无脑使用unordered_map来保存数字出现次数，然后按照出现次数排序，如果出现次数相等，则数值大的排在前面。

class Solution {
public:
    vector<int> frequencySort(vector<int>& nums) {
        unordered_map<int,int> cnt;
        for(auto num:nums){
            cnt[num]++;
        }
        
        sort(nums.begin(),nums.end(),[&](int a,int b){
            return cnt[a] == cnt[b] ? a>b:cnt[a]<cnt[b];
        });
        
        return nums;
    }
};

可以使用优化后的代码，这里数组由于下标不能为负，因此做了一个特殊处理。

class Solution {
public:
    vector<int> frequencySort(vector<int>& nums) {
        int map[201] = {0};
        for(auto num:nums)
            map[num+100]++;
        
        sort(nums.begin(),nums.end(),[&](int a,int b){
            return map[a+100] == map[b+100] ? a>b:map[a+100] < map[b+100]; 
        });
        
        return nums;
    }
};

运行结果:

Runtime: 4 ms, faster than 97.46% of C++ online submissions for Sort Array by Increasing Frequency.
Memory Usage: 10.8 MB, less than 98.99% of C++ online submissions for Sort Array by Increasing Frequency.