Top K Frequent - Implementation

Here's the solution with a min-heap of size $k$:

function topKFrequent(nums, k)
    // Phase 1: Count frequencies
    freq := empty hash map
    for each num in nums
        freq[num] := freq[num] + 1

    // Phase 2: Find top k by frequency
    heap := empty min-heap of (frequency, element) pairs

    for each (element, count) in freq
        if size of heap < k
            insert (count, element) into heap
        else if count > frequency at heap root
            extract min from heap
            insert (count, element) into heap

    return all elements from heap

Time: $O(n \log k)$. Space: $O(n)$.