Deepak Cherian

What do we do for group `0` when a user says `grouped_scan(np.array([1, 2, 3], by=[0, 1, 2], expected_groups=[1, 2])`?

After #442 we will have some support for reducing `sparse.COO` arrays by a dense array. - [ ] Support for reducing GCXS arrays - [ ] support for datetimes (would...

https://github.com/xarray-contrib/flox/blob/b20545b3bd6856529107496b3793134ada57f2c1/flox/core.py#L1949-L1961 Many copies of `cohort_index` are embedded in the graph. Instead, we should just embed it under its own key, and then add references to the graph layer.

The current one could be "Climate" since it show cases pretty typical stuff for "Meteorology/Oceanography/etc." From the SciPy tutorial, it looks like @ianhi is already 50% of the way there...

I'm not sure it does so currently. It's an important concept, so we should cover it. https://tutorial.xarray.dev/fundamentals/02.3_aligning_data_objects.html#alignment-putting-data-on-the-same-grid