此解决方案实现了一个生成器，以避免在内存中保留所有排列：

def permutations (orig_list):
    if not isinstance(orig_list, list):
        orig_list = list(orig_list)

    yield orig_list

    if len(orig_list) == 1:
        return

    for n in sorted(orig_list):
        new_list = orig_list[:]
        pos = new_list.index(n)
        del(new_list[pos])
        new_list.insert(0, n)
        for resto in permutations(new_list[1:]):
            if new_list[:1] + resto <> orig_list:
                yield new_list[:1] + resto

使用标准库中的itertools.permutations：

import itertools
list(itertools.permutations([1, 2, 3]))

从这里改编的是itertools.permutations如何实现的演示：

def permutations(elements):
    if len(elements) <= 1:
        yield elements
        return
    for perm in permutations(elements[1:]):
        for i in range(len(elements)):
            # nb elements[0:1] works in both string and list contexts
            yield perm[:i] + elements[0:1] + perm[i:]

itertools.permutations文档中列出了两种替代方法

def permutations(iterable, r=None):
    # permutations('ABCD', 2) --> AB AC AD BA BC BD CA CB CD DA DB DC
    # permutations(range(3)) --> 012 021 102 120 201 210
    pool = tuple(iterable)
    n = len(pool)
    r = n if r is None else r
    if r > n:
        return
    indices = range(n)
    cycles = range(n, n-r, -1)
    yield tuple(pool[i] for i in indices[:r])
    while n:
        for i in reversed(range(r)):
            cycles[i] -= 1
            if cycles[i] == 0:
                indices[i:] = indices[i+1:] + indices[i:i+1]
                cycles[i] = n - i
            else:
                j = cycles[i]
                indices[i], indices[-j] = indices[-j], indices[i]
                yield tuple(pool[i] for i in indices[:r])
                break
        else:
            return

另一个基于itertools.product：

def permutations(iterable, r=None):
    pool = tuple(iterable)
    n = len(pool)
    r = n if r is None else r
    for indices in product(range(n), repeat=r):
        if len(set(indices)) == r:
            yield tuple(pool[i] for i in indices)

对于Python 2.6及以上版本：

import itertools
itertools.permutations([1, 2, 3])

这将作为生成器返回。使用列表（排列（xs））作为列表返回。

以下代码是给定列表的就地排列，作为生成器实现。由于它只返回对列表的引用，因此不应在生成器外部修改列表。该解决方案是非递归的，因此使用了低内存。还可以很好地处理输入列表中元素的多个副本。

def permute_in_place(a):
    a.sort()
    yield list(a)

    if len(a) <= 1:
        return

    first = 0
    last = len(a)
    while 1:
        i = last - 1

        while 1:
            i = i - 1
            if a[i] < a[i+1]:
                j = last - 1
                while not (a[i] < a[j]):
                    j = j - 1
                a[i], a[j] = a[j], a[i] # swap the values
                r = a[i+1:last]
                r.reverse()
                a[i+1:last] = r
                yield list(a)
                break
            if i == first:
                a.reverse()
                return

if __name__ == '__main__':
    for n in range(5):
        for a in permute_in_place(range(1, n+1)):
            print a
        print

    for a in permute_in_place([0, 0, 1, 1, 1]):
        print a
    print

首先，导入itertools：

import itertools

排列（顺序重要）：

print(list(itertools.permutations([1,2,3,4], 2)))

[(1, 2), (1, 3), (1, 4),
(2, 1), (2, 3), (2, 4),
(3, 1), (3, 2), (3, 4),
(4, 1), (4, 2), (4, 3)]

组合（顺序无关紧要）：

print(list(itertools.combinations('123', 2)))

[('1', '2'), ('1', '3'), ('2', '3')]

笛卡尔积（具有多个可迭代项）：

print(list(itertools.product([1,2,3], [4,5,6])))

[(1, 4), (1, 5), (1, 6),
(2, 4), (2, 5), (2, 6),
(3, 4), (3, 5), (3, 6)]

