For each log entry, I have "user identifier" and "date" fields. I would like to find out the number of unique dates that a "user identifier" appears in.
Eg. For a given user identifier "1234567", if it appeared in different log entries of 5 different dates, then I would like the histogram to show "1234567": 5
If I have a million user identifiers then I would have a million counts to compute.
Is this possible? I ran into an error saying that I exceeded the bucket limit of 10000. I foresee that this problem will not be solved just by increasing the bucket limit..