Should a cube be considered equal to itself?

[bouweandela]

✨ Feature Request

Motivation

At the moment, a cube where the data contains a NaN is not considered equal to itself or to another cube with the same data. Conversely, coordinates/cell measures/ancillary variables with NaNs are equal to themselves. Maybe it would be nicer, and more consistent, to introduce the same behaviour for cubes.

See #5691 (comment) for some previous discussion.

[trexfeathers]

Xref: #4681

[pp-mo]

@SciTools/peloton discussed this.
General feeling is that, although a breaking change, it's hard to see how making cubes with NaNs equate could cause anyone a real problem. So, I think we can support this.
@bouweandela thanks a lot for raising this -- would you like to propose an actual PR to change ?

[bouweandela]

Thanks for getting back on this. I would be happy to make a PR.