Extrapolating using Pandas and Curve_fit error func() takes 3 positional arguments but 4 were given

def func(x,a,b):

    return   a*x + b

guess = (0.5,0.5,0.5)


fit_df = df.dropna()

col_params = {}

for col in fit_df.columns:

    x = fit_df.index.astype(float).values

    y = fit_df[col].values

    params = curve_fit(func,x,y,guess)

    col_params[col] = params[0] 
for col in df.columns:

    x = df[pd.isnull(df[col])].index.astype(float).values

    df[col][x] = func(x,*col_params[col])
print("Extrapolated data")
print(df)   

I'm using the code from another post to extrapolate values. I changed the func() so that it is linear not cubic however I get an error "func() takes 3 positional arguments but 4 were give"

Extrapolate values in Pandas DataFrame is where I got the original code. My question is how would I change it so it works with a linear relationship

When using: guess = (0.5,0.5) you should be able to make it run.

You have the parameters a, b while the original example had the parameters a, b, c, d. The initial guess is for the parameters a, b in your function, not for x.

Full code used to make your interpolation function run:

import pandas as pd
from io import StringIO
from scipy.optimize import curve_fit

df = pd.read_table(StringIO('''
                neg       neu       pos       avg
    0           NaN       NaN       NaN       NaN
    250    0.508475  0.527027  0.641292  0.558931
    500         NaN       NaN       NaN       NaN
    1000   0.650000  0.571429  0.653983  0.625137
    2000        NaN       NaN       NaN       NaN
    3000   0.619718  0.663158  0.665468  0.649448
    4000        NaN       NaN       NaN       NaN
    6000        NaN       NaN       NaN       NaN
    8000        NaN       NaN       NaN       NaN
    10000       NaN       NaN       NaN       NaN
    20000       NaN       NaN       NaN       NaN
    30000       NaN       NaN       NaN       NaN
    50000       NaN       NaN       NaN       NaN'''), sep='\s+')

# Do the original interpolation
df.interpolate(method='nearest', xis=0, inplace=True)

# Display result
print ('Interpolated data:')
print (df)
print ()

def func(x,a,b):

    return   a*x + b

guess = (0.5,0.5)


fit_df = df.dropna()

col_params = {}

for col in fit_df.columns:

    x = fit_df.index.astype(float).values

    y = fit_df[col].values

    params = curve_fit(func,x,y,guess)

    col_params[col] = params[0] 
for col in df.columns:

    x = df[pd.isnull(df[col])].index.astype(float).values

    df[col][x] = func(x,*col_params[col])
print("Extrapolated data")
print(df)