flatten() and ravel() for 1D conversion in NumPy - Step-by-Step Execution

Concept Flow - flatten() and ravel() for 1D conversion

Start with 2D array

↓

Choose method: flatten() or ravel()

↓

flatten()

returns a copy→1D array copy

↓

ravel()

returns a view if possible→1D array view or copy

↓

Use 1D array for further processing

Start with a 2D array, then use flatten() to get a copy as 1D, or ravel() to get a view or copy as 1D, then use the 1D array.

Execution Sample

NumPy

import numpy as np
arr = np.array([[1, 2], [3, 4]])
flat = arr.flatten()
rav = arr.ravel()
flat[0] = 100
rav[1] = 200
print(arr)
print(flat)
print(rav)

This code converts a 2D array to 1D using flatten() and ravel(), modifies elements, and prints all arrays to show differences.

Execution Table

Step	Action	Array State	Notes
1	Create arr = [[1, 2], [3, 4]]	[[1, 2], [3, 4]]	Original 2D array
2	flat = arr.flatten()	[1, 2, 3, 4]	flatten() returns a copy as 1D
3	rav = arr.ravel()	[1, 2, 3, 4]	ravel() returns a view as 1D
4	flat[0] = 100	[100, 2, 3, 4]	Change in flat only, arr unchanged
5	rav[1] = 200	[1, 200, 3, 4]	Change in rav affects arr because rav is a view
6	Print arr	[[1, 200], [3, 4]]	arr changed due to rav modification
7	Print flat	[100, 2, 3, 4]	flat unchanged after modification
8	Print rav	[1, 200, 3, 4]	rav shows updated values
9	End	-	Execution complete

💡 All steps executed; flatten() copy unaffected by changes, ravel() view reflects changes in original array.

Variable Tracker

Variable	Start	After Step 2	After Step 3	After Step 4	After Step 5	Final
arr	[[1, 2], [3, 4]]	[[1, 2], [3, 4]]	[[1, 2], [3, 4]]	[[1, 2], [3, 4]]	[[1, 200], [3, 4]]	[[1, 200], [3, 4]]
flat	N/A	[1, 2, 3, 4]	[1, 2, 3, 4]	[100, 2, 3, 4]	[100, 2, 3, 4]	[100, 2, 3, 4]
rav	N/A	N/A	[1, 2, 3, 4]	[1, 2, 3, 4]	[1, 200, 3, 4]	[1, 200, 3, 4]

Key Moments - 3 Insights

Why does changing flat[0] not affect the original array arr?

Because flatten() returns a copy of the array, changes to flat do not affect arr. See execution_table step 4 and 6 where arr remains unchanged after flat modification.

Why does changing rav[1] affect the original array arr?

Because ravel() returns a view (not a copy) when possible, changes to rav reflect in arr. See execution_table step 5 and 6 where arr changes after modifying rav.

Are flat and rav always independent arrays?

No. flat is always a copy, independent from arr. rav is a view if possible, so it may share data with arr. See variable_tracker for their states after modifications.

Visual Quiz - 3 Questions

Look at the execution_table at step 5. What is the value of arr after rav[1] = 200?

A [[1, 200], [3, 4]]

B [[1, 2], [3, 4]]

C [[100, 2], [3, 4]]

D [[1, 2], [200, 4]]

💡 Hint

Check execution_table row 5 and 6 showing arr changes after modifying rav.

According to variable_tracker, what is the value of flat after step 4?

A [1, 2, 3, 4]

B [100, 2, 3, 4]

C [1, 200, 3, 4]

D [[100, 2], [3, 4]]

💡 Hint

Look at variable_tracker row for flat after step 4.

If we modify arr[0,0] = 500 after step 5, what happens to rav?

A rav remains unchanged

B flat[0] becomes 500

C rav[0] becomes 500

D flat and rav both become 500

💡 Hint

Recall rav is a view of arr, so changes in arr reflect in rav.

Concept Snapshot

flatten() and ravel() convert arrays to 1D.
flatten() returns a copy; changes do not affect original.
ravel() returns a view if possible; changes affect original.
Use flatten() to keep original safe.
Use ravel() for memory efficiency.

Full Transcript

We start with a 2D numpy array. Using flatten(), we get a new 1D array copy. Using ravel(), we get a 1D view if possible. Changing the flattened copy does not change the original array. Changing the raveled view changes the original array because they share data. This shows the difference between copy and view in numpy. Remember flatten() is safe for independent changes, ravel() is efficient but linked to original.