Querying Graphs with Cypher

The Property-Graph Query Model

Cypher is the declarative query language for property graphs, originally created for Neo4j and now standardized as the basis of GQL (ISO/IEC 39075). Instead of describing how to fetch data step by step, you describe the shape of the pattern you want and let the engine find every place it occurs.

A Tiny Example Graph

We'll use this small social-and-work graph throughout the lesson:

• People: Alice, Bob, Carol, Dave
• Companies: Acme, Globex
• Alice and Bob WORKS_AT Acme; Carol WORKS_AT Globex
• Alice KNOWS Bob; Bob KNOWS Carol; Carol KNOWS Dave
• Alice CREATED the Widget product; Acme PRODUCES Widget

Declarative, Not Procedural

A relational query says join these tables on these keys. A Cypher query instead draws the relationship you're after:

MATCH (a:Person)-[:WORKS_AT]->(c:Company)
RETURN a.name, c.name

You never wrote a join condition. The arrow is the join, and the engine resolves it by following stored relationships.

Labels, Types, and Properties

• A node label such as :Person groups nodes of the same kind.
• A relationship type such as :WORKS_AT names the kind of edge.
• Both nodes and relationships hold properties — {name: "Alice"}, {since: 2021}.

Cypher is case-sensitive for labels, types, and property keys, and keywords are conventionally written in uppercase.