Let' say I have a DataFrame with a lot of columns df, and different functions fn(...) that can be applied to those columns to compute different results. Hereby, the column names fit the arguments of the functions:
import pandas as pd
def f1(A, C):
return(A + C)
def f2(B, C):
return(C - B)
df = pd.DataFrame([[1,2,3],[4,5,6],[7,8,9]], columns = ['A','B','C'])
To apply a function to each row, I can use the following line of code:
df['result'] = df.apply(lambda row: f1(row['A'], row['C']), axis = 'columns')
However, as different functions take different input arguments, it is kind of annoying to specify them explicitly, so I am looking for a way so that apply recognizes which value to pass to the function based on the column name.
In R, the above line is much shorter and I find this extremly convenient:
df$result = pmap(df, f1)
Is there anything similar in Python?
Btw, I know that apply is not recommended, but the functions are so complex that they are not vertorizable anyways (at least for me), and the above function-definition makes debugging much easier.
Following chrslg's answer, I think the shortest way to accomplish this is to unpack row and rewrite f1 such that unused arguments are passed to **kwargs:
import pandas as pd
def f1(A, C, **kwargs):
return(A + C)
df = pd.DataFrame([[1,2,3],[4,5,6],[7,8,9]], columns = ['A','B','C'])
df['result'] = df.apply(lambda row: f1(**row), axis = 'columns')