Python 如何生成列表的所有排列?
如何在Python中生成列表的所有排列,而不依赖于该列表中的元素类型 例如:Python 如何生成列表的所有排列?,python,algorithm,permutation,combinatorics,python-2.5,Python,Algorithm,Permutation,Combinatorics,Python 2.5,如何在Python中生成列表的所有排列,而不依赖于该列表中的元素类型 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] 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
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]
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
def排列(原始列表):
如果不存在(原始列表,列表):
原始列表=列表(原始列表)
产量来源表
如果len(原始列表)=1:
返回
对于已排序的n(原始列表):
新列表=原始列表[:]
pos=新列表索引(n)
del(新列表[位置])
新列表。插入(0,n)
对于置换中的resto(新列表[1:]):
如果新列表[:1]+恢复原始列表:
产生新的_列表[:1]+恢复
标准库中有一个函数用于此:
如果出于某种原因,您想自己实现它,或者只是想知道它是如何工作的,那么这里有一个很好的方法,来自:
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)
及以后:
(作为生成器返回。使用
列表(排列(l))
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
def permute_位置(a):
a、 排序()
产量表(a)
如果len(a)仅使用Python 2.6及更高版本编写以下代码
首先,导入itertools
:
import itertools
排列(顺序事项):
组合(顺序无关紧要):
笛卡尔积(有几个可比项):
笛卡尔积(带一个iterable和自身):
在我看来,一个非常明显的方式可能是:
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
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]
]
[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 2, 1]
[3, 1, 2]
[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 1, 2]
[3, 2, 1]
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]
称为:
permutations('abc')
我们确实可以迭代每个排列的第一个元素,就像tzwenn的答案一样。但是,以这种方式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
def所有元素(元素):
if len(元素)
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]
]
[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 2, 1]
[3, 1, 2]
[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 1, 2]
[3, 2, 1]
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]
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,perm(list(“ball”)
将起作用,perm(“ball”)
将不起作用,因为您无法更改字符串
此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。请注意,此算法具有n阶乘
时间复杂度,其中n
是输入列表的长度
在运行时打印结果:
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()
例如:
permutation([1,2,3])
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]
]
[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 2, 1]
[3, 1, 2]
[1, 2, 3]
[1, 3, 2]
[2, 1, 3]
[2, 3, 1]
[3, 1, 2]
[3, 2, 1]
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]
from\uuuuu future\uuuuu导入打印功能
def perm(n):
p=[]
对于范围(0,n+1)内的i:
p、 附加(i)
尽管如此:
对于范围(1,n+1)内的i:
打印(p[i],结束=“”)
打印(“”)
i=n-1
找到=0
而(未找到且i>0):
如果p[i]p[k]:
k=k-1
aux=p[i]
p[i]=p[k]
p[k]=aux
对于范围(1,(n-i)/2+1内的j:
aux=p[i+j]
p[i+j]=p[n-j+1]
p[n-j+1]=aux
如果未找到:
打破
烫发(5)
以功能性风格
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)])
结果是:
[[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,它在一个列表上工作,而不创建新的中间列表,类似于Ber的解决方案
您可以在这里自己尝试代码:我使用了一种基于的算法-对于长度为n的列表,您可以逐项组合每个排列,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的数字作为索引。这样,数字0到n-1以字典顺序对应所有可能的排列
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]]
这个方法是非递归的,但是在我的计算机上它稍微慢一点,当n!太大,无法转换为C长整数(对于我,n=13)。当我需要它时,它已经足够了,但它不是itertools.permutations,这是一个很长的过程。这是受到Haskell实现的启发,它使用列表理解:
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']
此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):
结果=[]
对于序号中的项目:
对于范围内的i(长度(项目)+1):
结果.追加(项[i::+c+项[:i])
返回结果
def perm(线路):
seq=[c代表直线中的c]
如果len(seq)为了提高性能,一个numpy解决方案的灵感来自(第22页):
复制大内存块可节省时间-
它比列表(itertools.permutations(范围(n))快20倍
生成所有可能的排列
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]]
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)
我在这些递归函数中看到了很多迭代,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,这里有一个粗俗的、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])
def all_insert(x,e,i=0):
如果i另一个解决方案,则返回[x[0:i]+[e]+x[i:]+all_insert(x,e,i+1):
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") )
为了节省大家可能的搜索和实验时间,这里是Python中的非递归置换解决方案,它也适用于Numba(从v.0.41开始):
要给人留下关于绩效的印象:
%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,仅当您必须从njitted函数调用此版本时,才使用此版本,否则更喜欢itertools实现。另一种方法(不带libs)
输入可以是字符串或列表
print(permutation('abcd'))
print(permutation(['a', 'b', 'c', 'd']))
免责声明:不成形插头由软件包作者提供。:)
该包不同于大多数实现,因为它生成的伪列表实际上并不包含置换,而是描述置换和映射之间的映射
lst = [1, 2, 3, 4]
#lst = ["yellow", "magenta", "white", "blue"]
seq = 2
final = calcperm(lst, seq)
print(len(final))
print(final)
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])
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])
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") )
@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]]
%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
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
print(permutation('abcd'))
print(permutation(['a', 'b', 'c', 'd']))
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]
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
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]
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]]