spring_boot_memgraph

Spring Boot + Spring Data Neo4j with Memgraph

This example shows that a Spring Boot application built with Spring Data Neo4j works as a drop-in replacement against Memgraph. Memgraph speaks the Bolt protocol and is wire-compatible with the Neo4j Java driver, so the only thing that changes versus a Neo4j project is the connection URI.

It is modeled on the official Spring Data Neo4j getting-started guide and reuses the same Movie / Person domain so you can compare them side by side.

🧠 What This Example Does

1. Connects Spring Data Neo4j to Memgraph over Bolt (bolt://localhost:7687).
2. Maps :Movie and :Person nodes (and the ACTED_IN / DIRECTED relationships) with the standard @Node, @Id, @Property and @Relationship annotations.
3. On startup, SchemaInitializer.java creates label indexes (:Movie, :Person) and label-property indexes (:Movie(title), :Person(name)), then a CommandLineRunner (DataLoader.java) loads a small movie graph and queries it — both a derived query and an explicit @Query Cypher statement.
4. Exposes a small REST API (MovieController.java) to load and query data over HTTP.

📦 Project Structure

src/main/java/com/memgraph/bestpractices/springboot/
├── SpringBootMemgraphApplication.java   # Spring Boot entry point
├── SchemaInitializer.java               # creates indexes on startup
├── DataLoader.java                      # loads + queries data on startup
├── entities/
│   ├── MovieEntity.java
│   └── PersonEntity.java
├── repositories/
│   └── MovieRepository.java
└── controllers/
    └── MovieController.java
src/main/resources/
└── application.properties               # <-- the only Memgraph-specific config

🚀 How to Run Memgraph

A docker-compose.yml is provided with Memgraph (MAGE) and Memgraph Lab:

docker compose up -d

This starts Memgraph on Bolt port 7687 and Memgraph Lab on http://localhost:3000.

Prefer a single container? You can also run:

docker run -it --rm -p 7687:7687 memgraph/memgraph-mage:3.10.1

🛠 Requirements

• Java 17+ (set in pom.xml via java.version)
• Maven 3.6+ (or use the Maven wrapper if you generate one)

Dependencies are declared in pom.xml — the standard Spring Boot starters spring-boot-starter-data-neo4j and spring-boot-starter-web. No Memgraph-specific library is needed.

🧪 How to Run the App

1. Start Memgraph: docker compose up -d
2. Run the app: mvn spring-boot:run

On startup you should see log lines like:

Loaded 1 movie(s) into Memgraph
Found movie 'The Matrix' with 2 actor(s) and tagline "Welcome to the Real World"
Keanu Reeves acted in: The Matrix

Try the REST API

# List all movies
curl http://localhost:8080/movies

# Fetch one movie by title
curl "http://localhost:8080/movies/by-title?title=The%20Matrix"

# Movies a given actor played in
curl "http://localhost:8080/movies/by-actor?name=Keanu%20Reeves"

# Create / update a movie
curl -X PUT http://localhost:8080/movies \
  -H "Content-Type: application/json" \
  -d '{"title":"John Wick","description":"People keep asking if I am back"}'

# Delete a movie by title
curl -X DELETE "http://localhost:8080/movies/John%20Wick"

🔁 Migrating an existing Neo4j project

The whole point: switching a Spring Data Neo4j app from Neo4j to Memgraph is a config change. In application.properties:

# Neo4j Aura
spring.neo4j.uri=neo4j+s://<instance>.databases.neo4j.io

# Memgraph
spring.neo4j.uri=bolt://localhost:7687

A few Memgraph-specific notes:

• Authentication is off by default in Memgraph, so the username/password properties are left blank. Fill them in if you created a user.
• Do not set spring.data.neo4j.database. Memgraph has a single database; setting it makes the driver issue USE <db>, which Memgraph rejects.
• Prefer an assigned business key for @Id (here title / name) rather than relying on internal node ids.
• Create indexes — Memgraph creates none out of the box. You need both a label index (CREATE INDEX ON :Movie) so plain MATCH (m:Movie) scans from findAll/count/deleteAll don't sweep the whole store, and a label-property index (CREATE INDEX ON :Movie(title)) for the @Id lookups SDN's generated Cypher uses. Without them you'll see PlanHinting sequential-scan warnings. This example creates both automatically in SchemaInitializer.java.

🔖 Version Compatibility

This example was built and tested with:

• Memgraph v3.10.1 (memgraph/memgraph-mage:3.10.1)
• Spring Boot 3.3.4 / Spring Data Neo4j 7.3.4
• Java 17

⚠️ Use Memgraph 3.10.1 or newer. On older versions (e.g. 3.2) the relationship-persistence query that Spring Data Neo4j generates could terminate the server connection — this was verified fixed on 3.10.1.

If you run into any issues or have questions, feel free to reach out on the Memgraph Discord server. We're happy to help!

🏢 Enterprise or Community?

✅ This example runs on Memgraph Community Edition.