Lets say I have this data

import pandas as pd

data = {
    'country':['USA', 'China', 'Japan', 'Germany', 'UK', 'India', 'USA', 'India']
    ,'foo':[1, 2, 3, 4, 5, 6, 7, 8]
    ,'bar':[11, 22, 33, 44, 55, 66, 77, 88]
}

df = pd.DataFrame(data, columns = ['country', 'GDP', 'population'])

With df.groupby('country').agg(['count']) I get the count(), groupby the country.

          GDP population
        count      count
country
China       1          1
Germany     1          1
India       2          2
Japan       1          1
UK          1          1
USA         2          2

How would I select the row with the max() value in the column bar or in the column foo? How to get the two max() (here India and USA) values back?

CodePudding user response：

Try:

# find the rows with a maximum either in foo or bar
mask = (counts == counts.values.max(0)).any(1)
res = counts[mask]
print(res)

Output

        foo        bar
        count      count
country                 
India       2          2
USA         2          2