如何生成列表的所有排列?例如:

permutations([])
[]

permutations([1])
[1]

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

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

当前回答

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

其他回答

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

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)

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

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

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]