Replace only first row for each user

I have a table that has users, food, and which one is their favorite.

The same user appears in several rows. I need to set exactly 1 of the rows per user as favorite (is_favorite=True):

Now every user has exactly 1 favorite food.

I successfully got exactly 1 row for each user, but can’t apply it to my initial df. I’m pretty sure it’s something simple I’m missing, but I don’t know pandas that well. It also feels like this is the wrong way to do it:

import pandas as pd

df = pd.DataFrame(
    dict(
        user=[1, 1, 3, 3, 3, 4],
        food=['Beef', 'Pork', 'Pork', 'Beef', 'Potatoes', 'Beef'],
        is_favorite=[False, False, False, False, False, False]))

# This works. It gives me exactly 1 row per user
first_food_per_user = df.groupby('user').nth(0).reset_index()

# This doesn't work
for _, row in first_food_per_user.iterrows():
    df['is_favorite'].loc[
        (df['user'] == row['user'])
        &
        df['food'] == row['food'],
    ] = True

>Solution :

No need to groupby, just use duplicated and boolean indexing:

df.loc[~df['user'].duplicated(), 'is_favorite'] = True

Output:

   user      food  is_favorite
0     1      Beef         True
1     1      Pork        False
2     3      Pork         True
3     3      Beef        False
4     3  Potatoes        False
5     4      Beef         True

If you want to set a random row use groupby.sample:

idx = df.groupby('user')['is_favorite'].sample(n=1).index
df.loc[idx, 'is_favorite'] = True

Example:

   user      food  is_favorite
0     1      Beef         True
1     1      Pork        False
2     3      Pork        False
3     3      Beef         True
4     3  Potatoes        False
5     4      Beef         True