NumPy Splitting Arrays – Complete Guide to split(), array_split(), hsplit(), vsplit(), and dsplit()

🐍 NumPy – Splitting Arrays

In data analysis and scientific computing, large datasets are often divided into smaller parts for easier processing.

NumPy provides powerful functions for splitting arrays into multiple sub-arrays without manually slicing data.

Array splitting is commonly used in:

• Data preprocessing
• Machine learning
• Batch processing
• Dataset partitioning
• Scientific computing

By mastering array splitting, you can efficiently organize and process large amounts of data.

---

What is Array Splitting?

Array splitting means dividing a large array into multiple smaller arrays.

Example:

Original array:

[1, 2, 3, 4, 5, 6]

Split into 3 parts:

[1, 2]
[3, 4]
[5, 6]

NumPy provides several functions for this purpose.

---

Using np.split()

The split() function divides an array into equal-sized sub-arrays.

Syntax

np.split(array, sections, axis=0)

Where:

• array = source array
• sections = number of splits or split indexes
• axis = dimension to split

---

Splitting a 1D Array

import numpy as np

arr = np.array([1, 2, 3, 4, 5, 6])

result = np.split(arr, 3)

print(result)

Output:

[array([1, 2]),
 array([3, 4]),
 array([5, 6])]

The array is divided into three equal parts.

---

Understanding Equal Division Requirement

split() requires equal division.

Example:

arr = np.array([1, 2, 3, 4, 5])

np.split(arr, 2)

Output:

ValueError

Because:

5 ÷ 2 = not equal

NumPy cannot create equal-sized parts.

---

Using np.array_split()

The array_split() function can split arrays even when equal division is not possible.

Example

import numpy as np

arr = np.array([1, 2, 3, 4, 5])

result = np.array_split(arr, 2)

print(result)

Output:

[array([1, 2, 3]),
 array([4, 5])]

Notice that NumPy distributes elements automatically.

---

Splitting Using Index Positions

Instead of specifying the number of sections, you can define split points.

import numpy as np

arr = np.array([10, 20, 30, 40, 50, 60])

result = np.split(arr, [2, 4])

print(result)

Output:

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

Split positions:

Index 2
Index 4

---

Splitting 2D Arrays

Consider the following matrix:

import numpy as np

arr = np.array([
    [1, 2],
    [3, 4],
    [5, 6],
    [7, 8]
])

---

Vertical Splitting

result = np.split(arr, 2)

print(result)

Output:

[array([[1, 2],
        [3, 4]]),

 array([[5, 6],
        [7, 8]])]

Rows are divided into two groups.

---

Splitting Along Columns

Use axis=1.

import numpy as np

arr = np.array([
    [1, 2],
    [3, 4],
    [5, 6]
])

result = np.split(arr, 2, axis=1)

print(result)

Output:

[array([[1],
        [3],
        [5]]),

 array([[2],
        [4],
        [6]])]

---

Using hsplit()

The hsplit() function performs horizontal splitting.

Example

import numpy as np

arr = np.array([
    [1, 2, 3, 4],
    [5, 6, 7, 8]
])

result = np.hsplit(arr, 2)

print(result)

Output:

[array([[1, 2],
        [5, 6]]),

 array([[3, 4],
        [7, 8]])]

Columns are divided.

---

Using vsplit()

The vsplit() function performs vertical splitting.

Example

import numpy as np

arr = np.array([
    [1, 2],
    [3, 4],
    [5, 6],
    [7, 8]
])

result = np.vsplit(arr, 2)

print(result)

Output:

[array([[1, 2],
        [3, 4]]),

 array([[5, 6],
        [7, 8]])]

Rows are divided.

---

Using dsplit()

The dsplit() function splits along the third axis.

Example:

import numpy as np

arr = np.arange(16).reshape(2, 2, 4)

result = np.dsplit(arr, 2)

print(result)

Output:

Two 3D arrays

This is useful when working with:

• RGB images
• Video frames
• Tensor data

---

Understanding Array Dimensions

Example:

arr.shape

Output:

(2, 3)

Meaning:

• 2 rows
• 3 columns

Splitting depends on which axis is selected.

---

Common Splitting Functions

Function	Purpose
split()	Equal splitting
array_split()	Unequal splitting
hsplit()	Horizontal split
vsplit()	Vertical split
dsplit()	Depth split

---

Split vs Array Split

Feature	split()	array_split()
Equal division required	Yes	No
Flexible sizes	No	Yes
Error on uneven split	Yes	No
Most commonly used	Sometimes	Frequently

---

Real-World Example

Splitting a dataset into training and testing sets.

import numpy as np

data = np.arange(100)

train, test = np.split(data, [80])

print(train.shape)
print(test.shape)

Output:

(80,)
(20,)

This approach is commonly used in machine learning projects.

---

Image Processing Example

RGB images contain three color channels:

Red
Green
Blue

Using dsplit():

red, green, blue = np.dsplit(image, 3)

Each channel becomes a separate array.

---

Performance Considerations

Splitting arrays is efficient because NumPy often creates views rather than copying data.

Benefits:

• Lower memory usage
• Faster execution
• Better scalability

---

Best Practices

• Use array_split() when sizes may not divide evenly.
• Use hsplit() for column operations.
• Use vsplit() for row operations.
• Verify array shape before splitting.
• Use views when possible to save memory.

---

Common Errors

Unequal Split Error

arr = np.array([1,2,3,4,5])

np.split(arr, 2)

Output:

ValueError:
array split does not result in an equal division

Solution:

np.array_split(arr, 2)

---

Summary

NumPy provides several powerful array splitting functions:

• split()
• array_split()
• hsplit()
• vsplit()
• dsplit()

These functions allow you to divide large arrays into smaller, manageable pieces for analysis and processing.

---

Conclusion

Array splitting is an essential NumPy skill for data manipulation and preprocessing. Whether you're preparing machine learning datasets, handling image channels, or organizing scientific data, NumPy's splitting functions provide efficient and flexible solutions.

Understanding when to use split(), array_split(), hsplit(), vsplit(), and dsplit() will help you write cleaner and more efficient Python code.