Repository navigation
Slicing GroupsHDU #6688
Description
Discovered in #6664 it seems like when creating a GroupsHDU from scratch one cannot slice it:
import numpy as np
imdata = np.arange(150.0).reshape(10, 1, 1, 3, 5)
pdata1 = np.arange(10) + 0.1
pdata2 = 42
x = fits.GroupData(imdata, bitpix=-32,
parnames=['abc', 'xyz', 'xyz'],
pardata=[pdata1, pdata2, pdata2])
hdu = fits.GroupsHDU(x)
hdu.data[:2]
throws a TypeError:
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
9 hdu = fits.GroupsHDU(x)
10 hdu.writeto(tempfile.TemporaryFile(mode='wb'))
---> 11 hdu.data[:2]
\astropy\io\fits\hdu\groups.py in __getitem__(self, key)
210
211 def __getitem__(self, key):
--> 212 out = super().__getitem__(key)
213 if isinstance(out, GroupData):
214 out.parnames = self.parnames
\astropy\io\fits\fitsrec.py in __getitem__(self, key)
494 # Store the new arrays for the _coldefs object
495 #
--> 496 arrays.append(self._coldefs._arrays[idx][key])
497
498 # Ensure that the sliced FITS_rec will view the same scaled
TypeError: 'NoneType' object is not subscriptable
But it works if one writes it to a file and then slices the read file:
hdu.writeto('tmp_groups.fits')
with fits.open('tmp_groups.fits') as hdul:
hdul[0].data[:2]
Activity
pllim commentedon Oct 9, 2017
Looks like a bug to me. Also, when I append
hduto aHDUListwithout writing to a file, theinfo()is different from the one read in from a file:saimn commentedon Dec 12, 2017
Yep, looks real, the code probably needs to be updated to cope with the
ColDefs->_arraysreferences and follow what has been done forhdu/table.py.