Duplicate DataFrame Rows based on Column values within each cell

Question

I have a Dataframe as shown Below, I have to duplicate rows using the comma as a delimiter. It's easier to understand once you see the dataframes below!:

ID      Fruit
10000   Apple, Orange, Pear
10001   Apple, Banana

I want to Dataframe below:

ID      Fruit
10000   Apple 
10000   Orange
10000   Pear
10001   Apple 
10001   Banana

Does this answer your question? [Pandas add new columns based on splitting another column](https://stackoverflow.com/questions/38956778/pandas-add-new-columns-based-on-splitting-another-column) — Grzegorz Skibinski, Mar 20 '20 at 12:34
I'm not sure it's the same, I'm not splitting the column into more columns, but more rows? — , Mar 20 '20 at 12:40
Sorry, you're right - it's a case for ```explode``` - check my answer below. — Grzegorz Skibinski, Mar 20 '20 at 13:02

score 1 · Answer 1 · answered Mar 20 '20 at 13:01

1

Try:

df['Fruit']=df['Fruit'].str.split(", ")
df=df.explode('Fruit')

Outputs:

      ID   Fruit
0  10000   Apple
0  10000  Orange
0  10000    Pear
1  10001   Apple
1  10001  Banana

answered Mar 20 '20 at 13:01

Grzegorz Skibinski

12,624
2
11
34

Jaroslav Bezděk · Accepted Answer · 2020-03-20T12:46:49.487

0

If df looks like this:

>>> df = pd.DataFrame({'ID': [10000, 10001], 'Fruit': ['Apple, Orange, Pear', 'Apple, Banana']})
>>> print(df)
      ID                Fruit
0  10000  Apple, Orange, Pear
1  10001        Apple, Banana

you can use the pandas.DataFrame.apply() method to make a new column of lists consisting of dictionaries with new rows. And after that, you can concatenate these lists in order to make a new data frame out of them. The code is following:

>>> df['new'] = df.apply(lambda row: [{'ID': row.ID, 'Fruit': item} for item in row.Fruit.split(', ')], axis=1)
>>> df_new = pd.DataFrame(df.new.sum())
>>> print(df_new)
      ID   Fruit
0  10000   Apple
1  10000  Orange
2  10000    Pear
3  10001   Apple
4  10001  Banana

edited Mar 20 '20 at 12:46

answered Mar 20 '20 at 12:41

Jaroslav Bezděk

6,967
6
29
46

1

Thanks, that's really useful. – Mar 20 '20 at 12:53
That's bad approach- you should rather avoid using ```.apply(...)```, unless you really have to: https://stackoverflow.com/a/54432584/11610186 – Grzegorz Skibinski Mar 20 '20 at 13:05

Duplicate DataFrame Rows based on Column values within each cell

2 Answers2