Saving, Updating, and Removing Documents

MongoTemplate / ReactiveMongoTemplatge let you save, update, and delete your domain objects and map those objects to documents stored in MongoDB. The API signatures of the imperative and reactive API are mainly the same only differing in their return types. While the synchronous API uses void, single Object and List the reactive counterpart consists of Mono<Void>, Mono<Object> and Flux.

Consider the following class:

public class Person {

	private String id;
	private String name;
	private int age;

	public Person(String name, int age) {
		this.name = name;
		this.age = age;
	}

	public String getId() {
		return id;
	}

	public String getName() {
		return name;
	}

	public int getAge() {
		return age;
	}

	@Override
	public String toString() {
		return "Person [id=" + id + ", name=" + name + ", age=" + age + "]";
	}
}

Given the Person class in the preceding example, you can save, update and delete the object, as the following example shows:

  • Imperative

  • Reactive

public class MongoApplication {

  private static final Log log = LogFactory.getLog(MongoApplication.class);

  public static void main(String[] args) {

    MongoOperations template = new MongoTemplate(new SimpleMongoClientDbFactory(MongoClients.create(), "database"));

    Person p = new Person("Joe", 34);

    // Insert is used to initially store the object into the database.
    template.insert(p);
    log.info("Insert: " + p);

    // Find
    p = template.findById(p.getId(), Person.class);
    log.info("Found: " + p);

    // Update
    template.updateFirst(query(where("name").is("Joe")), update("age", 35), Person.class);
    p = template.findOne(query(where("name").is("Joe")), Person.class);
    log.info("Updated: " + p);

    // Delete
    template.remove(p);

    // Check that deletion worked
    List<Person> people =  template.findAll(Person.class);
    log.info("Number of people = : " + people.size());


    template.dropCollection(Person.class);
  }
}

The preceding example would produce the following log output (including debug messages from MongoTemplate):

DEBUG apping.MongoPersistentEntityIndexCreator:  80 - Analyzing class class org.spring.example.Person for index information.
DEBUG work.data.mongodb.core.MongoTemplate: 632 - insert Document containing fields: [_class, age, name] in collection: person
INFO               org.spring.example.MongoApp:  30 - Insert: Person [id=4ddc6e784ce5b1eba3ceaf5c, name=Joe, age=34]
DEBUG work.data.mongodb.core.MongoTemplate:1246 - findOne using query: { "_id" : { "$oid" : "4ddc6e784ce5b1eba3ceaf5c"}} in db.collection: database.person
INFO               org.spring.example.MongoApp:  34 - Found: Person [id=4ddc6e784ce5b1eba3ceaf5c, name=Joe, age=34]
DEBUG work.data.mongodb.core.MongoTemplate: 778 - calling update using query: { "name" : "Joe"} and update: { "$set" : { "age" : 35}} in collection: person
DEBUG work.data.mongodb.core.MongoTemplate:1246 - findOne using query: { "name" : "Joe"} in db.collection: database.person
INFO               org.spring.example.MongoApp:  39 - Updated: Person [id=4ddc6e784ce5b1eba3ceaf5c, name=Joe, age=35]
DEBUG work.data.mongodb.core.MongoTemplate: 823 - remove using query: { "id" : "4ddc6e784ce5b1eba3ceaf5c"} in collection: person
INFO               org.spring.example.MongoApp:  46 - Number of people = : 0
DEBUG work.data.mongodb.core.MongoTemplate: 376 - Dropped collection [database.person]
public class ReactiveMongoApplication {

  private static final Logger log = LoggerFactory.getLogger(ReactiveMongoApplication.class);

  public static void main(String[] args) throws Exception {

    CountDownLatch latch = new CountDownLatch(1);

    ReactiveMongoTemplate template = new ReactiveMongoTemplate(MongoClients.create(), "database");

    template.insert(new Person("Joe", 34)).doOnNext(person -> log.info("Insert: " + person))
      .flatMap(person -> template.findById(person.getId(), Person.class))
      .doOnNext(person -> log.info("Found: " + person))
      .zipWith(person -> template.updateFirst(query(where("name").is("Joe")), update("age", 35), Person.class))
      .flatMap(tuple -> template.remove(tuple.getT1())).flatMap(deleteResult -> template.findAll(Person.class))
      .count().doOnSuccess(count -> {
        log.info("Number of people: " + count);
        latch.countDown();
      })

      .subscribe();

    latch.await();
  }
}

MongoConverter caused implicit conversion between a String and an ObjectId stored in the database by recognizing (through convention) the Id property name.

