Pandas df: exponentiate whole row when a condition is true-CodePudding

I am new in Python (I am using Python 3.6).

I have a pandas dataframe that looks like following:

import pandas as pd
data_exp = [['no', 2,1.5,3], ['exp', 2,1.5,3], ['no', 2,1.5,5]]
df = pd.DataFrame(data, columns = ['Type', 'V1','V2','V3'])
df

What I want to do is the following:

if type in the first column of the database is equal to exp, then I want to exp() all the values on the same row from column 2 to the last one. If else, simply do not change the values in the other columns.

So the output would be:

data = [[2,1.5,3], [7.39,4.48,20.09], [2,1.5,5]]

Any idea how to do it? The function np.exp(df.iloc[i,1:]) for each value of i in the row numbering does not work.

CodePudding user response：

Use DataFrame.iloc for both sides with boolean mask - compare values by exp in Type column, converting mask to numpy array is for prevent NotImplementedError: iLocation based boolean indexing on an integer type is not available:

m = df['Type'].eq('exp').to_numpy()
#for oldier pandas version
m = df['Type'].eq('exp').values
df.iloc[m,1:] = np.exp(df.iloc[m,1:])
print(df)
  Type        V1        V2         V3
0   no  2.000000  1.500000   3.000000
1  exp  7.389056  4.481689  20.085537
2   no  2.000000  1.500000   5.000000