np.split() for dividing arrays in NumPy

Introduction

We use np.split() to break a big array into smaller parts. This helps us work with pieces of data separately.

You want to divide a list of daily temperatures into weeks.

You have a long list of sales numbers and want to analyze each quarter separately.

You want to split a dataset into training and testing parts for a machine learning model.

You want to process parts of an image separately by splitting its pixel array.

Syntax

NumPy

import numpy as np

# Split array into multiple sub-arrays
np.split(array, indices_or_sections, axis=0)

array is the original array you want to split.

indices_or_sections can be an integer (number of equal parts) or a list of indices where to split.

Examples

Splits the array into 3 equal parts.

NumPy

import numpy as np

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

# Split into 3 equal parts
parts = np.split(array, 3)
print(parts)

Splits the array at positions 2 and 4, creating 3 parts.

NumPy

import numpy as np

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

# Split at indices 2 and 4
parts = np.split(array, [2, 4])
print(parts)

Splitting an empty array returns empty arrays.

NumPy

import numpy as np

array = np.array([])

# Split empty array into 2 parts
parts = np.array_split(array, 2)
print(parts)

Splitting a single element array into 1 part returns the array itself.

NumPy

import numpy as np

array = np.array([10])

# Split single element array into 1 part
parts = np.split(array, 1)
print(parts)

Sample Program

This program creates an array of 12 numbers and splits it into 4 equal parts. It prints the original array and each part after splitting.

NumPy

import numpy as np

# Create an array of 12 numbers
numbers = np.array([10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120])

print("Original array:", numbers)

# Split into 4 equal parts
split_parts = np.split(numbers, 4)

print("\nArray split into 4 parts:")
for index, part in enumerate(split_parts, start=1):
    print(f"Part {index}:", part)

Important Notes

Time complexity is O(n) because the array elements are copied into new arrays.

Space complexity is O(n) since new arrays are created for each part.

A common mistake is trying to split into parts that do not evenly divide the array length, which causes an error.

Use np.array_split() if you want to split into unequal parts without error.

Summary

np.split() breaks an array into smaller arrays at specified points or into equal parts.

It requires the array to be divisible evenly if using an integer number of parts.

Useful for dividing data into chunks for separate analysis or processing.

Practice

1. What does the np.split() function do in NumPy?

easy

A. It merges multiple arrays into one large array.

B. It divides an array into multiple smaller arrays at specified indices or into equal parts.

C. It sorts the elements of an array in ascending order.

D. It reshapes an array into a different shape without changing data.

Solution

1. Step 1: Understand the purpose of np.split()

   The function is designed to break a large array into smaller arrays either by specifying split points or by dividing into equal parts.

2. Step 2: Compare with other array functions

   Unlike merging, sorting, or reshaping, np.split specifically divides arrays.

3. Final Answer:

   It divides an array into multiple smaller arrays at specified indices or into equal parts. -> Option B

4. Quick Check:

   np.split() = divides array [OK]

Hint: Remember: split means break into parts [OK]

Common Mistakes:

• Confusing split with merge
• Thinking it sorts arrays
• Mixing with reshape

2. Which of the following is the correct syntax to split a 1D NumPy array arr into 3 equal parts using np.split()?

easy

A. np.split(arr, 3)

B. np.split(arr, [3])

C. np.split(arr, indices=3)

D. np.split(arr, sections=3)

Solution

1. Step 1: Recall np.split() syntax

   The function signature is np.split(array, indices_or_sections), where indices_or_sections can be an integer for equal parts.

2. Step 2: Identify correct usage

   Passing 3 as the second argument splits the array into 3 equal parts. Keyword arguments like 'indices_or_sections' are not used explicitly in common syntax.

3. Final Answer:

   np.split(arr, 3) -> Option A

4. Quick Check:

   np.split(arr, 3) splits into 3 parts [OK]

Hint: Use integer directly for equal parts [OK]

Common Mistakes:

• Using list instead of integer for equal parts
• Wrong keyword arguments
• Confusing indices with number of parts

3. What is the output of the following code?

import numpy as np
arr = np.array([10, 20, 30, 40, 50, 60])
splits = np.split(arr, [2, 4])
print(splits)

medium

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

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

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

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

Solution

1. Step 1: Understand split indices

   The list [2, 4] means split at index 2 and then at index 4.

2. Step 2: Divide the array accordingly

   arr[:2] = [10, 20], arr[2:4] = [30, 40], arr[4:] = [50, 60]

3. Final Answer:

   [array([10, 20]), array([30, 40]), array([50, 60])] -> Option C

4. Quick Check:

   Split at 2 and 4 gives 3 parts [OK]

Hint: Split indices mark where to cut array [OK]

Common Mistakes:

• Misunderstanding split points
• Off-by-one errors in slicing
• Confusing output format

4. The following code throws an error. What is the cause?

import numpy as np
arr = np.array([1, 2, 3, 4, 5])
splits = np.split(arr, 3)
print(splits)

medium

A. The array length is not divisible by 3, so equal splits are impossible.

B. np.split() requires a list of indices, not an integer.

C. The array must be 2D to use np.split().

D. The print statement syntax is incorrect.

Solution

1. Step 1: Check array length and split parts

   The array length is 5, but we want to split into 3 equal parts.

2. Step 2: Understand np.split() behavior

   np.split() requires the array length to be divisible by the number of parts when splitting equally, otherwise it raises an error.

3. Final Answer:

   The array length is not divisible by 3, so equal splits are impossible. -> Option A

4. Quick Check:

   Length 5 not divisible by 3 causes error [OK]

Hint: Check divisibility of array length by parts [OK]

Common Mistakes:

• Using integer splits on incompatible length
• Confusing np.split with np.array_split
• Assuming print error

5. You have a 2D NumPy array data with shape (6, 4). You want to split it into 3 equal parts along the first axis (rows) for separate analysis. Which code correctly does this using np.split()?

hard

A. np.split(data, [1, 3, 5], axis=1)

B. np.split(data, 3, axis=1)

C. np.split(data, [2, 4], axis=1)

D. np.split(data, 3, axis=0)

Solution

1. Step 1: Understand array shape and axis

   data has 6 rows and 4 columns. Splitting into 3 equal parts along rows means each part has 2 rows.

2. Step 2: Use np.split() with correct axis and parts

   np.split(data, 3, axis=0) splits rows into 3 equal parts. axis=1 would split columns, which is not desired.

3. Final Answer:

   np.split(data, 3, axis=0) -> Option D

4. Quick Check:

   Split rows into 3 parts with axis=0 [OK]

Hint: Use axis=0 to split rows, axis=1 for columns [OK]

Common Mistakes:

• Splitting along wrong axis
• Using indices instead of number for equal parts
• Ignoring array shape