pivot groupby 和 sum pandas dataframe

Question

This should be simple, but i'm new to working in python. Any suggestions please?

#original dataframe
df = pd.DataFrame({'year':[1,1,1,1,1],
                   'month':[4,4,4,4,4],
                   'mode': ['a','b','a','a','b']},
                  columns=['year','month','mode'])

#pivot/groupby etc
# df2=df.pivot(columns=('year','month'), values=('mode')).count()

#create this dataframe
df2 = pd.DataFrame({'year':[1],
                   'month':[4],
                   'a': [3],
                   'b':[2]},
                  columns=['year','month','a','b'])

I'm working in Koalas Apache Spark environment ( documentation ), so solution should work on it.

Answer 1

df.pivot_table(index=['year','month'], aggfunc='size', columns='mode')

Answer 2

Alternatively You can use pd.get_dummies() :

pd.get_dummies(df).groupby(['year','month']).sum()

result:

            mode_a  mode_b
year month                
1    4           3       2

Note: I'm not sure that works in Koalas Apache Spark environment.