我需要一个滚动窗口(又名滑动窗口)可迭代的序列/迭代器/生成器。(默认的Python迭代可以被认为是一种特殊情况,其中窗口长度为1。)我目前正在使用以下代码。我怎样才能做得更优雅和/或更有效?

def rolling_window(seq, window_size):
    it = iter(seq)
    win = [it.next() for cnt in xrange(window_size)] # First window
    yield win
    for e in it: # Subsequent windows
        win[:-1] = win[1:]
        win[-1] = e
        yield win

if __name__=="__main__":
    for w in rolling_window(xrange(6), 3):
        print w

"""Example output:   
   [0, 1, 2]
   [1, 2, 3]
   [2, 3, 4]
   [3, 4, 5]
"""

对于window_size == 2的特定情况(即,在序列中迭代相邻的重叠对),请参见如何从列表中迭代重叠(当前,下一个)值对?


当前回答

尝试我的部分,简单,一行,使用islice的python方式。但是,可能不是最佳效率。

from itertools import islice
array = range(0, 10)
window_size = 4
map(lambda i: list(islice(array, i, i + window_size)), range(0, len(array) - window_size + 1))
# output = [[0, 1, 2, 3], [1, 2, 3, 4], [2, 3, 4, 5], [3, 4, 5, 6], [4, 5, 6, 7], [5, 6, 7, 8], [6, 7, 8, 9]]

解释: 通过使用window_size的islice创建窗口,并在所有数组上使用map迭代此操作。

其他回答

>>> n, m = 6, 3
>>> k = n - m+1
>>> print ('{}\n'*(k)).format(*[range(i, i+m) for i in xrange(k)])
[0, 1, 2]
[1, 2, 3]
[2, 3, 4]
[3, 4, 5]

另一种从列表生成固定长度窗口的简单方法

from collections import deque

def window(ls,window_size=3):
    window = deque(maxlen=window_size)

    for element in ls:
        
        if len(window)==window_size:
            yield list(window)
        window.append(element)

ls = [0,1,2,3,4,5]

for w in window(ls):
    print(w)

除了我想到的解决方案,我还测试了一些解决方案。我发现我想出的是最快的,所以我想分享我的python3实现。

import itertools
import sys

def windowed(l, stride):
    return zip(*[itertools.islice(l, i, sys.maxsize) for i in range(stride)])
def GetShiftingWindows(thelist, size):
    return [ thelist[x:x+size] for x in range( len(thelist) - size + 1 ) ]

>> a = [1, 2, 3, 4, 5]
>> GetShiftingWindows(a, 3)
[ [1, 2, 3], [2, 3, 4], [3, 4, 5] ]
#Importing the numpy library
import numpy as np
arr = np.arange(6) #Sequence
window_size = 3
np.lib.stride_tricks.as_strided(arr, shape= (len(arr) - window_size +1, window_size), 
strides = arr.strides*2)

"""Example output:

  [0, 1, 2]
  [1, 2, 3]
  [2, 3, 4]
  [3, 4, 5]

"""