Python islice 类未按预期工作

Question

我正在编写一个函数，它将列表划分为（几乎）相等的“n”个分布。我希望这个函数返回一个生成器，但生成生成器似乎存在问题。该函数适用于可迭代对象。看看这个sn-p：

import itertools


def divide_list(array, n, gen_length=None):
    """
    :param array: some iterable that you wish to divide
    :param n: the number of lists you would like to return
    :param gen_length: The length of the generator if array is a generator. Not necessary for lists and tuples.
    :return: a generator of the divided list

    Example:
    In: list(divide_list([1, 2, 3, 4, 5, 6, 7, 8, 9], 4))
    Out: [[1, 2, 3], [4, 5], [6, 7], [8, 9]]
    """
    if isinstance(array, (list, tuple)):
        floor, rem = divmod(len(array), n)

        items_index = (0, floor)
        for _ in range(n):
            prev, next_ = items_index[0], items_index[1] + 1 if rem > 0 else items_index[1]
            yield array[prev:next_]
            items_index = (next_, next_ + floor)
            rem -= 1
    else:
        floor, rem = divmod(gen_length, n)

        items_index = (0, floor)
        for _ in range(n):
            prev, next_ = items_index[0], items_index[1] + 1 if rem > 0 else items_index[1]
            yield itertools.islice(array, prev, next_)
            items_index = (next_, next_ + floor)
            rem -= 1


if __name__ == '__main__':
    array_ = iter([12, 7, 9, 31, 13, 11, 7, 3])

    print('Generator:')
    print('----------')
    for value in divide_list(array_, 3, gen_length=8):
        print(list(value))

    print('')
    array_ = [12, 7, 9, 31, 13, 11, 7, 3]

    print('List:')
    print('-----')
    for value in divide_list(array_, 3):
        print(value)

这是输出：

Generator:
----------
[12, 7, 9]
[7, 3]
[]

List:
-----
[12, 7, 9]
[31, 13, 11]
[7, 3]

为什么最后一台发电机耗尽？有时，它会耗尽最后两个生成器。

Answer 1

为什么这不起作用的解释是因为您在提供非零起点时使用islice 跳过元素。这里的关键问题是您应该将迭代器提前一个数量，而不是在每次产量时跳过任何一个。这与序列案例不同，在序列案例中，您为每个案例指定明确的索引。

但是，请注意，您不需要以不同的方式处理这些情况。这是一种处理这两种情况的超级简单方法 - 关键是始终使用迭代器：

def divide(iterable, n, length=None):
    if length is None:
        length = len(iterable)
    it = iter(iterable)
    floor, rem = divmod(length, n)
    while result := list(islice(it, floor + bool(rem))):
        yield result
        rem = max(rem - 1, 0)

在 REPL 中：

>>> from itertools import islice
>>> def divide(iterable, n, length=None):
...     if length is None:
...         length = len(iterable)
...     it = iter(iterable)
...     floor, rem = divmod(length, n)
...     while result := list(islice(it, floor + bool(rem))):
...         yield result
...         rem = max(rem - 1, 0)
...
>>> list(divide([1, 2, 3, 4, 5, 6, 7, 8, 9], 4))
[[1, 2, 3], [4, 5], [6, 7], [8, 9]]
>>> list(divide(iter([1, 2, 3, 4, 5, 6, 7, 8, 9]), 4, length=9))
[[1, 2, 3], [4, 5], [6, 7], [8, 9]]

Answer 2

问题是你没有考虑到你已经使用了迭代器

>>> import itertools
>>> array_ = iter([12, 7, 9, 31, 13, 11, 7, 3])
>>> list(itertools.islice(array_,0,3))
[12, 7, 9]
>>> list(itertools.islice(array_,3,6)) #where are 31,13 and 11? you skipped them bacause, see below 
[7, 3]
>>> array_ = iter([12, 7, 9, 31, 13, 11, 7, 3])
>>> list(itertools.islice(array_,0,3))
[12, 7, 9]
>>> list(array_) #this is what remains in the iterator
[31, 13, 11, 7, 3]
>>> 

Answer 3

感谢@KellyBundy，我能够修改代码以获得预期的结果。我不知道islice 正在切断生成器，然后将这些值移回索引 0 处的开头。我将其视为在我切断的索引处留下了一个空值。这是修改后的代码：

import itertools


def divide_list(array, n, gen_length=None):
    """
    :param array: some iterable that you wish to divide
    :param n: the number of lists you would like to return
    :param gen_length: The length of the generator if array is a generator. Not necessary for lists and tuples.
    :return: a generator of the divided list

    Example:
    In: list(divide_list([1, 2, 3, 4, 5, 6, 7, 8, 9], 4))
    Out: [[1, 2, 3], [4, 5], [6, 7], [8, 9]]
    """
    if isinstance(array, (list, tuple)):
        floor, rem = divmod(len(array), n)

        items_index = (0, floor)
        for _ in range(n):
            prev, next_ = items_index[0], items_index[1] + 1 if rem > 0 else items_index[1]
            yield array[prev:next_]
            items_index = (next_, next_ + floor)
            rem -= 1
    else:
        floor, rem = divmod(gen_length, n)

        up_to = floor
        for _ in range(n):
            up_to = up_to + 1 if rem > 0 else up_to
            yield itertools.islice(array, up_to)
            up_to = floor
            rem -= 1


if __name__ == '__main__':
    array_ = iter([12, 7, 9, 31, 13, 11, 7, 3])

    print('Generator:')
    print('----------')
    for value in divide_list(array_, 3, gen_length=8):
        print(list(value))

    print('')
    array_ = [12, 7, 9, 31, 13, 11, 7, 3]

    print('List:')
    print('-----')
    for value in divide_list(array_, 3):
        print(value)

这将输出预期的结果：

Generator:
----------
[12, 7, 9]
[31, 13, 11]
[7, 3]

List:
-----
[12, 7, 9]
[31, 13, 11]
[7, 3]