笛卡尔积（具有一个可迭代的和自身）：

print(list(itertools.product([1,2], repeat=3)))

[(1, 1, 1), (1, 1, 2), (1, 2, 1), (1, 2, 2),
(2, 1, 1), (2, 1, 2), (2, 2, 1), (2, 2, 2)]

在我看来，一个很明显的方式可能是：

def permutList(l):
    if not l:
            return [[]]
    res = []
    for e in l:
            temp = l[:]
            temp.remove(e)
            res.extend([[e] + r for r in permutList(temp)])

    return res

list2Perm = [1, 2.0, 'three']
listPerm = [[a, b, c]
            for a in list2Perm
            for b in list2Perm
            for c in list2Perm
            if ( a != b and b != c and a != c )
            ]
print listPerm

输出：

[
    [1, 2.0, 'three'], 
    [1, 'three', 2.0], 
    [2.0, 1, 'three'], 
    [2.0, 'three', 1], 
    ['three', 1, 2.0], 
    ['three', 2.0, 1]
]

def permutations(head, tail=''):
    if len(head) == 0:
        print(tail)
    else:
        for i in range(len(head)):
            permutations(head[:i] + head[i+1:], tail + head[i])

称为：

permutations('abc')

人们确实可以对每个排列的第一个元素进行迭代，正如tzwen的答案。然而，这样编写此解决方案更有效：

def all_perms(elements):
    if len(elements) <= 1:
        yield elements  # Only permutation possible = no permutation
    else:
        # Iteration over the first element in the result permutation:
        for (index, first_elmt) in enumerate(elements):
            other_elmts = elements[:index]+elements[index+1:]
            for permutation in all_perms(other_elmts): 
                yield [first_elmt] + permutation

这个解决方案大约快了30%，显然是因为递归以len（元素）<=1而不是0结尾。它的内存效率也高得多，因为它使用了一个生成器函数（通过yield），就像Riccardo Reyes的解决方案一样。

#!/usr/bin/env python

def perm(a, k=0):
   if k == len(a):
      print a
   else:
      for i in xrange(k, len(a)):
         a[k], a[i] = a[i] ,a[k]
         perm(a, k+1)
         a[k], a[i] = a[i], a[k]

perm([1,2,3])

输出：

[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 2, 1]
[3, 1, 2]

当我交换列表的内容时，需要一个可变的序列类型作为输入。例如，烫发（list（“ball”）会起作用，而烫发（“ball”）不会起作用，因为你不能更改字符串。

这种Python实现的灵感来自Horowitz、Sahni和Rajasekeran在《计算机算法》一书中提出的算法。

注意，该算法具有n个阶乘时间复杂度，其中n是输入列表的长度

打印跑步结果：

global result
result = [] 

def permutation(li):
if li == [] or li == None:
    return

if len(li) == 1:
    result.append(li[0])
    print result
    result.pop()
    return

for i in range(0,len(li)):
    result.append(li[i])
    permutation(li[:i] + li[i+1:])
    result.pop()    

例子：

permutation([1,2,3])

输出：

[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 1, 2]
[3, 2, 1]

from __future__ import print_function

def perm(n):
    p = []
    for i in range(0,n+1):
        p.append(i)
    while True:
        for i in range(1,n+1):
            print(p[i], end=' ')
        print("")
        i = n - 1
        found = 0
        while (not found and i>0):
            if p[i]<p[i+1]:
                found = 1
            else:
                i = i - 1
        k = n
        while p[i]>p[k]:
            k = k - 1
        aux = p[i]
        p[i] = p[k]
        p[k] = aux
        for j in range(1,(n-i)/2+1):
            aux = p[i+j]
            p[i+j] = p[n-j+1]
            p[n-j+1] = aux
        if not found:
            break

perm(5)

功能性风格

def addperm(x,l):
    return [ l[0:i] + [x] + l[i:]  for i in range(len(l)+1) ]

def perm(l):
    if len(l) == 0:
        return [[]]
    return [x for y in perm(l[1:]) for x in addperm(l[0],y) ]

print perm([ i for i in range(3)])

