Surprising `.groupby` behavior with float index

[max-sixty]

What is your issue?

We raise an error on grouping without supplying dims, but not for float indexes — is this intentional or an oversight?

This is without flox installed

da =  xr.tutorial.open_dataset("air_temperature")['air']

da.drop_vars('lat').groupby('lat').sum()

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[8], line 1
----> 1 da.drop_vars('lat').groupby('lat').sum()
...
ValueError: cannot reduce over dimensions ['lat']. expected either '...' to reduce over all dimensions or one or more of ('time', 'lon').

But with a float index, we don't raise:

da.groupby('lat').sum()

...returns the original array:

Out[15]:
<xarray.DataArray 'air' (time: 2920, lat: 25, lon: 53)>
array([[[296.29   , 296.79   , 297.1    , ..., 296.9    , 296.79   ,
         296.6    ],
        [295.9    , 296.19998, 296.79   , ..., 295.9    , 295.9    ,
         295.19998],
        [296.6    , 296.19998, 296.4    , ..., 295.4    , 295.1    ,
         294.69998],
...

And if we try this with a non-float index, we get the error again:

da.groupby('time').sum()

ValueError: cannot reduce over dimensions ['time']. expected either '...' to reduce over all dimensions or one or more of ('lat', 'lon').