DatasetDataFrameRDD
Spark v1.6 in 2015Spark v1.3 in 2013Spark v1.0 in 2011
Type-safeNot type-safeType-safe ??
High level APIHigh level APILow level API
EncodersCatalystTungsten?
OptimizationCatalyst OptimizerNo Optimization
OOPS Style APISQL Style APIOOPS Style API
What to do ApproachWhat to do ApproachHow to do Approach
Scala, JavaScala, Java, Python, RScala, Java, Python, R
Structured SchemaStructured SchemaNo Schema
Compile Time ErrorRun Time ErrorCompile Time Error
Serialization can be avoidedSerialization can be avoidedSerialization can’t be avoided