Indexing returns views not copies in NumPy - Step-by-Step Execution

Concept Flow - Indexing returns views not copies

Create original array

↓

Index array to get subset

↓

Subset is a view (shares data)

↓

Modify subset

↓

Original array changes too

↓

If copy made, original unchanged

Indexing an array returns a view that shares data with the original array. Changing the view changes the original.

Execution Sample

NumPy

import numpy as np
arr = np.array([1, 2, 3, 4])
view = arr[1:3]
view[0] = 20
print(arr)

Create an array, get a slice view, modify the view, and print original array to see change.

Execution Table

Step	Action	Array state	View state	Note
1	Create arr	[1, 2, 3, 4]	N/A	Original array created
2	Create view = arr[1:3]	[1, 2, 3, 4]	[2, 3]	View shares data with arr
3	Modify view[0] = 20	[1, 20, 3, 4]	[20, 3]	Change in view affects arr
4	Print arr	[1, 20, 3, 4]	N/A	Original array shows updated value

💡 Execution ends after printing arr showing shared data change

Variable Tracker

Variable	Start	After Step 2	After Step 3	Final
arr	[1, 2, 3, 4]	[1, 2, 3, 4]	[1, 20, 3, 4]	[1, 20, 3, 4]
view	N/A	[2, 3]	[20, 3]	[20, 3]

Key Moments - 2 Insights

Why does changing the view also change the original array?

Because the view shares the same data buffer as the original array, so modifying the view modifies the original. See execution_table step 3.

Is the view a separate copy of the data?

No, the view is not a copy but a window into the original array's data. This is why changes reflect back. See execution_table step 2.

Visual Quiz - 3 Questions

Look at the execution table at step 3, what is the value of arr after modifying view?

A [1, 2, 3, 4]

B [20, 3, 3, 4]

C [1, 20, 3, 4]

D [1, 3, 3, 4]

💡 Hint

Check the 'Array state' column at step 3 in execution_table

At which step does the view get created?

A Step 2

B Step 3

C Step 1

D Step 4

💡 Hint

Look at the 'Action' column for when view = arr[1:3] happens

If we changed view[0] without the view sharing data, what would happen to arr?

A arr would also change

B arr would remain unchanged

C arr would become empty

D arr would raise an error

💡 Hint

Refer to the concept that indexing returns views, not copies, in variable_tracker

Concept Snapshot

Indexing a NumPy array returns a view, not a copy.
Modifying the view changes the original array.
Slices share the same data buffer.
To avoid this, use .copy() to make a separate copy.
Remember: view = window, copy = separate data.

Full Transcript

We start by creating a NumPy array arr with values [1, 2, 3, 4]. Then we create a view by slicing arr from index 1 to 3, which gives us [2, 3]. This view shares the same data as arr. When we modify the first element of the view to 20, the original array arr also changes at that position, becoming [1, 20, 3, 4]. Finally, printing arr shows the updated array. This happens because the view is not a copy but a window into the original data. To avoid this, one must explicitly create a copy.