replace a dim with a coordinate from another dataset

I often want a function that takes a dataarray / dataset and replaces a dimension with a coordinate from a different dataset.

@shoyer proposed the following simple solution.

def replace_dim(da, olddim, newdim):
    renamed = da.rename({olddim: newdim.name})

    # note that alignment along a dimension is skipped when you are overriding
    # the relevant coordinate values
    renamed.coords[newdim.name] = newdim
    return renamed

Is this of broad enough interest to add a build in method for?

Start by reviewing the DataArray and Dataset APIs alongside the proposed replace_dim function, focusing on how a dimension and coordinate from another dataset should be represented. Decide whether this belongs as a built-in method and define the expected behavior; done means the supported operation is documented and its behavior is covered by tests.