我从来不想填充我的块，所以这个要求是必要的。我发现在任何可迭代对象上工作的能力也是必需的。鉴于此，我决定扩展公认的答案，https://stackoverflow.com/a/434411/1074659。

如果由于需要比较和筛选填充值而不需要填充，则这种方法的性能会受到轻微的影响。然而，对于大块大小，这个实用程序是非常高性能的。

#!/usr/bin/env python3
from itertools import zip_longest


_UNDEFINED = object()


def chunker(iterable, chunksize, fillvalue=_UNDEFINED):
    """
    Collect data into chunks and optionally pad it.

    Performance worsens as `chunksize` approaches 1.

    Inspired by:
        https://docs.python.org/3/library/itertools.html#itertools-recipes

    """
    args = [iter(iterable)] * chunksize
    chunks = zip_longest(*args, fillvalue=fillvalue)
    yield from (
        filter(lambda val: val is not _UNDEFINED, chunk)
        if chunk[-1] is _UNDEFINED
        else chunk
        for chunk in chunks
    ) if fillvalue is _UNDEFINED else chunks

如果你不介意使用外部包，你可以使用iteration_utilities。Grouper from iteration_utilities它支持所有可迭代对象(不仅仅是序列):

from iteration_utilities import grouper
seq = list(range(20))
for group in grouper(seq, 4):
    print(group)

打印:

(0, 1, 2, 3)
(4, 5, 6, 7)
(8, 9, 10, 11)
(12, 13, 14, 15)
(16, 17, 18, 19)

如果长度不是组大小的倍数，它还支持填充(不完整的最后一组)或截断(丢弃不完整的最后一组)最后一个:

from iteration_utilities import grouper
seq = list(range(17))
for group in grouper(seq, 4):
    print(group)
# (0, 1, 2, 3)
# (4, 5, 6, 7)
# (8, 9, 10, 11)
# (12, 13, 14, 15)
# (16,)

for group in grouper(seq, 4, fillvalue=None):
    print(group)
# (0, 1, 2, 3)
# (4, 5, 6, 7)
# (8, 9, 10, 11)
# (12, 13, 14, 15)
# (16, None, None, None)

for group in grouper(seq, 4, truncate=True):
    print(group)
# (0, 1, 2, 3)
# (4, 5, 6, 7)
# (8, 9, 10, 11)
# (12, 13, 14, 15)

基准

我还决定比较上面提到的几种方法的运行时间。这是一个对数-对数图，根据不同大小的列表将“10”个元素分组。对于定性结果:较低意味着更快:

至少在这个基准测试中iteration_utilities。石斑鱼表现最好。接着是Craz。

基准是用simple_benchmark1创建的。运行这个基准测试的代码是:

import iteration_utilities
import itertools
from itertools import zip_longest

def consume_all(it):
    return iteration_utilities.consume(it, None)

import simple_benchmark
b = simple_benchmark.BenchmarkBuilder()

@b.add_function()
def grouper(l, n):
    return consume_all(iteration_utilities.grouper(l, n))

def Craz_inner(iterable, n, fillvalue=None):
    args = [iter(iterable)] * n
    return zip_longest(*args, fillvalue=fillvalue)

@b.add_function()
def Craz(iterable, n, fillvalue=None):
    return consume_all(Craz_inner(iterable, n, fillvalue))

def nosklo_inner(seq, size):
    return (seq[pos:pos + size] for pos in range(0, len(seq), size))

@b.add_function()
def nosklo(seq, size):
    return consume_all(nosklo_inner(seq, size))

def SLott_inner(ints, chunk_size):
    for i in range(0, len(ints), chunk_size):
        yield ints[i:i+chunk_size]

@b.add_function()
def SLott(ints, chunk_size):
    return consume_all(SLott_inner(ints, chunk_size))

def MarkusJarderot1_inner(iterable,size):
    it = iter(iterable)
    chunk = tuple(itertools.islice(it,size))
    while chunk:
        yield chunk
        chunk = tuple(itertools.islice(it,size))

@b.add_function()
def MarkusJarderot1(iterable,size):
    return consume_all(MarkusJarderot1_inner(iterable,size))

def MarkusJarderot2_inner(iterable,size,filler=None):
    it = itertools.chain(iterable,itertools.repeat(filler,size-1))
    chunk = tuple(itertools.islice(it,size))
    while len(chunk) == size:
        yield chunk
        chunk = tuple(itertools.islice(it,size))

