Spark Core How to fetch max n rows of an RDD function without using Rdd.max()

Question

I have an RDD having below elements:

(\'09\', [25, 66, 67])
(\'17\', [66, 67, 39])
(\'04\', [25])
(\'08\', [120, 122])
(\'28\', [25, 67])
(\'30\', [1