Often, you need to create a copy of an array to make changes without affecting the original array.

Simple Assignment

First, we'll discuss why we can't simply create another variable using array_2 = array_1, where array_1 is our original array. Let's look at an example:

We changed the value of the first element of array_2 to 10, but this assignment also changed the value of the first element of array_1 to 10.

Note

With array_2 = array_1, you are not creating a new array; instead, you are creating a reference to the same array in memory. Therefore, any changes made to array_2 will also affect array_1.

To solve this problem, we could write array_2 = np.array([1, 2, 3]), but that would mean writing the same code twice. Remember the key principle in coding: Don't repeat yourself.

ndarray.copy() Method

Luckily, NumPy has an ndarray.copy() method as a solution to this problem. Let's see it in action:

Now, we have created a new array for array_2 with the same elements as array_1.

For 2D arrays, the copying procedure is exactly the same.

numpy.copy() Function

Instead of the .copy() method, we can also use the copy() function, which takes the array as its parameter: array_2 = np.copy(array_1).

Both the function and the method work the same; however, there is one nuance. They both have the order parameter, which specifies the memory layout of the array, but their default values are different. If you want to read more about it, here is the documentation: ndarray.copy, numpy.copy.