I am trying to manipulate a copy of a matrix, as follows.
import numpy as np
A = np.matrix([[4, 1, -1, 1], [1, 4, -1, -1], [-1, -1, 5, 1], [1, -1, 1, 3]])
B = A
B[0, 0] = 0
This would change 00-entry of A to 0 as well, but I don't want this to happen. I tried it with np.array() as well, but the same thing happens again. I think this is because A and B are pointing to the same spot in the memory.
I'd appreciate it if anyone let's me know how to prevent A from changing, while B is changed.
You are completely right, A and B point to the same memory allocation in your case.
numpy has a copy function for that what you want to achieve:
B = np.copy(A)
When you do B = A, B is just a reference to the already existing array A. NumPy doesn't copy straight ahead for efficiency reasons.
So, you should use the arr.copy() to copy the contents of the array as in:
In [9]: B = A.copy()
In [10]: A
Out[10]:
matrix([[ 4, 1, -1, 1],
[ 1, 4, -1, -1],
[-1, -1, 5, 1],
[ 1, -1, 1, 3]])
# update an entry in array `B`
In [11]: B[0, 0] = 0
# A is unaffected
In [12]: A
Out[12]:
matrix([[ 4, 1, -1, 1],
[ 1, 4, -1, -1],
[-1, -1, 5, 1],
[ 1, -1, 1, 3]])
# change happens only in array B
In [13]: B
Out[13]:
matrix([[ 0, 1, -1, 1],
[ 1, 4, -1, -1],
[-1, -1, 5, 1],
[ 1, -1, 1, 3]])
Try this:
B[:] = A
This will make a deep copy
You need to create the copy of the object. You can do this with numpy.copy since you are working with a numpy object.
So, your code will be like so:
B = np.copy(A)
There are two kinds of copy used with python.
A shallow copy constructs a new compound object and then (to the extent possible) inserts references into it to the objects found in the original.
A deep copy constructs a new compound object and then, recursively, inserts copies into it of the objects found in the original.
They work independent of object type.
from copy import copy, deepcopy
# Shallow copy
b= copy(a)
# Deep copy of object
b= deepcopy(a)
You are correct that right now your B and A are referencing the same spot in the memory. For your problem I would recommend
numpy.matrix.copy
check out the documentation here: https://docs.scipy.org/doc/numpy-1.10.4/reference/generated/numpy.matrix.copy.html
There is an example there with exactly what you are looking for:
x = np.array([[1,2,3],[4,5,6]], order='F')
y = x.copy()
x.fill(0)
x = array([[0, 0, 0],[0, 0, 0]])
y = array([[1, 2, 3],[4, 5, 6]])
B = Anever creates a copy in python no matter what objectsAandBrefer to