Group two categories as others-CodePudding

I would like to group C and D in my dataframe

   Category  count
0     A    327
1     B    20
2     C    30
3     D    302

   Category  count
0     A    327
1     B    20
2     NOT A and B    332

I can replace the value of C and D and then group but is there a better way to do so?

CodePudding user response：

You can use concat after boolean masking.

m = (df['Category'].ne('A')) & (df['Category'].ne('B'))
df = pd.concat([
    df[~m],
    pd.DataFrame({
        'Category': ['NOT A and B'],
        'count': [df[m]['count'].sum()]
    })
], ignore_index=True)

print(df):

      Category  count
0            A    327
1            B     20
2  NOT A and B    332

CodePudding user response：

Other options are:

import pandas as pd

d = pd.DataFrame({'Category': list('ABDC'), 'count': [327, 20, 30, 302]})

d['Category'] = d.Category.map({x:x for x in ['A', 'B']}).fillna('NOT A or B')
d.groupby('Category').agg({'count': sum})

Or:

d['Category'] = np.where(d.Category.isin(['A', 'B']), d.Category, 'NOT A or B')
d.groupby('Category').agg(sum)