Groupby does not allow grouping of lat-lon pair

[alterapars]

What happened:
xarray returns error message when grouping on latitude/longitude pairs

What you expected to happen:
Seamless execution of the groupby method providing a latitude/longitude pair

Minimal Complete Verifiable Example:

import numpy as np
import xarray as xr

xr_dataframe = xr.tutorial.load_dataset("air_temperature")

xr_dataframe.air.groupby('lon', 'lat')

returns the following error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-33-e32cb3023749> in <module>
----> 1 ds.air.groupby('lon', 'lat')
      2 

~\AppData\Local\Continuum\anaconda3\envs\xarray\lib\site-packages\xarray\core\common.py in groupby(self, group, squeeze, restore_coord_dims)
    690         # backward-compat issues.
    691         if not isinstance(squeeze, bool):
--> 692             raise TypeError(
    693                 f"`squeeze` must be True or False, but {squeeze} was supplied"
    694             )

TypeError: `squeeze` must be True or False, but lat was supplied

if I state squeeze explicitely:

xr_dataframe.air.groupby('lon', 'lat', squeeze=False).sum()

it returns the error message:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-27-988979b87a81> in <module>
----> 1 MODIS_burned_area.burned_area.groupby('lon', 'lat', squeeze=False).sum()

TypeError: groupby() got multiple values for argument 'squeeze'

handing over an array similar to the pd notation:

xr_dataframe.groupby(['lon', 'lat']).sum()

returns the following error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-37-4ae9ac839e02> in <module>
----> 1 ds.groupby(['lon', 'lat'], squeeze=False).sum()

~\AppData\Local\Continuum\anaconda3\envs\xarray\lib\site-packages\xarray\core\common.py in groupby(self, group, squeeze, restore_coord_dims)
    694             )
    695 
--> 696         return self._groupby_cls(
    697             self, group, squeeze=squeeze, restore_coord_dims=restore_coord_dims
    698         )

~\AppData\Local\Continuum\anaconda3\envs\xarray\lib\site-packages\xarray\core\groupby.py in __init__(self, obj, group, squeeze, grouper, bins, restore_coord_dims, cut_kwargs)
    308         if not isinstance(group, (DataArray, IndexVariable)):
    309             if not hashable(group):
--> 310                 raise TypeError(
    311                     "`group` must be an xarray.DataArray or the "
    312                     "name of an xarray variable or dimension."

TypeError: `group` must be an xarray.DataArray or the name of an xarray variable or dimension.Received ['lon', 'lat'] instead.

Describe alternatives you've considered:

I chose to use the pandas groupby method in the meantime which works with lat-lon pairs.

Environment:

INSTALLED VERSIONS ------------------ commit: None python: 3.9.1 | packaged by conda-forge | (default, Dec 21 2020, 22:05:45) [MSC v.1916 64 bit (AMD64)] python-bits: 64 OS: Windows OS-release: 10 machine: AMD64 processor: Intel64 Family 6 Model 142 Stepping 12, GenuineIntel byteorder: little LC_ALL: None LANG: None LOCALE: English_United States.1252 libhdf5: 1.10.6 libnetcdf: 4.7.3

xarray: 0.16.2
pandas: 1.1.5
numpy: 1.19.4
scipy: 1.5.3
netCDF4: 1.5.5.1
pydap: None
h5netcdf: None
h5py: 3.1.0
Nio: None
zarr: None
cftime: 1.3.0
nc_time_axis: None
PseudoNetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: 1.3.2
dask: 2020.12.0
distributed: 2020.12.0
matplotlib: 3.3.3
cartopy: None
seaborn: 0.11.1
numbagg: None
pint: None
setuptools: 49.6.0.post20201009
pip: 20.3.3
conda: None
pytest: None
IPython: 7.19.0
sphinx: None

[mathause]

Yes, this is a long-standing issue, #1056 What may work is to use groupby with a MuliIndex (da.stack(latlon=("lat", "lon")).groupby("latlon"))

However, what is your use case? Because every lat/ lon pair is distinct, so it wouldn't actually do anything (it would sum over the only data value there is in each group). Unless I am missing something, which I probably do.

[alterapars]

Thank you @mathause , I see. My use cases would be

• time series analysis at each location
• visualization of e.g. standard deviation over time for each grid cell on my regional dataset

[mathause]

You should be able to do this without groupby. If you have a dataset with dimensions time x lat x lon you can calculte the mean over time at each location using xr_dataframe.air.mean("time") or if you want the mean at each time you can use xr_dataframe.air.mean(("lon", "lat")). The visualisation should work via xr_dataframe.air.std("time").plot().

[mathause]

I close this as duplicate of #1056 - feel free to re-open if you have more questions on this.