What is your issue?

I'm trying to bundle some code using Xarray with PyInstaller.

As a MRE, consider the following file:

```python

test.py

import pandas as pd import xarray as xr print(xr.version) pd.DataFrame({'a': range(10)}).to_xarray() and the following command-line to build and run the Pyinstaller executable:shell pyinstaller test.py && dist/test/test ```

Ideally, I would have expected the printed version to be 2023.8.0 (as when running the script without Pyinstaller). The fallback 9999 from https://github.com/pydata/xarray/blob/b08a9d868febc0c6aa0944d654ec92d833a70c37/xarray/init.py#L55 would also have done the job.

Unfortunately, the actual version number is '999' which make the following pandas code to fail with: Traceback (most recent call last): File "test.py", line 4, in <module> pd.DataFrame({'a': range(10)}).to_xarray() File "pandas/core/generic.py", line 3243, in to_xarray File "pandas/compat/_optional.py", line 161, in import_optional_dependency ImportError: Pandas requires version '2022.03.0' or newer of 'xarray' (version '999' currently installed).

As a workaround, I added the line python xr.__version__ = "9999" in my test.py file...

Let me consider a DataArray with a dimension containing a single element:

```python In [1]: import numpy as np

In [2]: import xarray as xr

In [3]: da = xr.DataArray(np.linspace(0.0, 100.0, 101).reshape(1, 101), coords=[('x', [1.0]), ('y', np.linspace(0.0, 1.0, 101))]) ```

Asking x=1.0 in interp fails.

More precisely, the 2D interpolation returns a NaN python In [5]: da.interp(x=1.0, y=0.4) /opt/anaconda3/lib/python3.7/site-packages/scipy/interpolate/interpolate.py:2533: RuntimeWarning: invalid value encountered in true_divide (grid[i + 1] - grid[i])) Out[5]: <xarray.DataArray ()> array(nan) Coordinates: x float64 1.0 y float64 0.4 and the 1D interpolation returns a ValueError: ```python In [6]: da.interp(x=1.0)

ValueError Traceback (most recent call last) <ipython-input-6-efd4490bc97e> in <module> ----> 1 da.interp(x=1.0) ... ValueError: x and y arrays must have at least 2 entries ```

I understand why the interpolation is impossible in an array with a single element.

However, this behavior breaks my assumption that interp is a superset of sel, meaning that it gives the same result as sel for all inputs that are valid inputs for sel (except for a possible int to float type conversion).

Here, I would expect da.interp(x=1.0) to give me the same result as da.sel(x=1.0).

What do you think of this corner case?

I have two DataArrays with the same dimension, but the index is in a different order. Adding them with A + B works fine, but the in-place addition fails.

