On Tue, Oct 16, 2012 at 4:36 PM, Byron Blay <byronblay@googlemail.com> wrote:
Copying / deepcopying an array does not copy the writeable flag:
import numpy from copy import deepcopy a = numpy.array([1,2,3,4,5]) a.flags.writeable = False b = deepcopy(a) b.flags.writeable True
Is this a bug? Should I raise an issue on github?
I think this is intended? numpy doesn't really think of writeable and non-writeable arrays as being different sorts of objects; flags.writeable just records a fact about the underlying memory that the array refers to. If you make a new copy, then (by definition!) you get a nice new chunk of freshly allocated memory, and the memory is writeable. That said, this might not be the most useful approach. Can you elaborate on your situation where this is causing a problem? -n