Python:NumPy .view()

The .view() method in NumPy returns a new array object that views the same data as the original array. Both arrays share the same underlying memory block, meaning that any modification to the data in one will directly affect the other. Only the array’s metadata (such as data type or shape) may differ.

Creating a view is much faster and more memory-efficient than creating a copy, especially when working with large arrays.

Syntax

ndarray.view([dtype][, type])

Parameters:

• dtype (data-type): The desired data type for the new array view. Changing the dtype changes how the data bytes are interpreted, not the data itself.
• type (type): The desired type for the resulting object (e.g., np.matrix).

Return value:

Returns a new ndarray object that shares the data of the original array.

Example

In this example, modifying the data in the view (view_array) also changes the original array (original_array) because both share the same memory:

import numpy as np

# Create the original array
original_array = np.array([1, 2, 3, 4, 5])

# Create a view of the original array
view_array = original_array.view()

print(f"Original Array before modification: {original_array}")
print(f"View Array before modification: {view_array}")

# Modify a single element in the view
view_array[0] = 99

print(f"\nOriginal Array after modifying the view: {original_array}")

Copy to clipboard
Copy to clipboard

The output of this code is:

Original Array before modification: [1 2 3 4 5]
View Array before modification: [1 2 3 4 5]

Original Array after modifying the view: [99  2  3  4  5]

Copy to clipboard
Copy to clipboard

Codebyte Example

In this example, changing the original array also updates its view, since both reference the same data in memory:

Visit us
Visit us
Hide code
Code
Output
Hide output
Loading...
Copy to your clipboard
Copy Codebyte
Run