How to group dataframe parts of rows into list in pandas groupby

Question

I saw this page, How to group dataframe rows into list in pandas groupby but, that's not what I need.

Let my datatable example see please.

index	column1	column2
0	apple	red
1	banana	a
1	banana	b
2	grape	wow
2	grape	that's
2	grape	great
2	grape	fruits!
3	melon	oh
3	melon	no
...a lot of data...	...a lot of data...	... a lot of data...

and I want to groupby ONLY index 2~3 (I need to range because my real data is hudge.)

so I want to this table

.	column1	column2
0	apple	red
1	banana	a
1	banana	b
2	grape	wow that's great grape fruits!
3	melon	oh no
...a lot of data...	...a lot of data...	... a lot of data...

How can I get this?

Answer 1

Let's filter the dataframe based on index column, then groupby on the filtered dataframe.

m = df['index'].isin([2,3])

out = (pd.concat([df[~m],
                  (df[m].groupby('column1', as_index=False)
                   .agg({'index': 'first', 'column2': ' '.join}))], ignore_index=True)
       .sort_values('index', ignore_index=True))

print(out)

   index column1                   column2
0      0   apple                       red
1      1  banana                         a
2      1  banana                         b
3      2   grape  wow that's great fruits!
4      3   melon                     oh no