结果：

[[0, 1, 2], [1, 0, 2], [1, 2, 0], [0, 2, 1], [2, 0, 1], [2, 1, 0]]

这里有一个算法，它在不创建新的中间列表的情况下处理列表，类似于Ber在https://stackoverflow.com/a/108651/184528.

def permute(xs, low=0):
    if low + 1 >= len(xs):
        yield xs
    else:
        for p in permute(xs, low + 1):
            yield p        
        for i in range(low + 1, len(xs)):        
            xs[low], xs[i] = xs[i], xs[low]
            for p in permute(xs, low + 1):
                yield p        
            xs[low], xs[i] = xs[i], xs[low]

for p in permute([1, 2, 3, 4]):
    print p

您可以在这里亲自尝试代码：http://repl.it/J9v

我使用了一种基于阶乘数系统的算法——对于长度为n的列表，您可以逐项组装每个排列，从每个阶段留下的项目中进行选择。第一项有n个选项，第二项有n-1个选项，最后一项只有一个选项，因此可以使用阶乘数系统中数字的数字作为索引。这是数字0到n-1对应于词典顺序中的所有可能的排列。

from math import factorial
def permutations(l):
    permutations=[]
    length=len(l)
    for x in xrange(factorial(length)):
        available=list(l)
        newPermutation=[]
        for radix in xrange(length, 0, -1):
            placeValue=factorial(radix-1)
            index=x/placeValue
            newPermutation.append(available.pop(index))
            x-=index*placeValue
        permutations.append(newPermutation)
    return permutations

permutations(range(3))

输出：

[[0, 1, 2], [0, 2, 1], [1, 0, 2], [1, 2, 0], [2, 0, 1], [2, 1, 0]]

此方法是非递归的，但在我的计算机上速度稍慢，xrange在n！太大，无法转换为C长整数（我的n=13）。当我需要它的时候，它已经足够了，但它远没有itertools.permutations。

这是受Haskell实现使用列表理解的启发：

def permutation(list):
    if len(list) == 0:
        return [[]]
    else:
        return [[x] + ys for x in list for ys in permutation(delete(list, x))]

def delete(list, item):
    lc = list[:]
    lc.remove(item)
    return lc

递归之美：

>>> import copy
>>> def perm(prefix,rest):
...      for e in rest:
...              new_rest=copy.copy(rest)
...              new_prefix=copy.copy(prefix)
...              new_prefix.append(e)
...              new_rest.remove(e)
...              if len(new_rest) == 0:
...                      print new_prefix + new_rest
...                      continue
...              perm(new_prefix,new_rest)
... 
>>> perm([],['a','b','c','d'])
['a', 'b', 'c', 'd']
['a', 'b', 'd', 'c']
['a', 'c', 'b', 'd']
['a', 'c', 'd', 'b']
['a', 'd', 'b', 'c']
['a', 'd', 'c', 'b']
['b', 'a', 'c', 'd']
['b', 'a', 'd', 'c']
['b', 'c', 'a', 'd']
['b', 'c', 'd', 'a']
['b', 'd', 'a', 'c']
['b', 'd', 'c', 'a']
['c', 'a', 'b', 'd']
['c', 'a', 'd', 'b']
['c', 'b', 'a', 'd']
['c', 'b', 'd', 'a']
['c', 'd', 'a', 'b']
['c', 'd', 'b', 'a']
['d', 'a', 'b', 'c']
['d', 'a', 'c', 'b']
['d', 'b', 'a', 'c']
['d', 'b', 'c', 'a']
['d', 'c', 'a', 'b']
['d', 'c', 'b', 'a']

该算法是最有效的算法，它避免了递归调用中的数组传递和操作，适用于Python 2、3：

def permute(items):
    length = len(items)
    def inner(ix=[]):
        do_yield = len(ix) == length - 1
        for i in range(0, length):
            if i in ix: #avoid duplicates
                continue
            if do_yield:
                yield tuple([items[y] for y in ix + [i]])
            else:
                for p in inner(ix + [i]):
                    yield p
    return inner()

