The error occurs because the astype('timedelta64[D]') step is unnecessary and incorrectly handled by pandas. Instead use the following code: import pandas as pd import numpy as np

Example DataFrame

df = pd.DataFrame({'designation_date': ['2021-01-01', '2021-01-02']})

Calculate recency

df['recency'] = pd.Timestamp('today') - pd.to_datetime(df['designation_date'])

Convert to days

df['recency'] = df['recency'] / np.timedelta64(1, 'D') print(df)