[HUDI-113]: Use Pair over # delimited string

[pseudomuto]

@vinothchandar

As outlined in HUDI-113, I've updated the bloom index to use a Pair rather than "#" as the separator.

[vinothchandar]

Caused by: org.apache.spark.SparkException: 
Job aborted due to stage failure: Task 1 in stage 40.0 failed 1 times, most recent failure: Lost task 1.0 in stage 40.0 (TID 56, localhost, executor driver): java.lang.ClassCastException: scala.Tuple2 cannot be cast to java.lang.String
	at com.uber.hoodie.index.bloom.BucketizedBloomCheckPartitioner.getPartition(BucketizedBloomCheckPartitioner.java:142)
	at org.apache.spark.shuffle.sort.BypassMergeSortShuffleWriter.write(BypassMergeSortShuffleWriter.java:152)
	at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:96)
	at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:53)

See some errors due to casting mismatch..

High level question... I see that sorting a List already provides records ordering by the first and second level (in that order). So, can we just go back to Tuple2<String, HoodieKey> or if we have to use a custom comparator, can it just work off that?

[vinothchandar]

Essentially, I saying the following should be sufficient information?

new Tuple2<>(matchingFile,  new HoodieKey(recordKey, partitionPath));

This will also reduce shuffled data.. a nice bonus..

[vinothchandar]

Minor style nits. lgtm otherwise. Can merge into the next release if you can confirm these

[vinothchandar]

nice clean up

[pseudomuto]

Apologies for the number of rebases. At this point, I think this is good. Settled on using Pair<String, String> and updated the partitioner accordingly.