this one is throwing an error in CI:

=================================== FAILURES ===================================
__________________ [doctest] pandas.core.apply.relabel_result __________________
1324     order: New order for relabelling
1325 
1326     Examples:
1327     ---------
1328     >>> result = DataFrame({"A": [np.nan, 2, np.nan],
1329     ...       "C": [6, np.nan, np.nan], "B": [np.nan, 4, 2.5]})  # doctest: +SKIP
1330     >>> funcs = {"A": ["max"], "C": ["max"], "B": ["mean", "min"]}
1331     >>> columns = ("foo", "aab", "bar", "dat")
1332     >>> order = [0, 1, 2, 3]
1333     >>> relabel_result(result, funcs, columns, order)
UNEXPECTED EXCEPTION: NameError("name 'result' is not defined")
Traceback (most recent call last):
  File "/home/runner/micromamba/envs/test/lib/python3.8/doctest.py", line 1336, in __run
    exec(compile(example.source, filename, "single",
  File "<doctest pandas.core.apply.relabel_result[4]>", line 1, in <module>
NameError: name 'result' is not defined

I think you need to remove # doctest: +SKIP from a few lines above

Fixed the error!

Thanks @kkangs0226! ... Is there any chance that you could run this command? :

./scripts/validate_docstrings.py pandas.core.apply.relabel_result

Here is the output, it still flags as failed example but I think @MarcoGorelli has mentioned that this error is unrelated. It seems to be one of F821 undefined errors.

sorry I meant that the Ubuntu / Numpy Dev (pull_request) CI job's failure is unrelated

this one looks related, you may need to put something like from pandas.core.apply import relabel_result at the top of the doctest

thanks for trying - sure, I'll take a look tomorrow, hopefully there's a simple fix

The docstring of relabel_result has many errors. I see 16:
The summary should be in one line, the parameters and even the subtitles are incorrect. (They shouldn't have ":", it should be Examples, not Examples:) etc.

But at the beginning of the function it's written "Internal function...", and this is not on the documentation website.
Should we test internal functions docstrings?! @MarcoGorelli

Thank you @kkangs0226... We'll get there. :-)

I don't think we should fix the whole docstring here. Let's merge this when all the problems to the original PR scope are fixed, and open follow up PRs for anything else we want to fix.

@kkangs0226 can you provide the whole traceback for the error please? With only the error and the line you shared I don't know what the problem can be, and more context would be helpful.

Here is the entire traceback, hope it is able to provide more context.

Line 21, in pandas.core.apply.relabel_result
Failed example:
    relabel_result(result, funcs, columns, order)
Exception raised:
    Traceback (most recent call last):
      File "/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/doctest.py", line 1336, in __run
        exec(compile(example.source, filename, "single",
      File "<doctest pandas.core.apply.relabel_result[6]>", line 1, in <module>
        relabel_result(result, funcs, columns, order)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/apply.py", line 1377, in relabel_result
        s = s[col_idx_order]
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/series.py", line 968, in __getitem__
        return self._get_with(key)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/series.py", line 1003, in _get_with
        return self.loc[key]
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/indexing.py", line 1100, in __getitem__
        return self._getitem_axis(maybe_callable, axis=axis)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/indexing.py", line 1329, in _getitem_axis
        return self._getitem_iterable(key, axis=axis)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/indexing.py", line 1269, in _getitem_iterable
        keyarr, indexer = self._get_listlike_indexer(key, axis)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/indexing.py", line 1459, in _get_listlike_indexer
        keyarr, indexer = ax._get_indexer_strict(key, axis_name)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/indexes/base.py", line 5851, in _get_indexer_strict
        self._raise_if_missing(keyarr, indexer, axis_name)
      File "/Users/kingbob/Desktop/NUS/3281:3282/pandas/pandas/core/indexes/base.py", line 5910, in _raise_if_missing
        raise KeyError(f"None of [{key}] are in the [{axis_name}]")
    KeyError: "None of [Index([-1], dtype='int64')] are in the [index]"```

ok, got it - I'll post a new review comment showing how I'd go about fixing this (as it's a bit long and easy to get lost in a thread)

I don't think we should fix the whole docstring here. Let's merge this when all the problems to the original PR scope are fixed, and open follow up PRs for anything else we want to fix.

@kkangs0226 can you provide the whole traceback for the error please? With only the error and the line you shared I don't know what the problem can be, and more context would be helpful.

Yes, the parentheses are to avoid a flake8 error. But the examples should look as if you were using the interpreter. You wouldn’t be using parentheses in that case.

Does this comment refer to the parenthesis wrapping this whole block? If that's the case, I don't think this is true. The original syntax is much more common and readable than this in my opinion. This is what we've been doing in the docs that I know, it's how code was chained in the original post about method chaining, it's also the syntax that Matt Harrison uses in the Effective pandas book, and it's what I would write.

To be clear, I'm referring to using

(df.pipe(foo)
   .pipe(bar)
   .pipe(foobar)
)

instead of

df.pipe(
    foo
).pipe(
    bar
).pipe(
    foobar
)

I'm fine with whatever (even if I personally found the former much more readable), but I wouldn't ask people to change the the latter because the former is not what someone would write, because I think many people would actually write that. :)