ERR: missing stacklevel in to_dict UserWarning

[chris-b1]

Code Sample, a copy-pastable example if possible

In [4]: df = pd.DataFrame([[1,2,3]], columns=['a', 'a', 'b'])

In [5]: df
Out[5]:
   a  a  b
0  1  2  3

In [6]: df.to_dict()
C:\<snip>\pandas\core\frame.py:891:
UserWarning: DataFrame columns are not unique, some columns will be omitted.
  "columns will be omitted.", UserWarning)
Out[6]: {'a': {0: 2}, 'b': {0: 3}}

Problem description

warning should specific stacklevel to warn at calling site, xref #9584

[gfyoung]

In the future, we'll probably have to go back and find any other hanging warnings like these.

[chris-b1]

Yep, probably some others, a sufficiently clever grep might be able to turn them up. (basically whenever we call warn(...) without a stacklevel passed.

[jreback]

a lot of these prob don't have the correct stack_level; note that on most assert_raises_warnings we are passing check_stacklevel=False (so its not checking). certainly these could/should be changed.

[gfyoung]

That's true, we probably would need to go back and comb through those a bit.

[imih]

Should the @deprecate decorator be used for this?

Additionally, we might want to add a keyword stacklevel to it.

[TomAugspurger]

@imih no I don't think so. That's for when we change the name of a keyword argument.

[imih]

@TomAugspurger You're refering to @deprecate_kwarg. @deprecate is something else. https://github.com/pandas-dev/pandas/blob/master/pandas/util/_decorators.py#L9 and it's not currently in use in pandas-dev.