MATLAB file I/O (loadmat, savemat) in SciPy - Step-by-Step Execution

Concept Flow - MATLAB file I/O (loadmat, savemat)

Start: Prepare data in Python

↓

Use savemat() to save data to .mat file

↓

File saved on disk

↓

Use loadmat() to read .mat file

↓

Data loaded back into Python

↓

Use data for analysis or display

↓

End

This flow shows saving Python data to a MATLAB .mat file and then loading it back into Python using scipy functions.

Execution Sample

SciPy

from scipy.io import savemat, loadmat

# Save data
data = {'x': [1, 2, 3], 'y': [4, 5, 6]}
savemat('datafile.mat', data)

# Load data
loaded = loadmat('datafile.mat')
print(loaded)

This code saves a dictionary to a .mat file and then loads it back, printing the loaded data.

Execution Table

Step	Action	Input/Variable	Result/Output
1	Prepare data dictionary	data = {'x': [1,2,3], 'y': [4,5,6]}	{'x': [1,2,3], 'y': [4,5,6]}
2	Call savemat()	savemat('datafile.mat', data)	File 'datafile.mat' created with variables x and y
3	Call loadmat()	loadmat('datafile.mat')	Dictionary with keys: '__header__', '__version__', '__globals__', 'x', 'y'
4	Print loaded data	print(loaded)	{'__header__': ..., '__version__': ..., '__globals__': ..., 'x': array([[1], [2], [3]]), 'y': array([[4], [5], [6]])}
5	End	N/A	Data loaded successfully from .mat file

💡 Data saved and loaded successfully; loadmat adds metadata keys automatically

Variable Tracker

Variable	Start	After savemat	After loadmat	Final
data	{'x': [1,2,3], 'y': [4,5,6]}	No change	No change	No change
loaded	N/A	N/A	{'__header__': ..., 'x': array([[1], [2], [3]]), 'y': array([[4], [5], [6]])}	Same as after loadmat

Key Moments - 2 Insights

Why does loadmat output include keys like '__header__' and '__version__'?

loadmat always adds metadata keys to the loaded dictionary to store file info. Your saved variables appear as keys alongside these.

Why are the loaded arrays wrapped in 2D arrays like array([[1], [2], [3]])?

MATLAB stores arrays as 2D by default, so loadmat returns them as 2D numpy arrays even if originally 1D lists.

Visual Quiz - 3 Questions

Look at the execution table, what does savemat() do at step 2?

A Loads data from a .mat file

B Prints the data dictionary

C Creates a .mat file with the given variables

D Deletes the .mat file

💡 Hint

See step 2 in execution_table where savemat creates the file

At which step does loadmat() add metadata keys like '__header__'?

A Step 3

B Step 2

C Step 1

D Step 4

💡 Hint

Check step 3 in execution_table where loadmat returns dictionary with metadata keys

If the original data had a variable 'z', what would happen in the loaded dictionary?

A The loaded dictionary would exclude 'z'

B The loaded dictionary would include key 'z' with its data

C The file would fail to save

D The loaded dictionary would only have metadata keys

💡 Hint

savemat saves all keys; loadmat loads all saved variables as keys

Concept Snapshot

Use savemat(filename, dict) to save Python dict data to a MATLAB .mat file.
Use loadmat(filename) to load .mat file data back into Python as a dict.
loadmat adds metadata keys like '__header__' automatically.
Arrays load as numpy arrays, often 2D by default.
This allows easy data exchange between Python and MATLAB.

Full Transcript

This visual execution trace shows how to save Python data to a MATLAB .mat file using scipy's savemat function, and then load it back using loadmat. First, a Python dictionary with lists is prepared. Then savemat writes this data to a file named 'datafile.mat'. Next, loadmat reads this file and returns a dictionary that includes the saved variables plus some metadata keys like '__header__'. The loaded arrays appear as 2D numpy arrays. This process allows sharing data between Python and MATLAB easily.