Implement `top_k_frequent(values, k)`. Return the `k` distinct integers that occur most often. Order the result by decreasing frequency; when two values have the same frequency, place the value whose first occurrence in the input is earlier first. `k` is positive and does not exceed the number of distinct values. Target `O(n)` time by using the fact that no value can occur more than `n` times. ```hint Use frequency as an array index Count values in a hash map, then place each distinct value into a bucket whose index is its frequency. ``` ```hint Make tied buckets deterministic After counting, scan the original input once more. The first time each value is encountered, append it to its frequency bucket; equal-frequency values will then retain first-occurrence order without a comparison sort. ``` ### Discussion Extensions - Compare bucket selection with full sorting and a size-`k` heap. - For a continuous stream, what additional structure would let frequency changes update a maintained top set without rebuilding all buckets?