@b.add_function()
def MarkusJarderot2(iterable,size):
    return consume_all(MarkusJarderot2_inner(iterable,size))

@b.add_arguments()
def argument_provider():
    for exp in range(2, 20):
        size = 2**exp
        yield size, simple_benchmark.MultiArgument([[0] * size, 10])

r = b.run()

1免责声明:我是iteration_utilities和simple_benchmark库的作者。

如果列表很大，执行效率最高的方法是使用生成器:

def get_chunk(iterable, chunk_size):
    result = []
    for item in iterable:
        result.append(item)
        if len(result) == chunk_size:
            yield tuple(result)
            result = []
    if len(result) > 0:
        yield tuple(result)

for x in get_chunk([1,2,3,4,5,6,7,8,9,10], 3):
    print x

(1, 2, 3)
(4, 5, 6)
(7, 8, 9)
(10,)

要避免所有到列表的转换，请导入itertools和:

>>> for k, g in itertools.groupby(xrange(35), lambda x: x/10):
...     list(g)

生产:

... 
0 [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
1 [10, 11, 12, 13, 14, 15, 16, 17, 18, 19]
2 [20, 21, 22, 23, 24, 25, 26, 27, 28, 29]
3 [30, 31, 32, 33, 34]
>>> 

我检查了groupby，它不转换为列表或使用len，所以我(认为)这将延迟每个值的解析，直到它实际使用。不幸的是，没有一个现成的答案(在这个时候)似乎提供了这种变化。

显然，如果你需要依次处理每一项，在g上嵌套一个for循环:

for k,g in itertools.groupby(xrange(35), lambda x: x/10):
    for i in g:
       # do what you need to do with individual items
    # now do what you need to do with the whole group

我对此特别感兴趣的是需要消耗一个生成器，以批量提交最多1000个更改到gmail API:

    messages = a_generator_which_would_not_be_smart_as_a_list
    for idx, batch in groupby(messages, lambda x: x/1000):
        batch_request = BatchHttpRequest()
        for message in batch:
            batch_request.add(self.service.users().messages().modify(userId='me', id=message['id'], body=msg_labels))
        http = httplib2.Http()
        self.credentials.authorize(http)
        batch_request.execute(http=http)

还有另一个答案，它的优点是:

1)容易理解 2)适用于任何可迭代对象，而不仅仅是序列(上面的一些答案会阻塞文件句柄) 3)不立即将数据块加载到内存 4)不会在内存中生成对同一迭代器的块长的引用列表 5)在列表的末尾没有填充填充值

话虽如此，我还没有计算它的时间，所以它可能比一些更聪明的方法慢，而且考虑到用例，一些优势可能是无关紧要的。

def chunkiter(iterable, size):
  def inneriter(first, iterator, size):
    yield first
    for _ in xrange(size - 1): 
      yield iterator.next()
  it = iter(iterable)
  while True:
    yield inneriter(it.next(), it, size)

In [2]: i = chunkiter('abcdefgh', 3)
In [3]: for ii in i:                                                
          for c in ii:
            print c,
          print ''
        ...:     
        a b c 
        d e f 
        g h 

Update: A couple of drawbacks due to the fact the inner and outer loops are pulling values from the same iterator: 1) continue doesn't work as expected in the outer loop - it just continues on to the next item rather than skipping a chunk. However, this doesn't seem like a problem as there's nothing to test in the outer loop. 2) break doesn't work as expected in the inner loop - control will wind up in the inner loop again with the next item in the iterator. To skip whole chunks, either wrap the inner iterator (ii above) in a tuple, e.g. for c in tuple(ii), or set a flag and exhaust the iterator.

另一种方法是使用双参数形式的iter:

from itertools import islice

def group(it, size):
    it = iter(it)
    return iter(lambda: tuple(islice(it, size)), ())

这可以很容易地适应使用填充(这类似于Markus Jarderot的答案):

from itertools import islice, chain, repeat

def group_pad(it, size, pad=None):
    it = chain(iter(it), repeat(pad))
    return iter(lambda: tuple(islice(it, size)), (pad,) * size)

这些甚至可以组合为可选的填充:

_no_pad = object()
def group(it, size, pad=_no_pad):
    if pad == _no_pad:
        it = iter(it)
        sentinel = ()
    else:
        it = chain(iter(it), repeat(pad))
        sentinel = (pad,) * size
    return iter(lambda: tuple(islice(it, size)), sentinel)