almanac provides tools for working with recurrence rules, the fundamental building blocks used to identify calendar “events”, such as weekends or holidays.
Install the released version of almanac from CRAN with:
Install the development version from GitHub with:
Mac and Windows users should not have any problems installing almanac. Linux users need libv8 to install the dependency R package, V8. See the V8 installation instructions for more information. almanac uses ES5 JavaScript, so it does not require any “modern” JavaScript features and should work with the “old” V8 engine provided by Ubuntu versions before 19.04.
Constructing recurrence rules looks like this:
# Thanksgiving = "The fourth Thursday in November"
on_thanksgiving <- yearly() %>%
recur_on_ymonth("November") %>%
recur_on_wday("Thursday", nth = 4)
on_thanksgiving
#> <rrule[yearly / 1900-01-01 / 2100-01-01]>
#> - ymonth: Nov
#> - wday: Thu[4]After constructing a recurrence rule, it can be used to generate dates that are in the “event set”. For example, you can search for all Thanksgivings between 2000-2006.
alma_search("2000-01-01", "2006-12-31", on_thanksgiving)
#> [1] "2000-11-23" "2001-11-22" "2002-11-28" "2003-11-27" "2004-11-25"
#> [6] "2005-11-24" "2006-11-23"Determine if a particular date is a part of the event set with alma_in().
You can also shift an existing sequence of dates, “stepping over” dates that are part of the event set.
wednesday_before_thanksgiving <- as.Date("2000-11-22")
# Thanksgiving was on 2000-11-23.
# This steps over Thanksgiving to 2000-11-24.
# Then steps 1 more day to 2000-11-25.
alma_step(wednesday_before_thanksgiving, n = 2, on_thanksgiving)
#> [1] "2000-11-25"There is an additional “stepper” object you can create for more intuitive stepping. Combine it with %s+% to perform the same step done by alma_step(). Create a stepper function with stepper(), and then use it by supplying the number of days to step.
step_over_thanksgiving <- stepper(on_thanksgiving)
wednesday_before_thanksgiving %s+% step_over_thanksgiving(2)
#> [1] "2000-11-25"The above example just scratches the surface of what almanac can do. Practically speaking, you’ll probably have multiple holidays and events that you’d like to combine into one big recurrence object. This is known as a recurrence bundle.
This example creates recurrence rules for weekends and Christmas, and bundles them together along with the Thanksgiving rule in such a way that we get the union of the underlying event sets.
on_weekends <- weekly() %>%
recur_on_weekends()
on_christmas <- yearly() %>%
recur_on_mday(25) %>%
recur_on_ymonth("Dec")
bundle <- runion() %>%
add_rschedule(on_weekends) %>%
add_rschedule(on_christmas) %>%
add_rschedule(on_thanksgiving)
bundle
#> <runion[3 rschedules / 0 rdates / 0 exdates]>We can create a stepper that steps over all of the events in the bundle. If these two holidays were the only ones that your company celebrated, the stepper could be viewed as a way to step forward by a “business day”.
For example, Christmas was on a Monday in 2006. If you wanted to step 1 business day forward from the Friday before Christmas, you’d probably like it to step over the weekend and the Christmas Monday to Tuesday. The bundle lets you do exactly that!
business_day <- stepper(bundle)
# Christmas was on a Monday in 2006.
# This is the Friday before Christmas
friday <- as.Date("2006-12-22")
# Step forward 1 business day, going over the weekend and Christmas
friday %s+% business_day(1)
#> [1] "2006-12-26"View the vignettes on the website to learn more about how to use almanac.
vignette("almanac")
vignette("adjust-and-shift")
vignette("quarterly")
vignette("icalendar")
almanac has developed as a composite of ideas from multiple different libraries.
First off, it directly embeds the amazing JavaScript library rrule for the core event set calculations. To do this, it uses the equally awesome R package, V8, from Jeroen Ooms.
The date shifting / adjusting functions are modeled after similar functions in QuantLib.
The fast binary search based implementations of alma_next() and alma_step() are inspired by Pandas and the implementation of Numpy’s busday_offset().
The author of gs, James Laird-Smith, has been a great collaborator as we have bounced ideas off of each other. gs attempts to solve a similar problem, but with a slightly different implementation.