MCVE Code Sample

```python import numpy as np import xarray as xr

n = 5

d1 = np.arange(n) np.random.shuffle(d1) A = xr.DataArray(np.ones(n), coords=[('dim', d1)])

d2 = np.arange(n) np.random.shuffle(d2) B = xr.DataArray(np.ones(n), coords=[('dim', d2)])

print(A + B) A += B ```

Expected Output

A = A + B is working fine. I would expect A += B to do the same.

Problem Description

The in-place addition A += B fails: Traceback (most recent call last): File "/home/matthieu/xarray-test.py", line 15, in <module> A += B File "/opt/anaconda3/envs/xarray-tests/lib/python3.8/site-packages/xarray/core/dataarray.py", line 2619, in func with self.coords._merge_inplace(other_coords): File "/opt/anaconda3/envs/xarray-tests/lib/python3.8/contextlib.py", line 113, in __enter__ return next(self.gen) File "/opt/anaconda3/envs/xarray-tests/lib/python3.8/site-packages/xarray/core/coordinates.py", line 140, in _merge_inplace variables, indexes = merge_coordinates_without_align( File "/opt/anaconda3/envs/xarray-tests/lib/python3.8/site-packages/xarray/core/merge.py", line 328, in merge_coordinates_withou t_align return merge_collected(filtered, prioritized) File "/opt/anaconda3/envs/xarray-tests/lib/python3.8/site-packages/xarray/core/merge.py", line 210, in merge_collected raise MergeError( xarray.core.merge.MergeError: conflicting values for index 'dim' on objects to be combined: first value: Int64Index([1, 2, 0, 3, 4], dtype='int64', name='dim') second value: Int64Index([4, 2, 3, 1, 0], dtype='int64', name='dim')

Versions

Trying to make error message slightly more informative when the user might expect automatic alignment with in-place operation.

• [X] Closes #3910
• [ ] Tests added
• [X] Passes isort -rc . && black . && mypy . && flake8
• [ ] Fully documented, including whats-new.rst for all changes and api.rst for new API

Building a dataset from pandas with a multi-index with categorical values: ```python import pandas as pd

cat = pd.CategoricalDtype(categories=['foo', 'bar', 'baz']) i1 = pd.Series(['foo', 'bar'], dtype=cat) i2 = pd.Series(['bar', 'bar'], dtype=cat)

df = pd.DataFrame({'i1': i1, 'i2': i2, 'values': [1, 2]}) ds = df.set_index(['i1', 'i2']).to_xarray()

print(ds) ```

Expected output: <xarray.Dataset> Dimensions: (i1: 2, i2: 1) Coordinates: * i1 (i1) object 'foo' 'bar' * i2 (i2) object 'bar' Data variables: values (i1, i2) int64 1 2

Actual output: <xarray.Dataset> Dimensions: (i1: 3, i2: 3) Coordinates: * i1 (i1) object 'foo' 'bar' 'baz' * i2 (i2) object 'foo' 'bar' 'baz' Data variables: values (i1, i2) float64 nan 1.0 nan nan 2.0 nan nan nan nan

It is not wrong, but it is inconsistent with the non-categorical case (which gives the expected output above) and the single-index case (no filling with NaNs for single index).

Output of xr.show_versions()

@fujiisoup Would you mind reviewing this PR? The bug was actually straightforward.

• [X] Closes #3674
• [X] Tests added
• [X] Passes isort -rc . && black . && mypy . && flake8
• [X] Fully documented, including whats-new.rst for all changes and api.rst for new API

Dear xarray team,

Thank you very much for your work on this useful package. Here is a bug I just found in my code.

MCVE Code Sample

Creating a Dataset from pandas when the coordinate is a categorical series: ```python import pandas as pd

ind = pd.Series(['foo', 'bar'], dtype='category') df = pd.DataFrame({'ind': ind, 'values': [1, 2]}) df = df.set_index('ind')

ds = df.to_xarray()

print(ds.sel(ind='foo')) ```

Expected Output

When ind is not categorical, the code returns the expected output: <xarray.Dataset> Dimensions: () Coordinates: ind <U3 'foo' Data variables: values int64 1

Problem Description

When ind is categorical, it fails and gives the following traceback: Traceback (most recent call last): File "/home/matthieu/foo.py", line 9, in <module> print(ds.sel(ind='foo')) File "/opt/anaconda3/envs/issue/lib/python3.8/site-packages/xarray/core/dataset.py", line 2013, in sel pos_indexers, new_indexes = remap_label_indexers( File "/opt/anaconda3/envs/issue/lib/python3.8/site-packages/xarray/core/coordinates.py", line 39 1, in remap_label_indexers pos_indexers, new_indexes = indexing.remap_label_indexers( File "/opt/anaconda3/envs/issue/lib/python3.8/site-packages/xarray/core/indexing.py", line 260, in remap_label_indexers idxr, new_idx = convert_label_indexer(index, label, dim, method, tolerance) File "/opt/anaconda3/envs/issue/lib/python3.8/site-packages/xarray/core/indexing.py", line 179, in convert_label_indexer indexer = index.get_loc( TypeError: get_loc() got an unexpected keyword argument 'tolerance'

Output of xr.show_versions()