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

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}")


其他回答

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

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
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')
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)

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']