Using MongoDB data inside Hadoop with the help of Morphia

Question

I've been playing with the MongoInputFormat that allows having all documents in a MongoDB collection put through a MapReduce job written in Hadoop.

As you can see in the provided examples (this, this and this) the type the document is in that is provided to the mapper is a BSONObject (an Interface in Java).

Now I also like Morphia very much which allows mapping the raw data from MongoDB into POJOs that are much easier to use.

Since I can only get a BSONObject as input I thought about using the method described at the bottom of this page of the Morphia wiki:

BlogEntry blogEntry = morphia.fromDBObject(BlogEntry.class, blogEntryDbObj);

My problem is that this method requires a DBObject instead of a BSONObject. A DBObject is really:

public interface DBObject extends BSONObject

So as you can see I cannot simply cast from BSONObject to DBObject and call the provided method.

How do I handle this the best way?

Answer 1

You'll notice that the BSONObject and DBObject interfaces are very similar. Just because a converson doesn't exist doesn't mean it's not easy to create a trivial one:

class BSONDBObject extends BasicBSONObject implements DBObject {
    boolean ispartial = false;
    public BSONDBObject(BSONObject source) {
        this.putAll(source);
    }
    public boolean isPartialObject() {
        return ispartial;
    }
    public void markAsPartialObject() {
        this.ispartial = true;
    }
}

Now, you just need to

BSONObject bson; // Filled by the MongoInputFormat
BSONBDObject dbo = BSONDBObject(bson);
EntityCache = entityCache = new DefaultEntityCache();
BlogEntry blogEntry = morphia().fromDBObject(BlogEntry.class, dbo, entityCache);

Answer 2

I found a solution that works for me:

• First make it into a JSON text string
• Parse this into a DBObject
• Map that using Morphia to a useful instance.

Effectively I now have something like this:

BSONObject bson; // Filled by the MongoInputFormat

EntityCache = entityCache = new DefaultEntityCache();
String json = JSON.serialize(bson)    
DBObject dbObject = (DBObject) JSON.parse(json);
BlogEntry blogEntry = morphia().fromDBObject(BlogEntry.class, dbObject, entityCache);