C++ solution, beats ~65%

[vsingla2]

Runtime Complexity: O(n)
Space Complexity: O(m), where m is the number of duplicates

class Solution {
public:
    vector<int> findDuplicates(vector<int>& nums) {
        vector<int> sol;
        for(int i = 0; i < nums.size(); i++) {
            if(nums[abs(nums[i]) - 1] > 0) {
                nums[abs(nums[i]) - 1] = -nums[abs(nums[i]) - 1];
            }
            else {
                sol.push_back(abs(nums[i]));
            }
        }
        return sol;
    }
};