I’m trying to read a subfolder from .mat file in Python using hdf5 package. I could extract the integers and values from the datase, but couldn’t read the strings.
result=matfile[“/pro/tttt/rrr”]
Return is a series of lists:
[[<HDF5 object reference>] [<HDF5 object reference>] [<HDF5 object reference>]]
I need to extract these strings and combine them with another column as its header. Any suggestions?
An object reference points to another object (typically another dataset) in the file. You use it to get the referenced object. There are several SO answers with examples. Use this search to find them "[matlab] [h5py] object". Here is an answer I wrote earlier this year that is a good starting point: reading matlab data into numpy arrays. Here is a link to related h5py doc: Object and Region References.
I wrote a short example for your file. See code below. It gets the dataset referenced by the first object. Do the same for [1], [2], etc.
array_from_objref = matfile[ matfile[“/pro/tttt/rrr"][0] ][:].
Print the full path with this:
print(matfile[“/pro/tttt/rrr"][0].name)
array_from_objref = file[ file['/pro/tttt/rrr'][0] ][:] then I got the following error: AttributeError: 'numpy.ndarray' object has no attribute 'encode' do you know why? but it's true, those string are a reference (it's the header) for another array in that dataset
print(matfile[“/pro/tttt/rrr"][0].dtype)print( matfile[ matfile[“/pro/tttt/rrr"][0] ].dtype ) to get the dtype of the referenced object. Or, if you print the name, you can use that instead.>>> element=file[file["/pro/tttt/rrrrr"][0][0]] >>> elementStr=''.join(chr(i) for i in element[0]) >>> print(elementStr) S , if I run with [1], it prints the second character, etc,. not working with [:] neighter len(x) :/