Raise proper error for scalar array when coords is a dict

[griverat]

As explained here #3159 (comment) , when a user uses a scalar array to build a DataArray with coords given as a dictionary the error is not self explanatory.

>>> xr.DataArray(np.array(1), coords={'x': np.arange(4), 'y': 'a'}, dims=['x'])
...
KeyError: 'x'

This PR makes sure that when data is a scalar array and dims is not empty, it sets the shape to (0,) to make it fail with the proper raise message

>>> xr.DataArray(np.array(1), coords={'x': np.arange(4), 'y': 'a'}, dims=['x'])
...
ValueError: conflicting sizes for dimension 'x': length 0 on the data but length 4 on coordinate 'x'

• Test updated
• Passes black . && mypy . && flake8
• Fully documented, including whats-new.rst for all changes and api.rst for new API (is this needed for a change like this?)

[shoyer]

ValueError: conflicting sizes for dimension 'x': length 0 on the data but length 4 on coordinate 'x' seems like an improvement, but I'm still not sure it's the right message.

Maybe something like ValueError: different number of dimensions on data and dims: 0 vs 3 would be more precise?

I think this could be achieved by checking len(dims) == len(shape) in the case where dims was explicitly provided, i.e., around this line:

xarray/xarray/core/dataarray.py

Line 117 in aaeea62

for d in dims:

For reference, here's the current traceback for this error:

In [3]: xr.DataArray(np.array(1), coords={'x': np.arange(4), 'y': 'a'}, dims=['x'])
---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-3-68fe094298fd> in <module>
----> 1 xr.DataArray(np.array(1), coords={'x': np.arange(4), 'y': 'a'}, dims=['x'])

~/dev/xarray/xarray/core/dataarray.py in __init__(self, data, coords, dims, name, attrs, encoding, indexes, fastpath)
    344             data = _check_data_shape(data, coords, dims)
    345             data = as_compatible_data(data)
--> 346             coords, dims = _infer_coords_and_dims(data.shape, coords, dims)
    347             variable = Variable(dims, data, attrs, encoding, fastpath=True)
    348

~/dev/xarray/xarray/core/dataarray.py in _infer_coords_and_dims(shape, coords, dims)
    140
    141         for d, s in zip(v.dims, v.shape):
--> 142             if s != sizes[d]:
    143                 raise ValueError(
    144                     "conflicting sizes for dimension %r: "

KeyError: 'x'

[griverat]

I think this could be achieved by checking len(dims) == len(shape) in the case where dims was explicitly provided, i.e., around this line:

Thanks for the suggestion. The raise now happens if len(dims) != len(shape) near the line you suggested.

[shoyer]

Looks great, thank you!

(We don't really need release notes for every minor change, especially if nothing is changing from a user perspective.)