Wrapper for h5py allowing lazy indexing and straightfoward discovery of the keys.
This is useful for exploring files in the usual case where the number of different datasets isn't too large. This is especially convenient when the h5 file contains links, which aren't fully explored by the default h5py.visit and h5py.visititems methods.
from h5file import h5file
h = hfile('linked_file.h5')
Calling h or print(h) will show the files and keys.
Indexing returns datasets including partial matches string (provided they only match one key). Do
hdfdataset = h['data1']
or
data = h['data1'][()]
Within IPython, you can use tab completion to explore the keys using an expression like
h['dat+TAB
To get the data. The search function supports regular expressions.
The usual methods of
h.keys()
and
h.items()
return lists of keys and lists of (key, dataset) pairs, but one can additionally filter these by passing a getkey argument to the function as
h.keys(getkey='data1').