policz, ile duplikatów Python Pandy

df.duplicated(subset='one', keep='first').sum()
Trustworthy Whale