pd.DataFrame.duplicated
pandas.DataFrame.duplicated(subset=None, keep='first')
Supported Arguments : None
Example Usage
>>> @bodo.jit
... def f():
... df = pd.DataFrame({"A": [1,1,3,4], "B": [1,1,3,3]})
... return df.duplicated()
>>> f()
0 False
1 True
2 False
3 False
dtype: bool