Apache Spark Datasets are an extension of the DataFrame API that provide a type-safe object-oriented programming interface. This API was first introduced in the 1.6 release. Spark 2.0 version brought out unification of DataFrame and Dataset APIs. DataFrame becomes a generic, untyped Dataset; or a Dataset is a DataFrame with an added structure. The term "structure" in this context refers to a pattern or an organization of underlying data, more like a table schema in RDBMS parlance. The structure imposes a limit on what can be expressed or contained in the underlying data. This in turn enables better optimizations in memory organization as well as physical execution. Compile-time type checking leads to catching errors earlier than during runtime. For example, a type mismatch in a SQL comparison does not get caught until runtime, whereas it would be caught during compile time itself if it were expressed as a sequence of operations on Datasets. However, the inherent dynamic nature of...