Mongodb $lookup in Spring data mongo

Question

I\'m a new Mongodb and I have a problem with $lookup with java spring.

I would like to use this shell in Spring data

db.NewFeed.aggregate([
    {
           


        

Answer 1

Here is an example:

Collection posts

{
"_id" : ObjectId("5a198074ed31adaf5d79fe8a"),
"title" : "Post 1",
"authors" : [1, 2]
},
{
"_id" : ObjectId("5a198074ed31adaf5d79fe8d"),
"title" : "Post 2",
"authors" : [2]
}

Collection users

{
"_id" : ObjectId("5a18b483ed31ada08fd6ed82"),
"userId" : 1,
"name" : "Vinod Kumar"
},
{
"_id" : ObjectId("5a18b483ed31ada08fd6ed83"),
"userId" : 2,
"name" : "Jim Hazel"
},
{
"_id" : ObjectId("5a18b483ed31ada08fd6ed84"),
"userId" : 3,
"name" : "Alex Wong"
}

Mongodb query with lookup and match

db.users.aggregate([
{
  $lookup:
    {
      from: "users",
      localField: "userid",
      foreignField: "authors",
      as: "post"
    }
  },
  {
     $match: { "post": { $ne: [] } }
  }
]).pretty()

Spring Mongoopration syntax

LookupOperation lookupOperation = LookupOperation.newLookup().
            from("posts").
            localField("userid").
            foreignField("authors").
            as("post");

AggregationOperation match = Aggregation.match(Criteria.where("post").size(1));


Aggregation aggregation = Aggregation.newAggregation(lookupOperation, match);

List<BasicDBObject> results = mongoOperation.aggregate(aggregation, "users", BasicDBObject.class).getMappedResults();

Answer 2

Can use below to join 3 Collections

MongoClient mongoClient = new MongoClient( "localhost" , 27017 );

DB db = mongoClient.getDB( "DBname" );

BasicDBObject query = BasicDBObject.parse("{$match:{_id:61188}},\n" +
        "   {\n" +
        "     $lookup:\n" +
        "       {\n" +
        "         from: \"CustomerDetails\",\n" +
        "         localField: \"supplierId\",\n" +
        "         foreignField: \"supplierId\",\n" +
        "         as: \"newfield\"\n" +
        "       }\n" +
        "  }\n" +
        "  ,  {\n" +
        "     $lookup:\n" +
        "       {\n" +
        "         from: \"ItemDetails\",\n" +
        "         localField: \"supplierId\",\n" +
        "         foreignField: \"supplierId\",\n" +
        "         as: \"newfield\"\n" +
        "       }\n" +
        "  }");

AggregationOutput dumps = db.getCollection("HeaderInfo").aggregate(query);

System.out.println("result="+dumps.results());

Answer 3

Its too late to answer this, but it might helps other who are facing the same issue. If you are using spring-boot-data-mongodb-2.0 or above version then there is a easy way to implement this.

AggregationOperation match = Aggregation.match(Criteria.where("username").is("user001")));
AggregationOperation query = Aggregation.lookup("NewfeedContent", "content.contentId", "_id", "NewfeedContent");
// If you want to unwind
//AggregationOperation unwind = Aggregation.unwind("Patient");
Aggregation agr = Aggregation.newAggregation(query, match, unwind);
AggregationResults<Document> result = springTemplate.aggregate(agr, "CollectionName", Document.class);

Answer 4

Not every "new" feature makes it immediately into abstraction layers such as spring-mongo.

So instead, all you need do is define a class that uses the AggregationOperation interface, which will instead take a BSON Object specified directly as it's content:

public class CustomAggregationOperation implements AggregationOperation {
    private DBObject operation;

    public CustomAggregationOperation (DBObject operation) {
        this.operation = operation;
    }

    @Override
    public DBObject toDBObject(AggregationOperationContext context) {
        return context.getMappedObject(operation);
    }
}

Then you can use in your aggregation like this:

Aggregation aggregation = newAggregation(
    match(
        Criteria.where("username").is("user001")
    ),
    new CustomAggregationOperation(
        new BasicDBObject(
            "$lookup",
            new BasicDBObject("from", "NewFeedContent")
                .append("localField","content.contentId")
                .append("foreignField", "_id")
                .append("as", "NewFeedContent")
        )
    )
)

Which shows the custom class mixed with the built in match() pipeline helper.

All that happens underneath each helper is that they serialize to a BSON representation such as with DBObject anyway. So the constructor here just takes the object directly, and returns it directly from .toDBObject(), which is the standard method on the interface that will be called when serializing the pipline contents.

Answer 5

Joining Two Collections with Spring Data MongoDB

Employee Class

class Employee {
    private String _id;
    private String name;
    private String dept_id;
}

Department Class

class Department {
    private String _id;
    private String dept_name;
}

Employee Result Class

public class EmpDeptResult {

    private String _id;
    private String name;
    private List<Object> departments;
}

EmployeeService Class

public class EmployeeService {

    @Autowired
    private MongoTemplate mongoTemplate;

    private Logger LOGGER = LoggerFactory.getLogger(EmployeeService.class);

    public void lookupOperation(){
    LookupOperation lookupOperation = LookupOperation.newLookup()
                        .from("Department")
                        .localField("dept_id")
                        .foreignField("_id")
                        .as("departments");

    Aggregation aggregation = Aggregation.newAggregation(Aggregation.match(Criteria.where("_id").is("1")) , lookupOperation);
        List<EmpDeptResult> results = mongoTemplate.aggregate(aggregation, "Employee", EmpDeptResult.class).getMappedResults();
        LOGGER.info("Obj Size " +results.size());
    }
}