Valentin has come a long way from C++ programming on little machines with 4 MB of unmanaged memory to big clusters with terabytes of RAM. Along the way, he progressively abandoned his old imperative habits to embrace the functional way—a journey that seems to never come to an end.
When validating data with Spark, or read/writing it to Kafka topics, the goto solution is to write a Scala case class or a Java Bean.
But what if you had only 5 developers, 10000+ data structures and only a few months to ship your project?
Let me show you how the power of hylomorphisms combined with expressive schemas allowed us to write the code that validates and transforms data from dozens of tables from hundred of data sources, and ship our project in time and on budget.Slides