python分割成数组_查找序列并使用Python将其拆分为数组

关于什么

>>> ddata = data[1:] - data[:-1] # faster than np.diff(data)

>>> ddata

array([ 1, 1, 1, 1, 1, 1, 1, 1, 1, 2631, 1,

1, 1, 1, 1, 1, 1, 1, 1, 1996, 1, 1,

1, 12, 1, 1, 1, 1, 1, 1, 1, 1, 1,

971, 1, 1, 1, 1, 1, 1, 1, 1, 1, 40,

1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1])

最后

>>> sequences = np.split(data, np.argwhere(ddata>1).flatten() + 1)

>>> sequences[0]

array([347, 348, 349, 350, 351, 352, 353, 354, 355, 356])

>>> sequences[2]

array([4992, 4993, 4994, 4995])

>>> sequences[-1]

array([6036, 6037, 6038, 6039, 6040, 6041, 6042, 6043, 6044, 6045, 6046,

6047])

执行时间：data [1：]-data [：-1] VS np.diff(data)-tested via repl.it

setup_="""

import numpy as np

MULT = 100

data = np.array(MULT*[ 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 2987,

2988, 2989, 2990, 2991, 2992, 2993, 2994, 2995, 2996, 4992, 4993,

4994, 4995, 5007, 5008, 5009, 5010, 5011, 5012, 5013, 5014, 5015,

5016, 5987, 5988, 5989, 5990, 5991, 5992, 5993, 5994, 5995, 5996,

6036, 6037, 6038, 6039, 6040, 6041, 6042, 6043, 6044, 6045, 6046,

6047])

"""

import timeit

然后

Python 3.6.1 (default, Dec 2015, 13:05:11)

[GCC 4.8.2] on linux

>>> timeit.Timer('data[1:] - data[:-1]', setup=setup_).timeit()

35.17186516011134

>>> timeit.Timer('np.diff(data)', setup=setup_).timeit()

63.88404295803048