Hi guys, I am wondering if it's possible to estimate the number of distinct keys and their distribution in a way or another.
More concretely, for every stage, it is possible to determine the number of distinct keys and for each key the number of values before the data is actually processed? Thanks,Robert
