Loop over rows

For example, you may have a dataset where one column contains a comma-separated list—for instance, the hobbies column:

id	name	hobbies
1	virdio	listen to music, reading, sleeping
2	juan	sleeping, fishing, eating
3	dani	reading, ski, gym

and your goal is to transform that table into a more structured format like this:

id	name	hobbies
1	virdio	listen to music
1	virdio	reading
1	virdio	sleeping
2	juan	fishing
2	juan	eating
3	dani	reading
3	dani	ski
3	dani	gym

Two approaches you can use

Vectorized operations

Using .apply(func), we can process each row without writing an explicit for loop:

data['hobbies']=data['hobbies'].str.split(',').apply(
    lambda items:
    list(
        map(lambda item:item.strip(),items)
    )
)

This is generally the best practice.

Looping with a for loop

data['hobbies'] = data['hobbies'].astype(object)

for idx in data.index:
    items = str(data.at[idx, 'hobbies']).split(',')
    cleaned = [i.strip() for i in items]
    data.at[idx, 'hobbies'] = cleaned

This approach is more time-consuming and often redundant.

Both work, but the final step is to explode the list into rows:

data = data.explode('hobbies')

This creates a more granular and specific table.