Indexing returns views not copies in NumPy - Cheat Sheet & Quick Revision

Recall & Review

beginner

What does it mean that numpy indexing returns views, not copies?

It means when you select part of a numpy array using indexing, you get a view that points to the original data. Changing the view changes the original array too.

beginner

How can you create a copy of a numpy array slice instead of a view?

Use the .copy() method on the slice. For example, slice.copy() creates a new array independent of the original.

intermediate

Why is it useful that numpy indexing returns views?

Views save memory and time because they don’t duplicate data. You can work with parts of large arrays efficiently.

beginner

What happens if you modify a view created by numpy indexing?

The original array changes too because the view shares the same data buffer.

beginner

Give an example of numpy indexing that returns a view.

Example:<br><code>import numpy as np<br>arr = np.array([1, 2, 3, 4])<br>view = arr[1:3]<br>view[0] = 10<br>print(arr)</code><br>Output:<br><code>[ 1 10  3  4]</code><br>This shows the original array changed when the view was modified.

What does numpy indexing return by default?

A A new unrelated array

B A copy of the original array

C A list

D A view of the original array

Indexing in numpy returns a view, which shares data with the original array.

How do you create a copy of a numpy array slice?

A Use <code>.copy()</code> method

B Use <code>.view()</code> method

C Use <code>list()</code> function

D Use <code>slice()</code> function

The .copy() method creates a new array independent of the original.

If you change a value in a numpy view, what happens?

A The original array changes too

B Only the view changes

C Both arrays become disconnected

D An error occurs

Views share data with the original array, so changes affect both.

Why might numpy use views instead of copies?

A To confuse users

B Because copies are impossible

C To save memory and improve speed

D To prevent changes

Views avoid copying data, saving memory and making operations faster.

Which method returns a new array that does NOT affect the original when changed?

A Slicing with colon

B <code>.copy()</code>

C Indexing with brackets

D <code>.view()</code>

.copy() creates a separate array independent of the original.

Explain in your own words what happens when you index a numpy array. How does this affect memory and data changes?

Think about whether the data is copied or shared.

Describe how to safely modify a part of a numpy array without changing the original array.

Consider how to break the link between the slice and original.