name of array disappear when apply "where" filter with Datarray but not with np.array

[svicario]

I was trying to select by non dimensional coordinate, and I found that "where" works well, but it erase the name of my datarray only if filtering is done with another datarray as a non dimensional coordinate

import xarray as xr
temp = 15 + 8 * np.random.randn(2, 2, 3)

precip = 10 * np.random.rand(2, 2, 3)

lon = [[-99.83, -99.32], [-99.79, -99.23]]

lat = [[42.25, 42.21], [42.63, 42.59]]
ds = xr.Dataset({'temperature': (['x', 'y', 'time'],  temp)},
                 coords={'lon': (['x', 'y'], lon),
                         'lat': (['x', 'y'], lat),
                         'precipitation': (['x', 'y', 'time'], precip),
                         'time': pd.date_range('2014-09-06', periods=3),
                         'reference_time': pd.Timestamp('2014-09-05')})

ds["temperature"].where(ds.precipitation>5).sum(axis=2).plot()

this line would produce a label near color bar
but only using a numpy array this happens as here

ds["temperature"].where(ds.precipitation.values>5).sum(axis=2).plot()

Expected Output

I think that selecting with a datarray should not change the label of the data.
Further I was expecting that as for "sel", "where" should change the title of the graph
so that I know from the title that I performed a selection with where with a non dimensional coordinate

[shoyer]

We currently drop names because that's our default behavior for xarray operations with multiple named arguments. But the where method uses apply_ufunc() with keep_attrs=True, which indicates that attributes from the first object should be kept. It would probably make sense to keep the name of the first object, too, when using keep_attrs=True. That would fix this issue and also fix the problem for other xarray methods (e.g., fillna) that use the same logic internally.

[max-sixty]

Confirmed. To show the specific result:

ds['temperature'].name
# 'temperature'

ds["temperature"].where(ds.precipitation>5).name
# None