import json
table = sc.textFile("../messages.json")#.sample(False,0.0001,12345)
dataset = table.map(json.loads)
dataset.persist()
dataset.take(1)
sub_auth = dataset.map(lambda d: (d['subreddit'],d['author'])).distinct()
auth_occ = sub_auth.map(lambda r: (r[1],1)).reduceByKey(lambda a,b: a+b)
auth_occ.sortBy(lambda a: a[1]).take(50)