LeetCode 90. Subsets II

Approach 1 (Proposed by Hickey)

class Solution {
    public List<List<Integer>> subsetsWithDup(int[] nums) {
        List<List<Integer>> result = new ArrayList<>();
		Arrays.sort(nums);
		List<Integer> path = new ArrayList<>();
		backtrack(nums, 0, result, path);
		return result;
    }
    private void backtrack(int[] nums, int level, List<List<Integer>> result, List<Integer> path) {
		if (level == nums.length) {
			result.add(new ArrayList<>(path));
			return;
		}
		path.add(nums[level]);
		backtrack(nums, level + 1, result, path);
		path.remove(path.size() - 1);
		// 一定要放在选之后，没选之前
		while (level + 1 < nums.length && nums[level] == nums[level + 1]) {
			level++;
		}
		backtrack(nums, level + 1, result, path);
	}
}

https://www.laioffer.com/en/news/2017-06-02-multiple-ways-to-solve-subsets-ii/

---

Approach 2

需要注意的是 nums[i - 1] == nums[i] 的作用是在递归树的同一层上的， 也就是防止 nums = [1, 2, 2'] 出现 [1, 2]， [1, 2'] 两个重复的解 （即： 通过 nums[1] == nums[2]: continue 进行的剪枝）
而解 [1, 2, 2'] 实际中的 2, 2' 是位于递归树的不同层级的， 也就是剪枝要从每一层的第二个元素开始， 所以还要增加一个 i > start 或者 i != start

class Solution {
    public List<List<Integer>> subsetsWithDup(int[] nums) {
        List<List<Integer>> res = new ArrayList<>();
        Arrays.sort(nums);
        backtrack(nums, 0, new ArrayList<>(), res);
        return res;
    }
    private void backtrack(int[] nums, int smallestCandidate, List<Integer> path, List<List<Integer>> res){
        res.add(new ArrayList<>(path));
        for(int i = smallestCandidate; i < nums.length; i++){
            if(i > smallestCandidate && nums[i] == nums[i - 1]){
                continue;
            }
            path.add(nums[i]);
            backtrack(nums, i + 1, path, res);
            path.remove(path.size() - 1);
        }
    }
}

https://leetcode-cn.com/problems/subsets-ii/solution/jian-zhi-qu-zhong-de-si-lu-by-mrsate/