Splitting Arrays in NumPy

Splitting Arrays Equally with np.split()

The np.split() function is used to divide an array into multiple equal-sized parts. This only works if the array can be split evenly — otherwise, NumPy will raise an error.

Here’s an example of splitting a 1D array into 3 equal parts:

import numpy as np

# Create a 1D array with 6 elements
arr = np.array([10, 20, 30, 40, 50, 60])

# Split it into 3 equal parts
result = np.split(arr, 3)

print(result)

import numpy as np

# Create a 1D array with 6 elements
arr = np.array([10, 20, 30, 40, 50, 60])

# Split it into 3 equal parts
result = np.split(arr, 3)

print(result)

How It Works:

• The array has 6 elements and is split into 3 parts → each part has 2 elements.
• np.split(arr, 3) means “split into 3 equal sub-arrays”.
• If the array size wasn’t divisible by 3, this would raise a ValueError.
• You can also use keyword arguments like np.split(arr, indices_or_sections=3) for clarity.

Output

[array([10, 20]), array([30, 40]), array([50, 60])]

[array([10, 20]), array([30, 40]), array([50, 60])]

💡 Tip: Use np.split() only when you're sure the array can be divided evenly. Otherwise, use np.array_split() to avoid errors.

Splitting Arrays Flexibly with np.array_split()

The np.array_split() function works like np.split(), but it allows you to split arrays even when they can’t be divided evenly.

If the array size doesn't divide perfectly, NumPy will create sub-arrays of slightly different sizes — no error is raised.

Here’s an example of splitting a 1D array with 5 elements into 3 parts:

import numpy as np

# Create a 1D array with 5 elements
arr = np.array([10, 20, 30, 40, 50])

# Split into 3 parts (not evenly divisible)
result = np.array_split(arr, 3)

print(result)

import numpy as np

# Create a 1D array with 5 elements
arr = np.array([10, 20, 30, 40, 50])

# Split into 3 parts (not evenly divisible)
result = np.array_split(arr, 3)

print(result)

How It Works:

• The array has 5 elements, but we ask for 3 splits.
• NumPy creates 3 sub-arrays: two with 2 elements, one with 1 element.
• This function avoids errors and distributes the elements as evenly as possible.

Output

[array([10, 20]), array([30, 40]), array([50])]

[array([10, 20]), array([30, 40]), array([50])]

The result is a list of NumPy arrays. Each sub-array contains a chunk of the original data.

💡 Tip: Use array_split() when you're not sure the array can be divided evenly. It's more flexible than split() and won’t raise an error.

Axis-Based Splitting in NumPy

When working with 2D arrays (like tables or matrices), NumPy offers specific functions to split along rows or columns:

• np.hsplit() – splits horizontally (by columns)
• np.vsplit() – splits vertically (by rows)

Here’s an example using a 2D array with shape (2, 4):

import numpy as np

# Create a 2D array (2 rows, 4 columns)
arr = np.array([[10, 20, 30, 40],
                [50, 60, 70, 80]])

# Split into 2 parts along columns (horizontal split)
h_result = np.hsplit(arr, 2)

# Split into 2 parts along rows (vertical split)
v_result = np.vsplit(arr, 2)

print("Horizontal split:")
print(h_result)

print("Vertical split:")
print(v_result)

import numpy as np

# Create a 2D array (2 rows, 4 columns)
arr = np.array([[10, 20, 30, 40],
                [50, 60, 70, 80]])

# Split into 2 parts along columns (horizontal split)
h_result = np.hsplit(arr, 2)

# Split into 2 parts along rows (vertical split)
v_result = np.vsplit(arr, 2)

print("Horizontal split:")
print(h_result)

print("Vertical split:")
print(v_result)

How It Works:

• hsplit() splits along axis 1 (columns), dividing each row into chunks.
• vsplit() splits along axis 0 (rows), dividing the array into separate rows.
• These functions are shorthand versions of np.split() with a fixed axis.

Output

Horizontal split:
[array([[10, 20],
        [50, 60]]), 
 array([[30, 40],
        [70, 80]])]

Vertical split:
[array([[10, 20, 30, 40]]), 
 array([[50, 60, 70, 80]])]

Horizontal split:
[array([[10, 20],
        [50, 60]]), 
 array([[30, 40],
        [70, 80]])]

Vertical split:
[array([[10, 20, 30, 40]]), 
 array([[50, 60, 70, 80]])]

💡 Tip: Use hsplit() and vsplit() when working with 2D arrays and want quick column-wise or row-wise splitting without manually specifying the axis.

Frequently Asked Questions