org.apache.spark
(Since version 0.9.0) use combineCombinersByKey with TaskContext argument
(Since version 0.9.0) use combineValuesByKey with TaskContext argument