用法：

for p in permute((1,2,3)):
    print(p)

(1, 2, 3)
(1, 3, 2)
(2, 1, 3)
(2, 3, 1)
(3, 1, 2)
(3, 2, 1)

def pzip(c, seq):
    result = []
    for item in seq:
        for i in range(len(item)+1):
            result.append(item[i:]+c+item[:i])
    return result


def perm(line):
    seq = [c for c in line]
    if len(seq) <=1 :
        return seq
    else:
        return pzip(seq[0], perm(seq[1:]))

对于性能，一个由Knuth启发的numpy解决方案（第22页）：

from numpy import empty, uint8
from math import factorial

def perms(n):
    f = 1
    p = empty((2*n-1, factorial(n)), uint8)
    for i in range(n):
        p[i, :f] = i
        p[i+1:2*i+1, :f] = p[:i, :f]  # constitution de blocs
        for j in range(i):
            p[:i+1, f*(j+1):f*(j+2)] = p[j+1:j+i+2, :f]  # copie de blocs
        f = f*(i+1)
    return p[:n, :]

复制大量内存可节省时间-它比列表（itertools.permutations（range（n））快20倍：

In [1]: %timeit -n10 list(permutations(range(10)))
10 loops, best of 3: 815 ms per loop

In [2]: %timeit -n100 perms(10) 
100 loops, best of 3: 40 ms per loop

对于Python，我们可以使用itertools并导入排列和组合来解决问题

from itertools import product, permutations
A = ([1,2,3])
print (list(permutations(sorted(A),2)))

生成所有可能的排列

我正在使用python3.4：

def calcperm(arr, size):
    result = set([()])
    for dummy_idx in range(size):
        temp = set()
        for dummy_lst in result:
            for dummy_outcome in arr:
                if dummy_outcome not in dummy_lst:
                    new_seq = list(dummy_lst)
                    new_seq.append(dummy_outcome)
                    temp.add(tuple(new_seq))
        result = temp
    return result

测试用例：

lst = [1, 2, 3, 4]
#lst = ["yellow", "magenta", "white", "blue"]
seq = 2
final = calcperm(lst, seq)
print(len(final))
print(final)

我看到在这些递归函数中进行了很多迭代，而不是纯粹的递归。。。

所以对于那些连一个循环都不能遵守的人来说，这里有一个粗略的、完全不必要的完全递归的解决方案

def all_insert(x, e, i=0):
    return [x[0:i]+[e]+x[i:]] + all_insert(x,e,i+1) if i<len(x)+1 else []

def for_each(X, e):
    return all_insert(X[0], e) + for_each(X[1:],e) if X else []

def permute(x):
    return [x] if len(x) < 2 else for_each( permute(x[1:]) , x[0])


perms = permute([1,2,3])

另一种解决方案：

def permutation(flag, k =1 ):
    N = len(flag)
    for i in xrange(0, N):
        if flag[i] != 0:
            continue
        flag[i] = k 
        if k == N:
            print flag
        permutation(flag, k+1)
        flag[i] = 0

permutation([0, 0, 0])

我的Python解决方案：

def permutes(input,offset):
    if( len(input) == offset ):
        return [''.join(input)]

    result=[]        
    for i in range( offset, len(input) ):
         input[offset], input[i] = input[i], input[offset]
         result = result + permutes(input,offset+1)
         input[offset], input[i] = input[i], input[offset]
    return result

# input is a "string"
# return value is a list of strings
def permutations(input):
    return permutes( list(input), 0 )

# Main Program
print( permutations("wxyz") )

def permutation(word, first_char=None):
    if word == None or len(word) == 0: return []
    if len(word) == 1: return [word]

    result = []
    first_char = word[0]
    for sub_word in permutation(word[1:], first_char):
        result += insert(first_char, sub_word)
    return sorted(result)

def insert(ch, sub_word):
    arr = [ch + sub_word]
    for i in range(len(sub_word)):
        arr.append(sub_word[i:] + ch + sub_word[:i])
    return arr


assert permutation(None) == []
assert permutation('') == []
assert permutation('1')  == ['1']
assert permutation('12') == ['12', '21']

print permutation('abc')

