NumPy多维数组的mask掩码索引

参考链接: Boolean or “mask” index arrays

说明:通过逻辑判断符号可以得到一个mask掩码,其值是True或者False,并且维度和原来的多维数组相同.使用mask来索引得到的是一个一维数组,筛选出True位置对应的元素.通过mask掩码就可以对原多维数组进行修改.但是如果将mask筛选出的内容赋值给一个变量的话,无法通过这个变量来修改原来的多维数组,他们互不影响.
实验1:

(base) PS C:\Users\chenxuqi> python
Python 3.7.4 (default, Aug  9 2019, 18:34:13) [MSC v.1915 64 bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> z = np.arange(12).reshape(3,4)
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> mask4cxq = z > 6
>>> mask4cxq
array([[False, False, False, False],
       [False, False, False,  True],
       [ True,  True,  True,  True]])
>>> z[mask4cxq]
array([ 7,  8,  9, 10, 11])
>>> a = z[mask4cxq]
>>> a
array([ 7,  8,  9, 10, 11])
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> a[0] = 20200910
>>> a
array([20200910,        8,        9,       10,       11])
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> z[mask4cxq] = 8888888
>>> z
array([[      0,       1,       2,       3],
       [      4,       5,       6, 8888888],
       [8888888, 8888888, 8888888, 8888888]])
>>> a
array([20200910,        8,        9,       10,       11])
>>>
>>>  

实验2: 掩码mask与整数索引混合使用

Python 3.7.4 (tags/v3.7.4:e09359112e, Jul  8 2019, 20:34:20) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license()" for more information.
>>> import numpy as np
>>> z = np.arange(12).reshape(3,4)
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> mask4cxq = z > 6
>>> mask4cxq
array([[False, False, False, False],
       [False, False, False,  True],
       [ True,  True,  True,  True]])
>>> z[mask4cxq]
array([ 7,  8,  9, 10, 11])
>>> z[mask4cxq] = 20200910
>>> z
array([[       0,        1,        2,        3],
       [       4,        5,        6, 20200910],
       [20200910, 20200910, 20200910, 20200910]])
>>> z[z < 4] = 888
>>> z
array([[     888,      888,      888,      888],
       [       4,        5,        6, 20200910],
       [20200910, 20200910, 20200910, 20200910]])
>>> 
>>> 
>>> 
>>> z = np.arange(12).reshape(3,4)
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> z[...,0][z[...,0] > 1] = 20200910 # 只修改某一列
>>> z
array([[       0,        1,        2,        3],
       [20200910,        5,        6,        7],
       [20200910,        9,       10,       11]])
>>> 
>>> 
>>> z = np.arange(12).reshape(3,4)
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> z[z[...,1]>4]
array([[ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> z[z[...,1]>4,1]
array([5, 9])
>>> # 只修改某一列
>>> z[z[...,1]>4,1] = 20200910
>>> z
array([[       0,        1,        2,        3],
       [       4, 20200910,        6,        7],
       [       8, 20200910,       10,       11]])
>>> 
>>> z = np.arange(12).reshape(3,4)
>>> z
array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])
>>> z[..., 1:][z[..., 1:]<10] = 20200910
>>> z
array([[       0, 20200910, 20200910, 20200910],
       [       4, 20200910, 20200910, 20200910],
       [       8, 20200910,       10,       11]])
>>> 
>>> 
>>>