Python: may we sort a dataframe by a "function" of a row?

Question

With a pandas dataframe, we can sort the dataframe based on the value of a row

dfObj = dfObj.sort_values(by='b', axis=1, ascending=False)

May we do it based on the absolute value (or more generally, a function) of a row? I expect something like

dfObj = dfObj.sort_values(by=abs('b'), axis=1, ascending=False)

If it can't work, how can we produce a row equivalent to the absolute value of row 'b' and then sort?

Thanks!!

Does [this](https://stackoverflow.com/questions/13838405/custom-sorting-in-pandas-dataframe) help you? — rpanai, Apr 20 '20 at 01:46
Thanks @rpanai ```df['m'] = pd.Categorical(df['m'], ["March", "April", "Dec"])``` How can I add a row instead of a column? — user2644687, Apr 20 '20 at 01:51
Check [this](https://stackoverflow.com/questions/30486263/sorting-by-absolute-value-without-changing-the-data) — tidakdiinginkan, Apr 20 '20 at 02:17

score 0 · Answer 1 · answered Apr 20 '20 at 02:33

0

You are very close, you can do:

dfObj = dfObj.sort_values(dfObj.loc['b'].abs(), axis=1, ascending=False)

answered Apr 20 '20 at 02:33

Allen Qin

1 Answers1