algorithm - DataStructure for Intersection Counts -


we have requirement have maintain distinct counts every hour of day of month, various combinations(user meeting criteria). thinking of using hyperloglog it, 1 of other requirements provide counts of union , intersection matching conditions(criteria).

we have these operations on day/week/month. far have read unions supported via hyperloglog. intersections more 2 hyperloglog seems have high error rates. there other data structure intersections meeting low space requirements high cardinality or supports intersection , union counting large distinct occurrences ?

any pointers helpful. thanks!!

check out augmenting hyperloglog minhash.


Comments