Saving and loading data (scipy.io) - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of loading a saved .mat file?

Consider the following Python code that saves a dictionary to a .mat file and then loads it back. What will be the output of the print statement?

SciPy

import numpy as np
from scipy.io import savemat, loadmat

my_dict = {'array': np.array([1, 2, 3])}
savemat('test.mat', my_dict)
loaded = loadmat('test.mat')
print(loaded['array'])

A [1 2 3]

B

[[1]
 [2]
 [3]]

C [[1 2 3]]

D KeyError

💡 Hint

Remember how numpy arrays are stored and loaded in MATLAB format. The shape might change.

When saving a 1D numpy array with savemat, it is stored as a 2D array with shape (n, 1). Loading it back returns a 2D array. Here, it is loaded as a column vector with shape (3,1), so printing shows [[1] [2] [3]].

❓ data_output

intermediate

How many keys are in the loaded .mat file dictionary?

After running the following code, how many keys does the loaded dictionary contain?

SciPy

from scipy.io import savemat, loadmat

data = {'x': [1, 2], 'y': [3, 4]}
savemat('data.mat', data)
loaded_data = loadmat('data.mat')
print(len(loaded_data.keys()))

A 2

B 3

C 5

D 4

💡 Hint

loadmat adds some default keys starting with '__'.

loadmat returns a dictionary with the saved keys plus some metadata keys like '__header__', '__version__', and '__globals__'. So total keys are original 2 plus 3 metadata keys = 5 keys.

🔧 Debug

advanced

Why does this code raise a FileNotFoundError?

Examine the code below. Why does it raise a FileNotFoundError?

SciPy

from scipy.io import loadmat

loaded = loadmat('nonexistent_file.mat')

A The loadmat function is deprecated and cannot open files.

B loadmat requires the file to be in .npy format, not .mat.

C The file path must be absolute, relative paths cause errors.

D The file 'nonexistent_file.mat' does not exist in the current directory.

💡 Hint

Check if the file path is correct and the file exists.

FileNotFoundError occurs because the specified file does not exist in the directory where the code runs. loadmat expects the file to exist.

🚀 Application

advanced

Which option correctly saves multiple numpy arrays to a .mat file?

You have two numpy arrays, a and b. Which code snippet correctly saves both arrays into a single .mat file named 'arrays.mat'?

SciPy

import numpy as np
from scipy.io import savemat

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

A savemat('arrays.mat', {'a': a, 'b': b})

B savemat('arrays.mat', [a, b])

C savemat('arrays.mat', a=a, b=b)

D savemat('arrays.mat', {'arrays': [a, b]})

💡 Hint

savemat expects a dictionary mapping variable names to arrays.

savemat requires a dictionary where keys are variable names and values are numpy arrays. Option A correctly passes a dictionary with keys 'a' and 'b'. Option A passes a list, which is invalid. Option A uses keyword arguments which savemat does not accept. Option A nests arrays inside a list under one key, which saves only one variable.

🧠 Conceptual

expert

What is the main limitation of using scipy.io.loadmat for loading MATLAB files?

Which of the following is a key limitation when using scipy.io.loadmat to load MATLAB .mat files?

A It cannot load MATLAB files saved in version 7.3 or later because they use HDF5 format.

B It only loads files smaller than 1MB due to memory constraints.

C It converts all numeric arrays to Python lists automatically.

D It requires MATLAB to be installed on the system to work.

💡 Hint

Check the file format versions supported by scipy.io.loadmat.

scipy.io.loadmat cannot read MATLAB .mat files saved in version 7.3 or later because those files use the HDF5 format, which loadmat does not support. For those files, h5py or MATLAB engine is needed.