What is this

This is an exercise about creating a UDF in Java for Apache Spark to read a column of type binary in a Dataframe that encodes a Java Class. More details can be found in this Medium article.

Compile and Run

The application is as self-contained as possible. You just need to have Java and Maven installed on your system. You then need to run:

mvn clean install

and:

mvn exec:java

That's it. This will create a local Apache Spark cluster, create a dummy DataSet of MyWonderfulClass objects and then run the newly created UDF on it.