The preceding example is meant to show the use of save, update, and remove operations on MongoTemplate / ReactiveMongoTemplate and not to show complex mapping functionality. The query syntax used in the preceding example is explained in more detail in the section “Querying Documents”.

MongoDB requires that you have an _id field for all documents. Please refer to the ID handling section for details on the special treatment of this field.
MongoDB collections can contain documents that represent instances of a variety of types. Please refer to the type mapping for details.

Insert / Save

There are several convenient methods on MongoTemplate for saving and inserting your objects. To have more fine-grained control over the conversion process, you can register Spring converters with the MappingMongoConverter — for example Converter<Person, Document> and Converter<Document, Person>.

The difference between insert and save operations is that a save operation performs an insert if the object is not already present.

The simple case of using the save operation is to save a POJO. In this case, the collection name is determined by name (not fully qualified) of the class. You may also call the save operation with a specific collection name. You can use mapping metadata to override the collection in which to store the object.

When inserting or saving, if the Id property is not set, the assumption is that its value will be auto-generated by the database. Consequently, for auto-generation of an ObjectId to succeed, the type of the Id property or field in your class must be a String, an ObjectId, or a BigInteger.

The following example shows how to save a document and retrieving its contents:

Inserting and retrieving documents using the MongoTemplate
  • Imperative

  • Reactive

import static org.springframework.data.mongodb.core.query.Criteria.where;
import static org.springframework.data.mongodb.core.query.Criteria.query;

//...

template.insert(new Person("Bob", 33));

Person person = template.query(Person.class)
    .matching(query(where("age").is(33)))
    .oneValue();
import static org.springframework.data.mongodb.core.query.Criteria.where;
import static org.springframework.data.mongodb.core.query.Criteria.query;

//...

Mono<Person> person = mongoTemplate.insert(new Person("Bob", 33))
    .then(mongoTemplate.query(Person.class)
        .matching(query(where("age").is(33)))
        .one());

The following insert and save operations are available:

  • void save (Object objectToSave): Save the object to the default collection.

  • void save (Object objectToSave, String collectionName): Save the object to the specified collection.

A similar set of insert operations is also available:

  • void insert (Object objectToSave): Insert the object to the default collection.

  • void insert (Object objectToSave, String collectionName): Insert the object to the specified collection.

How the _id Field is Handled in the Mapping Layer

MongoDB requires that you have an _id field for all documents. If you do not provide one, the driver assigns an ObjectId with a generated value without considering your domain model as the server isn’t aware of your identifier type. When you use the MappingMongoConverter, certain rules govern how properties from the Java class are mapped to this _id field:

  1. A property or field annotated with @Id (org.springframework.data.annotation.Id) maps to the _id field.

  2. A property or field without an annotation but named id maps to the _id field.

The following outlines what type conversion, if any, is done on the property mapped to the _id document field when using the MappingMongoConverter (the default for MongoTemplate).

  1. If possible, an id property or field declared as a String in the Java class is converted to and stored as an ObjectId by using a Spring Converter<String, ObjectId>. Valid conversion rules are delegated to the MongoDB Java driver. If it cannot be converted to an ObjectId, then the value is stored as a string in the database.

  2. An id property or field declared as BigInteger in the Java class is converted to and stored as an ObjectId by using a Spring Converter<BigInteger, ObjectId>.

If no field or property specified in the previous sets of rules is present in the Java class, an implicit _id file is generated by the driver but not mapped to a property or field of the Java class.

When querying and updating, MongoTemplate uses the converter that corresponds to the preceding rules for saving documents so that field names and types used in your queries can match what is in your domain classes.

Some environments require a customized approach to map Id values such as data stored in MongoDB that did not run through the Spring Data mapping layer. Documents can contain _id values that can be represented either as ObjectId or as String. Reading documents from the store back to the domain type works just fine. Querying for documents via their id can be cumbersome due to the implicit ObjectId conversion. Therefore documents cannot be retrieved that way. For those cases @MongoId provides more control over the actual id mapping attempts.

