How to convert pandas date column to numeric [duplicate]

Question

Let say I have following code

import pandas as pd
df = pd.DataFrame([[pd.to_datetime('2019-01-01').to_period('M'), pd.to_datetime('2019-01-01').to_period('M')], [pd.to_datetime('2019-01-01').to_period('M'), pd.to_datetime('2019-01-01').to_period('M')], [pd.to_datetime('2019-01-01').to_period('M'), pd.to_datetime('2019-01-01').to_period('M')]], columns = ['date1', 'date2'])
df['dat3'] = df['date1'] - df['date2']
print(df)
     date1    date2             dat3
0  2019-01  2019-01  <0 * MonthEnds>
1  2019-01  2019-01  <0 * MonthEnds>
2  2019-01  2019-01  <0 * MonthEnds>

Is there any way to have the elements of dat3 as number/integer so that I can perform numerical calculation with that column?