issues

What happened?

After using set_index to create a MultiIndex, calling groupby on a MultiIndex level and then mean raises an error.

What did you expect to happen?

Apply mean to groups, no error.

Minimal Complete Verifiable Example

Python d = DataArray( data=[ [0, 1, 2, 3, 4, 5, 6], [7, 8, 9, 10, 11, 12, 13], [14, 15, 16, 17, 18, 19, 20] ], coords={ "greek": ("a", ['alpha', 'beta', 'gamma']), "colors": ("a", ['red', 'green', 'blue']), "compass": ("b", ['north', 'south', 'east', 'west', 'northeast', 'southeast', 'southwest']), "integer": ("b", [0, 1, 2, 3, 4, 5, 6]), }, dims=("a", "b") ) d = d.set_index(a=['greek', 'colors'], b=['compass', 'integer']) g = d.groupby('greek') m = g.mean(...)

MVCE confirmation

• [X] Minimal example — the example is as focused as reasonably possible to demonstrate the underlying issue in xarray.
• [X] Complete example — the example is self-contained, including all data and the text of any traceback.
• [X] Verifiable example — the example copy & pastes into an IPython prompt or Binder notebook, returning the result.
• [X] New issue — a search of GitHub Issues suggests this is not a duplicate.

Relevant log output

Python Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/local/lib/python3.10/site-packages/xarray/core/_aggregations.py", line 5698, in mean return self.reduce( File "/usr/local/lib/python3.10/site-packages/xarray/core/groupby.py", line 1201, in reduce return self.map(reduce_array, shortcut=shortcut) File "/usr/local/lib/python3.10/site-packages/xarray/core/groupby.py", line 1104, in map return self._combine(applied, shortcut=shortcut) File "/usr/local/lib/python3.10/site-packages/xarray/core/groupby.py", line 1136, in _combine index, index_vars = create_default_index_implicit(coord) File "/usr/local/lib/python3.10/site-packages/xarray/core/indexes.py", line 1045, in create_default_index_implicit index = PandasMultiIndex(array, name) File "/usr/local/lib/python3.10/site-packages/xarray/core/indexes.py", line 615, in __init__ raise ValueError( ValueError: conflicting multi-index level name 'greek' with dimension 'greek'

Anything else we need to know?

No response

Environment

http://xarray.pydata.org/en/stable/groupby.html says:

xarray supports “group by” operations with the same API as pandas

but when I supply the level keyword argument as described at https://pandas.pydata.org/pandas-docs/stable/groupby.html#groupby-with-multiindex, I get:
``` TypeError Traceback (most recent call last) <ipython-input-12-566fc67c0151> in <module>() ----> 1 hvm_it_v6_obj = hvm_it_v6.groupby(level=["category","obj"]).mean(dim="presentation") 2 hvm_it_v6_obj

TypeError: groupby() got an unexpected keyword argument 'level' ```

Code Sample, a copy-pastable example if possible

The problem occurs for me in the midst of a data-processing pipeline that starts with some ~40MB netCDF files. I've tried to create pasteable code that reproduces the behavior from scratch, but I haven't succeeded.

Problem description

I pass two DataArrays to xr.align() with join="outer". The DataArrays are dtype float64, and contain a mix of NaNs and floats. They are 2D and have MultiIndexes with some numeric and some string levels.

The tuple of DataArrays returned by align() have the correct shape and expected indexes, but the contents of the arrays are all NaNs. The original float values are gone. np.nonzero(~np.isnan(da)) returns an empty array.

I've set breakpoints and delved into the code. On line 656 in xarray.core.variable.Variable._getitem_with_mask, self contains non-NaN values, but the data returned by as_indexable(self._data)[actual_indexer] evaluates as all NaNs. However, data.array at that point (which is xarray.backends.netCDF4_.NetCDF4ArrayWrapper) has non-NaNs. So it's some sort of masking caused by the indexing that makes it look like data is all NaNs.

Expected Output

A tuple of DataArrays which contain some non-NaN values.

Output of xr.show_versions()