Sorting in Multi-Layers – Pandas pivot_table

I am looking for a solution to sort a pandas pivot table, as it is possible in Excel pivot tables, according to the aggregated values and thereby determine the order of the first level.
I have found a solution that uses pd.append.
A warning tells me that pd.append will not be available in the future and that I should therefore use pd.concat.

I have not managed to solve my problem with pd.concat.

This is the solution with pd.append

>Solution :

You should use concat this way:

out = pd.concat([
      df_pivot.iloc[:-1]
              .assign(sort=lambda x: x['Totals'].groupby(level=0).transform('sum'))
              .sort_values(['sort','Name','Totals'], 
                           ascending=[False,True,False], kind='mergesort')
              .drop('sort', axis=1),
      df_pivot.iloc[[-1]] # slicing as DataFrame, not Series
           ])

Output:

Year            2019  2020  Totals
Name   A   B                      
Brian  foo one    20     0      20
       zz  xy     10     0      10
       bar two     0     9       9
           one     0     8       8
       zz  zz      0     5       5
Peter  foo one    20     0      20
       bar two     0     9       9
           one     6     0       6
Amy    foo two    11     0      11
           one     4     0       4
Totals            71    31     102