Auto-Generated Field for MongoDB using Spring Boot

1. Overview

In this tutorial, we’re going to learn how to implement a sequential, auto-generated field for MongoDB in Spring Boot.

When we’re using MongoDB as the database for a Spring Boot application, we can’t use @GeneratedValue annotation in our models as it’s not available. Hence we need a method to produce the same effect as we’ll have if we’re using JPA and an SQL database.

The general solution to this problem is simple. We’ll create a collection (table) that’ll store the generated sequence for other collections. During the creation of a new record, we’ll use it to fetch the next value.

2. Dependencies

Let’s add the following spring-boot starters to our pom.xml:

<dependencies>
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-web</artifactId>
        <versionId>2.1.0.RELEASE</versionId>
    </dependency>
    <dependency>
        <groupId>org.springframework.boot</groupId>
        <artifactId>spring-boot-starter-data-mongodb</artifactId>
        <versionId>2.1.0.RELEASE</versionId>
    </dependency>
</dependencies>

The latest version for the dependencies is managed by spring-boot-starter-parent.

3. Collections

As discussed in the overview, we’ll create a collection that’ll store the auto-incremented sequence for other collections. We’ll call this collection database_sequences. It can be created using either the mongo shell or MongoDB Compass. Let’s create a corresponding model class:

@Document(collection = "database_sequences")
public class DatabaseSequence {

    @Id
    private String id;

    private long seq;

    //getters and setters omitted
}

Let’s then create a users collection, and a corresponding model object, that’ll store the details of people that are using our system:

@Document(collection = "users")
public class User {

    @Transient
    public static final String SEQUENCE_NAME = "users_sequence";

    @Id
    private long id;

    private String email;

    //getters and setters omitted
}

In the User model created above, we added a static field SEQUENCE_NAME, which is a unique reference to the auto-incremented sequence for the users collection.

We also annotate it with the @Transient to prevent it from being persisted alongside other properties of the model.

4. Creating a New Record

So far, we’ve created the required collections and models. Now, we’ll create a service that’ll generate the auto-incremented value that can be used as id for our entities.

Let’s create a SequenceGeneratorService that has generateSequence():

public long generateSequence(String seqName) {
    DatabaseSequence counter = mongoOperations.findAndModify(query(where("_id").is(seqName)),
      new Update().inc("seq",1), options().returnNew(true).upsert(true),
      DatabaseSequence.class);
    return !Objects.isNull(counter) ? counter.getSeq() : 1;
}

Now, we can use the generateSequence() while creating a new record:

User user = new User();
user.setId(sequenceGenerator.generateSequence(User.SEQUENCE_NAME));
user.setEmail("john.doe@example.com");
userRepository.save(user);

To list all the users, we’ll use the UserRepository:

List<User> storedUsers = userRepository.findAll();
storedUsers.forEach(System.out::println);

As it is now, we have to set the id field every time we create a new instance of our model. We can circumvent this process by creating a listener for Spring Data MongoDB lifecycle events.

To do that, we’ll create a UserModelListener that extends AbstractMongoEventListener<User> and then we’ll override the onBeforeConvert():

@Override
public void onBeforeConvert(BeforeConvertEvent<User> event) {
    event.getSource().setId(sequenceGenerator.generateSequence(User.SEQUENCE_NAME));
}

Now, every time we save a new User, the id will be set automatically.

5. Conclusion

In conclusion, we’ve seen how to generate sequential, auto-incremented values for the id field and simulate the same behavior as seen in SQL databases.

Hibernate uses a similar method for generating auto-incremented values by default.

As usual, the complete source code is available over on Github.