DataArray.isin does not accept sets intuitively

What is your issue?

This is halfway between an feature request and a bug report. I do not find the way that DataArray.isin handles sets is very intuitive:

Example:

np_arr = np.arange(5)
dask_arr = da.arange(5)
xr_arr = xr.DataArray(np_arr)
pd_series = pd.Series(np_arr)

s = {1, 2}
print("numpy", np.isin(np_arr, s))
print("dask", da.isin(dask_arr, s).compute())
print("xarray", xr_arr.isin(s))
print("pandas", pd_series.isin(s))

Which results in:

numpy [False False False False False]
dask [False False False False False]
xarray <xarray.DataArray (dim_0: 5)> Size: 5B
array([False, False, False, False, False])
Dimensions without coordinates: dim_0
pandas 0    False
1     True
2     True
3    False
4    False
dtype: bool

Personally, I find only pandas handles this in an intuitive way. But I also understand, that you might want to stay consistent with numpy and dask.

I'd be interested to see what you think!

Start by reproducing the DataArray.isin example in the issue and compare its set handling with the shown NumPy, Dask, and pandas results. The issue does not name files or tests, and work is not ready until the desired behavior and consistency decision are established.