Add index=False option to_markdown()

[chrisjcameron]

Problem description

This is a feature request to add an optional index=False argument to the new to_markdown() dataframe method. This would produce a markdown table without the initial unnamed index column.

[dsaxton]

This is actually possible by passing showindex=False, but could probably be better documented (it's one of the tabulate kwargs):

In [1]: import pandas as pd       
                                                                                                                                                                                                        
In [2]: df = pd.DataFrame({"a": [1, 2, 3]})              
                                                                                                                                                                                 
In [3]: print(df.to_markdown(showindex=False))                                                                                                                                                                                            
|   a |
|----:|
|   1 |
|   2 |
|   3 |

[MarcoGorelli]

TBH I'm not sure it's worth documenting tabulate-specific kwargs, but adding an example showing the behaviour of df.to_markdown(showindex=False) would be nice.

@chrisjcameron are you interested in submitting a PR? If so, see the contributing guide for how to get started

[jreback]

adding index=False here would be ok (and translating to tabulate)

PR welcome

[quangngd]

take

[quangngd]

What should the behaviour be if user set both showindex and index? Which one should take the priority or should I raise an ValueError

[mproszewska]

Does it make sense to add the index parameter if the showindex parameter already exists and will be used in the same way?
I can submit PR with added examples

[chrisjcameron]

It would be more consistent with the other pandas to_* methods if the parameter to control including the index was called index rather than showindex.

[chrisjcameron]

What should the behaviour be if user set both showindex and index? Which one should take the priority or should I raise an ValueError

I think most users will be expecting pandas-like keywords to work, so I would prioritize those. An alternative might be a warning about "When both index and showindex are set, one value will be used but which is undefined". This would at least give users a clue that these need not both be specified but gives a lot of latitude for future refinement that changes the behavior.

[franasa]

thumbs up for this enhancement to the docs. l landed here looking exactly for that

pandas/pandas/core/frame.py

Lines 1996 to 2024 in 3adf334

	@Appender(
	"""
	Examples
	--------
	>>> df = pd.DataFrame(
	... data={"animal_1": ["elk", "pig"], "animal_2": ["dog", "quetzal"]}
	... )
	>>> print(df.to_markdown())
	| | animal_1 | animal_2 |
	|---:|:-----------|:-----------|
	| 0 | elk | dog |
	| 1 | pig | quetzal |
	"""
	)
	@Substitution(klass="DataFrame")
	@Appender(_shared_docs["to_markdown"])
	def to_markdown(
	self, buf: Optional[IO[str]] = None, mode: Optional[str] = None, **kwargs
	) -> Optional[str]:
	kwargs.setdefault("headers", "keys")
	kwargs.setdefault("tablefmt", "pipe")
	tabulate = import_optional_dependency("tabulate")
	result = tabulate.tabulate(self, **kwargs)
	if buf is None:
	return result
	buf, _, _, _ = get_filepath_or_buffer(buf, mode=mode)
	assert buf is not None # Help mypy.
	buf.writelines(result)
	return None