Major Description
currently, when user set dozens of precise count distinct metrics in one cube, we put all the count distinct metrics column in one columnFamily. Which result in HBase scan become slow because the one KeyValue is too big. we could set more columnFamily to speed up the HBase scan in this scenario.