doc: DataFrame.apply can return several columns

[ohadle]

Might be worth mentioning in the docs that if "func" returns a Series, the result is a DataFrame.
It took me a while to figure this out (tried with tuples and a dict).

[jtratner]

If you wanted to operate rows wise, you could also select two columns firsf (df[[a,b]]) and then apply off of that with a function that returns 2 - tuples / two element lists?

If you want to create a new DataFrame of a different size, you can use zip or izip in a list comprehension to the DataFrame constructor.

DataFrame([my_apply_func(a,  b) for a, b in zip(df[a], df[b])], columns=['a', 'b', 'c'])

[jtratner]

And I guess the title of this changed since I posted this... Whoops.

[jreback]

@ohadle this has been in the docs for a while: http://pandas.pydata.org/pandas-docs/dev/groupby.html#flexible-apply

[ohadle]

Sorry, I kept looking over here
http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.apply.html
and missed that part of the docs. Thanks for the pointers.
Might still be worth adding to the API docs?

[jreback]

you mean a reference to the apply section...sure that would be nice, Pull-Request?

[jorisvandenbossche]

This is now in the docstring of apply: http://pandas-docs.github.io/pandas-docs-travis/generated/pandas.DataFrame.apply.html