Python Pandas select column by value by column of colnames-CodePudding

I have a column containing the names of certain columns of a dataframe.

import pandas as pd
d = {'col1': [1, 2], 'col2': [3, 4], 'col3':['col1', 'col2']}
df = pd.DataFrame(data=d)

Now I would like to create a col4 with values from either col1 (first row) or col2 (second row), based on the names in col3.

CodePudding user response：

Use indexing lookup:

idx, cols = pd.factorize(df['col3'])
df['col4'] = df.reindex(cols, axis=1).to_numpy()[np.arange(len(df)), idx]

Output:

   col1  col2  col3  col4
0     1     3  col1     1
1     2     4  col2     4

CodePudding user response：

you can use df.eval

df['col4']=df.apply(lambda row:df.eval(df.col3)[row.name][row.name],axis=1)
df
Out[233]: 
   col1  col2  col3  col4
0     1     3  col1     1
1     2     4  col2     4