Example 1. @MongoId mapping
public class PlainStringId {
  @MongoId String id; (1)
}

public class PlainObjectId {
  @MongoId ObjectId id; (2)
}

public class StringToObjectId {
  @MongoId(FieldType.OBJECT_ID) String id; (3)
}
1 The id is treated as String without further conversion.
2 The id is treated as ObjectId.
3 The id is treated as ObjectId if the given String is a valid ObjectId hex, otherwise as String. Corresponds to @Id usage.

Into Which Collection Are My Documents Saved?

There are two ways to manage the collection name that is used for the documents. The default collection name that is used is the class name changed to start with a lower-case letter. So a com.test.Person class is stored in the person collection. You can customize this by providing a different collection name with the @Document annotation. You can also override the collection name by providing your own collection name as the last parameter for the selected MongoTemplate method calls.

Inserting or Saving Individual Objects

The MongoDB driver supports inserting a collection of documents in a single operation. The following methods in the MongoOperations interface support this functionality:

  • insert: Inserts an object. If there is an existing document with the same id, an error is generated.

  • insertAll: Takes a Collection of objects as the first parameter. This method inspects each object and inserts it into the appropriate collection, based on the rules specified earlier.

  • save: Saves the object, overwriting any object that might have the same id.

Inserting Several Objects in a Batch

The MongoDB driver supports inserting a collection of documents in one operation. The following methods in the MongoOperations interface support this functionality via insert or a dedicated BulkOperations interface.

Batch Insert
  • Imperative

  • Reactive

Collection<Person> inserted = template.insert(List.of(...), Person.class);
Flux<Person> inserted = template.insert(List.of(...), Person.class);
Bulk Insert
  • Imperative

  • Reactive

BulkWriteResult result = template.bulkOps(BulkMode.ORDERED, Person.class)
    .insert(List.of(...))
    .execute();
Mono<BulkWriteResult> result = template.bulkOps(BulkMode.ORDERED, Person.class)
    .insert(List.of(...))
    .execute();

Server performance of batch and bulk is identical. However bulk operations do not publish lifecycle events.

Any @Version property that has not been set prior to calling insert will be auto initialized with 1 (in case of a simple type like int) or 0 for wrapper types (eg. Integer).
Read more in the see Optimistic Locking section.

Update

For updates, you can update the first document found by using MongoOperation.updateFirst or you can update all documents that were found to match the query by using the MongoOperation.updateMulti method or all on the fluent API. The following example shows an update of all SAVINGS accounts where we are adding a one-time $50.00 bonus to the balance by using the $inc operator:

Updating documents by using the MongoTemplate / ReactiveMongoTemplate
  • Imperative

  • Reactive

import static org.springframework.data.mongodb.core.query.Criteria.where;
import org.springframework.data.mongodb.core.query.Update;

// ...

UpdateResult result = template.update(Account.class)
    .matching(where("accounts.accountType").is(Type.SAVINGS))
    .apply(new Update().inc("accounts.$.balance", 50.00))
    .all();
import static org.springframework.data.mongodb.core.query.Criteria.where;
import org.springframework.data.mongodb.core.query.Update;

// ...

Mono<UpdateResult> result = template.update(Account.class)
    .matching(where("accounts.accountType").is(Type.SAVINGS))
    .apply(new Update().inc("accounts.$.balance", 50.00))
    .all();

In addition to the Query discussed earlier, we provide the update definition by using an Update object. The Update class has methods that match the update modifiers available for MongoDB. Most methods return the Update object to provide a fluent style for the API.

@Version properties if not included in the Update will be automatically incremented. Read more in the see Optimistic Locking section.

Methods for Running Updates for Documents

  • updateFirst: Updates the first document that matches the query document criteria with the updated document.

  • updateMulti: Updates all objects that match the query document criteria with the updated document.

updateFirst does not support ordering. Please use findAndModify to apply Sort.
Index hints for the update operation can be provided via Query.withHint(…​).

Methods in the Update Class

You can use a little "'syntax sugar'" with the Update class, as its methods are meant to be chained together. Also, you can kick-start the creation of a new Update instance by using public static Update update(String key, Object value) and using static imports.

