Cypher

Cypher uses ASCII-art syntax to visually represent graph patterns, making queries intuitive and readable. Instead of describing how to traverse, you declare what pattern to find.

• Nodes are represented with parentheses: (node:Label)
• Relationships use arrows and brackets: -[rel:TYPE]->
• Patterns compose naturally: (a:Person)-[:KNOWS]->(b:Person)
• The engine optimizes traversal behind the scenes

This declarative approach separates query intent from execution strategy, allowing the query planner to choose optimal traversal algorithms.

Cypher provides specialized clauses designed for graph operations that go far beyond SQL's relational capabilities.

• MATCH: Finds patterns in the graph using index lookups and traversal
• MERGE: Creates patterns only if they don't exist, preventing duplicates
• CREATE: Adds nodes and relationships atomically
• DETACH DELETE: Removes nodes and all connected relationships safely
• CALL: Invokes procedures for graph algorithms like PageRank or community detection

Variable-length path queries like (a)-[*2..5]->(b) find paths of any length between nodes, enabling friend-of-friend queries natively.

Originally developed for Neo4j, Cypher became an open standard through the openCypher project in 2015, enabling adoption across the graph ecosystem.

• GQL (Graph Query Language): ISO/IEC 39075:2024 standard heavily influenced by Cypher
• Cypher for Apache Spark: Enables graph queries on distributed data
• Cypher for Gremlin: Bridges property graph and RDF/SPARQL worlds
• Amazon Neptune, Memgraph, RedisGraph, and AgensGraph all support Cypher

This standardization ensures query portability and prevents vendor lock-in for enterprise knowledge graph deployments.

Cypher includes built-in functions for computing metrics over graph structures, essential for analytics and feature engineering.

• collect(): Gathers matched entities into lists for aggregation
• count(), avg(), sum(): Standard statistical operations on node properties
• shortestPath(): Finds the minimum-hop path between two nodes using bidirectional BFS
• allShortestPaths(): Returns all paths with the minimum length
• path length(): Computes the number of hops in a matched path

These functions enable complex analytical queries like finding central nodes or computing graph-level statistics directly in the query language.

Cypher operates on the labeled property graph model, which provides schema flexibility while supporting optional constraints for data integrity.

• Nodes and relationships can have arbitrary key-value properties without predefined schemas
• UNIQUE constraints enforce property uniqueness on labeled nodes
• EXISTENCE constraints require specific properties on nodes or relationships
• NODE KEY constraints define composite keys for entity identity
• Relationship type constraints restrict which node labels can connect

This balance enables agile graph modeling for evolving knowledge graphs while maintaining data quality where needed.

Modern Cypher supports subqueries for modular query composition and complex multi-part analysis.

• CALL subquery: Executes a nested query and unions results with the outer query
• EXISTS subquery: Tests pattern existence for conditional logic
• COUNT subquery: Evaluates pattern cardinality inline
• Post-UNION processing: Applies additional filtering and aggregation after combining results

These features enable sophisticated graph analytics like comparing subgraph densities or computing conditional traversals within a single query execution.