Save and load NumPy arrays with MATLAB .mat files using SciPy

Last update on March 26 2025 13:16:44 (UTC/GMT +8 hours)

17. MATLAB .mat File Round-trip

Write a NumPy program that saves a NumPy array to a MATLAB .mat file using scipy.io.savemat and then reads it back using scipy.io.loadmat.

Sample Solution:

Python Code:

import numpy as np
import scipy.io

# Create a NumPy array
array_to_save = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]])

# Define the file name to save the .mat file
mat_file_name = 'array_data.mat'

# Save the NumPy array to a .mat file
scipy.io.savemat(mat_file_name, {'array': array_to_save})

# Load the data back from the .mat file
loaded_data = scipy.io.loadmat(mat_file_name)

# Retrieve the array from the loaded data
loaded_array = loaded_data['array']

# Print the loaded array to verify it matches the original array
print('Loaded array:\n', loaded_array)

Output:

Loaded array:
 [[1 2 3]
 [4 5 6]
 [7 8 9]]

Explanation:

• Import Libraries:
• Imported numpy as np for array creation.
• Imported scipy.io for saving and loading .mat files.
• Create a NumPy Array:
• Created a 3x3 NumPy array named array_to_save.|
• Define File Name:
• Set the .mat file name to array_data.mat.
• Save Array to .mat File:
• Used scipy.io.savemat to save array_to_save to the specified .mat file, under the variable name 'array'.
• Load Data from .mat File:
• Used scipy.io.loadmat to read the data from the .mat file into a dictionary named loaded_data.
• Retrieve Array from Loaded Data:
• Extracted the array from loaded_data using the key 'array'.
• Print the Loaded Array:
• Printed the loaded array to verify it matches the original array.

For more Practice: Solve these Related Problems:

• Write a Numpy program to save an array to a MATLAB .mat file using scipy.io.savemat and then load it back, verifying all keys and shapes.
• Write a Numpy program to export a structured array to a .mat file and then convert specific fields to a different type upon reloading.
• Write a Numpy program to save multiple arrays into a .mat file and then extract one array for further MATLAB-compatible processing.
• Write a Numpy program to write an array to a .mat file, modify the file externally, and then load it back into Python to check for consistency.

Python-Numpy Code Editor:

Previous: How to write a NumPy array to a SQLite database and read it back?
Next: Read PNG image data into NumPy array using Python.