The Update class contains the following methods:

  • Update addToSet (String key, Object value) Update using the $addToSet update modifier

  • Update currentDate (String key) Update using the $currentDate update modifier

  • Update currentTimestamp (String key) Update using the $currentDate update modifier with $type timestamp

  • Update inc (String key, Number inc) Update using the $inc update modifier

  • Update max (String key, Object max) Update using the $max update modifier

  • Update min (String key, Object min) Update using the $min update modifier

  • Update multiply (String key, Number multiplier) Update using the $mul update modifier

  • Update pop (String key, Update.Position pos) Update using the $pop update modifier

  • Update pull (String key, Object value) Update using the $pull update modifier

  • Update pullAll (String key, Object[] values) Update using the $pullAll update modifier

  • Update push (String key, Object value) Update using the $push update modifier

  • Update pushAll (String key, Object[] values) Update using the $pushAll update modifier

  • Update rename (String oldName, String newName) Update using the $rename update modifier

  • Update set (String key, Object value) Update using the $set update modifier

  • Update setOnInsert (String key, Object value) Update using the $setOnInsert update modifier

  • Update unset (String key) Update using the $unset update modifier

Some update modifiers, such as $push and $addToSet, allow nesting of additional operators.

// { $push : { "category" : { "$each" : [ "spring" , "data" ] } } }
new Update().push("category").each("spring", "data")

// { $push : { "key" : { "$position" : 0 , "$each" : [ "Arya" , "Arry" , "Weasel" ] } } }
new Update().push("key").atPosition(Position.FIRST).each(Arrays.asList("Arya", "Arry", "Weasel"));

// { $push : { "key" : { "$slice" : 5 , "$each" : [ "Arya" , "Arry" , "Weasel" ] } } }
new Update().push("key").slice(5).each(Arrays.asList("Arya", "Arry", "Weasel"));

// { $addToSet : { "values" : { "$each" : [ "spring" , "data" , "mongodb" ] } } }
new Update().addToSet("values").each("spring", "data", "mongodb");

Aggregation Pipeline Updates

Update methods exposed by MongoOperations and ReactiveMongoOperations also accept an Aggregation Pipeline via AggregationUpdate. Using AggregationUpdate allows leveraging MongoDB 4.2 aggregations in an update operation. Using aggregations in an update allows updating one or more fields by expressing multiple stages and multiple conditions with a single operation.

The update can consist of the following stages:

  • AggregationUpdate.set(…​).toValue(…​)$set : { …​ }

  • AggregationUpdate.unset(…​)$unset : [ …​ ]

  • AggregationUpdate.replaceWith(…​)$replaceWith : { …​ }

Example 2. Update Aggregation
AggregationUpdate update = Aggregation.newUpdate()
    .set("average").toValue(ArithmeticOperators.valueOf("tests").avg())     (1)
    .set("grade").toValue(ConditionalOperators.switchCases(                 (2)
        when(valueOf("average").greaterThanEqualToValue(90)).then("A"),
        when(valueOf("average").greaterThanEqualToValue(80)).then("B"),
        when(valueOf("average").greaterThanEqualToValue(70)).then("C"),
        when(valueOf("average").greaterThanEqualToValue(60)).then("D"))
        .defaultTo("F")
    );

template.update(Student.class)                                              (3)
    .apply(update)
    .all();                                                                 (4)
db.students.update(                                                         (3)
   { },
   [
     { $set: { average : { $avg: "$tests" } } },                            (1)
     { $set: { grade: { $switch: {                                          (2)
                           branches: [
                               { case: { $gte: [ "$average", 90 ] }, then: "A" },
                               { case: { $gte: [ "$average", 80 ] }, then: "B" },
                               { case: { $gte: [ "$average", 70 ] }, then: "C" },
                               { case: { $gte: [ "$average", 60 ] }, then: "D" }
                           ],
                           default: "F"
     } } } }
   ],
   { multi: true }                                                          (4)
)
1 The 1st $set stage calculates a new field average based on the average of the tests field.
2 The 2nd $set stage calculates a new field grade based on the average field calculated by the first aggregation stage.
3 The pipeline is run on the students collection and uses Student for the aggregation field mapping.
4 Apply the update to all matching documents in the collection.

