>>> pd.Series(['foo', 'fuz', np.nan]).str.replace('f.', 'ba', regex=True) 0 bao 1 baz 2 NaN dtype: object