like multiplicities, cardinalities, data types and more.

The properties on Types and Traits can either be of a base type or link to other Types, Traits or Enum values and data structures. Base data properties, Traits and Enums are attached directly to the Record, while Type references are directional links to other records.

Composition of Traits and Types means that property definitions from their parents get inherited down a hierarchy. The same property name may not be inherited from multiple parents, that is called a prevented diamond property violation.

Types (records/documents)

All records have a type that defines the data structure of the record. Types have properties that may be composed (inherited) from one or multiple types.

Traits (data structures/subdocuments)

Traits are reusable data structures that can be attached to types as properties, and may be composed by one or multiple traits. By defining a data structure as a trait, instead of as a type, it can be reused across types to build complex data structures.

Some traits have special meaning in DFRNT.tech and are used to build data structures and taxonomies that have semantic meaning, as part of the collaborative semantic knowledge graph.

Enum constants (named entries)

Enumerations (list of constant values) can be used to constrain properties to certain values. They are frequently used for categories and other sets of information.

    • Related Articles

    • How can I learn more about DFRNT

      Learning about DFRNT is closely connected to leveraging the open source TerminusDB data product engine, for which DFRNT is a UI and hosting provider. DFRNT is a user interface and hosting provider for the open source TerminusDB data product ...
    • Use TerminusCMS data products in DFRNT

      Additionally to DFRNT hosted data products, you can connect and collaborate directly on data products hosted at TerminusCMS from DFRNT, using TerminusDB API keys. DFRNT.com helps you manage data products that are either hosted in DFRNT through the ...
    • The DFRNT Data Product Builder

      DFRNT is a data product builder with knowledge graph visualisation, modelling and management, with version control, API access and a datalog engine. With DFRNT data products, digital architects, cybersecurity professionals, data practitioners and ...
    • Render and browse records as Markdown documents, together with inbound and outbound links

      Data in DFRNT can be browsed like a wiki, complete with inbound and outbound links so that data can be navigated freely and for exploration with Markdown for rich text. Document contents can be viewed by opening a record through the canvas, or by ...
    • Custom class metadata configuration

      Colours, link directions, class labels, the property display order can be altered by setting custom class metadata. DFRNT provides a special section called "metadata" to define custom behaviours for types and traits. Although these behaviours cannot ...