As we all know, dates can be a harrowing/boarderline traumatic experience. Here I give a condensed and hopefully rapid overview of dates in R.
There are multiple date classes in R. The simplest is the Date class, but there are also POSIXt dates and a relatively recent Lubridate package.
Date Class
These are the `easy’ ones to deal with and are stored internally as integers from 1970-01-01.
as.Date("2017-01-12")
as.Date("2017-24-01", format = "%Y-%d-%m")
# Alternative origin - i.e. Excel
as.Date(400, origin = "1900-01-01")
difftime(as.Date("2018-01-01"), as.Date("2017-01-01"), units = "days")
POSIXt - datetime variants
Portable Operating System Interface (POSIX) is an interoperability standard (that not many people seem to have much faith in). In R there is POSIXlt and POSIXct for local and for calendar time respecitively. The former is a list, the latter is a number (seconds since origin):
> unclass(as.POSIXlt(Sys.time()))
$sec
[1] 30.34591
$min
[1] 35
$hour
[1] 12
$mday
[1] 25
$mon
[1] 3
$year
[1] 118
$wday
[1] 3
$yday
[1] 114
$isdst
[1] 0
$zone
[1] "AEST"
$gmtoff
[1] 36000
attr(,"tzone")
[1] "" "AEST" "AEDT"
# POSIXct
unclass(Sys.time())
[1] 1524623797
You can display (or store) dates in different formats by using the format command or option:
as.POSIXct("101221 10:22", format = "%y%m%d %H:%M")
[1] "2010-12-21 10:22:00 AEDT"
as.character(as.POSIXct("101221 10:22",
format = "%y%m%d %H:%M"), format = "%m-%d-%y %H:%M")
[1] "12-21-10 10:22"
Data frames do weird things to dates. Specifically, when you store a POSIXlt variable in a data.frame it is converted to a POSIXct. However, if you assign a POSIXlt to a data.frame field using $ then the lt class is retained!
hotdate <- "20081101 01:20:00"
ct <- as.POSIXct(hotdate, format = "%Y%m%d %H:%M:%S")
lt <- as.POSIXlt(hotdate, format = "%Y%m%d %H:%M:%S")
df <- data.frame(orig = hotdate,
ct = ct,
lt = lt)
# POSIXlt converted to ct
> str(df)
'data.frame': 1 obs. of 3 variables:
$ orig: Factor w/ 1 level "20081101 01:20:00": 1
$ ct : POSIXct, format: "2008-11-01 01:20:00"
$ lt : POSIXct, format: "2008-11-01 01:20:00"
# POSIXlt NOT converted to ct!!!!
df$lt2 <- as.POSIXlt(hotdate, format = "%Y%m%d %H:%M:%S")
str(df)
'data.frame': 1 obs. of 4 variables:
$ orig: Factor w/ 1 level "20081101 01:20:00": 1
$ ct : POSIXct, format: "2008-11-01 01:20:00"
$ lt : POSIXct, format: "2008-11-01 01:20:00"
$ lt2 : POSIXlt, format: "2008-11-01 01:20:00"
# Rounding
# NOPE
df[, "ct"] <- round(df[, "ct"], units = "hours")
# FORCE
df[, "ct"] <- as.POSIXct(round(ct, units = "hours"))
# NAH
df[, "lt"] <- round(lt, units = "hours")
Warning message:
In `[<-.data.frame`(`*tmp*`, , "lt",
value = list(sec = 0, min = 0L, :
provided 11 variables to replace 1 variables
# MAGIC via $ !
df$lt <- round(lt, units = "hours")
Lubridate
Makes life a bit easier (sometimes) and represents time via different objects. Important concepts are:
- instants
- intervals
- durations
- periods
ymd_hms("2012-12-31 23:59:59")
## [1] "2012-12-31 23:59:59 UTC"
month(ldate) <- 8
# An _instant_ in time:
lubridate::now()
# Last day of the current month
lubridate::ceiling_date(now(), unit = "month") - lubridate::days(1)
# The time between two instants is an _interval_
# Interval constructed using _duration_ i.e. dyears
# dyears record the exact number of seconds in a time span and deals with
# leap years, daylight savings etc so you need to be careful in its use.
lubridate::ymd_hms(hotdate) - lubridate::dyears(1)
# versus
# Interval constructed using _period_ i.e. years
lubridate::ymd_hms(hotdate) - lubridate::years(1)
# Nice
now() %within% interval(ymd("2012-07-01"), ymd("2019-06-30"))
Painless?