How can I create a Hdf5 dataset file from several h5 files?

I want to create a Hdf5 dataset by compressing more than hundred h5 files. How can I do that?

[per isakson]

I'm not sure I understand the meaning of "Hdf5 dataset".

See Work with HDF5 Virtual Datasets (VDS)

[Gulfam Saju]

I mean, I want to create a dataset, where the format is hdf5 by compressing several h5 files.

Answer 1

[Ayush Modi]

Hi Gulfam,

As per my understanding, you want to merge multiple h5 files into a single hdf5 file and apply a compression on the hdf5 file. You can achieve this using the functions “h5read”, “h5create” and “h5write”. Here is an example to demonstrate how you can accomplish this:

% List of .h5 files to be combined
h5_files = {'sample1.h5', 'sample2.h5'}; % Add your file names here

% Name of the new combined HDF5 file
combined_hdf5_file = 'combined_data3.hdf5';

% Dataset name within the .h5 files that you want to combine
dataset_name = '/myDataset1';

% Initialize a variable to hold all the data
all_data = [];

% Loop through each .h5 file
for i = 1:length(h5_files)
    % Read the dataset from the .h5 file
    data = h5read(h5_files{i}, dataset_name);
    
    % Concatenate the data
    all_data = cat(1, all_data, data); % Adjust the dimension as necessary
end

% Create the new .hdf5 file and write the combined dataset
h5create(combined_hdf5_file, dataset_name, size(all_data), 'ChunkSize',[size(all_data)], 'Deflate',9);
h5write(combined_hdf5_file, dataset_name, all_data);

Please refer to the following MathWorks documentation for more information on “h5create” function:

https://www.mathworks.com/help/matlab/ref/h5create.html#d126e644037

I hope this resolves the issue you were facing.