From: Andre' Walker-L. <wal...@gm...> - 2012-10-06 17:42:49
|
Hi All, I have a bunch of hdf5 files I am using to create one hdf5 file. Each individual file has many different pieces of data, and they are all the same shape in each file. I am using createEArray to make the large array in the final file. if the data files in the individual h5 files are of shape (256,1,2), then I have to use createEArray('/path/','name',tables.floatAtom64(),(0,256,1,2),expectedrows=len(data_files)) if the np array I have grabbed from an individual file to append to my EArray is defined as data, is there a way to use data.shape to create the shape of my EArray? In spirit, I want to do something like (0,data.shape) but this does not work. I have been scouring the numpy manual to see how to convert data.shape (256,1,2) to (0,256,1,2) but failed to figure this out (if I don't know ahead of time the shape of data - in which case I can manually reshape). Thanks, Andre |