Justus Magin

I don't think there's support for this right now (so you'd have to call your own function to format), but this would be a use case for the custom quantity...

thanks, @jthielen, those are some good points. I've tried to address them in the most recent commits, though I suspect to be really useful we'd need to rewrite / restructure...

> Seems like some tests with xarray are making CI fail. Do we know what's triggering this ? It seems `xarray` is using `zeros_like` in `isnull` to construct a all-`False`...

another option would be to avoid wrapping `bool` dtypes, but that might make the code much more complicated

this has become tricky... see pydata/xarray#7395 for more details. The most recent commit fixes a bug in `full_like` when called on a `pint` quantity wrapping a `dask` array (`dask` chose...

As mentioned above, we've been discussing this a bit in the `pint-xarray` issue above. In short, `xarray` uses a few `numpy` protocols and the array properties `dtype`, `shape`, and `ndim`...

`shapely` 2.0 apparently will be exactly that: `numpy` arrays of geometries with spatial operations as `numpy` ufuncs (though those arrays will still have the `object` dtype). See also [this discussion...

I believe the reason is the `fill_value`. At the moment, `float*` is one of the few data types that can have missing values (using `nan`), while `int*` can't represent missing...

`zarr`'s `fill_value` is translated to the `_FillValue` attribute. The masking is applied without checking the actual values (which is potentially expensive) using `where`, and the mask value and the promoted...

you can work around this by manually opening and json-decoding the files, and then passing the resulting list of dict to `MultiZarrToZarr`: ```python fs, _ = fsspec.core.url_to_fs(urls[0], **target_options) refs =...