Mongo DB + java grouping documents by feild on selected date range

Question

I am new to MongoDB, using it in Java using mongoJavaDriver.jar.

I have more documents in my collection, below is a part to sample it.

{
    "_id" : ObjectId("57ee4767d782023f80bd4f97"),
    "USER_NAME" : "abc123xgvdfbvdf@einrot.com",
    "LOGIN" : ISODate("2016-09-29T18:30:00Z"),
    "LOGOUT" : ISODate("2016-09-30T11:07:19.598Z"),
    "LONGITUDE" : "long",
    "LATITUDE" : "lat",
    "SOURCE" : "Web",
    "LAST_UPDATED" : ISODate("2016-09-30T11:07:19.598Z")
}
{
    "_id" : ObjectId("57ee4767d782023f80bd4f98"),
    "USER_NAME" : "abc123xgvdfbvdf@fleckens.hu",
    "LOGIN" : ISODate("2016-09-29T18:30:00Z"),
    "LOGOUT" : ISODate("2016-09-30T11:07:19.601Z"),
    "LONGITUDE" : "long",
    "LATITUDE" : "lat",
    "SOURCE" : "Web",
    "LAST_UPDATED" : ISODate("2016-09-30T11:07:19.601Z")
}
{
    "_id" : ObjectId("57ee4767d782023f80bd4f99"),
    "USER_NAME" : "prashanthtfxgvdfbvdf@gmail.com",
    "LOGIN" : ISODate("2016-09-29T18:30:00Z"),
    "LOGOUT" : ISODate("2016-09-30T11:07:19.603Z"),
    "LONGITUDE" : "long",
    "LATITUDE" : "lat",
    "SOURCE" : "Web",
    "LAST_UPDATED" : ISODate("2016-09-30T11:07:19.603Z")
}
{
    "_id" : ObjectId("57ee4767d782023f80bd4f9e"),
    "USER_NAME" : "57ee4767d782023f80bd4f9f@jourrapide.com",
    "LOGIN" : ISODate("2016-09-29T18:30:00Z"),
    "LOGOUT" : ISODate("2016-09-30T11:07:19.608Z"),
    "LONGITUDE" : "long",
    "LATITUDE" : "lat",
    "SOURCE" : "Web",
    "LAST_UPDATED" : ISODate("2016-09-30T11:07:19.608Z")
}
{
    "_id" : ObjectId("57ee4767d782023f80bd4f9f"),
    "USER_NAME" : "zxcxgvdfbvdf@jourrapide.com",
    "LOGIN" : ISODate("2016-09-29T18:30:00Z"),
    "LOGOUT" : ISODate("2016-09-30T11:07:19.609Z"),
    "LONGITUDE" : "long",
    "LATITUDE" : "lat",
    "SOURCE" : "Web",
    "LAST_UPDATED" : ISODate("2016-09-30T11:07:19.609Z")
}

I am able to run query to find documents on selected date range and it returns some value.

My Code:

    Date current = new Date();
    Date current2  = new Date(current .getYear(), current.getMonth(), current.getDate()-days);
    current2.setHours(0);
    current2.setMinutes(0);
    BasicDBObject dateRange = new BasicDBObject ("$gte",current2 );
    dateRange.put("$lt", new Date(current.getYear(), current.getMonth(), current.getDate()+1));
    BasicDBObject query = new BasicDBObject("LAST_UPDATED", dateRange);
    System.out.println(collection.find().count());
    System.out.println(collection.find(query).count());

Total documents count in this collection is 155.

Count based on the selected date is 37.

In these selected 37 documents, there are redundant values in username. I want to group them based on the USER_NAME field. so that this count will become less than 37. How can I do that?

Answer 1

You can use the aggregation framework to carry out the needed aggregation. Consider running the following aggregation framework pipeline in mongo shell which basically uses the $match filter to restrict documents getting into the pipeline for processing based on the date range query supplied, then groups the documents by the USER_NAME field using $group and count the distinct values using $sum :

var now = new Date(),
    days = 4,
    start = new Date(now.getYear(), now.getMonth(), now.getDate()-days),
    end = new Date(now.getYear(), now.getMonth(), now.getDate()+1);
end.setHours(0);
end.setMinutes(0);
var pipeline = [
    { "$match": { "LAST_UPDATED": { "$gte": start, "$lte": end } } },
    {
        "$group": {
            "_id": "$USER_NAME",
            "count": { "$sum": 1 }
        }
    }
];
db.collection.aggregate(pipeline);

---

Converting the above to Java becomes:

public class JavaAggregation {
    public static void main(String args[]) throws UnknownHostException {

        MongoClient mongo = new MongoClient();
        DB db = mongo.getDB("test"); // your database name

        DBCollection coll = db.getCollection("collectionName"); // your collection name

        // create the pipeline operations, first with the $match
        Date now = new Date();
        Date start = new Date(now.getYear(), now.getMonth(), now.getDate()-days);
        Date end  = new Date(now.getYear(), now.getMonth(), now.getDate()+1);
        end.setHours(0);
        end.setMinutes(0);
        DBObject match = new BasicDBObject("$match",
                            new BasicDBObject("LAST_UPDATED",
                                new BasicDBObject("$gte", start).append("$lt", end)
                            )
                        );

        // build the $group operations
        DBObject groupFields = new BasicDBObject( "_id", "$USER_NAME");
        groupFields.put("count", new BasicDBObject( "$sum", 1));

        DBObject group = new BasicDBObject("$group", groupFields);
        List<DBObject> pipeline = Arrays.asList(match, group);

        AggregationOutput output = coll.aggregate(pipeline);

        for (DBObject result : output.results()) {
            System.out.println(result);
        }
    }
}