Graph Database overview

Database purpose-built for storing and traversing graph structures — nodes (entities) and edges (relationships) with properties on both. Graph databases excel at multi-hop relationship queries that would require expensive recursive joins in a relational database. Common use cases: fraud detection, recommendation engines, knowledge graphs, identity and access management, and social networks. Neo4j uses the Cypher query language; Amazon Neptune supports Gremlin and SPARQL; Dgraph uses GraphQL+DQL. Property graphs (Neo4j, Neptune) and RDF triplestores (Stardog, Amazon Neptune/SPARQL) are the two dominant paradigms. Vector + graph hybrids are emerging for AI knowledge graph applications.