pandas dataframe map using lambda function with multiple input arguments | AttributeError: 'DataFrame' object has no attribute 'map'

Given:

import numpy as np
import pandas as pd

df = pd.DataFrame(data={'user_ip': ["u1", "u2", "u3", "u4", "u5"],
                        'a':  [1, np.nan, 8, 2, 0], 
                        'b':  [2, 5, 1, np.nan, 0], 
                        'c':  [3, 0, np.nan, 0, 7],
                        'd':  [0, 2, 1, 2, 9],
                    },
                  )

 user_ip   a     b       c      d
0   u1    1.0   2.0     3.0     0
1   u2    NaN   5.0     0.0     2
2   u3    8.0   1.0     NaN     1
3   u4    2.0   NaN     0.0     2
4   u5    0.0   0.0     7.0     9

Goal:

I’d like to loop through each row to get a new column using my custom defined function with input arguments (including DataFrame and its column) as follows:

def fcn(df, col, x, y):
    return x*df[col] + y

df["new_col_apply"] = df.apply(lambda inp_df: fcn(inp_df, col="b", x=2, y=10), axis=1)

My solution works fine but apply() method seems quite slow for my original dataframe containing more than 900K rows.

I am aware of map() but since DataFrame doesn’t have map() transformation and I specifically need to input my DataFrame and its column (col) as input to my function fcn, my following snippet:

df["new_col_map"] = df.map(lambda inp_df: fcn(inp_df, col="b", x=2, y=10), na_action="ignore")

ends up in AttributeError as bellow:

---------------------------------------------------------------------------

AttributeError                            Traceback (most recent call last)

<ipython-input-14-dfe11c4bff87> in <cell line: 1>()
----> 1 df["new_col_map"] = df.map(lambda inp_df: fcn(inp_df, col="b", x=2, y=10), na_action="ignore")
      2 df

/usr/local/lib/python3.10/dist-packages/pandas/core/generic.py in __getattr__(self, name)
   5900         ):
   5901             return self[name]
-> 5902         return object.__getattribute__(self, name)
   5903 
   5904     def __setattr__(self, name: str, value) -> None:

AttributeError: 'DataFrame' object has no attribute 'map'

Is there any better and faster alternative than apply() transformation to loop through large pandas DataFrame with custom defined functions with several arguments?

Cheers,

>Solution :

Try directly invoking the function:

df["new_col"] = fcn(df, col="b", x=2, y=10)