if row in multiple column contains 1 how to add new column in dataframe containing column names where value is 1

Question

my dataframe looks like below

My question is under Locations column how can i add column names where row value is 1, for example against Japan/US, Newyork,Osaka should be printed under Locations column....Pls advice how to solve this in Python?

Answer 1

Try this:

import pandas as pd


DF = pd.DataFrame
S = pd.Series


def construct_df() -> DF:
    data = {
        "Countries": ["Japan/US", "Australia & NZ", "America & India"],
        "Portugal": [0, 0, 0],
        "Newyork": [1, 0, 1],
        "Delhi": [0, 0, 0],
        "Osaka": [1, 0, 1],
        "Bangalore": [0, 0, 0],
        "Sydney": [0, 0, 0],
        "Mexico": [0, 0, 0],
    }
    return pd.DataFrame(data)


def calc_locations(x: DF) -> S:
    x__location_cols_only = x.select_dtypes("integer")
    x__ones_as_location_col_name = x__location_cols_only.apply(
        lambda ser: ser.replace({0: "", 1: ser.name})
    )
    location_cols = x__location_cols_only.columns.tolist()
    ret = x__ones_as_location_col_name[location_cols[0]]
    for colname in location_cols[1:]:
        col = x__ones_as_location_col_name[colname]
        ret = ret.str.cat(col, sep=",")
    ret = ret.str.replace(r",+", ",").str.strip(",")
    return ret


df_final = construct_df().assign(Locations=calc_locations)

assert df_final["Locations"].tolist() == ["Newyork,Osaka", "", "Newyork,Osaka"]

Answer 2

Lets say your data is like below

import pandas as pd

data = {'Countries':  ['JP/US', 'Aus/NZ', 'America/India'],
         'Portugal': [0, 0, 0],
         'Newyork': [1, 0, 1],
         'Delhi': [0, 0, 1],
         'Osaka': [1, 0, 0],
         'Sydney': [0, 0, 0],
         'Mexico': [0, 0, 0],
        }
data_df = pd.DataFrame(data)

DF looks like (request you to provide above set of data to build a DF for us to provide you results) :

       Countries  Delhi  Mexico  Newyork  Osaka  Portugal  Sydney
0          JP/US      0       0        1      1         0       0
1         Aus/NZ      0       0        0      0         0       0
2  America/India      1       0        1      0         0       0

If you execute below statements

data_df = data_df.set_index('Countries')
data_df['Locations'] = data_df.apply(lambda x: ", ".join(x[x!=0].index.tolist()), axis=1)

Your output will look like below

               Delhi  Mexico  Newyork  Osaka  Portugal  Sydney       Locations
Countries                                                                     
JP/US              0       0        1      1         0       0  Newyork, Osaka
Aus/NZ             0       0        0      0         0       0                
America/India      1       0        1      0         0       0  Delhi, Newyork

Answer 3

You can do:

data = {
        "Countries": ["Japan/US", "Australia & NZ", "America & India"],
        "Portugal": [0, 0, 0],
        "Newyork": [1, 0, 1],
        "Delhi": [0, 0, 0],
        "Osaka": [1, 0, 1],
        "Bangalore": [0, 0, 0],
        "Sydney": [0, 0, 0],
        "Mexico": [0, 0, 0],
    }
df=pd.DataFrame(data)
df1=df.iloc[:,1:]*df.iloc[:,1:].columns
df['Location']=df1.values.tolist()
df['Location']=df['Location'].apply(lambda x:','.join([y for y in x if len(y)>1]))