Python pandas sort multiple columns based on their values and other rows-CodePudding

I would like to sort columns like that as now it just fills in from 0 to make sure there are none

    websiteName         0      1       2      3
0   websiteName1    TLSv1.0 TLSv1.1 TLSv1.2 None  
1   websiteName1    TLSv1.2 None    None    None
2   websiteName2    TLSv1.2 TLSv1.3 None    None
3   websiteName3    TLSv1.0 TLSv1.2 TLSv1.3 None
4   websiteName4    TLSv1.2 TLSv1.3 None    None

Expected result

    websiteName         0      1       2      3
0   websiteName1    TLSv1.0 TLSv1.1 TLSv1.2 None
1   websiteName1    None    None    TLSv1.2 None
2   websiteName2    None    None    TLSv1.2 TLSv1.3 
3   websiteName3    TLSv1.0 None    TLSv1.2 TLSv1.3
4   websiteName4    None    None    TLSv1.2 TLSv1.3

CodePudding user response：

You can melt, then pivot:

(df
 .replace({'None': pd.NA})   # assuming 'None' strings
 .melt(id_vars='websiteName')
 .assign(TLS=lambda d: d['value'])
 .pivot_table(index='websiteName', columns='TLS', values='value', aggfunc='first')
 .set_axis(range(len(df.columns)-1), axis=1) # comment to keep TLSvx.x names
 .reset_index()
)

output:

    websiteName        0        1        2        3
0  websiteName1  TLSv1.0  TLSv1.1  TLSv1.2      NaN
1  websiteName2      NaN      NaN  TLSv1.2  TLSv1.3
2  websiteName3  TLSv1.0      NaN  TLSv1.2  TLSv1.3
3  websiteName4      NaN      NaN  TLSv1.2  TLSv1.3