How can sort DataSeries based values and index?-CodePudding

I have a dataseries (mean_ratings) like below:

A. Morin                  39203.340909
AMMA                      53476.250000
Acalli                    43798.333333
Adi                       41277.500000
Aequare (Gianduja)        45185.000000
                              ...     
Zokoko                    44615.000000
Zotter                    46138.326531
hello cocoa               26162.333333
hexx                      35043.333333
twenty-four blackbirds    36877.666667

I sort this series with this code:

best_ratings = mean_ratings.sort_values(0,ascending=False) 

Idilio (Felchlin)                  70040.000000
Dole (Guittard)                    65662.500000
Chocola'te                         64725.000000
Christopher Morel (Felchlin)       63787.500000
Madecasse (Cinagra)                62454.166667
Salgado                            61285.000000
Compania de Chocolate (Salgado)    61285.000000
Chloe Chocolat                     61285.000000
Oialla by Bojessen (Malmo)         61285.000000
Rozsavolgyi                        61285.000000

But I want to sort this, first based on values , and then for equal values based on alphabetic indexes. How can I do this simultaneously?

CodePudding user response：

You have to use kind='mergesort' on the second sort. Something like:

df.sort_index().sort_values(0, ascending=False, kind='mergesort')

CodePudding user response：

Pass .sort_values() a list of the column names you wish to sort by to sort at multiple levels!

https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.sort_values.html

>>> df = pd.DataFrame({'foo':['a', 'a', 'b', 'c'], 'bar':[2,1,4,3]})
>>> df
  foo  bar
0   a    2
1   a    1
2   b    4
3   c    3
>>> df.sort_values(["foo", "bar"])
  foo  bar
1   a    1
0   a    2
2   b    4
3   c    3
>>> df.sort_values(["bar", "foo"])
  foo  bar
1   a    1
0   a    2
3   c    3
2   b    4