BUG operations involving MaskedArray with output given do not return output #8416
Conversation
|
I turned this into a PR to fix the issue that |
| @@ -2988,8 +2988,11 @@ def __array_wrap__(self, obj, context=None): | |||
| Wraps the numpy array and sets the mask according to context. | |||
|
|
|||
| """ | |||
| result = obj.view(type(self)) | |||
| result._update_from(self) | |||
| if obj is self: # for in-place operations | |||
There was a problem hiding this comment.
Note: probably this is better done in __array_prepare__, but given that we hopefully soon have __array_func__ to work with, I felt it not quite worth it to try to define that...
numpy/ma/core.py
Outdated
| result = result.copy() | ||
| if result is not self: | ||
| # MHvK: Not sure why this copy is wanted. | ||
| result = result.copy() |
There was a problem hiding this comment.
I just spent a little bit of effort trying to figure out if this copy had any purpose, but I can't see any. It's not related to the copying of masks discussed in #5580, for example.
Want to just remove it?
Without this commit, after an operation like ``` a = np.ma.MaskedArray([1., 2., 3.]) b = np.add(a, 1., out=a) ``` one would have `b is not a`. With this PR, it is guaranteed that `b is a`.
mhvk
force-pushed
the
ma/ensure-given-output-is-returned
branch
from
6640387
to
7927b3c
Compare
|
@ahaldane - OK, I agree it made little sence so just removed the strange copy (and rebased). Hopefully, tests will still pass. |
|
Looks good to me otherwise, merging. Thanks @mhvk |
unnecessary since numpygh-8416
EDIT -- turned into a PR.
Fixes an inconsistency found in #8414:
Here,
d is bshould beTrue. That it isFalses suggested__array_wrap__is not implemented quite correctly, probably missing a special path forif self is obj-- this PR fixes that.