Access multiple datasets within a h5 group simultaneously

within my root group of an hdf5 file I have several data sets (4 dimensional) of equal shape and size (1 channel images to be precise). Now I want to access the elements at index x for every of those data sets and stack the respective 3-dimensional data onto each other in a single tensor, resulting in a data layout

#Datasets x 1 x H x W

What is the best and most efficient way of doing so with h5py?

An explicit loop over all the data sets seems infeasible…