DObject[T] can be transformed to a DList[T]with a single element and to a DList[Iterable[S]] if T <:< Iterable[S]groupByKeyWith method to explicitly specify the Grouping instance to useinmemory argument to run the Scoobi code with Scala collections instead of Hadoop in order to faster check the application logic(A, B) has a Grouping instance if A and B does$HADOOP_CLASSPATH variablelibjars directory in the cluster if its size has changedlibjars directory can be changed with a system variabletranspose method for matricesGrouping, Ordering and WireFormatinstances for tagged typesScoobiAppWireFormat instances for Shapeless tagged typesIterable returned by materialise lazyWireFormats with up to 22 arguments in tuples and case classesDataSink api to declare that outputs must be compressedDMatrix can be multiplied by a function/tmp/scoobi-{user.name}/by default nowmapreduce.task.classpath.user.precedence=trueDList.materialize is deprecated, useDList.materialise instead