REST APIs with Java, Spring Boot, and MongoDB
If you want to write REST APIs in Java at the speed of light, I have what you need. I wrote this template to get you started. I have tried to solve as many problems as possible in it.
So if you want to start writing REST APIs in Java, clone this project, and you will be up to speed in no time.
1 git clone https://github.com/mongodb-developer/java-spring-boot-mongodb-starter
That’s all folks! All you need is in this repository. Below I will explain a few of the features and details about this template, but feel free to skip what is not necessary for your understanding.
All the extra information and commands you need to get this project going are in the
README.md
file which you can read in GitHub.1 package com.mongodb.starter; 2 3 import [...] 4 5 import static org.bson.codecs.configuration.CodecRegistries.fromProviders; 6 import static org.bson.codecs.configuration.CodecRegistries.fromRegistries; 7 8 9 public class MongoDBConfiguration { 10 11 12 private String connectionString; 13 14 15 public MongoClient mongoClient() { 16 CodecRegistry pojoCodecRegistry = fromProviders(PojoCodecProvider.builder().automatic(true).build()); 17 CodecRegistry codecRegistry = fromRegistries(MongoClientSettings.getDefaultCodecRegistry(), pojoCodecRegistry); 18 return MongoClients.create(MongoClientSettings.builder() 19 .applyConnectionString(new ConnectionString(connectionString)) 20 .codecRegistry(codecRegistry) 21 .build()); 22 } 23 24 }
The important section here is the MongoDB configuration, of course. Firstly, you will notice the connection string is automatically retrieved from the
application.properties
file, and secondly, you will notice the configuration of the MongoClient
bean.A
Codec
is the interface that abstracts the processes of decoding a BSON value into a Java object and encoding a Java object into a BSON value.A
CodecRegistry
contains a set of Codec
instances that are accessed according to the Java classes that they encode from and decode to.The MongoDB driver is capable of encoding and decoding BSON for us, so we do not have to take care of this anymore. All the configuration we need for this project to run is here and nowhere else.
Just for the sake of it, I also used multi-document ACID transactions in a few methods where it could potentially make sense to use ACID transactions. You can check all the code in the
MongoDBPersonRepository
class.Here is an example:
1 private static final TransactionOptions txnOptions = TransactionOptions.builder() 2 .readPreference(ReadPreference.primary()) 3 .readConcern(ReadConcern.MAJORITY) 4 .writeConcern(WriteConcern.MAJORITY) 5 .build(); 6 7 8 public List<PersonEntity> saveAll(List<PersonEntity> personEntities) { 9 try (ClientSession clientSession = client.startSession()) { 10 return clientSession.withTransaction(() -> { 11 personEntities.forEach(p -> p.setId(new ObjectId())); 12 personCollection.insertMany(clientSession, personEntities); 13 return personEntities; 14 }, txnOptions); 15 } 16 }
As you can see, I’m using an auto-closeable try-with-resources which will automatically close the client session at the end. This helps me to keep the code clean and simple.
Some of you may argue that it is actually too simple because transactions (and write operations, in general) can throw exceptions, and I’m not handling any of them here… You are absolutely right and this is an excellent transition to the next part of this article.
Transactions in MongoDB can raise exceptions for various reasons, and I don’t want to go into the details too much here, but since MongoDB 3.6, any write operation that fails can be automatically retried once. And the transactions are no different. See the documentation for retryWrites.
If retryable writes are disabled or if a write operation fails twice, then MongoDB will send a MongoException (extends RuntimeException) which should be handled properly.
Luckily, Spring provides the annotation
ExceptionHandler
to help us do that. See the code in my controller PersonController
. Of course, you will need to adapt and enhance this in your real project, but you have the main idea here.1 2 public final ResponseEntity<Exception> handleAllExceptions(RuntimeException e) { 3 logger.error("Internal server error.", e); 4 return new ResponseEntity<>(e, HttpStatus.INTERNAL_SERVER_ERROR); 5 }
MongoDB's aggregation pipeline is a very powerful and efficient way to run your complex queries as close as possible to your data for maximum efficiency. Using it can ease the computational load on your application.
Just to give you a small example, I implemented the
/api/persons/averageAge
route to show you how I can retrieve the average age of the persons in my collection.1 2 public double getAverageAge() { 3 List<Bson> pipeline = List.of(group(new BsonNull(), avg("averageAge", "$age")), project(excludeId())); 4 return personCollection.aggregate(pipeline, AverageAgeDTO.class).first().averageAge(); 5 }
Also, you can note here that I’m using the
personCollection
which was initially instantiated like this:1 private MongoCollection<PersonEntity> personCollection; 2 3 4 void init() { 5 personCollection = client.getDatabase("test").getCollection("persons", PersonEntity.class); 6 }
Normally, my personCollection should encode and decode
PersonEntity
object only, but you can overwrite the type of object your collection is manipulating to return something different — in my case, AverageAgeDTO.class
as I’m not expecting a PersonEntity
class here but a POJO that contains only the average age of my "persons".Swagger is the tool you need to document your REST APIs. You have nothing to do — the configuration is completely automated. Just run the server and navigate to http://localhost:8080/swagger-ui.html. the interface will be waiting for you.
You can test the REST APIs from this web page and explore the models. Don’t forget to disable it in production. ;-)
That's all I did to make it work:
1 <dependency> 2 <groupId>org.springdoc</groupId> 3 <artifactId>springdoc-openapi-starter-webmvc-ui</artifactId> 4 <version>2.2.0</version> 5 </dependency>
Yes, there is a Nyan Cat section in this post. Nyan Cat is love, and you need some Nyan Cat in your projects. :-)
Did you know that you can replace the Spring Boot logo in the logs with pretty much anything you want?
Well, now you know. You are welcome.
Have a look at the
banner.txt
file if you want to replace this awesome Nyan Cat with your own custom logo.I like to use patorjk and the "Epic" font for each project name. It's easier to identify which log file I am currently reading.
I hope you like my template, and I hope I will help you be more productive with MongoDB and the Java stack.
If you see something which can be improved, please feel free to open a GitHub issue or directly submit a pull request. They are very welcome. :-)
If you are new to MongoDB Atlas, give our Quick Start post a try to get up to speed with MongoDB Atlas in no time.