Coord name not set when `concat`ing along a DataArray

[darikg]

from xarray import DataArray, concat
a = DataArray([0], dims='a')
out = concat([a, a], dim=DataArray([0, 1], dims='b'))
print(out.coords)

Coordinates:

None (b) int32 0 1

I would've expected the name of the new coordinate to be the name of the DataArray variable but instead it's None.

Environment:

Output of xr.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.9.2 | packaged by conda-forge | (default, Feb 21 2021, 04:59:43) [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: en_US.UTF-8
LOCALE: English_United States.1252
libhdf5: None
libnetcdf: None

xarray: 0.17.0
pandas: 1.2.4
numpy: 1.20.2
scipy: None
netCDF4: None
pydap: None
h5netcdf: None
h5py: None
Nio: None
zarr: None
cftime: None
nc_time_axis: None
PseudoNetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: None
dask: None
distributed: None
matplotlib: None
cartopy: None
seaborn: None
numbagg: None
pint: None
setuptools: 49.6.0.post20210108
pip: 21.1.1
conda: None
pytest: None
IPython: 7.23.0
sphinx: None

[keewis]

would've expected the name of the new coordinate to be the name of the DataArray variable

that's exactly what it's doing: it uses the name of DataArray([0, 1], dims='b'), which is None. To fix that, you can assign a name to the DataArray:

In [3]: a = xr.DataArray([0], dims='a')
   ...: out = xr.concat([a, a], dim=xr.DataArray([0, 1], dims='b', name="x"))
   ...: out
Out[3]: 
<xarray.DataArray (b: 2, a: 1)>
array([[0],
       [0]])
Coordinates:
    x        (b) int64 0 1
Dimensions without coordinates: b, a

Since None is not a particularly good name, we could raise for unnamed DataArray objects (not sure if falling back to dims for 1D DataArray objects would be a good idea)

[darikg]

Thanks @keewis! I was confusing dimension names and variable names. I would support raising or falling back to a reasonably sane default -- the reason I stumbled on this was having a None coord name was breaking the _repr_html in Jupyter and causing a much more confusing error message

[keewis]

The repr will be fixed by #5149, but we should probably still fix that. Since dim is the dimension to concatenate along I guess we should fall back to that (not sure though). Thoughts, @pydata/xarray?

For reference, the conversion happens in _calc_concat_dim_coord ("else"):

xarray/xarray/core/concat.py

Lines 244 to 265 in 04acabb

	def _calc_concat_dim_coord(dim):
	"""
	Infer the dimension name and 1d coordinate variable (if appropriate)
	for concatenating along the new dimension.
	"""
	from .dataarray import DataArray
	if isinstance(dim, str):
	coord = None
	elif not isinstance(dim, (DataArray, Variable)):
	dim_name = getattr(dim, "name", None)
	if dim_name is None:
	dim_name = "concat_dim"
	coord = IndexVariable(dim_name, dim)
	dim = dim_name
	elif not isinstance(dim, DataArray):
	coord = as_variable(dim).to_index_variable()
	(dim,) = coord.dims
	else:
	coord = dim
	(dim,) = coord.dims
	return dim, coord

[max-sixty]

Agree that this should coerce to the name of the dim if the array has no name.

Technically should the kwarg of concat be coord rather than dim? I generally supply dim names there, rather than coords vars. (though I'm not suggesting we actually change it given the backward-compat).

Note that we already do this in the DataArray constructor:

In [12]: a = xr.DataArray([0,1], coords=[xr.DataArray([1,2])], dims='b')   # No name on the coord

In [13]: a
Out[13]:
<xarray.DataArray (b: 2)>
array([0, 1])
Coordinates:
  * b        (b) int64 1 2    # Yes name on the coord