8

Following this question which asks (and answers) how to read .mat files that were created in Matlab using Scipy, I want to know how to access the fields in the imported structs.

I have a file in Matlab from which I can import a struct:

>> load bla % imports a struct called G
>> G

G = 

         Inp: [40x40x2016 uint8]
         Tgt: [8x2016 double]
         Ltr: [1x2016 double]
    Relevant: [1 2 3 4 5 6 7 8]

Now I want to do the same in Python:

x = scipy.io.loadmat('bla.mat')
>>> x
{'__version__': '1.0', '__header__': 'MATLAB 5.0 MAT-file, Platform: PCWIN, Created on: Wed Jun 07 21:17:24 2006', 'G': array([[<scipy.io.matlab.mio5.mat_struct object at 0x0191F230>]], dtype=object), '__globals__': []}
>>> x['G']
array([[<scipy.io.matlab.mio5.mat_struct object at 0x0191F230>]], dtype=object)
>>> G = x['G']
>>> G
array([[<scipy.io.matlab.mio5.mat_struct object at 0x0191F230>]], dtype=object)

The question is, how can I access the members of the struct G: Inp, Tgt, Ltr and Relevant, the way I can in Matlab?

Community
  • 1
  • 1
Nathan Fellman
  • 122,701
  • 101
  • 260
  • 319
  • What do help(G) and dir(G) say (and same for G[0]) – Kimvais Dec 31 '09 at 10:08
  • What happens if you use `x = scipy.io.loadmat('bla.mat',struct_as_record=True)`? See http://docs.scipy.org/doc/scipy/reference/generated/scipy.io.loadmat.html – unutbu Dec 31 '09 at 12:00

1 Answers1

6

First, I'd recommend to upgrade to Scipy svn if possible - there has been active development of the matlab io with some really dramatic speed ups recently.

Also as mentioned it might be worth trying with struct_as_record=True. But otherwise you should be able to get it out by playing around interactively.

Your G is an array of mio struct objects - you can check G.shape for example. In this case I think G = x['G'][0,0] should give the object you want. Then you should be able to access G.Inp etc.

robince
  • 10,826
  • 3
  • 35
  • 48