In DavisVaughan/timewarp: Group Dates

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

warp

The goal of warp is to provide tooling to group dates by a variety of periods, such as: yearly, monthly, by second, by week of the month, and more.

library(warp)

Installation

You can install the release version from CRAN with:

install.package("warp")

You can install the development version from GitHub with:

# install.packages("pak")
pak::pak("DavisVaughan/warp")

Example

One of the core functions in warp is warp_distance(), which allows you to provide a date time vector and compute the "distance" from an origin. For example, this computes the number of months from the unix epoch.

x <- as.Date("1970-01-01") + -2:2
x

warp_distance(x, period = "month")

The values that warp_distance() returns correspond to the distance from x to the origin, in units defined by the period and the width defined by every. The origin defaults to the unix epoch of 1970-01-01 00:00:00 in the time zone of x, but you can change that. In this case the distances are saying that, for example, "1970-01-02" is in the same month as the origin, and "1969-12-31" is 1 month group away.

You can also compute daily distances. Rather than grouping by 1 day, let's lump every 2 days together, starting from the default origin.

# Groups 1970-01-01 and 1970-01-02 together
warp_distance(x, period = "day", every = 2)

You will often want to set your own origin date. Let's shift it forward 1 to 1970-01-02.

origin <- as.Date("1970-01-02")
origin

# Groups 1970-01-02 and 1970-01-03 together
warp_distance(x, period = "day", every = 2, origin = origin)

Another interesting period to group by is the "mweek", i.e. the week of the month. Notice that days 1-7 of January 1970 are grouped into the same bucket. Also note that days 29-31 of December 1969 fell at the end of their corresponding month. This irregular week of size 3 is treated as the 5th week of that month, but the offset value of -1 is still the number of week buckets from the origin of 1970-01-01.

y <- as.Date("1969-12-28") + 0:14

tibble::tibble(
  y = y,
  mweek = warp_distance(y, "mweek")
)

Inspiration

The algorithm for warp_distance() was inspired by xts::endpoints().

DavisVaughan/timewarp documentation built on Nov. 3, 2023, 5:36 p.m.