Read .mat files in Python

Question

Is it possible to read binary MATLAB .mat files in Python?

I've seen that SciPy has alleged support for reading .mat files, but I'm unsuccessful with it. I installed SciPy version 0.7.0, and I can't find the loadmat() method.

Solution 1

An import is required, import scipy.io...

import scipy.io
mat = scipy.io.loadmat('file.mat')

Solution 2

Neither scipy.io.savemat, nor scipy.io.loadmat work for MATLAB arrays version 7.3. But the good part is that MATLAB version 7.3 files are hdf5 datasets. So they can be read using a number of tools, including NumPy.

For Python, you will need the h5py extension, which requires HDF5 on your system.

import numpy as np
import h5py
f = h5py.File('somefile.mat','r')
data = f.get('data/variable1')
data = np.array(data) # For converting to a NumPy array

Solution 3

First save the .mat file as:

save('test.mat', '-v7')

After that, in Python, use the usual loadmat function:

import scipy.io as sio
test = sio.loadmat('test.mat')

Solution 4

There is a nice package called mat4py which can easily be installed using

pip install mat4py

It is straightforward to use (from the website):

Load data from a MAT-file

The function loadmat loads all variables stored in the MAT-file into a simple Python data structure, using only Python’s dict and list objects. Numeric and cell arrays are converted to row-ordered nested lists. Arrays are squeezed to eliminate arrays with only one element. The resulting data structure is composed of simple types that are compatible with the JSON format.

Example: Load a MAT-file into a Python data structure:

from mat4py import loadmat

data = loadmat('datafile.mat')

The variable data is a dict with the variables and values contained in the MAT-file.

Save a Python data structure to a MAT-file

Python data can be saved to a MAT-file, with the function savemat. Data has to be structured in the same way as for loadmat, i.e. it should be composed of simple data types, like dict, list, str, int, and float.

Example: Save a Python data structure to a MAT-file:

from mat4py import savemat

savemat('datafile.mat', data)

The parameter data shall be a dict with the variables.

Solution 5

Having MATLAB 2014b or newer installed, the MATLAB engine for Python could be used:

import matlab.engine
eng = matlab.engine.start_matlab()
content = eng.load("example.mat", nargout=1)

View more solutions

Comments

[texnic]

scipy does not support v7.3 mat-files (see notes here). See the answer by vikrantt for solution.

[watsonic]

however, you can save mat-files as earlier versions. see: mathworks.com/help/matlab/import_export/mat-file-versions.ht‌​ml (header: 'Save to Nondefault MAT-File Version')

[watsonic]

e.g. save('myfile.mat','-v7')

[chipaudette]

This works fine, if you use the '-v7.3' flag in Matlab when saving out your data. Using the default save (at least in Matlab R2014b) results in a file that cannot be read using the technique above. If you do use the '-v7.3' flag, the numeric data can be read just fine.

[vikrantt]

Yes, that's what I said in my post. You need to use -v7.3 while saving in Matlab. You should do that anyways as it uses a better/more supported/standardized format.

[Daniel]

That's a somehow crazy documentation provided by mathworks. 40 pages explaining the format, without mentioning that it is a subset of HDF5.

[heracho]

Could you please explain what is the relation between f and data in your example? How can I move f to a numpy array?

[Kevin Katzke]

Save a variable with this command from the prompt: save('filename', '-v7.3', 'var1');

[VimNing]

I got this error: ModuleNotFoundError: No module named 'pylab'.

[Daniel]

You got the error when trying this answers? That is odd, it does not use pylab.

[denis]

Note that mat4py gives you a json-like tree of dicts, lists, lists of lists ... -- no numpy at all. (mat4py/cmd.py my.mat writes my.json, 1 long line.)

[Cleb]

@denis: Yes, that's also stated above. But a good point indeed: I usually like this structure, e.g. in web applications as numpy arrays are not JSON serializable.

[s2t2]

Encountered: mat4py.loadmat.ParseError: Can only read from Matlab level 5 MAT-files

[Cleb]

@s2t2: never ran into this issue before. What matlab version and which scipy version are you using?

[Aleksejs Fomins]

ParseError: Unexpected field name length: 43

[Cleb]

@AleksejsFomins: Probably best to open a new question and link to this thread; then it will be easier to help (make sure to provide all info needed to reproduce the error).

[devspartan]

How would i even know that it contains data under data/variable1 ??

[Chachni]

@s2t2 mat4py cannot read the new matlab format from version 7.3, which is encoded as hdf5.

[Packard CPW]

@devSpartan f.keys() will show you what you can access

[Ramin Melikov]

"Unable to open file (file signature not found)"

[ThatNewGuy]

@heracho f is the file object that lets you read data from the hdf5 file. It's similar to using f = open('myfile.txt') to read a text file.

[ThatNewGuy]

@vikrantt This is a great solution for reading a struct but it does not work for a Matlab table. To read tables, you can't use the save function. Instead, you need to utilize the h5create and h5write functions.

[emilaz]

At least for .mat files generated with MATLAB, this will result in a UnicodeDecodeError.

[MrCrHaM]

@emilaz This is expected. np.loadtxt is only meant for .mat files generated by Octave.

[emilaz]

Yes, I just put it there as clarification for future people looking at this :)

[SjonTeflon]

Also note that mat4py does not allow to read complex valued .mat files

[nKandel]

The 2 things that it did better are: it preserve the mat object dimension and the loaded object was in NumPy array. Thanks

[Abrar_11648]

I faced this issue: "mat4py.loadmat.ParseError: Can only read from Matlab level 5 MAT-files"

[panter]

I'm glad it helped :)

[ZaydH]

Updated link to the SciPy.io tutorial docs.scipy.org/doc/scipy/tutorial/io.html @FranckDernoncourt

[LightCC]

For the record, this answer requires a valid Matlab installation and license - it runs Matlab in the background to accomplish the read. And there may be limitations on what format you get the items in that need further work to make them readable. For example, Simulink.Bus objects come in as a "matlab object" and must be processed further, with issues if you want to extract the Bus Element objects.