How to create two columns from single one in Pandas dataframe, so that we have pairs in lines?-CodePudding

I have a Pandas dataframe with a column with elements of type string. I need to get 2 columns for future pairwise comparison.

Example: We had: Abc Bdf Ftp ...

What I need to get: (2 different columns) Abc, Bdf Abc, Ftp Bdf, Ftp ...

I've searched a lot of different sources and what I came to is that I need to use itertools. But how?

CodePudding user response：

use shift

The dataframe:

df = pd.DataFrame(columns = ['col1'], data = ['Abc','Bdf', 'Ftp'])

and the answer

df['col2'] = df['col1'].shift()
df.fillna(method = 'ffill', axis=1)

output:


    col1    col2
0   Abc     Abc
1   Bdf     Abc
2   Ftp     Bdf

CodePudding user response：

from itertools import combinations

df2 = pd.DataFrame(combinations(df['col1'], 2),
                   columns=('col1', 'col2'))

Output:

  col1 col2
0  Abc  Bdf
1  Abc  Ftp
2  Bdf  Ftp