Automatic naming causes trouble

[randolf-scholz]

Code Sample, a copy-pastable example if possible

A "Minimal, Complete and Verifiable Example" will make it much easier for maintainers to help you:
http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports

from xarray import Dataset, DataArray
ds = Dataset(coords={'foo':[1,2,3], 'bar':[1,2,3]})
da = DataArray([0], dims='test')
da.name = 'abc'
ds[(1,2)] = da
print(da.name)
print(ds[(1,2)].name)
print(type(ds[(1,2)].name))
print(ds[(1,2)])              # throws error

Problem description

The name gets overwritten with the tuple key, but is should stay intact?!?
Secondly the print statement throws an error because internally name_str = '%r ' % arr.name is called. But here arr.name is a tuple !!

Instead, it should be repr(arr.name) or F"{arr.name}"

Expected Output

The name should not change.

Output of xr.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.6.8 |Anaconda, Inc.| (default, Dec 30 2018, 01:22:34) [GCC 7.3.0] python-bits: 64 OS: Linux OS-release: 4.18.0-16-generic machine: x86_64 processor: x86_64 byteorder: little LC_ALL: None LANG: en_US.UTF-8 LOCALE: en_US.UTF-8 libhdf5: 1.10.4 libnetcdf: 4.6.1

xarray: 0.12.0
pandas: 0.24.2
numpy: 1.16.2
scipy: 1.2.1
netCDF4: 1.4.2
pydap: None
h5netcdf: None
h5py: 2.9.0
Nio: None
zarr: None
cftime: 1.0.3.4
nc_time_axis: None
PseudonetCDF: None
rasterio: None
cfgrib: None
iris: None
bottleneck: 1.2.1
dask: None
distributed: None
matplotlib: 3.0.3
cartopy: None
seaborn: None
setuptools: 40.8.0
pip: 19.0.3
conda: 4.6.8
pytest: None
IPython: 7.3.0
sphinx: 1.8.5

[shoyer]

Thanks for the report.

Yes, this looks like a bug. A pull request to fix this would be very welcome.

[randolf-scholz]

@shoyer I had a look at xarray/core/formatting.py and it seems like there are a couple of places where keys get formatted this way. That is %-style formatting is used, ignoring the possibility that a dict key may be of type tuple. Personally I am really a big fan of f-strings, as they are faster and more readable than %-style formatting. They would in my opinion be one of the best ways fix this error and prevent it from happening again in the future.

In fact, the exact error that happened here is listed as one of the rationales for f-strings in the PEP: https://www.python.org/dev/peps/pep-0498/#rationale

Using them would of course mean dropping python 3.5 support.

[shoyer]

I also love fstrings, but I don't think we're ready to drop Python 3.5 support yet. Let's try explicit calls to .format() for now.

[keewis]

the formatting of the tuple-typed key works now (#2906), so I'm closing this.

The name of da is still dropped, but that is intended behavior (it also happens with ds["def"] = da). For a discussion about what to do with tuple-typed keys see #4821.