不调用len（），这对大型列表很有用：

def splitter(l, n):
    i = 0
    chunk = l[:n]
    while chunk:
        yield chunk
        i += n
        chunk = l[i:i+n]

这是可迭代的：

def isplitter(l, n):
    l = iter(l)
    chunk = list(islice(l, n))
    while chunk:
        yield chunk
        chunk = list(islice(l, n))

上述产品的功能风味：

def isplitter2(l, n):
    return takewhile(bool,
                     (tuple(islice(start, n))
                            for start in repeat(iter(l))))

OR:

def chunks_gen_sentinel(n, seq):
    continuous_slices = imap(islice, repeat(iter(seq)), repeat(0), repeat(n))
    return iter(imap(tuple, continuous_slices).next,())

OR:

def chunks_gen_filter(n, seq):
    continuous_slices = imap(islice, repeat(iter(seq)), repeat(0), repeat(n))
    return takewhile(bool,imap(tuple, continuous_slices))

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。

我想我没有看到这个选项，所以只需添加另一个：）：

def chunks(iterable, chunk_size):
  i = 0;
  while i < len(iterable):
    yield iterable[i:i+chunk_size]
    i += chunk_size

senderle答案的一个线性版本：

from itertools import islice
from functools import partial

seq = [1,2,3,4,5,6,7]
size = 3
result = list(iter(partial(lambda it: tuple(islice(it, size)), iter(seq)), ()))
assert result == [(1, 2, 3), (4, 5, 6), (7,)]

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。

我很好奇不同方法的性能，这里是：

在Python 3.5.1上测试

import time
batch_size = 7
arr_len = 298937

#---------slice-------------

print("\r\nslice")
start = time.time()
arr = [i for i in range(0, arr_len)]
while True:
    if not arr:
        break

    tmp = arr[0:batch_size]
    arr = arr[batch_size:-1]
print(time.time() - start)

#-----------index-----------

print("\r\nindex")
arr = [i for i in range(0, arr_len)]
start = time.time()
for i in range(0, round(len(arr) / batch_size + 1)):
    tmp = arr[batch_size * i : batch_size * (i + 1)]
print(time.time() - start)

#----------batches 1------------

def batch(iterable, n=1):
    l = len(iterable)
    for ndx in range(0, l, n):
        yield iterable[ndx:min(ndx + n, l)]

print("\r\nbatches 1")
arr = [i for i in range(0, arr_len)]
start = time.time()
for x in batch(arr, batch_size):
    tmp = x
print(time.time() - start)

#----------batches 2------------

from itertools import islice, chain

def batch(iterable, size):
    sourceiter = iter(iterable)
    while True:
        batchiter = islice(sourceiter, size)
        yield chain([next(batchiter)], batchiter)


print("\r\nbatches 2")
arr = [i for i in range(0, arr_len)]
start = time.time()
for x in batch(arr, batch_size):
    tmp = x
print(time.time() - start)

#---------chunks-------------
def chunks(l, n):
    """Yield successive n-sized chunks from l."""
    for i in range(0, len(l), n):
        yield l[i:i + n]
print("\r\nchunks")
arr = [i for i in range(0, arr_len)]
start = time.time()
for x in chunks(arr, batch_size):
    tmp = x
print(time.time() - start)

#-----------grouper-----------

from itertools import zip_longest # for Python 3.x
#from six.moves import zip_longest # for both (uses the six compat library)

def grouper(iterable, n, padvalue=None):
    "grouper(3, 'abcdefg', 'x') --> ('a','b','c'), ('d','e','f'), ('g','x','x')"
    return zip_longest(*[iter(iterable)]*n, fillvalue=padvalue)

arr = [i for i in range(0, arr_len)]
print("\r\ngrouper")
start = time.time()
for x in grouper(arr, batch_size):
    tmp = x
print(time.time() - start)

结果：

slice
31.18285083770752

index
0.02184295654296875

batches 1
0.03503894805908203

batches 2
0.22681021690368652

chunks
0.019841909408569336

grouper
0.006506919860839844