Keep index dimension when selecting only a single coord

MCVE Code Sample

# Your code here
import numpy as np
import xarray as xr

data = np.zeros((10, 4))
example_xr = xr.DataArray(data, coords=[range(10), ["idx0", "idx1", "idx2", "dim3"]], dims=["rows", "cols"])

# desired behavior
subset = example_xr[:, 1:2]
subset.shape

# inclusive indexing means both idx1 and idx2 kept
subset_named1 = example_xr.loc[:, "idx1":"idx2"]
subset_named1.shape

# slicing behavior means that 2nd dimension is dropped
subset_named2 = example_xr.loc[:, "idx1"]
subset_named2.shape

Expected Output

I'd like to be able to use named .loc indexing to select only a single named coord from one dimension, but not have that dimension collapse when subsetting.

Problem Description

I looked, but wasn't able to find anything in the documentation about how to perform this same action using named coords. It works with integer-based slicing.

Output of xr.show_versions()

Run the MCVE with Python, NumPy, and xarray, comparing the shapes of integer slicing and the two named .loc selections. Trace the .loc selection entry point and add regression coverage showing that selecting one named coordinate preserves its dimension, with the expected shape as the completion criterion.