Question

I am very new to Python. I need to have a 3 dimensional matrix, in order to save 8 by 8 matrix in some length. Let’s call 530. The problem is that I used np.array since matrix cannot have more than 2 dimensions as numpy argues.
R = zeros([8,8,530],float)
I calculated my 8 by 8 matrix as a np.matrix
R[:,:,ii] = smallR
And, then I try to save it in mat file as scipy claims to do so.
sio.savemat('R.mat',R)
However, error says ‘numpy.ndarray’ object has no attribute ‘items’

/usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio.py:266: FutureWarning: Using oned_as default value ('column') This will change to 'row' in future versions
oned_as=oned_as)
Traceback (most recent call last):
File "ClassName.py", line 83, in <module>
print (buildR()[1])
File "ClassName.py", line 81, in buildR
sio.savemat('R.mat',R)
File "/usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio.py", line 269, in savemat
MW.put_variables(mdict)
File "/usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio5.py", line 827, in put_variables
for name, var in mdict.items():
AttributeError: 'numpy.ndarray' object has no attribute 'items'

Answer 1

If you type help(sio.savemat), you see:

savemat(file_name, mdict, appendmat=True, format='5', long_field_names=False, do_compression=False, oned_as=None)
    Save a dictionary of names and arrays into a MATLAB-style .mat file.
[...]
    mdict : dict
        Dictionary from which to save matfile variables.

and so even if you don’t recognize .items() as a dictionary method, it’s clear we’re going to need to use a dictionary (a set of key, value pairs; google “python dictionary tutorial” if necessary).

In this case:

>>> from numpy import zeros
>>> from scipy import io as sio
>>> 
>>> R = zeros([8,8,530],float)
>>> R += 12.3
>>> 
>>> sio.savemat('R.mat', {'R': R})
>>> 
>>> S = sio.loadmat('R.mat')
>>> S
{'R': array([[[ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        ..., 

        ..., 
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3]]]), '__version__': '1.0', '__header__': 'MATLAB 5.0 MAT-file Platform: posix, Created on: Sat Feb 25 18:16:02 2012', '__globals__': []}
>>> S['R']
array([[[ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        ..., 

        ..., 
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3],
        [ 12.3,  12.3,  12.3, ...,  12.3,  12.3,  12.3]]])

Basically, a dictionary is used so that the arrays can be named, as you can store multiple objects in one .mat file.