Ticket #1312 (new wish)

Opened 16 months ago

Last modified 16 months ago

Add HDF5 file support to Orange

Reported by: echlebek Owned by:
Milestone: Future Component: library
Severity: minor Keywords:
Cc: anze Blocking:
Blocked By:


 HDF5 is used by many organizations for storing large datasets. Adding HDF5 support to Orange can be accomplished through the use of the h5py library, which reads HDF5 datasets into numpy arrays.

1D and 2D datasets can easily map to Orange.data.Table objects, so supporting those datasets is a good starting point.

Change History

comment:1 Changed 16 months ago by echlebek

I have a working prototype of a widget that allows a user to select an HDF5 dataset from a tree of objects. After some further testing I'll push it to my fork of Orange. It would be nice if someone from Orange could review it and consider it for merging at some point.

Is it reasonable to add  h5py as an Orange dependency? My company has been using it for several years to store data and we have found it to be of very high quality.

comment:2 Changed 16 months ago by anze

  • Cc anze added

I can take a look, just post the link to your fork when the code is ready.

We have been trying lately to limit the number of external dependencies and moving functionality to add-ons. If your changes do not substantially modify existing orange objects, they should be able to function just fine as an add-on. We can discuss this further after I get a chance to see the code.

comment:3 Changed 16 months ago by echlebek

The code is self-contained, I would imagine it should function just fine as an addon. I haven't looked at your documentation for creating addons, but I'd imagine it shouldn't be too difficult for this task.

Thanks for taking a look at the code. You can find it  here.

Note: See TracTickets for help on using tickets.