如何在 Pandas 中的串联 dataframe 中找到列/行组合的最大值

Question

After concatenating four multindexed tables with yearly kg/ha data I end up with a dataframe containing 22617 rows and 144 columns. What I want to do is to find the maximum of each index/year combination to have a dataframe with 36 columns. Here is an example of the data with two columns of two of the initial dataframes:

                               Y1980      Y1981      Y1980      Y1981
FID_CATCHM CCA_2  GRIDCODE                     
0          1059.0 2         21.70426  22.058224   21.70426  22.058224 
                  3         21.70426  22.058224    0.00000   0.000000
                  4          0.00000   0.000000   21.70426  22.058224
1          1059.0 2          0.00000   0.000000   21.70426  22.058224
                  4         21.70426  22.058224   21.70426  22.058224
2          1001.0 2         20.71299  21.058432   20.71299  21.058432
                  3          0.00000   0.000000   20.71299  21.058432
           1054.0 2         20.25414  20.283833   20.25414  20.283833
                  4          0.00000   0.000000   20.25414  20.283833
           1059.0 2         21.70426  22.058224   21.70426  22.058224
                  3         21.70426  22.058224   21.70426  22.058224
                  4         21.70426  22.058224   21.70426  22.058224
3          1059.0 1         21.70426  22.058224    0.00000   0.000000
                  2         21.70426  22.058224   21.70426  22.058224
                  3         21.70426  22.058224   21.70426  22.058224
                  4         21.70426  22.058224   21.70426  22.058224
4          1058.0 1          0.00000   0.000000   23.79386  24.201496
                  2         23.79386  24.201496   23.79386  24.201496
                  3          0.00000   0.000000    0.00000   0.000000
                  4         23.79386  24.201496   23.79386  24.201496
                     

What I tried to do is to use a mask

df_max = (df
           .groupby(['FID_CATCHM',
               'CCA_2', 'GRIDCODE'])
           .max())
df_mask = df_max.max(axis=1).to_frame('maximum')

but the output is identical to the concatenated dataframe. How can this be done? I appreciate every help.

Answer 1

I think you need max per columns and if necessary then per MultiIndex:

df = df.max(level=0, axis=1).max(level=[0,1,2], axis=0)