输出：['abc'，'acb'，'bac'，'bca'，'cab'，'cba']

为了节省您可能的搜索和实验时间，下面是Python中的非递归置换解决方案，它也适用于Numba（从0.41版开始）：

@numba.njit()
def permutations(A, k):
    r = [[i for i in range(0)]]
    for i in range(k):
        r = [[a] + b for a in A for b in r if (a in b)==False]
    return r
permutations([1,2,3],3)
[[1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]]

要给人留下绩效印象：

%timeit permutations(np.arange(5),5)

243 µs ± 11.1 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)
time: 406 ms

%timeit list(itertools.permutations(np.arange(5),5))
15.9 µs ± 8.61 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
time: 12.9 s

因此，只有在必须从njit函数调用它时才使用此版本，否则更倾向于itertools实现。

使用计数器

from collections import Counter

def permutations(nums):
    ans = [[]]
    cache = Counter(nums)

    for idx, x in enumerate(nums):
        result = []
        for items in ans:
            cache1 = Counter(items)
            for id, n in enumerate(nums):
                if cache[n] != cache1[n] and items + [n] not in result:
                    result.append(items + [n])

        ans = result
    return ans
permutations([1, 2, 2])
> [[1, 2, 2], [2, 1, 2], [2, 2, 1]]

另一种方法（无libs）

def permutation(input):
    if len(input) == 1:
        return input if isinstance(input, list) else [input]

    result = []
    for i in range(len(input)):
        first = input[i]
        rest = input[:i] + input[i + 1:]
        rest_permutation = permutation(rest)
        for p in rest_permutation:
            result.append(first + p)
    return result

输入可以是字符串或列表

print(permutation('abcd'))
print(permutation(['a', 'b', 'c', 'd']))

免责声明：无耻的插件由包作者。：）

trotter包与大多数实现的不同之处在于，它生成的伪列表实际上不包含排列，而是描述排列与排序中各个位置之间的映射，从而可以处理非常大的排列“列表”，如本演示所示，它在一个包含字母表中所有字母排列的伪列表中执行相当即时的操作和查找，而不使用比典型网页更多的内存或处理。

在任何情况下，要生成排列列表，我们可以执行以下操作。

import trotter

my_permutations = trotter.Permutations(3, [1, 2, 3])

print(my_permutations)

for p in my_permutations:
    print(p)

输出：

A pseudo-list containing 6 3-permutations of [1, 2, 3].
[1, 2, 3]
[1, 3, 2]
[3, 1, 2]
[3, 2, 1]
[2, 3, 1]
[2, 1, 3]

常规实现（无收益-将在内存中完成所有操作）：

def getPermutations(array):
    if len(array) == 1:
        return [array]
    permutations = []
    for i in range(len(array)): 
        # get all perm's of subarray w/o current item
        perms = getPermutations(array[:i] + array[i+1:])  
        for p in perms:
            permutations.append([array[i], *p])
    return permutations

产量实施：

def getPermutations(array):
    if len(array) == 1:
        yield array
    else:
        for i in range(len(array)):
            perms = getPermutations(array[:i] + array[i+1:])
            for p in perms:
                yield [array[i], *p]

基本思想是在第一个位置遍历数组中的所有元素，然后在第二个位置遍历所有其他元素，而没有为第一个位置选择的元素，等等。您可以使用递归来实现这一点，其中停止条件是到达一个由1个元素组成的数组，在这种情况下，您返回该数组。

def permuteArray (arr):

    arraySize = len(arr)

    permutedList = []

    if arraySize == 1:
        return [arr]

    i = 0

    for item in arr:

        for elem in permuteArray(arr[:i] + arr[i + 1:]):
            permutedList.append([item] + elem)

        i = i + 1    

    return permutedList

我不打算在一个新的行中穷尽所有的可能性，以使它有点独特。

无论如何，我们可以使用sympy库，也支持多集合排列

import sympy
from sympy.utilities.iterables import multiset_permutations
t = [1,2,3]
p = list(multiset_permutations(t))
print(p)

