R help - Sep 2008 - Truncating dates (and other date-time manipulations)

Dear all,

I've been struggling to perform common operations on dates that I need
to be able to correct draw date-time scales - in particular I need to
be able to round/truncate/ceiling dates to arbitrary precision - e.g.
to weeks, months or years (or multiples thereof).  I haven't been able
to find anything to do this in base R (trunc.Date only truncates to
sub-day units), or in the date related contributed packages I've
looked in (e.g. chron).  Have I missed something?

Otherwise, I have been working on a bottom up set of functions to make
date manipulation easier - http://gist.github.com/10238.  These should
work for all date-time types and I've tried to use a consistent naming
scheme and consistent output conventions.  This seems like a rather
small bundle of functions to be worth turning into a package, so I was
wondering if anyone would be interested in adopting them into an
existing date-time related package.

Regards,

Hadley

-- 
http://had.co.nz/

I'm wrapping boost date_time into an R package.  I'll post it up to
cran shortly.

http://www.boost.org/doc/libs/1_36_0/doc/html/date_time.html

I'm not sure if that is what you are looking for, but there are a lot
of useful utilities in this library.

-Whit


On Thu, Sep 11, 2008 at 11:02 AM, hadley wickham <h.wickham at gmail.com>
wrote:
> Dear all,
>
> I've been struggling to perform common operations on dates that I need
> to be able to correct draw date-time scales - in particular I need to
> be able to round/truncate/ceiling dates to arbitrary precision - e.g.
> to weeks, months or years (or multiples thereof).  I haven't been able
> to find anything to do this in base R (trunc.Date only truncates to
> sub-day units), or in the date related contributed packages I've
> looked in (e.g. chron).  Have I missed something?
>
> Otherwise, I have been working on a bottom up set of functions to make
> date manipulation easier - http://gist.github.com/10238.  These should
> work for all date-time types and I've tried to use a consistent naming
> scheme and consistent output conventions.  This seems like a rather
> small bundle of functions to be worth turning into a package, so I was
> wondering if anyone would be interested in adopting them into an
> existing date-time related package.
>
> Regards,
>
> Hadley
>
> --
> http://had.co.nz/
>
> ______________________________________________
> R-help at 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.
>

See ?cut.Date

In the zoo package see:
?as.yearmon
?as.yearqtr
?aggregate.zoo

and the many examples in:
?plot.zoo
?xyplot.zoo
as well as the three zoo vignettes.

Also in the xts package look at
?to.period

For regularly spaced series the tis package supports a wide
variety of time bases and can convert among them.

There are as.zoo.tis and as.tis.zoo routines in zoo and tis; also,
xts is a subclass of zoo so all these packages can work together.


On Thu, Sep 11, 2008 at 11:02 AM, hadley wickham <h.wickham at gmail.com>
wrote:
> Dear all,
>
> I've been struggling to perform common operations on dates that I need
> to be able to correct draw date-time scales - in particular I need to
> be able to round/truncate/ceiling dates to arbitrary precision - e.g.
> to weeks, months or years (or multiples thereof).  I haven't been able
> to find anything to do this in base R (trunc.Date only truncates to
> sub-day units), or in the date related contributed packages I've
> looked in (e.g. chron).  Have I missed something?
>
> Otherwise, I have been working on a bottom up set of functions to make
> date manipulation easier - http://gist.github.com/10238.  These should
> work for all date-time types and I've tried to use a consistent naming
> scheme and consistent output conventions.  This seems like a rather
> small bundle of functions to be worth turning into a package, so I was
> wondering if anyone would be interested in adopting them into an
> existing date-time related package.
>
> Regards,
>
> Hadley
>
> --
> http://had.co.nz/
>
> ______________________________________________
> R-help at 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.
>

Look at the ti (Time Index) class in package tis.  Here's some examples I
just did:
> x <- Sys.Date()
> x
[1] "2008-09-11"
> ti(x, "wsaturday")       ## a ti for the week that x falls into
[1] 20080913
class: ti
> ti(x + 1, "wsaturday") - 1  ## ti for the latest complete week 
[1] 20080906
class: ti
> as.Date(ti(x + 1, "wsaturday") - 1)  ## as a Date
[1] "2008-09-06"
> ymd(ti(x + 1, "wsaturday") - 1)  ## or just the yyyymmdd number
[1] 20080906

You can do the same for "wsunday", if you want weeks that end on
Sunday.  In
all, ti supports 7 different weekly frequencies, for weeks ending on Monday,
Tuesday, ..., Sunday.  It also has daily and business day frequencies, 14
biweekly frequencies, semimonthly, monthly, bimonthly, and three quarterly
frequencies (for quarters that end in October, November and December).
Similarly, there are 12 annual frequencies and 6 semiannual ones, and a
biannual frequency.  Also supported are hourly, minutely and secondly
frequencies.

Gabor and I have done some work towards making zoo and ti objects compatible
with each other.



"hadley wickham" <h.wickham at gmail.com>
writes:
> I've been struggling to perform common operations on dates that I need
> to be able to correct draw date-time scales - in particular I need to
> be able to round/truncate/ceiling dates to arbitrary precision - e.g.
> to weeks, months or years (or multiples thereof).  I haven't been able
> to find anything to do this in base R (trunc.Date only truncates to
> sub-day units), or in the date related contributed packages I've
> looked in (e.g. chron).  Have I missed something?
-- 
Jeff