CODEWITHSUNDEEP
pythonpandas
Nemra Khalil
Nemra Khalil

Reputation: 111

How to create a new column from columns names that have True values

Suppose I have this dataframe

name   age  gender     travelled_abroad     for_work    for_entertainment      set_at_home   
John   32    male       True                  True            False              False
Ieren  52    female     False                 False           False              True
Kris   13    male       True                  False           True               False

I want to create a new column that has the columns that their values are True

name   age  gender     travelled_abroad     for_work    for_entertainment      set_at_home   final 
John   32    male       True                  True            False              False      travelled_Abroad, for_work
Ieren  52    female     False                 False           False              True       set_at_home
Kris   13    male       True                  False           True               False      travelled_abroad, for_entertainment

I've tried

def new_column(row):
    for col in df:
        if row == True:
            return ','.join(row)
df['final'] = df.applymap(new_column)

but I got nothing. any help?

Upvotes: 0

Views: 84

Answers (3)

Osama Hamdy
Osama Hamdy

Reputation: 103

Just as Mehdi has done with a tiny amendment:

data ={"name": ['John', 'Ieren', 'Kris', 'Houda'], "age": [32, 52, 13, 28], "travelled_abroad": [True, False, True, True], "for_work":[True, False, False, True], \
"for_entertainment": [False, False, True, True], "set_at_home": [False, True, False, False]}

df = pd.DataFrame(data)


cols = ['travelled_abroad','for_work','for_entertainment','set_at_home']
df = df.assign(final = lambda x: x[cols].apply(lambda s: ','.join(np.array(cols)[s.astype(bool)]),axis=1))
df

enter image description here

Another lengthier approach but tidier one from my point of view is to do some melting and merging to make the data more elegant:

  1. Create a travel status column

     df_travel_status = pd.melt(df, id_vars= ['name', 'age'], value_vars=['travelled_abroad', 'set_at_home'], var_name="travel_status", value_name="to_be_dropped" )
 df_travel_status = df_travel_status[df_travel_status['to_be_dropped']]
 df_travel_status

enter image description here

  1. Then a travel reason column:

    df_travel_reason = pd.melt(df, id_vars= ['name', 'age'], value_vars=['for_work', 'for_entertainment'], var_name='travel_reason', value_name="to_be_dropped_2" ) df_travel_reason = df_travel_reason[df_travel_reason["to_be_dropped_2"]] df_travel_reason

enter image description here

  1. Merging and dropping the extra columns

    df_modified = df_travel_status.merge(df_travel_reason, on=['name', 'age'], how='left') df_modified.drop(columns=['to_be_dropped', "to_be_dropped_2"], inplace=True) df_modified enter image description here

  2. Finally reaching to the intended dataframe and clearing any duplicates

    df_modified = df_modified.replace(np.nan, "") df_modified.groupby(['name', "age", "travel_status"])['travel_reason'].apply(", ".join).reset_index()

enter image description here

Upvotes: 2

Henry Yik
Henry Yik

Reputation: 22493

You can use list comprehension:

df["new"] = [", ".join(df.columns[3:][i]) for i in df.iloc[:, 3:].to_numpy()]

print (df)

    name  age  gender  travelled_abroad  for_work  for_entertainment  set_at_home                                  new
0   John   32    male              True      True              False        False           travelled_abroad, for_work
1  Ieren   52  female             False     False              False         True                          set_at_home
2   Kris   13    male              True     False               True        False  travelled_abroad, for_entertainment

Upvotes: 2

Mehdi Golzadeh
Mehdi Golzadeh

Reputation: 2583

Use assign function:

cols = ['travelled_abroad','for_work','for_entertainment','set_at_home']
df = df.assign(final = lambda x: x['travelled_abroad','for_work','for_entertainment','set_at_home'].apply(lambda s: ','.join(np.array(cols)[s.astype(bool)]),axis=1))

Upvotes: 3

Related Questions