As we have already observed, Hadoop is very good at reading through a dataset and calculating the analytics. However, we will often have to merge two datasets to analyze the data. This recipe will explain how to join two datasets using Hadoop.

As an example, this recipe will use the Tomcat developer archives dataset. A common belief among the open source community is that the more a developer is involved with the community (for example, by replying to e-mail threads in the project's mailing lists and helping others and so on), the more quickly they will receive responses to their queries. In this recipe, we will test this hypothesis using the Tomcat developer mailing list.

To test this hypothesis, we will run the MapReduce jobs as explained in the following figure:

We will use the MBOX-formatted e-mail archives and use the custom InputFormat and RecordReader explained in the earlier recipe to parse them. Map tasks will receive the sender of the e-mail (from...