# [[1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]]

答案的灵感来自获取numpy数组的所有排列

from typing import List
import time, random

def measure_time(func):
    def wrapper_time(*args, **kwargs):
        start_time = time.perf_counter()
        res = func(*args, **kwargs)
        end_time = time.perf_counter()
        return res, end_time - start_time

    return wrapper_time


class Solution:
    def permute(self, nums: List[int], method: int = 1) -> List[List[int]]:
        perms = []
        perm = []
        if method == 1:
            _, time_perm = self._permute_recur(nums, 0, len(nums) - 1, perms)
        elif method == 2:
            _, time_perm = self._permute_recur_agian(nums, perm, perms)
            print(perm)
        return perms, time_perm

    @measure_time
    def _permute_recur(self, nums: List[int], l: int, r: int, perms: List[List[int]]):
        # base case
        if l == r:
            perms.append(nums.copy())

        for i in range(l, r + 1):
            nums[l], nums[i] = nums[i], nums[l]
            self._permute_recur(nums, l + 1, r , perms)
            nums[l], nums[i] = nums[i], nums[l]

    @measure_time
    def _permute_recur_agian(self, nums: List[int], perm: List[int], perms_list: List[List[int]]):
        """
        The idea is similar to nestedForLoops visualized as a recursion tree.
        """
        if nums:
            for i in range(len(nums)):
                # perm.append(nums[i])  mistake, perm will be filled with all nums's elements.
                # Method1 perm_copy = copy.deepcopy(perm)
                # Method2 add in the parameter list using + (not in place)
                # caveat: list.append is in-place , which is useful for operating on global element perms_list
                # Note that:
                # perms_list pass by reference. shallow copy
                # perm + [nums[i]] pass by value instead of reference.
                self._permute_recur_agian(nums[:i] + nums[i+1:], perm + [nums[i]], perms_list)
        else:
            # Arrive at the last loop, i.e. leaf of the recursion tree.
            perms_list.append(perm)



