- class Solution:
- def findSubsequences(self, nums: List[int]) -> List[List[int]]:
- used = [False]*len(nums)
- result = []
- nums.sort()
- def backtracking(nums,path,startindex,used):
- nonlocal result
- if len(path)>1:
- result.append(path[:])
- for i in range(startindex,len(nums)):
- if i>0 and nums[i] == nums[i-1] and used[i-1]==False:
- continue
- else:
- path.append(nums[i])
- used[i] = True
- backtracking(nums,path,i+1,used)
- path.pop()
- used[i] = False
- backtracking(nums,[],0,used)
- return result
问题1:如何按照顺序?(一般是直接增序)
- class Solution:
- def findSubsequences(self, nums: List[int]) -> List[List[int]]:
- result = []
- def backtracking(nums,path,startindex):
- nonlocal result
- if len(path) > 1:
- result.append(path[:])
- #每进入新的一层循环,都自动清空
- useset = set()
- for i in range(startindex,len(nums)):
- #没有在useset合集中,新加的数要比path里面的最后一个元素大
- #len(path) 防止溢出
- if (len(path)>0 and nums[i]< path[-1]) or (nums[i] in useset):
- continue
- else:
- path.append(nums[i])
- useset.add(nums[i])
- backtracking(nums,path,i+1)
- path.pop()
- backtracking(nums,[],0)
- return result
重点:不能对集合进行排序,怎么做呢?
任何分析,都需要画个图出来观察

题目要求:
都是递增:所以有个条件就是path中最后一个数要比新加入的数小
删除同一层节点的重复元素?使用一个set存放,如果同一层里面的数存在,就跳过
视频讲解:回溯算法精讲,树层去重与树枝去重 | LeetCode:491.递增子序列_哔哩哔哩_bilibili
二刷(未ac)
- var findSubsequences = function(nums) {
- let result = []
- const isValid = function(arr){
- for(let i = 0;i<arr.length-1;i++){
- if(arr[i+1]<arr[i]){
- return false
- }
- }
- return true
- }
- const backtracking = function(nums,path,startindex){
- let used = []
- // when end?
- if(path.length>nums.length){
- return
- }
- if(path.length > 1){
- if(isValid(path)){
- result.push([...path])
- }
- }
- for(let i = startindex;i<nums.length;i++){
- if(used[nums[i]]){
- continue
- }
- used[nums[i]] = true
- path.push(nums[i])
- backtracking(nums,path,i+1)
- path.pop()
- }
- }
- backtracking(nums,[],0)
- return result
- };
组合和排列问题差别在:
排列不需要在意集合中的顺序,顺序不同也算不同
组合需要算集合中的顺序,顺序不同不算不同
所以在这里不需要节点去重。
但是需要使用used集合记录一下每个数是否取过

- class Solution:
- def permute(self, nums: List[int]) -> List[List[int]]:
- result = []
- used = [False]*len(nums)
- def backtracking(nums,path,used):
- nonlocal result
- #终止条件
- if len(nums) == len(path):
- result.append(path[:])
- return
- #单层遍历逻辑
- for i in range(0,len(nums)):
- #如果这个数使用过的话,就跳过
- if used[i] == True:
- continue
- else:
- path.append(nums[i])
- used[i] = True
- backtracking(nums,path,used)
- path.pop()
- used[i] = False
- backtracking(nums,[],used)
- return result
视频讲解:组合与排列的区别,回溯算法求解的时候,有何不同?| LeetCode:46.全排列_哔哩哔哩_bilibili
二刷(未ac)
- var permute = function(nums) {
- let result = []
- let used = new Array(nums.length).fill(0)
- const backtracking = function(nums,path,used){
- // 终止条件
- if(path.length === nums.length){
- result.push([...path])
- }
- for(let i = 0;i<nums.length;i++){
- if(used[i]===1){continue}
- path.push(nums[i])
- used[i] = 1
- backtracking(nums,path,used)
- used[i] = 0
- path.pop()
- }
- }
- backtracking(nums,[],used)
- return result
- };
本题 就是我们讲过的 40.组合总和II 去重逻辑 和 46.全排列 的结合,可以先自己做一下,然后重点看一下 文章中 我讲的拓展内容。 used[i - 1] == true 也行,used[i - 1] == false 也行

- class Solution:
- def permuteUnique(self, nums: List[int]) -> List[List[int]]:
- result = []
- used = [False]*len(nums)
- nums.sort()
- def backtracking(nums,used,path):
- nonlocal result
- if len(nums) == len(path):
- result.append(path[:])
- return
- #单层逻辑
- for i in range(0,len(nums)):
- if used[i] == True:
- continue
- elif (i>0 and nums[i] == nums[i-1] and used[i-1] == False):
- continue
- else:
- path.append(nums[i])
- used[i] = True
- backtracking(nums,used,path)
- used[i] = False
- path.pop()
- backtracking(nums,used,[])
- return result
重点:
记得排序啊,重复元素去重,太重要了,这个细节
视频讲解:回溯算法求解全排列,如何去重?| LeetCode:47.全排列 II_哔哩哔哩_bilibili
二刷(ac)
- var permuteUnique = function(nums) {
- nums.sort((a,b)=>(a-b))
- let result = []
- let used = new Array(nums.length).fill(0)
- const backtracking = function(nums,path,used){
- if(path.length === nums.length){
- result.push([...path])
- return
- }
- for(let i =0;i<nums.length;i++){
- if(i>0 && nums[i-1]===nums[i]&&used[i-1]===1){
- continue
- }
- if(used[i]===1){
- continue
- }
- path.push(nums[i])
- used[i]=1
- backtracking(nums,path,used)
- used[i]=0
- path.pop()
- }
- }
- backtracking(nums,[],used)
- return result
- };