BUG: Masked singleton can be reshaped to be non-scalar #10292
Conversation
|
This is less good a fix than intended, and still doesn't stop assigning the shape of |
numpy/ma/core.py
Outdated
| def __setattr__(self, attr, value): | ||
| if self is self.__singleton: | ||
| raise MaskError("Cannot change attributes of the masked singleton") | ||
| super(MaskedConstant, self).__setattr__(attr, value) |
There was a problem hiding this comment.
Is this functionality needed? Maybe a note.
There was a problem hiding this comment.
It stops it from breaking subclasses, if anyone else is brave enough to do that (@mhvk?)
There was a problem hiding this comment.
No, while we do subclass MaskedArray, even we don't subclass this (as far as I know).
There was a problem hiding this comment.
Update: yes, this is needed in order to reassign __class__ of duplicate instances in __array_finalize__
It's possible this is the cause of numpygh-10270, where it seems something is wrong with the shape
51ca981 to
8b1b7f1
Compare
| @@ -4981,6 +4981,10 @@ class Sub(type(np.ma.masked)): pass | |||
| assert_(a is not np.ma.masked) | |||
| assert_not_equal(repr(a), 'masked') | |||
|
|
|||
| def test_attributes_readonly(self): | |||
| assert_raises(AttributeError, setattr, np.ma.masked, 'shape', (1,)) | |||
| assert_raises(AttributeError, setattr, np.ma.masked, 'dtype', np.int64) | |||
There was a problem hiding this comment.
Previous commit had MaskError, but this seems more useful for code that does:
try:
x.shape = (1, 2)
except AttributeError:
x = array(x)
x.shape = (1, 2)
| @@ -6309,6 +6309,18 @@ def copy(self, *args, **kwargs): | |||
| # precedent for this with `np.bool_` scalars. | |||
| return self | |||
|
|
|||
| def __setattr__(self, attr, value): | |||
| if not self.__has_singleton(): | |||
There was a problem hiding this comment.
Can we combine the two to if self.__has_singleton() and self is self.__singleton? would imply longer comment but no duplicated code.
There was a problem hiding this comment.
I wrote it this way because the rationale for the two cases is different (even if it leads to the same result), and so the code duplication is perhaps justifiable - it may be desirable to only allow the __class__ attribute to be set in the last branch, for instance.
|
Thanks Eric. |
It's possible this is the cause of gh-10270, where it seems something is wrong with the shape