NumPy Array View or Shallow Copy. The following shows how to create a view or shallow copy of a NumPy array of the Python NumPy module. 

import numpy as np

In [7]:

a = np.arange(12)
b = a
b is a

Out[7]:

True

In [8]:

a

Out[8]:

array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11])

In [9]:

b.shape = 3,4
a.shape

Out[9]:

(3, 4)

In [10]:

a

Out[10]:

array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])

In [11]:

def f(x):
    print(id(x))

id(a)

Out[11]:

140043709957584

In [12]:

f(a)

Out[12]: 

140043709957584

View or Shallow Copy:

In [15]:

c = a.view()
c is a

Out[15]:

False

In [16]:

a

Out[16]:

array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])

In [17]:

c

Out[17]:

array([[ 0,  1,  2,  3],
       [ 4,  5,  6,  7],
       [ 8,  9, 10, 11]])

In [18]:

c.base is a

Out[18]:

True

In [19]:

c.flags.owndata

Out[19]:

False

In [20]:

a.flags.owndata

Out[20]:

True

In [21]:

c.shape = 2,6
a.shape

Out[21]:

(3, 4)

In [22]:

c[0,4] = 1234
a

Out[22]:

array([[   0,    1,    2,    3],
       [1234,    5,    6,    7],
       [   8,    9,   10,   11]])

In [23]:

s = a[ : , 1:3]
s[:] = 10
a

Out[23]:

array([[   0,   10,   10,    3],
       [1234,   10,   10,    7],
       [   8,   10,   10,   11]])

In [24]:

s

Out[24]:

array([[10, 10],
       [10, 10],
       [10, 10]])