Combining rows to 'others' in pandas

Question

I have a pandas dataframe like this:

  character  count
0         a    104
1         b     30
2         c    210
3         d     40
4         e    189
5         f     20
6         g     10

I want to have only the top 3 characters in the dataframe and the remaining are combined as others so table become:

  character  count
0         c    210
1         e    189
2         a    104
3    others    100

How can I achieve this?

Thank you.

Answer 1

we can use Series.nlargest() method:

In [31]: new = df.nlargest(3, columns='count')

In [32]: new = pd.concat(
    ...:         [new,
    ...:          pd.DataFrame({'character':['others'],
    ...:                        'count':df.drop(new.index)['count'].sum()})
    ...:         ], ignore_index=True)
    ...:

In [33]: new
Out[33]:
  character  count
0         c    210
1         e    189
2         a    104
3    others     60

or bit less idiomatic solution:

In [16]: new = df.nlargest(3, columns='count')

In [17]: new.loc[len(new)] = ['others', df.drop(new.index)['count'].sum()]

In [18]: new
Out[18]:
  character  count
2         c    210
4         e    189
0         a    104
3    others    100