issue_comments: 706766842

Following the data through combine_by_coords to _infer_concat_order_from_coords the following happens

https://github.com/pydata/xarray/blob/080caf4246fe2f4d6aa0c5dcb65a99b376fa669b/xarray/core/combine.py#L49

```python import pandas as pd

indexes = [da1.indexes["time"], da2.indexes["time"]]

first_items = pd.Index([index[0] for index in indexes]) series = first_items.to_series() series.rank(method="dense", ascending=ascending) which returnspython Series([], dtype: float64) thus it tries to iterate through an empty list. We probably need to setpython series.rank(method="dense", ascending=True, numeric_only=False) which raisespython <array_function internals> in lexsort(args, *kwargs)

cftime/_cftime.pyx in cftime._cftime.datetime.richcmp()

TypeError: cannot compare cftime.DatetimeGregorian(2000, 1, 1, 0, 0, 0, 0) and cftime.DatetimeProlepticGregorian(2001, 1, 1, 0, 0, 0, 0) (different calendars) ``` but I wonder what test failures this brings...