使用Java在MongoDB上使用Hadoop Mapreduce

Question

I'm working on a MongoDB structure that has a tree structure.. as given in this link ( see docs )

I would like to know what will be passed in the key in the map method. Is it the topmost parent in the tree ans should i iterate to get my value list.

I have a tree structure that has company as parent and company name, website,product etc as children. The product will in turn have product name as its child.

So, if i need the company name to be the key, how should i get it using java?

What will be the key here - company ?

If there is any sample program or documentation for doing mapreduce over mongoDB in java, can anyone please provide me asap?

Thanks

Answer 1

Thanks Trisha. The above link is doing the in-built map reduce thats available in MongoDB. I wanted an example for a Map Reduce using Hadoop and MongoDB. I have finally figured it out.

public static class MongoMapper extends
        Mapper<Object, BasicDBObject, Text, BSONWritable> {


    public void map(Object key, BasicDBObject value, Context context)
            throws IOException, InterruptedException {


        //This gives all the documents of the company
        BasicDBObject company = (BasicDBObject) value.get("company");

        // This gets the company name
        name = (String) company.get("name");

        //This gives the phone number 
        phone = (String) company.get("phone");

If we need to access the product list in company we can retrieve using a BasicDBList

// This will give the list of products and we can use product.get()
// to get the product name and other details. 
BasicDBList product = (BasicDBList) company.get("products");

I referred to this slideshare link.

http://www.slideshare.net/spf13/introduction-to-mongodb-and-hadoop