Understand python hdf5storage functions

Question

Is there anyone explain exact difference between hdf5storage.write() and hdf5storage.writes() functions. I've read documents but I did not understand it.

Answer 1

This is explained in the hdf5storage docs. Quoting:

• The main functions in this module are write() and read() which write a single Python variable to an HDF5 file (or reads and returns the read data).
• Version 0.1.10 added added two new functions: writes() and reads().
• They write and read more than one Python variable at once, although they can still work with a single variable.
• Calling write() opens and closes the HDF5 file. As a result, calling write() multiple times for multiple variables incurs a performance penalty. This is most noticeable with large HDF5 files.
• In addition, savemat() and loadmat() (to work with MATLAB data) now use writes() and reads() for improved performance.

Complete documentation is here for write() and writes().

For a Python variable named 'a', a simple write() call looks like:

hdf5storage.write(a, path='/a', filename='data.h5') 

The writes() call uses a dictionary where keys are HDF5 paths and values are data to write to the file. For a dictionary name mdict, the call looks like this:

hdf5storage.writes(mdict, filename='data.h5') 

Examples of each below for 3 arrays:

arr1 = np.arange(10)
arr2 = np.arange(10,20).reshape(5,2)
arr3 = np.arange(20,30).reshape(2,5)
hdf5storage.write(arr1, path='/arr1', filename='write_data.h5')
hdf5storage.write(arr2, path='/arr2', filename='write_data.h5')
hdf5storage.write(arr3, path='/arr3', filename='write_data.h5')

mdict = {'/arr1':arr1, '/arr2':arr2, '/arr3':arr3}
hdf5storage.writes(mdict, filename='writes_data.h5')

The resulting files should be the same.