Currently TimeSeriesDataFrame stores a time unit and a time column name internally. This is not consistent with the Scala implementation of TimeSeriesRDD and causes extra copying and complexity. We should just force the timeUnit and timeColumn to be nanoseconds and 'time'
Currently TimeSeriesDataFrame stores a time unit and a time column name internally. This is not consistent with the Scala implementation of TimeSeriesRDD and causes extra copying and complexity. We should just force the timeUnit and timeColumn to be nanoseconds and 'time'