2016 exam, Partioning
Hey,
In the 2016 exam Exercise 4 Question 3, I do not understand how it was decided that the partioning should be applied to the "pairs" RDD and not the "bossEvents" RDD. Wouldn't it make morse sense to partition the boss events since there can be multiple events for the same character id whereas for pairs there's exactly one element with that character id.
Another possibility is to partition both the RDDs but I imagine we don't do that as there is no guarantee as to whether two same ranges will end up on the same node ?
Could you please explain how the choice of what to partition is made ?
Thanks in advance!