Repository navigation
MaskArray.take does not work with scalar indices #7585
Description
Python 3.5, numpy 1.10.1
>>> m = np.ma.array([1, 2, 3, 4], mask=[True, 0, 0, 0])
>>> m[0]
masked
>>> m.take(0) #This is wrong
1
>>> m[[0,1]]
masked_array(data = [-- 2],
mask = [ True False],
fill_value = 999999)
>>> m.take([0, 1])
masked_array(data = [-- 2],
mask = [ True False],
fill_value = 999999)
Activity
[-]MaskArray.take does not work[/-][+]MaskArray.take does not work on scalars[/+][-]MaskArray.take does not work on scalars[/-][+]MaskArray.take does not work with scalar indices[/+]charris commentedon Apr 28, 2016
Interesting. @ahaldane Thoughts?
eric-wieser commentedon Apr 28, 2016
Fixed by adding a
np.asarrayand a[...], I thinkeric-wieser commentedon Apr 28, 2016
The underlying issue is that
np.int32(1).view(MaskedArray)does not return aMaskedArray!eric-wieser commentedon Apr 28, 2016
Workaround:
m.take(keys[np.newaxis])[0]in place ofm.take(keys)TST: Verify np.ma.take works on scalars
BUG: Ensure mask is preserved on scalars
charris commentedon May 2, 2016
Closed by #7586.
TST: Verify np.ma.take works on scalars
BUG: Ensure mask is preserved on scalars