TypeError: Expecting miINT32 as data typeД®rror: Command 'gcc -pthread -fno-strict-aliasing -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -fPIC -Iscipy/ndimage/src -I/USERPATH/_workspaces/python/test-envs/scipyloadmat/scipyfork/local/lib/python2.7/site-packages/numpy/core/include -I/ usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio5_utils.so in scipy.io.matlab.mio5_into_int32s (scipy/io/matlab/mio5_utils.c:4242)() usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio5_utils.so in scipy.io.matlab.mio5_header (scipy/io/matlab/mio5_utils.c:4730)() usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio5_utils.so in scipy.io.matlab.mio5_header (scipy/io/matlab/mio5_utils.c:4881)() > 228 header = self._matrix_reader.read_header() usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio5.pyc in read_var_header(self)Дў27 raise TypeError('Expecting miMATRIX type here, got %d' % mdtype) > 268 hdr, next_position = self.read_var_header() usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio5.pyc in get_variables(self, variable_names) > 126 matfile_dict = MR.get_variables(variable_names) usr/local/lib/python2.7/dist-packages/scipy/io/matlab/mio.pyc in loadmat(file_name, mdict, appendmat, *_kwargs)ДЎ24 variable_names = kwargs.pop('variable_names', None)ДЎ25 MR = mat_reader_factory(file_name, appendmat, *_kwargs) I also checked R.matlab readMat - it barfs on this file. I guess that is what matlab does to read it. To read it without error we'd have to allow miUINT32 as well as miINT32 for the array dimensions, and we'd have to allow miUTF8 (type 16) for the variable names instead of miINT (type 1). The data above appears to be a small data element length 4, type miINT32 (=5), data 16, giving the maximum length of the following field names. There follows the name 'RBR' and zero padding to 8 byte boundary -> 56 bytes. The miINT8 gives the data type of the variable name, so miUTF8 makes sense as well, but it's not in the standard as far as I can see. According to the standard, this should be (1, 3) meaning (miINT8, data length 3). Then come the matrix dimensions (1, 1) -> 40 bytesДЄfter comes the tag (16, 3). It should be (5, 8) - meaning ( miINT32, data length=8). Next comes the data (6, 8) meaning (miUINT32, data length=8).

(miUINT32, 8) -> 16 bytes (array flags tag).

(which are loaded as numpy ndarrays), recursing into the elements If isinstance(d, 5_params.mat_struct):ДЄ recursive function which constructs lists from cellarrays The following version will make lists of them instead, and continuing the recursion into the cell array elements if possible. Just an enhancement to mergen's answer, which unfortunately will stop recursing if it reaches a cell array of objects. If isinstance(elem, 5_params.mat_struct): If isinstance(dict, 5_params.mat_struct):ДЄ recursive function which constructs from matobjects nested dictionaries Todict is called to change them to nested dictionaries It calls the function check keys to cure all entriesДata = spio.loadmat(filename, struct_as_record=False, squeeze_me=True)Д¬hecks if entries in dictionary are mat-objects. This function should be called instead of direct spio.loadmatДЄs it cures the problem of not properly recovering python dictionariesДЇrom mat files.

Here are the functions, which reconstructs the dictionaries just use this loadmat instead of scipy.io's loadmat: import scipy.io as spio Now it was possible to access it by > v.c.d However, by using the option struct_as_record=False the field could be accessed: v=spio.loadmat('xy.mat',squeeze_me=True,struct_as_record=False) ValueError Traceback (most recent call last) If I now want to access the fields I get: > vigДЄrray(((array(3),),), dtype=)ДЄrray(array((3,), dtype=), dtype=object) I tried the following: vig=spio.loadmat('xy.mat',squeeze_me=True) Now I want to read the mat-File back into python.

To present the problem I run into in more detail, I give the following toy example: load scipy.io as spio mat files with scipy), I could not access deeper nested structures to recover them into dictionaries Using the given routines (how to load Matlab.

Scipy loadmat