Skip to content
FDE 101

Graph database

A graph database stores and queries connected data. In a property graph, domain entities are nodes, explicit connections are relationships, and both can hold properties.

Tiny example

Section titled “Tiny example”
(Asha:Person)-[:WORKED_AT]->(Acme:Company)
(Asha)-[:KNOWS]->(Maya:Person)

A traversal can follow KNOWS from Maya to Asha or WORKED_AT from Asha to Acme. The result depends on stored edges and the rules of the query.

Section titled “Graph search versus vector search”
  • Graph traversal answers questions about explicit connectivity: “Who worked with Maya?”
  • Vector search ranks learned similarity: “Whose profile resembles this role description?”
  • Structured filters enforce exact constraints: “Only profiles available in India.”

The graph inside an HNSW vector index is different. HNSW edges help the index navigate between nearby vectors; they do not represent business relationships.

See the complete vector search or graph search lesson.