Problem editorialmedium
Top K frequent words
Count everything, then rank by two rules at once: most frequent first, ties broken alphabetically. The real algorithm lives in the comparator.
15 min read·time O(N + U log K)·space O(U)
Count then rankComparator designBounded heap