Indexing via `sel`/`loc` fails on multiindex float coordinate

[ThorstenGroh]

What happened?

When I generate a stacked multiindex coordinate dataset

ds = xr.Dataset(
    data_vars={"d": (["x", "y"], np.random.randn(2, 3))},
    coords={"x": [3, 4], "y": [0.5, 1.2, 2.0]},
)
ds2 = ds.stack({"z": ["x", "y"]})

indexing the unstacked version of the data ds with both an integer as well as a float value

ds.sel(x=3)
ds.sel(y=1.2)

works without a problem. But if I try to index the stacked version ds2

ds2.sel(x=3)
ds2.sel(y=1.2)

the integer indexing works, but the float indexing ds2.sel(y=1.2) fails with an InvalidIndexError.

The pandas version works without any problem:

df = ds2["d"].to_pandas()
df[:, 1.2]

The problems seems to lie in xarray.core.utils.py in the maybe_cast_to_coords_dtype function

def maybe_cast_to_coords_dtype(label, coords_dtype):
    if coords_dtype.kind == "f" and not isinstance(label, slice):
        label = np.asarray(label, dtype=coords_dtype)
    return label

that explicitly calls np.asarray on the coordinate label. This then later fails in pandas.core.indexes via the _check_indexing_error method, where it is checked that the label is a scalar, which it was before maybe_cast_to_coords_dtype converted it. Why does this conversion happen there at all?

What did you expect to happen?

Expecting ds2.sel(x=3) to return:

<xarray.Dataset>
Dimensions:  (x: 2)
Coordinates:
  * x        (x) int64 3 4
Data variables:
    d        (x) int64 1 4

Minimal Complete Verifiable Example

import xarray as xr
ds = xr.Dataset(
    data_vars={"d": (["x", "y"], [[0, 1, 2], [3, 4, 5]])},
    coords={"x": [3, 4], "y": [0.5, 1.2, 2.0]},
)
ds2 = ds.stack({"z": ["x", "y"]})
ds2.sel(y=1.2)

MVCE confirmation

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

Relevant log output

Traceback (most recent call last):

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/base.py", line 3621, in get_loc
    return self._engine.get_loc(casted_key)

  File "pandas/_libs/index.pyx", line 136, in pandas._libs.index.IndexEngine.get_loc

  File "pandas/_libs/index.pyx", line 142, in pandas._libs.index.IndexEngine.get_loc

TypeError: '1.2' is an invalid key


During handling of the above exception, another exception occurred:

Traceback (most recent call last):

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/spyder_kernels/py3compat.py", line 356, in compat_exec
    exec(code, globals, locals)

  File "~/data_import.py", line 65, in <module>
    xr.Dataset(

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/xarray/core/dataset.py", line 2501, in sel
    pos_indexers, new_indexes = remap_label_indexers(

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/xarray/core/coordinates.py", line 421, in remap_label_indexers
    pos_indexers, new_indexes = indexing.remap_label_indexers(

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/xarray/core/indexing.py", line 121, in remap_label_indexers
    idxr, new_idx = index.query(labels, method=method, tolerance=tolerance)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/xarray/core/indexes.py", line 353, in query
    indexer, new_index = self.index.get_loc_level(

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/multi.py", line 2972, in get_loc_level
    loc, mi = self._get_loc_level(key, level=level)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/multi.py", line 3005, in _get_loc_level
    loc, new_index = self._get_loc_level(k, level=lev)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/multi.py", line 3113, in _get_loc_level
    indexer = self._get_level_indexer(key, level=level)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/multi.py", line 3222, in _get_level_indexer
    idx = self._get_loc_single_level_index(level_index, key)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/multi.py", line 2802, in _get_loc_single_level_index
    return level_index.get_loc(key)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/base.py", line 3628, in get_loc
    self._check_indexing_error(key)

  File "~/.miniconda3/envs/artiq/lib/python3.8/site-packages/pandas/core/indexes/base.py", line 5637, in _check_indexing_error
    raise InvalidIndexError(key)

InvalidIndexError: 1.2

Anything else we need to know?

No response

Environment

INSTALLED VERSIONS

commit: None
python: 3.8.13 | packaged by conda-forge | (default, Mar 25 2022, 06:04:10)
[GCC 10.3.0]
python-bits: 64
OS: Linux
OS-release: 5.17.5-76051705-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: ('en_US', 'UTF-8')
libhdf5: 1.12.1
libnetcdf: 4.8.1

xarray: 2022.3.0
pandas: 1.4.3
numpy: 1.23.1
scipy: 1.8.1
netCDF4: 1.6.0
pydap: None
h5netcdf: 1.0.1
h5py: 3.7.0
Nio: None
zarr: 2.12.0
cftime: 1.6.1
nc_time_axis: None
PseudoNetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: None
dask: None
distributed: None
matplotlib: 3.5.2
cartopy: None
seaborn: None
numbagg: None
fsspec: None
cupy: None
pint: None
sparse: None
setuptools: 63.2.0
pip: 22.1.2
conda: None
pytest: None
IPython: 7.33.0
sphinx: 5.0.2
/home/thorsten/.miniconda3/envs/artiq/lib/python3.8/site-packages/_distutils_hack/init.py:33: UserWarning: Setuptools is replacing distutils.
warnings.warn("Setuptools is replacing distutils.")

[ThorstenGroh]

The same holds for indexing via loc.

ds2.loc[{"y": 1.2}]

returns the same InvalidIndexError error message, while ds2.loc[{"x": 3}] works.

[benbovy]

Hi @ThorstenGroh, there has been several issues regarding indexes and float values which have been fixed in #5692. Your example runs fine with the last release candidate v2022.06.0rc0. I think we're close to a stable release.

[dcherian]

Closing since this seems to be fixed on main.

[ThorstenGroh]

Hi @ThorstenGroh, there has been several issues regarding indexes and float values which have been fixed in #5692. Your example runs fine with the last release candidate v2022.06.0rc0. I think we're close to a stable release.

Thanks. I missed that.