对于包含多个列表的列表,这里是一个重复的解决方案,为我工作,我希望它是正确的:

# Question 4
def flatten(input_ls=[]) -> []:
    res_ls = []
    res_ls = flatten_recursive(input_ls, res_ls)

    print("Final flatten list solution is: \n", res_ls)

    return res_ls


def flatten_recursive(input_ls=[], res_ls=[]) -> []:
    tmp_ls = []

    for i in input_ls:
        if isinstance(i, int):
            res_ls.append(i)
        else:
            tmp_ls = i
            tmp_ls.append(flatten_recursive(i, res_ls))

    print(res_ls)
    return res_ls


flatten([0, 1, [2, 3], 4, [5, 6]])  # test
flatten([0, [[[1]]], [[2, 3], [4, [[5, 6]]]]])

出口:

[0, 1, 2, 3]
[0, 1, 2, 3, 4, 5, 6]
[0, 1, 2, 3, 4, 5, 6]
Final flatten list solution is: 
 [0, 1, 2, 3, 4, 5, 6]
[0, 1]
[0, 1]
[0, 1]
[0, 1, 2, 3]
[0, 1, 2, 3, 4, 5, 6]
[0, 1, 2, 3, 4, 5, 6]
[0, 1, 2, 3, 4, 5, 6]
[0, 1, 2, 3, 4, 5, 6]
[0, 1, 2, 3, 4, 5, 6]
Final flatten list solution is: 
 [0, 1, 2, 3, 4, 5, 6]

matplotlib.cbook.flatten() 将为粘贴列表工作,即使它们比示例更深地粘贴。

import matplotlib
l = [[1, 2, 3], [4, 5, 6], [7], [8, 9]]
print(list(matplotlib.cbook.flatten(l)))
l2 = [[1, 2, 3], [4, 5, 6], [7], [8, [9, 10, [11, 12, [13]]]]]
print(list(matplotlib.cbook.flatten(l2)))

结果:

[1, 2, 3, 4, 5, 6, 7, 8, 9]
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13]

这比 underscore 快 18 倍。

Average time over 1000 trials of matplotlib.cbook.flatten: 2.55e-05 sec
Average time over 1000 trials of underscore._.flatten: 4.63e-04 sec
(time for underscore._)/(time for matplotlib.cbook) = 18.1233394636

def flatten(itr):
    for x in itr:
        try:
            yield from flatten(x)
        except TypeError:
            yield x

使用:这是一个发电机,通常你想将它插入一个不可分割的构建器,如列表()或<<<<<<<<或使用它在一个为旋转。

这个解决方案的好处是:

工作任何类型的无缝(即使是未来的!)工作任何组合和深度的无缝工作,如果顶级包含无依赖物品,快速和高效(你可以平滑无缝的部分,没有浪费时间的剩余部分你不需要)多样性(你可以使用它来构建一个无缝的你的选择或在旋转)

注:由于所有 iterables 都是浮动的,所以线条分为单个字符的序列. 如果您不喜欢/不喜欢这种行为,您可以使用下列版本,从浮动的 iterables 如线条和比特中进行过滤:

def flatten(itr):
    if type(itr) in (str,bytes):
        yield itr
    else:
        for x in itr:
            try:
                yield from flatten(x)
            except TypeError:
                yield x

要插入深厚的数据结构,请使用 iteration_utilities.deepflatten1:

>>> from iteration_utilities import deepflatten

>>> l = [[1, 2, 3], [4, 5, 6], [7], [8, 9]]
>>> list(deepflatten(l, depth=1))
[1, 2, 3, 4, 5, 6, 7, 8, 9]

>>> l = [[1, 2, 3], [4, [5, 6]], 7, [8, 9]]
>>> list(deepflatten(l))
[1, 2, 3, 4, 5, 6, 7, 8, 9]

这是一个发电机,所以你需要将结果投到列表中,或者明确地对其进行调解。

要单一的平面,如果每一个项目本身是不可分割的,你也可以使用 iteration_utilities.flatten 它本身只是一个薄的旋转器周围 itertools.chain.from_iterable:

>>> from iteration_utilities import flatten
>>> l = [[1, 2, 3], [4, 5, 6], [7], [8, 9]]
>>> list(flatten(l))
[1, 2, 3, 4, 5, 6, 7, 8, 9]

只需添加一些时间表(基于Nico Schlömer的答案,其中不包含此答案中的功能):

此分類上一篇

结果表明,如果 iterable 只包含几个内部 iterables 那么 总数将是最快的,但是,对于长期 iterables 只有 itertools.chain.from_iterable, iteration_utilities.deepflatten 或 nested 理解具有合理的性能, itertools.chain.from_iterable 是最快的(如 Nico Schlömer 已经注意到)。

from itertools import chain
from functools import reduce
from collections import Iterable  # or from collections.abc import Iterable
import operator
from iteration_utilities import deepflatten

def nested_list_comprehension(lsts):
    return [item for sublist in lsts for item in sublist]

def itertools_chain_from_iterable(lsts):
    return list(chain.from_iterable(lsts))

def pythons_sum(lsts):
    return sum(lsts, [])

def reduce_add(lsts):
    return reduce(lambda x, y: x + y, lsts)

def pylangs_flatten(lsts):
    return list(flatten(lsts))

def flatten(items):
    """Yield items from any nested iterable; see REF."""
    for x in items:
        if isinstance(x, Iterable) and not isinstance(x, (str, bytes)):
            yield from flatten(x)
        else:
            yield x

def reduce_concat(lsts):
    return reduce(operator.concat, lsts)

def iteration_utilities_deepflatten(lsts):
    return list(deepflatten(lsts, depth=1))


from simple_benchmark import benchmark

b = benchmark(
    [nested_list_comprehension, itertools_chain_from_iterable, pythons_sum, reduce_add,
     pylangs_flatten, reduce_concat, iteration_utilities_deepflatten],
    arguments={2**i: [[0]*5]*(2**i) for i in range(1, 13)},
    argument_name='number of inner lists'
)

b.plot()

1 Disclaimer:我是该图书馆的作者

你可以使用列表扩展方法. 它显示是最快的:

flat_list = []
for sublist in l:
    flat_list.extend(sublist)

表演:

import functools
import itertools
import numpy
import operator
import perfplot


def functools_reduce_iconcat(a):
    return functools.reduce(operator.iconcat, a, [])


def itertools_chain(a):
    return list(itertools.chain.from_iterable(a))


def numpy_flat(a):
    return list(numpy.array(a).flat)


def extend(a):
    n = []

    list(map(n.extend, a))

    return n


perfplot.show(
    setup = lambda n: [list(range(10))] * n,
    kernels = [
        functools_reduce_iconcat, extend, itertools_chain, numpy_flat
        ],
    n_range = [2**k for k in range(16)],
    xlabel = 'num lists',
    )

出口:

此分類上一篇

另一个不寻常的方法,适用于异常和均匀的整体列表:

from typing import List


def flatten(l: list) -> List[int]:
    """Flatten an arbitrary deep nested list of lists of integers.

    Examples:
        >>> flatten([1, 2, [1, [10]]])
        [1, 2, 1, 10]

    Args:
        l: Union[l, Union[int, List[int]]

    Returns:
        Flatted list of integer
    """
    return [int(i.strip('[ ]')) for i in str(l).split(',')]