Import struct with different data type to python dataframe

Hi,

I tried using python to load the data stored in a matlab struct - the struct i in a file attached.

Here is what I tried:

import scipy.io
import pandas as pd

mat = scipy.io.loadmat("file.mat")
data = mat["struct"]

Then, when I look to what is inside of 'data' I have this:

array([[(MatlabOpaque([(b'', b'MCOS', b'string', array([[3707764736],
                             [         2],
                             [         1],
                             [         1],
                             [         1],
                             [         1]], dtype=uint32))          ],
                     dtype=[('s0', 'O'), ('s1', 'O'), ('s2', 'O'), ('arr', 'O')]), MatlabOpaque([(b'', b'MCOS', b'datetime', array([[3707764736],
                             [         2],
                             [         1],
                             [         1],
                             [         2],
                             [         2]], dtype=uint32))            ],
                     dtype=[('s0', 'O'), ('s1', 'O'), ('s2', 'O'), ('arr', 'O')]), MatlabOpaque([(b'', b'MCOS', b'duration', array([[3707764736],
                             [         2],
                             [         1],
                             [         1],
                             [         3],
                             [         3]], dtype=uint32))            ],
                     dtype=[('s0', 'O'), ('s1', 'O'), ('s2', 'O'), ('arr', 'O')]), array([[1.0082, 1.0102, 1.01  ]]))                            ]],
      dtype=[('Name', 'O'), ('init', 'O'), ('age', 'O'), ('measure', 'O')])

Hoy can I get the data stored in the mat file then?

Thanks in advance for your help.

Answer 1

[Daniele Sportillo]

Hi,

first of all, I suggest you rename the variable struct in your file: struct is a MATLAB data type, and that may create problems with your code.

Then, you can use the MATLAB Engine API for Python (Call MATLAB from Python - MATLAB & Simulink (mathworks.com)) to load your .mat file.

import matlab.engine

eng = matlab.engine.start_matlab()
s = eng.load('file.mat')

You'll be able to directly use some of the fields in your structure. For other fields (e.g., datetime) you would need to first convert them

import matlab.engine
from datetime import datetime

eng = matlab.engine.start_matlab()
s = eng.load('file.mat')

print(s['struct']['measure'])
print(datetime.strptime(eng.string(s['struct']['init']),'%d-%b-%Y %H:%M:%S'))

Hope this helps.

Best,

Daniele