Date Comparison and Manipulation Functions?

[norseman]

W. eWatson wrote:
> I just tried the following code, and got an unexpected result.
> 
> from pyfdate import *
> t = Time()
> 
> ts = Time(2008, 8, 29,15,20,7)
> tnew = ts.plus(months=6)
> print "new date: ", tnew
> 
> Result:
> new date:  2009-02-28 15:20:07
> 
> I believe that should be April 1, 2009. If I use months = 1 and day =31, 
> I get Sept. 30, 2008 and not Oct. 1, 2008. Is there a way to get around 
> this?
> 

===================
Hummmm....


The old 'ditty' to help remember:
Thirty days have September, April, June and November
All the rest have 31 except February
Which has but 28 'til Leap Year gives it 29

and:
2008 08 29  DOY = 242    366 in year
2009 03 01  DOY =  60


366-242 = 124 + 60 = 184 - 1 (we don't count today) = 183
30 + 31 + 30 + 31 + 31 + 28 + 2 (left in aug.)      = 183
S    O    N    D    J    F    A

366+60-242-1(for today)=183

Puts us on 2009 Feb. 28

case #2: I think something is described other than as used.

08 29
01 31
-----
09 60

2008 09 01  DOY = 245
245 + 60 -1(today) = 304 which is DOY for 2008 10 30

or (depending on what the written meant)

08 31
01 00
-----
09 31

2008 09 00  DOY = 245
245 + 31 -1(today) = 274 =  2008 Sept. 30

pyfdate is using actual days per month and DOY to compute.
It's not quite the same as the 'rule of thumb' that is in general public 
use - but it is accurate. Most of us would just add the months and use 
the day or one close to it.




Steve
norseman at hughes.net