datetime subtraction in dict for loop

I have multiple data-frames where I need to loop start and end dates. The end dates are fixed; however, the start dates would varies between 1 to 10 years. For the sample below, all start dates are 6 years prior to the end dates. Therefore, is there a way to coding it where all k[i] are 6 years prior to l[i] instead of manually entering them for each df? I've tried using l[i] - 6, however, i got an error.

start dates:
k = {}
k[0]='2009-08-31'
k[1]='2010-08-31'
k[2]='2011-08-31'

end date:
l = {}
l[0]='2015-07-31'
l[1]='2016-07-31'
l[2]='2017-07-31'

t_l = len(l)

MAX_NUMBER = t_l
for i in range(MAX_NUMBER):
    df = df0.copy()
    tr_sta = k[i]
    tr_end = l[i]
    
    df = df[(df['date'] > tr_sta)]
    df = df[(df['date'] < tr_end)]

thanks in advance

CodePudding user response：

Let's say you first have a dataframe only with end dates

df = pd.DataFrame({'end date': [val for key, val in l.items()]})

Then, define start dates by using pd.offsets.DateOffset:

df['end date'] = pd.to_datetime(df['end date'], format='%Y-%m-%d')
df['start date'] = df['end date']   pd.offsets.DateOffset(years=6)

By the way, your start dates are actually 6 years prior and 1 month forward of the end dates. So you may also pass

df['start date'] = df['end date']   pd.offsets.DateOffset(years=6, months=1)