On Thu, Sep 11, 2008 at 10:43 PM, hadley wickham <[EMAIL PROTECTED]> wrote: >> The problem is trying to force everything into Date class. If you really >> want a monthly series you probably want yearmon rather than Date >> class > > Could you give a bit more detail about why you think this is the best > way to proceed? To me, it seems like it's easiest to have one date > class which you use regardless of the underlying accuracy of the > series (obviously in a date class the lowest level of discrimination > will be a single day).
That way all the methods that operate on it can specifically be taylored to that class. Working with monthly series is trivial if they are of yearmon class since the next month can be obtained by adding 1/12 instead of some weird number of days, the number of months between two dates is just 12 * (ym2 - ym1), etc. Even in the examples given in my last post in this thread look at how much simpler the yearmon examples were than the Date based examples. Working with dates when you really want months is just much needlessly complex and the benefit of having a single Date class just seems not worth it when R's oo capability is already in place to leverage. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
