I have the following DF:
Fecha | xG | xGA | Home | Away | Score |
---|---|---|---|---|---|
2022-05-01 | 1.53 | 0.45 | América | Cruz Azul | 0:0 |
2022-04-24 | 1.46 | 0.47 | Tigres UANL | América | 0:2 |
2022-04-21 | 1.40 | 0.43 | América | León | 2:0 |
2022-04-16 | 2.44 | 0.65 | Club Tijuana | América | 1:3 |
I want to create two new columns named HomeXG and AwayXG where the values are taken from the xG and xGA columns under the condition that if America is the home team the xG becomes HomeXG and if they are away xG is used for AwayXG.
Expected output:
Fecha | xG | xGA | Home | Away | Score | HomeXG | AwayXG |
---|---|---|---|---|---|---|---|
2022-05-01 | 1.53 | 0.45 | América | Cruz Azul | 0:0 | 1.53 | 0.45 |
2022-04-24 | 1.46 | 0.47 | Tigres UANL | América | 0:2 | 0.47 | 1.46 |
2022-04-21 | 1.40 | 0.43 | América | León | 2:0 | 1.40 | 0.43 |
2022-04-16 | 2.44 | 0.65 | Club Tijuana | América | 1:3 | 0.65 | 2.44 |
CodePudding user response:
You can use where
on a 2D slice of the DataFrame:
df[['HomexG', 'AwayxG']] = df[['xG', 'xGA']].where(df['Home'].eq('América'),
df[['xGA', 'xG']].values)
NB. the second argument of where
must be a numpy array to avoid index alignment!
output:
Fecha xG xGA Home Away Score HomexG AwayxG
0 2022-05-01 1.53 0.45 América Cruz Azul 0:0 1.53 0.45
1 2022-04-24 1.46 0.47 Tigres UANL América 0:2 0.47 1.46
2 2022-04-21 1.40 0.43 América León 2:0 1.40 0.43
3 2022-04-16 2.44 0.65 Club Tijuana América 1:3 0.65 2.44
CodePudding user response:
Use numpy.where
, solution testing only Home
column, it means América
should me in Away
for non matched rows:
df[['HomexG', 'AwayxG']] = np.where(df['Home'].eq('América').to_numpy()[:, None],
df[['xG','xGA']], df[['xGA','xG']])
print (df)
Fecha xG xGA Home Away Score HomexG AwayxG
0 2022-05-01 1.53 0.45 América Cruz Azul 0:0 1.53 0.45
1 2022-04-24 1.46 0.47 Tigres UANL América 0:2 0.47 1.46
2 2022-04-21 1.40 0.43 América León 2:0 1.40 0.43
3 2022-04-16 2.44 0.65 Club Tijuana América 1:3 0.65 2.44