这种方法可以很容易地移植到所有支持递归的编程语言中(没有itertools，没有yield，没有列表理解):

def combs(a):
    if len(a) == 0:
        return [[]]
    cs = []
    for c in combs(a[1:]):
        cs += [c, c+[a[0]]]
    return cs

>>> combs([1,2,3,4,5])
[[], [1], [2], [2, 1], [3], [3, 1], [3, 2], ..., [5, 4, 3, 2, 1]]

这个怎么样?使用字符串而不是列表，但同样的事情..string可以像Python中的列表一样处理:

def comb(s, res):
    if not s: return
    res.add(s)
    for i in range(0, len(s)):
        t = s[0:i] + s[i + 1:]
        comb(t, res)

res = set()
comb('game', res) 

print(res)

下面是一个惰性一行代码，同样使用itertools:

from itertools import compress, product

def combinations(items):
    return ( set(compress(items,mask)) for mask in product(*[[0,1]]*len(items)) )
    # alternative:                      ...in product([0,1], repeat=len(items)) )

这个答案背后的主要思想是:有2^N种组合——与长度为N的二进制字符串的数量相同。对于每个二进制字符串，您选择与“1”对应的所有元素。

items=abc * mask=###
 |
 V
000 -> 
001 ->   c
010 ->  b
011 ->  bc
100 -> a
101 -> a c
110 -> ab
111 -> abc

需要考虑的事情:

This requires that you can call len(...) on items (workaround: if items is something like an iterable like a generator, turn it into a list first with items=list(_itemsArg)) This requires that the order of iteration on items is not random (workaround: don't be insane) This requires that the items are unique, or else {2,2,1} and {2,1,1} will both collapse to {2,1} (workaround: use collections.Counter as a drop-in replacement for set; it's basically a multiset... though you may need to later use tuple(sorted(Counter(...).elements())) if you need it to be hashable)

Demo

>>> list(combinations(range(4)))
[set(), {3}, {2}, {2, 3}, {1}, {1, 3}, {1, 2}, {1, 2, 3}, {0}, {0, 3}, {0, 2}, {0, 2, 3}, {0, 1}, {0, 1, 3}, {0, 1, 2}, {0, 1, 2, 3}]

>>> list(combinations('abcd'))
[set(), {'d'}, {'c'}, {'c', 'd'}, {'b'}, {'b', 'd'}, {'c', 'b'}, {'c', 'b', 'd'}, {'a'}, {'a', 'd'}, {'a', 'c'}, {'a', 'c', 'd'}, {'a', 'b'}, {'a', 'b', 'd'}, {'a', 'c', 'b'}, {'a', 'c', 'b', 'd'}]

这里是另一个解决方案(一行程序)，涉及到使用itertools.combination函数，但这里我们使用了双链表理解式(而不是for循环或sum):

def combs(x):
    return [c for i in range(len(x)+1) for c in combinations(x,i)]

演示:

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

这个答案漏掉了一个方面:OP要求所有的组合……不仅仅是长度为r的组合。

所以你要么要遍历所有长度为L的循环:

import itertools

stuff = [1, 2, 3]
for L in range(len(stuff) + 1):
    for subset in itertools.combinations(stuff, L):
        print(subset)

或者——如果你想变得时髦(或者让那些在你之后阅读你的代码的人动脑筋)——你可以生成“组合()”生成器链，并遍历它:

from itertools import chain, combinations
def all_subsets(ss):
    return chain(*map(lambda x: combinations(ss, x), range(0, len(ss)+1)))

for subset in all_subsets(stuff):
    print(subset)

我知道使用itertools来获得所有的组合要实际得多，但是如果你碰巧想要，假设你想要编写很多代码，你可以只使用列表理解来部分实现这一点

对于两对组合:

lambda l: [(a, b) for i, a in enumerate(l) for b in l[i+1:]]

而且，对于三对组合，它是这样简单的:

lambda l: [(a, b, c) for i, a in enumerate(l) for ii, b in enumerate(l[i+1:]) for c in l[i+ii+2:]]

结果和使用itertools.combination是一样的:

import itertools
combs_3 = lambda l: [
    (a, b, c) for i, a in enumerate(l) 
    for ii, b in enumerate(l[i+1:]) 
    for c in l[i+ii+2:]
]
data = ((1, 2), 5, "a", None)
print("A:", list(itertools.combinations(data, 3)))
print("B:", combs_3(data))
# A: [((1, 2), 5, 'a'), ((1, 2), 5, None), ((1, 2), 'a', None), (5, 'a', None)]
# B: [((1, 2), 5, 'a'), ((1, 2), 5, None), ((1, 2), 'a', None), (5, 'a', None)]