hmrc / simple-reactivemongo

Simple case class serialization for Reactivemongo

Version Matrix


Join the chat at Apache-2.0 license Build Status Download Stories in Ready

Provides simple serialization for ReactiveMongo - reactive, asynchronous and non-blocking Scala driver for MongoDB.

This started as a fork of Play-ReactiveMongo as it seemed like a good idea to refactor out the coupling to a Play application.

With some minimal effort, as the ReactiveMongo people had already done the majority of the work, we felt that adding a base repository class creates a library without some of the issues the other simpler libraries have.

Upgrading from 5.x.x to 6.x.x?

With version 6.x.x of simple-reactivemongo, we are moving to the latest version of reactivemongo which comes with a few braking changes documented here:

You will most likely encounter some of the following issues. Please have a look on how we recommend to fix them.

No Json serializer as JsObject found

Due to keeping parity with upstream, companion objects for inner classes do not have the ImplicitBSONHandlers handlers. If you see an error like this: No Json serializer as JsObject found for type reactivemongo.bson.BSONDocument. Try to implement an implicit OWrites or OFormat for this type. Try adding the following import


WriteResult is no longer an Exception

The type hierarchy of the trait WriteResult has changed in new version of reactivemongo. It’s no longer an Exception. As it no longer represents errors in the public API, the following properties have been removed: errmsg, hasErrors, inError and message.

simple-reactivemongo previously exposed inError, something that is no longer possible.

Also, given that the LastError was part of the WriteResult, it used to be returned in a Future.Success, now it's returned in a Future.Failure.

If you have something like this in your code:

collection.doSomething([...]) map {
    case lastError if lastError.inError => // handle the error
    case _ => // return successfully

Now, you have to do this:

collection.doSomething([...]) map {
    case _ => // return successfully
} recover {
   case lastError => // handle the error

Main features

CASE CLASS <-> JSON <-> BSON conversion

Simple-reactivemongo uses Play Json to serialise/deserialise JSON to/from case classes and there is a specialized collection called JSONCollection that deals naturally with JSValue and JSObject instead of ReactiveMongo's BSONDocument.

Add simple-reactivemongo

In your project/build.sbt:

resolvers += Resolver.bintrayRepo("hmrc", "releases")

libraryDependencies ++= Seq(
  "" %% "simple-reactivemongo" % "[INSERT_VERSION]",
  "" %% "play-json" % "2.x.x"

For Java 7 and Play 2.3.x use versions <=4.1.0

Create a Repository class

Create a case class that represents to serialise to mongo.

Create JSON Read/Write converters. Or if you are doing nothing special create a companion object for the case class with an implicit member set by play.api.libs.json.Json.format[A]

Extend ReactiveRepository which will provide you with some commonly used functionality.

If the repository requires any indexes override indexes: Seq[Index] to provide a sequence of indexes that will be applied. Any errors will be logged should they fail.

If you prefer to drop the underscore for the 'id' field in the domain case class then wrap the domain formats in ReactiveMongoFormats.mongoEntity

case class TestObject(aField: String,
                      anotherField: Option[String] = None,
                      optionalCollection: Option[List[NestedModel]] = None,
                      nestedMapOfCollections: Map[String, List[Map[String, Seq[NestedModel]]]] = Map.empty,
                      modifiedDetails: CreationAndLastModifiedDetail = CreationAndLastModifiedDetail(),
                      jsValue: Option[JsValue] = None,
                      location : Tuple2[Double, Double] = (0.0, 0.0),
                      id: BSONObjectID = BSONObjectID.generate) {

  def markUpdated(implicit updatedTime: DateTime) = copy(
    modifiedDetails = modifiedDetails.updated(updatedTime)


object TestObject {

  import ReactiveMongoFormats.{objectIdFormats, mongoEntity}

  implicit val formats = mongoEntity {

    implicit val locationFormat = TupleFormats.tuple2Format[Double, Double]

    implicit val nestedModelformats = Json.format[NestedModel]


class SimpleTestRepository(implicit mc: MongoConnector)
  extends ReactiveRepository[TestObject, BSONObjectID]("simpleTestRepository", mc.db, TestObject.formats, ReactiveMongoFormats.objectIdFormats) {

  import reactivemongo.api.indexes.IndexType
  import reactivemongo.api.indexes.Index

  override def indexes: Seq[Index] = Seq(
    Index(Seq("aField" -> IndexType.Ascending), name = Some("aFieldUniqueIdx"), unique = true, sparse = true)

(See ReactiveRepositorySpec for example usage)

Built-in JSON converters (Formats) for often used types

Formats for BSONObjectId and Joda time classes are implemented (see ReactiveMongoFormats)

Configure underlying Akka system

ReactiveMongo loads it's configuration from the key mongo-async-driver

To change the log level (prevent dead-letter logging for example)

mongo-async-driver {
  akka {
    loglevel = WARNING

Create a repository class using AtomicsUpdate

The AtomicUpdate trait is a wrapper around the findAndModify command which modifies and returns a single document, using atomic operations. By default, the returned mongo document will include the modifications made on the update. Please refer to the mongo documentation concerning the commands which can be supplied to the update operations.

To include the trait AtomicsUpdate to your existing repository, follow the steps below.

  1. Update your repository to use the AtomicUpdate trait, passing the type of the object being read/written.
class SimpleTestRepository(implicit mc: MongoConnector)
  extends ReactiveRepository[TestObject, BSONObjectID]("simpleTestRepository", mc.db, TestObject.formats, ReactiveMongoFormats.objectIdFormats)
  with AtomicUpdate[TestObject]
  1. Include the below override in your class which is extending AtomicUpdate. This function is invoked by AtomicUpdate to decide if the update is either an upsert or an update.
    override def isInsertion(suppliedId: BSONObjectID, returned: AtomicTestObject): Boolean =

The two functions exposed from the AtomicUpdate trait are detailed below which both return the type Future[Option[DatabaseUpdate[T]]], where DatabaseUpdate encapsulates the update type which can be either Saved (new insert) or Updated (updated record). Please note the document with the modifications applied on the update command to mongo will be returned.

  1. def atomicUpsert(finder: BSONDocument, modifierBson: BSONDocument) This function will invoke atomicSaveOrUpdate passing the 'upsert' flag as true. Function used to insert a new record.

def atomicSaveOrUpdate(finder: BSONDocument, modifierBson: BSONDocument, upsert: Boolean)
 This function is used to override the upsert parameter.

The parameters for the functions above are defined below:-

  • finder - A BSON finder used to find an existing record.
  • modifierBson - The BSON modifier to be applied.
  • upsert - If the value is true, a BSONDocument will be added to the modifierBson parameter to generate the Id field on the collection using BSONObjectID.generate. If the value is false, no additional BSONDocument will be applied to modifierBson. If the 'finder' returns no document, then None will be returned. If unsure if the record already exists, then set the 'upsert' value to true.
Please refer to the unit test AtomicUpdateSpec for simple examples concerning using the trait AtomicUpdate.

For documentation, please refer to