neo4j-spring-data-skill

Annotate the class with @Node("Label") and give it an @Id field, typically with @GeneratedValue for internal IDs. Use @Property to rename fields and @Relationship with an explicit direction for associations to other entities.

Add a method annotated with @Query containing your Cypher statement, and bind parameters with $paramName placeholders matching the method arguments. Never concatenate user input into the query string, as that opens Cypher injection risks.

SDN loads relationships only up to a default depth of one hop, so deeper associations come back empty. Write an explicit @Query that returns the root node together with collect(r) and collect(p) so SDN can map the full subgraph.

Yes, extend ReactiveNeo4jRepository to get Mono and Flux return types, and add spring-boot-starter-webflux alongside the Neo4j starter. Do not mix imperative and reactive database access in the same application context.

Yes, the spring-ai-starter-vector-store-neo4j dependency provides Neo4jVectorStore, which reuses your spring.neo4j connection settings. Set initialize-schema to true and match embedding-dimension to your embedding model; Neo4j 5.15 or later is required.

Avoid SDN when you need raw driver-level control without Spring, when the task is pure Cypher authoring, or when running GDS algorithms. Those cases are better served by the Java driver directly or dedicated Cypher and GDS tooling.