What are the differences between a Graph database and a Triple store?



RDF Triple/Quad Stores or RDBMS engines that support RDF

  • Data is modeled as a collection of Entity Relationship Types
  • Entities are identified using IRIs (an open standard)
  • Entity Relationships are represented using subject->predicate->object sentence/statement collections — depictable as a Directed Graph
  • Data Definition operations are handled via RDF (and informed via shared vocabularies and ontologies e.g. RDF, RDF Schema, and OWL) or SPARQL (both W3C open standards)
  • Data Manipulation operations are handled via SPARQL (a W3C open standard)
  • Reasoning and Inference Rules are defined using SWRL (a W3C open standards) or SPARQL (courtesy of terms from SPIN or SHACL vocabularies)

Graph Databases

Everything is proprietary. Only sources of information are:

  • Use of phrase “Graph Database”
  • Confusing use of the term “simple”
  • Conversations laden with words (not terms) like “nodes” and “edges”
  • Loose mentions of JSON documents
  • Collateral laden with examples using proprietary languages (typically a language per product) re Data Definition and Manipulation operations.



OTOH, the whole TinkerPop ecosystem (Property Graph model) seems to be a solid option, depending on someone’s needs.