如何将任意长度的列表拆分为大小相等的块?


请参阅如果数据结果将直接用于循环,并且不需要存储,则如何以块形式遍历列表。

对于字符串输入的同一问题,请参见每n个字符拆分字符串?。相同的技术通常适用,但也有一些变化。


当前回答

我非常喜欢tzot和J.F.Sebastian提出的Python文档版本,但它有两个缺点:

它不是很明确我通常不希望在最后一个块中有填充值

我在代码中经常使用这个:

from itertools import islice

def chunks(n, iterable):
    iterable = iter(iterable)
    while True:
        yield tuple(islice(iterable, n)) or iterable.next()

更新:一个懒块版本:

from itertools import chain, islice

def chunks(n, iterable):
   iterable = iter(iterable)
   while True:
       yield chain([next(iterable)], islice(iterable, n-1))

其他回答

如果您知道列表大小:

def SplitList(mylist, chunk_size):
    return [mylist[offs:offs+chunk_size] for offs in range(0, len(mylist), chunk_size)]

如果没有(迭代器):

def IterChunks(sequence, chunk_size):
    res = []
    for item in sequence:
        res.append(item)
        if len(res) >= chunk_size:
            yield res
            res = []
    if res:
        yield res  # yield the last, incomplete, portion

在后一种情况下,如果您可以确保序列始终包含给定大小的整数个块(即没有不完整的最后一个块),则可以用更漂亮的方式重新表述。

一个简单的解决方案

OP已请求“相等大小的块”。我将“等尺寸”理解为“平衡”尺寸:如果尺寸不可能相等(例如,23/5),我们正在寻找尺寸大致相同的物品组。

这里的输入是:

项目列表:input_list(例如,23个数字的列表)要拆分这些项目的组数:n个组(例如5个)

输入:

input_list = list(range(23))
n_groups = 5

连续元素组:

approx_sizes = len(input_list)/n_groups 

groups_cont = [input_list[int(i*approx_sizes):int((i+1)*approx_sizes)] 
               for i in range(n_groups)]

“每N个”元素组:

groups_leap = [input_list[i::n_groups] 
               for i in range(n_groups)]

后果

print(len(input_list))

print('Contiguous elements lists:')
print(groups_cont)

print('Leap every "N" items lists:')
print(groups_leap)

将输出:23连续元素列表:[[0, 1, 2, 3], [4, 5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16, 17], [18, 19, 20, 21, 22]]跳过每“N”个项目列表:[[0, 5, 10, 15, 20], [1, 6, 11, 16, 21], [2, 7, 12, 17, 22], [3, 8, 13, 18], [4, 9, 14, 19]]

您可以使用numpy的array_split函数,例如np.array_split(np.array(data),20),将其拆分为20个大小几乎相等的块。

要确保块的大小完全相等,请使用np.split。

根据这个答案,得票最多的答案在最后留下一个“矮子”。这是我的解决方案,可以在没有矮子的情况下,尽可能地获得大小均匀的块。它基本上试图准确选择应该拆分列表的小数点,但只需将其舍入到最接近的整数:

from __future__ import division  # not needed in Python 3
def n_even_chunks(l, n):
    """Yield n as even chunks as possible from l."""
    last = 0
    for i in range(1, n+1):
        cur = int(round(i * (len(l) / n)))
        yield l[last:cur]
        last = cur

演示:

>>> pprint.pprint(list(n_even_chunks(list(range(100)), 9)))
[[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
 [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21],
 [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32],
 [33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43],
 [44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55],
 [56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66],
 [67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77],
 [78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88],
 [89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99]]
>>> pprint.pprint(list(n_even_chunks(list(range(100)), 11)))
[[0, 1, 2, 3, 4, 5, 6, 7, 8],
 [9, 10, 11, 12, 13, 14, 15, 16, 17],
 [18, 19, 20, 21, 22, 23, 24, 25, 26],
 [27, 28, 29, 30, 31, 32, 33, 34, 35],
 [36, 37, 38, 39, 40, 41, 42, 43, 44],
 [45, 46, 47, 48, 49, 50, 51, 52, 53, 54],
 [55, 56, 57, 58, 59, 60, 61, 62, 63],
 [64, 65, 66, 67, 68, 69, 70, 71, 72],
 [73, 74, 75, 76, 77, 78, 79, 80, 81],
 [82, 83, 84, 85, 86, 87, 88, 89, 90],
 [91, 92, 93, 94, 95, 96, 97, 98, 99]]

与排名前几的答案进行比较:

>>> pprint.pprint(list(chunks(list(range(100)), 100//9)))
[[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
 [11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21],
 [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32],
 [33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43],
 [44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54],
 [55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65],
 [66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76],
 [77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87],
 [88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98],
 [99]]
>>> pprint.pprint(list(chunks(list(range(100)), 100//11)))
[[0, 1, 2, 3, 4, 5, 6, 7, 8],
 [9, 10, 11, 12, 13, 14, 15, 16, 17],
 [18, 19, 20, 21, 22, 23, 24, 25, 26],
 [27, 28, 29, 30, 31, 32, 33, 34, 35],
 [36, 37, 38, 39, 40, 41, 42, 43, 44],
 [45, 46, 47, 48, 49, 50, 51, 52, 53],
 [54, 55, 56, 57, 58, 59, 60, 61, 62],
 [63, 64, 65, 66, 67, 68, 69, 70, 71],
 [72, 73, 74, 75, 76, 77, 78, 79, 80],
 [81, 82, 83, 84, 85, 86, 87, 88, 89],
 [90, 91, 92, 93, 94, 95, 96, 97, 98],
 [99]]

用户@tzot的解决方案zip_langest(*[iter(lst)]*n,fillvalue=padvalue)非常优雅,但如果lst的长度不能被n整除,它会填充最后一个子列表,以保持其长度与其他子列表的长度匹配。然而,如果这不可取,那么只需使用zip()生成类似的循环zip,并将lst的剩余元素(不能生成“完整”子列表)附加到输出即可。

输出示例为ABCDEFG,3->ABC DEF G。

单线版本(Python>=3.8):

list(map(list, zip(*[iter(lst)]*n))) + ([rest] if (rest:=lst[len(lst)//n*n : ]) else [])

A函数:

def chunkify(lst, chunk_size):
    nested = list(map(list, zip(*[iter(lst)]*chunk_size)))
    rest = lst[len(lst)//chunk_size*chunk_size: ]
    if rest:
        nested.append(rest)
    return nested

生成器(尽管每个批次都是一个元组):

def chunkify(lst, chunk_size):
    for tup in zip(*[iter(lst)]*chunk_size):
        yield tup
    rest = tuple(lst[len(lst)//chunk_size*chunk_size: ])
    if rest:
        yield rest

它比这里的一些最流行的答案产生相同的输出更快。

my_list, n = list(range(1_000_000)), 12

%timeit list(chunks(my_list, n))                                         # @Ned_Batchelder
# 36.4 ms ± 1.6 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

%timeit [my_list[i:i+n] for i in range(0, len(my_list), n)]              # @Ned_Batchelder
# 34.6 ms ± 1.12 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

%timeit it = iter(my_list); list(iter(lambda: list(islice(it, n)), []))  # @senderle
# 60.6 ms ± 5.36 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

%timeit list(mit.chunked(my_list, n))                                    # @pylang
# 59.4 ms ± 4.92 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

%timeit chunkify(my_list, n)
# 25.8 ms ± 1.84 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

同样,从Python 3.12开始,这个功能将作为itertools模块中的批处理方法来实现(目前是一个配方),因此这个答案很可能会被Python 3.12淘汰。