Save Python .npz as MATLAB .mat file

Question

I have a .npz file that I would like to save as a .mat file.

The file A.npz contains A['arr_0', 'arr_1', 'arr_2']

Currently, I am saving each one individually using the following:

import scipy.io as sio
a1 = A['arr_0']
sio.savemat('a1.mat', mdict={'a1': a1}

for each arr_i within A.npz.

Is there a simple way to save all the contents into a .mat (i.e. via cell format)? Or should I just create a script to simply save them all in a loop like fashion?

Answer 1

Create a npz:

In [147]: np.savez('test.npz',x=np.arange(3),y=np.ones((3,3)),z=np.array(3))

load; data is the dictionary of files/variables; load by indexing:

In [148]: data = np.load('test.npz')
In [149]: list(data.keys())
Out[149]: ['z', 'y', 'x']
In [150]: data['x']
Out[150]: array([0, 1, 2])

Save a similar dictionary to the .mat:

In [151]: io.savemat('a1.mat', mdict={'x':data['x'],'y':data['y'],'z':data['z']})
In [152]: io.loadmat('a1.mat')
Out[152]: 
{'__globals__': [],
 '__header__': b'MATLAB 5.0 MAT-file Platform: posix, Created on: Thu Feb 15 09:52:42 2018',
 '__version__': '1.0',
 'x': array([[0, 1, 2]]),
 'y': array([[1., 1., 1.],
        [1., 1., 1.],
        [1., 1., 1.]]),
 'z': array([[3]])}

alternative constructs:

io.savemat('a1.mat', mdict={key:data[key] for key in data.keys()})

or even:

io.savemat('a1.mat', mdict=data)

since data is a dictionary.

Answer 2

One-liner for shell scripts, etc.:

python -c "import pylab, scipy.io; scipy.io.savemat('tmp.mat', pylab.np.load('tmp.npz'))"