Search code examples
pythonpython-3.xpython-dateutil

Adding months to a date based on a comparison where dateutil.relativedelta is greater than a specified amount of time


My goal is to add a specified amount of months to a start_date based on the length of the time period between start_date and end_date. If the period lasted more than 10 years, 3 months should be added, otherwise 1 month. The calculation has to be precise and should be able to account for the specific numbers of days in a given month, as well as for leap years. I tried to accomplish this by using dateutil.relativedelta like so:

from dateutil.relativedelta import relativedelta
from dateutil import parser

# Extract date from string
start_date = parser.parse("2050-11-01")
end_date = parser.parse("1980-11-01")

# Calculate difference
delta = relativedelta(start_date, end_date)
print(delta)

if delta > relativedelta(years=+10):
    new_end_date = start_date + relativedelta(months=3)
else:
   new_end_date = start_date + relativedelta(months=1)

print(new_end_date)

However, this gives the following output:

relativedelta(years=+70)
TypeError: '>' not supported between instances of 'relativedelta' and 'relativedelta'

Searching for the error on SO made it clear that there is no simple way to compare dateutil.relativedelta. Does anyone know of a work-around for my use case?


Solution

  • Here's my suggestion:

    • Set the (relative) period you are testing against (10 years)
    • Add it to the earliest input date
    • Check if that new date is earlier than the latest input date

    This way, you still get the relative-ness of the relativedelta, but you get two hard dates to compare:

    from dateutil.relativedelta import relativedelta
    from dateutil import parser
    
    # Extract date from string
    start_date = parser.parse("2050-11-01")
    end_date = parser.parse("1980-11-01")
    
    # Set the size of the period
    period = relativedelta(years=10)
    
    # Test if adding the period to the earliest date makes it surpass the latest date
    if min(start_date, end_date) + period <= max(start_date, end_date):
        new_end_date = start_date + relativedelta(months=3)
    else:
       new_end_date = start_date + relativedelta(months=1)
    
    print(new_end_date)
    # 2051-02-01 00:00:00
    

    In this case, start & end do not imply earlier & later, so you have to do the min and max operations.