可以使用与groupby()和apply()类似的语法来完成此操作:

It can be done with similar syntax with groupby() and apply():

df['ratio'] = df.groupby(['a','b'], group_keys=False).apply(lambda g: g.c/(g.c * g.d).sum())