How to convert a sparse matrix to a dense matrix from .mat file in Python?

Question

import xlrd, xlwt
import scipy as sp
import scipy.io
import pandas as pd
import matplotlib.pyplot as plt
import pylab
from scipy.sparse import csc_matrix
%matplotlib inline

FileString = r'/content/drive/MyDrive/Thesis/EXIOBASE_3rx_aggLandUseExtensions_2015_pxp.mat'
MRIO = scipy.io.loadmat(FileString)

Regions = MRIO['IO']['A']

IN: Regions
OUT: array([[<42800x42800 sparse matrix of type '<class 'numpy.float64'>'
    with 5986549 stored elements in Compressed Sparse Column format>]],
      dtype=object)

IN: Regions.todense()
OUT: 
AttributeError                            Traceback (most recent call last)
<ipython-input-116-3ef413dd7ae9> in <module>()
----> 1 Regions.todense()

AttributeError: 'numpy.ndarray' object has no attribute 'todense'

I am trying to convert this sparse matrix from a MATLAB file into a dense matrix, thus I applied todense() function but I don't know why it doesn't work. Your help will be highly appreciated Thanks

Answer 1

You need to unpack:

array([[<42800x42800 sparse matrix of type '<class 'numpy.float64'>'
with 5986549 stored elements in Compressed Sparse Column format>]],
  dtype=object)

This is a ndarray as shown by the error. It has shape (1,1), object dtype. That one object is the sparse matrix. I deduced that from the display, the array, object and [[...]]`.

x[0,0].todense()
x[0,0].A        # for array

should work.

I'd suggest first displaying

 x[0,0]

I suspect in MATLAB this is a cell, with size (1,1). loadmat makes a liberal use of object dtype arrays to hold MATLAB cell and struct. Only a pure matrix becomes a numeric array.