Select one Series and drop it in the DataFrame in-place (see https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.pop.html?highlight=pop#pandas.DataFrame.pop)
HyukjinKwon
All 7 comments
cc @harupy and @itholic, are you interested in this?
HyukjinKwon
on 17 Sep 2019
i'm working on it
itholic
on 17 Sep 2019
@itholic thanks!
harupy
on 17 Sep 2019
The pandas doc says pop takes str, but it also can take a tuple of str when a dataframe has multi-index columns.
@HyukjinKwon
Do you think this should be covered in docstring and tests?
harupy
on 17 Sep 2019
That would be so nice if they are covered.
HyukjinKwon
on 17 Sep 2019
@harupy It sounds such a nice suggestion 馃槂
I just added DataFrame.pop with basic example at #791 ,
And will try to support with tuple for multi-index also as you commented.
itholic
on 17 Sep 2019
Thanks for considering the multi-index columns support.
I believe now that we have a basic infrastructure to support multi-index columns and tuple names for Series and Index using column_index. Actually managing Spark column names with data_columns is less useful now. We should rather first handle column_index properly.
I'm still working on fixing existing APIs. Hopefully we can think of multi-index columns from the beginning when we add new APIs.
ueshin
on 17 Sep 2019
Related issues
brookewenig
路
7Comments
HyukjinKwon
路
3Comments
timhughes
路
3Comments
fujikosu
路
3Comments
shril
路
5Comments
Most helpful comment
The pandas doc says
poptakesstr, but it also can take a tuple ofstrwhen a dataframe has multi-index columns.@HyukjinKwon
Do you think this should be covered in docstring and tests?