I agree this is annoying, but I don't think your diagnosis is correct here. The groups property isn't used by any internal routines AFAICT. The issue is that groups are sorted, but as text rather than ordered categorical -- notice that the labels are ordered alphabetically.

Hmmm, a mystery. I'll look into this a bit more.

For what it's worth, I don't think we have any good solutions short of
adding our own array type do to Categorical in xarray. We could set
sort=False in some cases when we call pd.factorize but that's not a great
alternative.

Ahh, so it's sorted, instead of keeping the original order.

I was expecting DataArray.groupby().reduce would work like np.apply_along_axis, and used the data of the result directly.

Just stumbled across this issue. Is there a recommended workaround?

I am usually doing this (specific to seasons):

import xarray as xr
ds = xr.tutorial.open_dataset('air_temperature')
airtemp_seasonal = ds.groupby('time.season').mean('time').sortby(xr.DataArray(['DJF','MAM','JJA', 'SON'],dims=['season']))

Thought this might help some folks who need to solve this problem.

I use reindex instead of sortby

Closing as stale, please reopen with an updated summary if relevant

Hehe, I'm so close to actually fixing this! :)