Upsert

Related to performing an updateFirst operation, you can also perform an upsert operation, which will perform an insert if no document is found that matches the query. The document that is inserted is a combination of the query document and the update document. The following example shows how to use the upsert method:

  • Imperative

  • Reactive

UpdateResult result = template.update(Person.class)
  .matching(query(where("ssn").is(1111).and("firstName").is("Joe").and("Fraizer").is("Update"))
  .apply(update("address", addr))
  .upsert();
Mono<UpdateResult> result = template.update(Person.class)
  .matching(query(where("ssn").is(1111).and("firstName").is("Joe").and("Fraizer").is("Update"))
  .apply(update("address", addr))
  .upsert();
upsert does not support ordering. Please use findAndModify to apply Sort.

@Version properties if not included in the Update will be automatically initialized. Read more in the see Optimistic Locking section.

Replacing Documents in a Collection

The various replace methods available via MongoTemplate allow to override the first matching Document. If no match is found a new one can be upserted (as outlined in the previous section) by providing ReplaceOptions with according configuration.

Replace one
Person tom = template.insert(new Person("Motte", 21)); (1)
Query query = Query.query(Criteria.where("firstName").is(tom.getFirstName())); (2)
tom.setFirstname("Tom"); (3)
template.replace(query, tom, ReplaceOptions.none()); (4)
1 Insert a new document.
2 The query used to identify the single document to replace.
3 Set up the replacement document which must hold either the same _id as the existing or no _id at all.
4 Run the replace operation. .Replace one with upsert
Person tom = new Person("id-123", "Tom", 21) (1)
Query query = Query.query(Criteria.where("firstName").is(tom.getFirstName()));
template.replace(query, tom, ReplaceOptions.replaceOptions().upsert()); (2)
1 The _id value needs to be present for upsert, otherwise MongoDB will create a new potentially with the domain type incompatible ObjectId. As MongoDB is not aware of your domain type, any @Field(targetType) hints are not considered and the resulting ObjectId might be not compatible with your domain model.
2 Use upsert to insert a new document if no match is found

It is not possible to change the _id of existing documents with a replace operation. On upsert MongoDB uses 2 ways of determining the new id for the entry: * The _id is used within the query as in {"_id" : 1234 } * The _id is present in the replacement document. If no _id is provided in either way, MongoDB will create a new ObjectId for the document. This may lead to mapping and data lookup malfunctions if the used domain types id property has a different type like e.g. Long.

Find and Modify

The findAndModify(…) method on MongoCollection can update a document and return either the old or newly updated document in a single operation. MongoTemplate provides four findAndModify overloaded methods that take Query and Update classes and converts from Document to your POJOs:

<T> T findAndModify(Query query, Update update, Class<T> entityClass);

<T> T findAndModify(Query query, Update update, Class<T> entityClass, String collectionName);

<T> T findAndModify(Query query, Update update, FindAndModifyOptions options, Class<T> entityClass);

<T> T findAndModify(Query query, Update update, FindAndModifyOptions options, Class<T> entityClass, String collectionName);

The following example inserts a few Person objects into the container and performs a findAndUpdate operation:

template.insert(new Person("Tom", 21));
template.insert(new Person("Dick", 22));
template.insert(new Person("Harry", 23));

Query query = new Query(Criteria.where("firstName").is("Harry"));
Update update = new Update().inc("age", 1);

Person oldValue = template.update(Person.class)
  .matching(query)
  .apply(update)
  .findAndModifyValue(); // oldValue.age == 23

Person newValue = template.query(Person.class)
  .matching(query)
  .findOneValue(); // newValye.age == 24

Person newestValue = template.update(Person.class)
  .matching(query)
  .apply(update)
  .withOptions(FindAndModifyOptions.options().returnNew(true)) // Now return the newly updated document when updating
  .findAndModifyValue(); // newestValue.age == 25

The FindAndModifyOptions method lets you set the options of returnNew, upsert, and remove. An example extending from the previous code snippet follows:

Person upserted = template.update(Person.class)
  .matching(new Query(Criteria.where("firstName").is("Mary")))
  .apply(update)
  .withOptions(FindAndModifyOptions.options().upsert(true).returnNew(true))
  .findAndModifyValue()

@Version properties if not included in the Update will be automatically incremented. Read more in the see Optimistic Locking section.

