Read HDF5 files with groups and subgroups.

Hey,

I have a HDF5 file with some groups and each of these groups contains a number of subgroups.

Is there a way to read all subgroups and save them in some normal arrays or so?

Answer 1

[Samay Sagar]

You can use the ”h5info” function to get information about the file structure of your input HDF5 file. Thereafter, you can loop through groups and subgroups to read datasets using “h5read” and store them in arrays.

Here’s how you can read HDF5 file to extract groups and subgroups data:

% Specify your HDF5 file name 
filename = 'yourfile.hdf5'; 
% Get info about the HDF5 file 
info = h5info(filename); 
% Preallocate a cell array to hold your data 
data = struct(); 
% Loop through each group 
for i = 1:length(info.Groups) 
    group_name = info.Groups(i).Name; 
    data.(matlab.lang.makeValidName(group_name)) = struct(); % Create a struct for the group 
    % Loop through each subgroup within the group 
    for j = 1:length(info.Groups(i).Groups) 
        subgroup_name = info.Groups(i).Groups(j).Name; 
        data.(matlab.lang.makeValidName(group_name)).(matlab.lang.makeValidName(subgroup_name)) = struct(); % Create a struct for the subgroup          
        % Loop through each dataset within the subgroup 
        for k = 1:length(info.Groups(i).Groups(j).Datasets) 
            dataset_name = info.Groups(i).Groups(j).Datasets(k).Name; 
            dataset_path = fullfile(subgroup_name, dataset_name); 
            % Read the dataset and store it in the corresponding struct 
            data.(matlab.lang.makeValidName(group_name)).(matlab.lang.makeValidName(subgroup_name)).(matlab.lang.makeValidName(dataset_name)) = h5read(filename, dataset_path); 
        end 
    end 
end 

Read more about “h5info” and “h5read” here:

https://www.mathworks.com/help/matlab/ref/h5info.html

https://www.mathworks.com/help/matlab/ref/h5read.html