CODEWITHSUNDEEP
pythonpandasdataframematlabmat
umich
umich

Reputation: 1

Convert matlab data in mat file containing many arrays into numpy arrays

I have a .mat file with many arrays and try to convert them to numpy arrays but was not successful so far.

from scipy.io import loadmat
import numpy as np
import pandas as pd

mat = loadmat("FanMap_data.mat")
print(mat.keys())


dict_keys(['__header__', '__version__', '__globals__', 'fanmap', 'fanmap1D'])

The file contains two sets of arrays 'fanmap' and 'fanmap1D' of float64 data.

fanmap consists of 1 array with shape(1,15), 2 arrays with shape (1943,1) and 26 arrays with shape(1943,15) fanmap1D consists of 22 arrays of shape(1,15).

I tried to use pd.DataFrame, but I do not know how to read so many arrays.

I expect to read all arrays and convert them to numpy arrays.

Upvotes: 0

Views: 68

Answers (2)

umich
umich

Reputation: 1

mat = loadmat("FanMap_data.mat")
fanmap1D_data = mat['fanmap1D']
fanmap1D_type = fanmap1D_data.dtype
print(fanmap1D_type) # to identify the available keys
fanmap1D_ndata = {n: fanmap1D_data[n][0,0] for n in fanmap1D_type.names}
fanmap1D_Npc = fanmap1D_ndata['Npc']

fanmap1D_Npc is a numpy array. All other arrays can be read in a similar way.

Upvotes: 0

Im Groot
Im Groot

Reputation: 639

You can iterate through the keys of the mat dictionary and convert each array to a NumPy array. One way to do is as below:

from scipy.io import loadmat
import numpy as np

mat = loadmat("FanMap_data.mat")
fanmap_arrays = {key: np.array(mat['fanmap'][0, 0][key]) for key in mat['fanmap'].dtype.names}
fanmap1D_arrays = {key: np.array(mat['fanmap1D'][0, 0][key]) for key in mat['fanmap1D'].dtype.names}

print("Arrays in 'fanmap':")
for key, value in fanmap_arrays.items():
    print(key, value.shape)

print("\nArrays in 'fanmap1D':")
for key, value in fanmap1D_arrays.items():
    print(key, value.shape)

Upvotes: 0

Related Questions