In the previous articles (here, and here) I gave the background to a project we did for a client, exploring the benefits of Spark-based ETL processing »