if __name__ == "__main__":
    array = [random.randint(-10, 10) for _ in range(3)]
    sol = Solution()
    # perms, time_perm = sol.permute(array, 1)
    perms2, time_perm2 = sol.permute(array, 2)
    print(perms2)
    # print(perms, perms2)
    # print(time_perm, time_perm2)
```

如果有人喜欢这个丑陋的单行线（虽然只适用于字符串）：

def p(a):
    return a if len(a) == 1 else [[a[i], *j] for i in range(len(a)) for j in p(a[:i] + a[i + 1:])]

这是初始排序后生成排列的渐近最优方式O（n*n！）。

有n个！最多进行一次置换，且具有下一次置换（..），以O（n）时间复杂度运行

在3个步骤中，

找到最大的j，使a[j]可以增加以最小可行量增加a[j]找到扩展新a[0..j]的字典最少方法

'''
Lexicographic permutation generation

consider example array state of [1,5,6,4,3,2] for sorted [1,2,3,4,5,6]
after 56432(treat as number) ->nothing larger than 6432(using 6,4,3,2) beginning with 5
so 6 is next larger and 2345(least using numbers other than 6)
so [1, 6,2,3,4,5]
'''
def hasNextPermutation(array, len):
    ' Base Condition '
    if(len ==1):
        return False
    '''
    Set j = last-2 and find first j such that a[j] < a[j+1]
    If no such j(j==-1) then we have visited all permutations
    after this step a[j+1]>=..>=a[len-1] and a[j]<a[j+1]

    a[j]=5 or j=1, 6>5>4>3>2
    '''
    j = len -2
    while (j >= 0 and array[j] >= array[j + 1]):
        j= j-1
    if(j==-1):
        return False
    # print(f"After step 2 for j {j}  {array}")
    '''
    decrease l (from n-1 to j) repeatedly until a[j]<a[l]
    Then swap a[j], a[l]
    a[l] is the smallest element > a[j] that can follow a[l]...a[j-1] in permutation
    before swap we have a[j+1]>=..>=a[l-1]>=a[l]>a[j]>=a[l+1]>=..>=a[len-1]
    after swap -> a[j+1]>=..>=a[l-1]>=a[j]>a[l]>=a[l+1]>=..>=a[len-1]

    a[l]=6 or l=2, j=1 just before swap [1, 5, 6, 4, 3, 2] 
    after swap [1, 6, 5, 4, 3, 2] a[l]=5, a[j]=6
    '''
    l = len -1
    while(array[j] >= array[l]):
        l = l-1
    # print(f"After step 3 for l={l}, j={j} before swap {array}")
    array[j], array[l] = array[l], array[j]
    # print(f"After step 3 for l={l} j={j} after swap {array}")
    '''
    Reverse a[j+1...len-1](both inclusive)

    after reversing [1, 6, 2, 3, 4, 5]
    '''
    array[j+1:len] = reversed(array[j+1:len])
    # print(f"After step 4 reversing {array}")
    return True

array = [1,2,4,4,5]
array.sort()
len = len(array)
count =1
print(array)
'''
The algorithm visits every permutation in lexicographic order
generating one by one
'''
while(hasNextPermutation(array, len)):
    print(array)
    count = count +1
# The number of permutations will be n! if no duplicates are present, else less than that
# [1,4,3,3,2] -> 5!/2!=60
print(f"Number of permutations: {count}")

如果不想使用内置方法，例如：

import itertools
list(itertools.permutations([1, 2, 3]))

你可以自己实现permute函数

from collections.abc import Iterable


def permute(iterable: Iterable[str]) -> set[str]:
    perms = set()

    if len(iterable) == 1:
        return {*iterable}

    for index, char in enumerate(iterable):
        perms.update([char + perm for perm in permute(iterable[:index] + iterable[index + 1:])])

    return perms


if __name__ == '__main__':
    print(permute('abc'))
    # {'bca', 'abc', 'cab', 'acb', 'cba', 'bac'}
    print(permute(['1', '2', '3']))
    # {'123', '312', '132', '321', '213', '231'}

def permutate(l):
    for i, x in enumerate(l):
        for y in l[i + 1:]:
            yield x, y


if __name__ == '__main__':
    print(list(permutate(list('abcd'))))
    print(list(permutate([1, 2, 3, 4])))

#[('a', 'b'), ('a', 'c'), ('a', 'd'), ('b', 'c'), ('b', 'd'), ('c', 'd')]
#[(1, 2), (1, 3), (1, 4), (2, 3), (2, 4), (3, 4)]

用递归求解，遍历元素，取第i个元素，然后问自己：“其余项目的排列是什么”，直到没有更多的元素。

我在这里解释了解决方案：https://www.youtube.com/watch?v=_7GE7psS2b4

class Solution:
    def permute(self,nums:List[int])->List[List[int]]:
        res=[]
        def dfs(nums,path):
            if len(nums)==0:
                res.append(path)
            for i in range(len(nums)):
                dfs(nums[:i]+nums[i+1:],path+[nums[i]])
        dfs(nums,[])
        return res

如果用户希望在列表中保留所有排列，可以使用以下代码：

def get_permutations(nums, p_list=[], temp_items=[]):
    if not nums:
        return
    elif len(nums) == 1:
        new_items = temp_items+[nums[0]]
        p_list.append(new_items)
        return
    else:
        for i in range(len(nums)):
            temp_nums = nums[:i]+nums[i+1:]
            new_temp_items = temp_items + [nums[i]]
            get_permutations(temp_nums, p_list, new_temp_items)

nums = [1,2,3]
p_list = []

get_permutations(nums, p_list)