Comment obtenir la valeur maximale de deux colonnes ou plus dans un dataframe pandas ?

Un exemple de comment obtenir la valeur maximale de deux colonnes ou plus dans un dataframe pandas

Créer une simple dataframe avec pandas

Créons une simple dataframe avec pandas

import pandas as pd
import numpy as np
import random

a = [random.choice(['a','b','c']) for i in range(20)]
c1 = np.random.uniform(0,1, size=20)
c2 = np.random.uniform(0,1, size=20)
c3 = np.random.uniform(0,1, size=20)

data = {'a':a,
        'c1':c1, 
        'c2':c2, 
        'c3':c3}

df = pd.DataFrame(data)

print(df)

donne par exemple

    a        c1        c2        c3
0   b  0.480835  0.108703  0.062893
1   b  0.469707  0.441162  0.584776
2   a  0.942992  0.254551  0.750689
3   c  0.634768  0.048194  0.810934
4   b  0.185650  0.551316  0.126457
5   c  0.661850  0.178760  0.321801
6   c  0.739558  0.878052  0.196450
7   a  0.344361  0.267702  0.401820
8   a  0.430181  0.300013  0.674959
9   b  0.412139  0.295223  0.799033
10  a  0.101919  0.810593  0.665577
11  c  0.036506  0.705999  0.183860
12  a  0.832732  0.288746  0.994192
13  b  0.474459  0.574548  0.539348
14  b  0.322733  0.015939  0.142288
15  c  0.108360  0.447548  0.474686
16  c  0.111654  0.048010  0.651970
17  c  0.360656  0.155344  0.854249
18  a  0.404405  0.034052  0.618728
19  b  0.234892  0.052290  0.889191

Trouvez la valeur maximale pour chaque ligne

Pour trouvez la valeur maximale pour chaque ligne entre les colonnes ['c1','c2','c3'] une solution est d'utiliser pandas.DataFrame.max:

df[['c1','c2','c3']].max(axis=1)

donne

0     0.480835
1     0.584776
2     0.942992
3     0.810934
4     0.551316
5     0.661850
6     0.878052
7     0.401820
8     0.674959
9     0.799033
10    0.810593
11    0.705999
12    0.994192
13    0.574548
14    0.322733
15    0.474686
16    0.651970
17    0.854249
18    0.618728
19    0.889191
dtype: float64

Pour ajouter une colonne contenant les valeurs max:

df['max value'] = df[['c1','c2','c3']].max(axis=1)

donne

    a        c1        c2        c3  max value
0   b  0.480835  0.108703  0.062893   0.480835
1   b  0.469707  0.441162  0.584776   0.584776
2   a  0.942992  0.254551  0.750689   0.942992
3   c  0.634768  0.048194  0.810934   0.810934
4   b  0.185650  0.551316  0.126457   0.551316
5   c  0.661850  0.178760  0.321801   0.661850
6   c  0.739558  0.878052  0.196450   0.878052
7   a  0.344361  0.267702  0.401820   0.401820
8   a  0.430181  0.300013  0.674959   0.674959
9   b  0.412139  0.295223  0.799033   0.799033
10  a  0.101919  0.810593  0.665577   0.810593
11  c  0.036506  0.705999  0.183860   0.705999
12  a  0.832732  0.288746  0.994192   0.994192
13  b  0.474459  0.574548  0.539348   0.574548
14  b  0.322733  0.015939  0.142288   0.322733
15  c  0.108360  0.447548  0.474686   0.474686
16  c  0.111654  0.048010  0.651970   0.651970
17  c  0.360656  0.155344  0.854249   0.854249
18  a  0.404405  0.034052  0.618728   0.618728
19  b  0.234892  0.052290  0.889191   0.889191

Trouver le nom de la colonne contenant la valeur maximale

Pour trouver le nom de la colonne contenant la valeur maximale pandas.DataFrame.idxmax

df[['c1','c2','c3']].idxmax(axis=1)

donne

0     c1
1     c3
2     c1
3     c3
4     c2
5     c1
6     c2
7     c3
8     c3
9     c3
10    c2
11    c2
12    c3
13    c2
14    c1
15    c3
16    c3
17    c3
18    c3
19    c3
dtype: object

Créez une nouvelle colonne pour stocker les résultats:

df['col name'] = df[['c1','c2','c3']].idxmax(axis=1)

donne

    a        c1        c2        c3  max value col name
0   b  0.480835  0.108703  0.062893   0.480835       c1
1   b  0.469707  0.441162  0.584776   0.584776       c3
2   a  0.942992  0.254551  0.750689   0.942992       c1
3   c  0.634768  0.048194  0.810934   0.810934       c3
4   b  0.185650  0.551316  0.126457   0.551316       c2
5   c  0.661850  0.178760  0.321801   0.661850       c1
6   c  0.739558  0.878052  0.196450   0.878052       c2
7   a  0.344361  0.267702  0.401820   0.401820       c3
8   a  0.430181  0.300013  0.674959   0.674959       c3
9   b  0.412139  0.295223  0.799033   0.799033       c3
10  a  0.101919  0.810593  0.665577   0.810593       c2
11  c  0.036506  0.705999  0.183860   0.705999       c2
12  a  0.832732  0.288746  0.994192   0.994192       c3
13  b  0.474459  0.574548  0.539348   0.574548       c2
14  b  0.322733  0.015939  0.142288   0.322733       c1
15  c  0.108360  0.447548  0.474686   0.474686       c3
16  c  0.111654  0.048010  0.651970   0.651970       c3
17  c  0.360656  0.155344  0.854249   0.854249       c3
18  a  0.404405  0.034052  0.618728   0.618728       c3
19  b  0.234892  0.052290  0.889191   0.889191       c3

Références

• Find the max of two or more columns with pandas
• Find the column name which has the maximum value for each row