There are several ways to solve this problem. The approach that we are going to focus on here is called Bucket Sort. For details of Bucket Sort, you can visit this website: 点击打开链接. Here is a visualization of Bucket Sort: 点击打开链接.
public List<Integer> topKFrequent(int[] nums, int k) {
List<Integer>[] bucket = new List[nums.length + 1];
Map<Integer, Integer> frequencyMap = new HashMap<Integer, Integer>();
for (int n : nums) {
frequencyMap.put(n, frequencyMap.getOrDefault(n, 0) + 1);
}
for (int key : frequencyMap.keySet()) {
int frequency = frequencyMap.get(key);
if (bucket[frequency] == null) {
bucket[frequency] = new ArrayList<>();
}
bucket[frequency].add(key);
}
List<Integer> res = new ArrayList<>();
for (int pos = bucket.length - 1; pos >= 0 && res.size() < k; pos--) {
if (bucket[pos] != null) {
res.addAll(bucket[pos]);
}
}
return res;
}
It is obvious that the running time for this approach is O(n).