Neil Vaytet
Neil Vaytet
Wow, it even had the `scippy` label at some point, it must be old!
Would simplifying `MultiIndex` mean that we could now implement negative strides more easily?
There is also `1 * sc.Unit(None)` which fails with ``` TypeError: __init__(): incompatible constructor arguments. The following argument types are supported: 1. scipp._scipp.core.Unit(arg0: str) Invoked with: None ``` Maybe that...
It's an interesting way of thining about the issue, but would this mean that the object given by `da.sel(time='s')` would not be useable until you apply some slicing to it?...
Sort of true, but we do use `da.bins` in a number of cases, such as `if da.bins is not None` or `da.bins.coords`?
Not sure what the output would be if you iterate over a 2d variable? So I would also raise for now.
As suggested in #3326, would a `split` method be useful? It could return a `DataGroup`? ```Py dg = da.split(dim='time') ```
I suggested `split` because I think it is quite obvious from the name what it's doing. I am not sure what `to_dataset` and `to_array` are actually doing from their names....
I think we should have an in-person meeting about this. This discussion is going on for too long.
I would be in favor of not allowing this. I don't think it's such a pain to have to convert manually, and I feel more in control of my operations,...