HDFS can be used for storing the data for long-term. MapReduce jobs submitted from either Oozie, Pig or Hive can be used to encode, improve and sample the data sets from HDFS into R. This helps to leverage complex analysis tasks on the subset of data prepared in R.
