Some of the Python functions return either a copy or a view of the input array. A Python copy stores the array in another location while a view uses the same memory content. This means copies are separate objects and treated as a deep copy in Python. Views are the original base array and are treated as a shallow copy. Here are some properties of copies and views:
- Modifications in a view affect the original data whereas modifications in a copy do not affect the original array.
- Views use the concept of shared memory.
- Copies require extra space compared to views.
- Copies are slower than views.
Let's understand the concept of copy and view using the following example:
# Create NumPy Array
arr = np.arange(1,5).reshape(2,2)
print(arr)
Output:
[[1, 2],
[3, 4]]
After creating a NumPy array, let's perform object copy operations:
# Create no copy only assignment
arr_no_copy=arr
# Create Deep Copy
arr_copy=arr.copy()
# Create shallow copy using View
arr_view=arr...
