pytutorial/numpy_mat_files/README.md

Numpy: Dealing with Matlab files

{:.no_toc}

* TOC {:toc}

The goal

We want to read and write Matlab files under Python.

Questions to David Rotermund

Reminder: Learning Python as Matlab user Please read NumPy for MATLAB users

Mat files under Python

MATLAB < 7.3 format mat files

This is a job for scipy.io

loadmat(file_name[, mdict, appendmat])	Load MATLAB file.
savemat(file_name, mdict[, appendmat, …])	Save a dictionary of names and arrays into a MATLAB-style .mat file.
whosmat(file_name[, appendmat])	List variables inside a MATLAB file.

Read

Under Matlab we create a test file

>> A = rand(10,100);
>> save -v7 Test_1.mat A

Under Python we look into the file for information:

import scipy.io as sio

Info = sio.whosmat("Test_1.mat")
print(Info) # --> [('A', (10, 100), 'double')]

And we can read the data: 

import numpy as np
import scipy.io as sio

mat_data = sio.loadmat("Test_1.mat")

print(
    mat_data.keys()
)  # -> dict_keys(['__header__', '__version__', '__globals__', 'A'])

a = mat_data["A"]
print(type(a))  # --> <class 'numpy.ndarray'>
print(a.dtype)  # --> float64
print(a.shape)  # --> (10, 100)

Write

Under Python we generate a .mat file:

import numpy as np
import scipy.io as sio

myrng = np.random.default_rng()

A: np.ndarray = myrng.random((10, 100), dtype=np.float64)
B: str = "Hellp world!"
mdic = {"A": A, "ImportantMessage": B}
sio.savemat("Test_2.mat", mdic)

And read it in under Matlab:

>> load Test_2.mat
>> whos
  Name                   Size             Bytes  Class     Attributes

  A                     10x100             8000  double              
  ImportantMessage       1x12                24  char                

>> ImportantMessage

ImportantMessage =

    'Hellp world!'

MATLAB == 7.3 format mat files