How to compare two date columns in a dataframe using pandas?

There's a easiest way.

You don't have to convert them with datetime. You can do it directly with pandas.

There's a api available with pandas as to_datetime

To substract between two days, first convert the string to datetime.

df3['dt_end_date'] = pd.to_datetime(df3['end_date']) df3['dt_start_date'] = pd.to_datetime(df3['start_date']) df3['difference'] = df3['dt_end_date'] - df3['dt_start_date'] 

That's it you will have the answer !!!

I think you're confusing how to use it dataframe here.

If you consider a normal value, you can use datetime module but you should use to_datetime() to convert a type of columne and dt attribute to count days.

For your goal, you can make another columne for diff to save timedelta, then you can check if there is negative value in the diff columns, as follows:

import pandas as pd df3 = pd.DataFrame({ 'start_date': ['2020-01-09', '2020-02-02'], 'end_date': ['2020-01-11', '2020-02-01'] }) df3['end_date'] = pd.to_datetime(df3['end_date'], format='%Y-%m-%d') df3['start_date'] = pd.to_datetime(df3['start_date'], format='%Y-%m-%d') df3['diff'] = df3['end_date'].dt.day - df3['start_date'].dt.day print(df3) # start_date end_date diff #0 2020-01-09 2020-01-11 2 #1 2020-02-02 2020-02-01 -1 if any(val < 0 for val in df3['diff'].values): # to check if negative value exists raise Exception("employee's leave 'start_date' cannot be greater than leave 'end_date'") 

You can raise Exception if there is any negative days, as follows:

Traceback (most recent call last): File "C:/Users/ABC/stackoverflow/main.py", line 16, in <module> raise Exception("employee's leave 'start_date' cannot be greater than leave 'end_date'") Exception: employee's leave 'start_date' cannot be greater than leave 'end_date'