LeetCode 46.Permutations

题目

求数组全排列

思路

代码

为了避免重复取出元素，设置used数组来标识是否用过这个数字。

class Solution:
    def permute(self, nums):
        res = []
        used = [0]*len(nums)
        if len(nums)==0:
            return res
        res = self.find(nums,0,[],res,used)
        return res
    
    def find(self, nums, index, p, res,used):
        if index == len(nums):
            res.append(p[:])
            print(p,'p is one ans')
            return res
        for i,n in enumerate(nums):
            if used[i] == 0:
                p.append(n)
                used[i] = 1
                print('use index',i,'value is',n,'p is',p,'used is',used)
                res = self.find(nums,index+1,p,res,used)               
                p.pop(-1)
                used[i] = 0
                print('use index',i,'value is',n,'complet','p is',p,'used is',used)
        return res

if __name__ == "__main__":
    a = Solution()
    print(a.permute([1,2,3]))

输出调试信息方便理解回溯的过程。每次使用完取出的元素，要放回，每次只取一个。
注意res.append§传入的是p的引用，所以后面弹出p中元素时会导致结果为[]，因此传入p的切片，就新建一个列表，不受后面的影响了。

use index 0 value is 1 p is [1] used is [1, 0, 0]
use index 1 value is 2 p is [1, 2] used is [1, 1, 0]
use index 2 value is 3 p is [1, 2, 3] used is [1, 1, 1]
[1, 2, 3] p is one ans
use index 2 value is 3 complet p is [1, 2] used is [1, 1, 0]
use index 1 value is 2 complet p is [1] used is [1, 0, 0]
use index 2 value is 3 p is [1, 3] used is [1, 0, 1]
use index 1 value is 2 p is [1, 3, 2] used is [1, 1, 1]
[1, 3, 2] p is one ans
use index 1 value is 2 complet p is [1, 3] used is [1, 0, 1]
use index 2 value is 3 complet p is [1] used is [1, 0, 0]
use index 0 value is 1 complet p is [] used is [0, 0, 0]
use index 1 value is 2 p is [2] used is [0, 1, 0]
use index 0 value is 1 p is [2, 1] used is [1, 1, 0]
use index 2 value is 3 p is [2, 1, 3] used is [1, 1, 1]
[2, 1, 3] p is one ans
use index 2 value is 3 complet p is [2, 1] used is [1, 1, 0]
use index 0 value is 1 complet p is [2] used is [0, 1, 0]
use index 2 value is 3 p is [2, 3] used is [0, 1, 1]
use index 0 value is 1 p is [2, 3, 1] used is [1, 1, 1]
[2, 3, 1] p is one ans
use index 0 value is 1 complet p is [2, 3] used is [0, 1, 1]
use index 2 value is 3 complet p is [2] used is [0, 1, 0]
use index 1 value is 2 complet p is [] used is [0, 0, 0]
use index 2 value is 3 p is [3] used is [0, 0, 1]
use index 0 value is 1 p is [3, 1] used is [1, 0, 1]
use index 1 value is 2 p is [3, 1, 2] used is [1, 1, 1]
[3, 1, 2] p is one ans
use index 1 value is 2 complet p is [3, 1] used is [1, 0, 1]
use index 0 value is 1 complet p is [3] used is [0, 0, 1]
use index 1 value is 2 p is [3, 2] used is [0, 1, 1]
use index 0 value is 1 p is [3, 2, 1] used is [1, 1, 1]
[3, 2, 1] p is one ans
use index 0 value is 1 complet p is [3, 2] used is [0, 1, 1]
use index 1 value is 2 complet p is [3] used is [0, 0, 1]
use index 2 value is 3 complet p is [] used is [0, 0, 0]
[[1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]]

简化版代码

class Solution(object):
    def permute(self, nums):
        """
        :type nums: List[int]
        :rtype: List[List[int]]
        """
        visited = [0] * len(nums)
        res = []
        
        def dfs(path):
            if len(path) == len(nums):
                res.append(path)
            else:
                for i in range(len(nums)):
                    if not visited[i]:
                        visited[i] = 1
                        dfs(path + [nums[i]]) #此处很妙，不用path.append，就不改变path的值，在函数调用返回后，path还是原来的值，就无需再pop了，为了使用加号合并两个列表，要把单个数字也加上[]变成列表。
                        visited[i] = 0
        
        dfs([])
        return res