Find and Replace

The most straight forward method of replacing an entire Document is via its id using the save method. However this might not always be feasible. findAndReplace offers an alternative that allows to identify the document to replace via a simple query.

Example 3. Find and Replace Documents
Optional<User> result = template.update(Person.class)      (1)
    .matching(query(where("firstame").is("Tom")))          (2)
    .replaceWith(new Person("Dick"))
    .withOptions(FindAndReplaceOptions.options().upsert()) (3)
    .as(User.class)                                        (4)
    .findAndReplace();                                     (5)
1 Use the fluent update API with the domain type given for mapping the query and deriving the collection name or just use MongoOperations#findAndReplace.
2 The actual match query mapped against the given domain type. Provide sort, fields and collation settings via the query.
3 Additional optional hook to provide options other than the defaults, like upsert.
4 An optional projection type used for mapping the operation result. If none given the initial domain type is used.
5 Trigger the actual processing. Use findAndReplaceValue to obtain the nullable result instead of an Optional.
Please note that the replacement must not hold an id itself as the id of the existing Document will be carried over to the replacement by the store itself. Also keep in mind that findAndReplace will only replace the first document matching the query criteria depending on a potentially given sort order.

Delete

You can use one of five overloaded methods to remove an object from the database:

template.remove(tywin, "GOT");                                              (1)

template.remove(query(where("lastname").is("lannister")), "GOT");           (2)

template.remove(new Query().limit(3), "GOT");                               (3)

template.findAllAndRemove(query(where("lastname").is("lannister"), "GOT");  (4)

template.findAllAndRemove(new Query().limit(3), "GOT");                     (5)
1 Remove a single entity specified by its _id from the associated collection.
2 Remove all documents that match the criteria of the query from the GOT collection.
3 Remove the first three documents in the GOT collection. Unlike <2>, the documents to remove are identified by their _id, running the given query, applying sort, limit, and skip options first, and then removing all at once in a separate step.
4 Remove all documents matching the criteria of the query from the GOT collection. Unlike <3>, documents do not get deleted in a batch but one by one.
5 Remove the first three documents in the GOT collection. Unlike <3>, documents do not get deleted in a batch but one by one.

Optimistic Locking

The @Version annotation provides syntax similar to that of JPA in the context of MongoDB and makes sure updates are only applied to documents with a matching version. Therefore, the actual value of the version property is added to the update query in such a way that the update does not have any effect if another operation altered the document in the meantime. In that case, an OptimisticLockingFailureException is thrown. The following example shows these features:

@Document
class Person {

  @Id String id;
  String firstname;
  String lastname;
  @Version Long version;
}

Person daenerys = template.insert(new Person("Daenerys"));                            (1)

Person tmp = template.findOne(query(where("id").is(daenerys.getId())), Person.class); (2)

daenerys.setLastname("Targaryen");
template.save(daenerys);                                                              (3)

template.save(tmp); // throws OptimisticLockingFailureException                       (4)
1 Intially insert document. version is set to 0.
2 Load the just inserted document. version is still 0.
3 Update the document with version = 0. Set the lastname and bump version to 1.
4 Try to update the previously loaded document that still has version = 0. The operation fails with an OptimisticLockingFailureException, as the current version is 1.

Only certain CRUD operations on MongoTemplate do consider and alter version properties. Please consult MongoOperations java doc for detailed information.

Optimistic Locking requires to set the WriteConcern to ACKNOWLEDGED. Otherwise OptimisticLockingFailureException can be silently swallowed.
As of Version 2.2 MongoOperations also includes the @Version property when removing an entity from the database. To remove a Document without version check use MongoOperations#remove(Query,…​) instead of MongoOperations#remove(Object).
As of Version 2.2 repositories check for the outcome of acknowledged deletes when removing versioned entities. An OptimisticLockingFailureException is raised if a versioned entity cannot be deleted through CrudRepository.delete(Object). In such case, the version was changed or the object was deleted in the meantime. Use CrudRepository.deleteById(ID) to bypass optimistic locking functionality and delete objects regardless of their version.