BUG: Define <, <=, >, >= for masked arrays
#21812
Conversation
|
Arg, no... this should not really be necessary, closing for now. |
|
Oops, I hadn't seen you closed it! Why would this not be necessary? It is true that in principle this could be done inside |
|
Yeah, I will reopen (and remove the release notes, I don't think they are actually necessary). |
This defines the comparison operators (other than `==` and `!=`) explicitly for masked arrays. The mask is ignored for the resuling `res._data` (unlike `==` and `!=` which take the mask into account. Closes numpygh-21770, although the way that masked arrays propagate the fill-value seems generally broken and error prone.
|
I guess my point was that the fill-value propagation seems broken for all ufuncs at least when the result dtype does not match the (first) input. The best solution is probably indeed EDIT: Maybe unclear... This is sufficient to fix all the issues. However, if ufuncs were not broken, this should not be necessary. For one, it leaves |
|
Yes, agreed that fill-value propagation is broken/tricky. But separate from this PR at least. |
<, <=, >, >= for masked arrays<, <=, >, >= for masked arrays
|
@seberg Are you happy with this? |
|
Yes, happy enough andnd it should fix (most) of astropy's current issue. We still need to figure out if there is something to do about gh-21784 or not (but it is distinct). |
|
Thanks Sebastian. |
This defines the comparison operators (other than
==and!=)explicitly for masked arrays.
The mask is ignored for the resuling
res._data(unlike==and!=which take the mask into account.Closes gh-21770, although the way that masked arrays propagate the
fill-value seems generally broken and error prone.