R for time data

Dr. Peng Zhao (✉ peng.zhao@xjtlu.edu.cn)

Department of Health and Environmental Sciences
Xi’an Jiaotong-Liverpool University

1 Learning objectives

  • Convert characters into time data
  • Calculate time data with basic R functions

2 Time format

Character:

d1 <- "2/11/1962"
d1
[1] "2/11/1962"
d1 + 1

Advantages:

  • Friendly to users

Disadvantages:

  • Confusing
  • Difficult to process

Numeric:

  • Excel: 1900-01-01
  • Igor: 1904-01-01

Advantages:

  • Clear meanings
  • Easy to calculate
  • Save memory

Disadvantages:

  • Hard to read for human
  • Inconsistency

Date/time format:

d2 <- Sys.Date()
d2
[1] "2023-03-02"
t2 <- Sys.time()
t2
[1] "2023-03-02 11:22:59 CST"
d2 + 1
[1] "2023-03-03"
t2 + 1
[1] "2023-03-02 11:23:00 CST"
class(d2)
[1] "Date"
class(t2)
[1] "POSIXct" "POSIXt"
  • Clear meanings
  • Easy to calculate
  • Friendly to human

3 Date

3.1 Getting date

Paired functions: as.Date() and format().

d3 <- as.Date("2/11/1962", format="%d/%m/%Y" )
as.numeric(d3)
d3 + 2617

d4 <- as.Date( "2/11/1962", format="%m/%d/%Y" )

format(d3, '%Y %m %d')
format(d3, "%A %d %B %Y")

3.2 Calculating date

library('Epi')
data("diet")
help(diet)
str(diet)

bdat <- diet$dox[1]
format(bdat, format="%A %d %B %Y")
diet$dox2 <- format(diet$dox, format="%A %d %B %Y")

bdat + 1
max(diet$dox)
range(diet$dox)
mean(diet$dox)
median(diet$dox)
diff(range(diet$dox))
difftime(min(diet$dox), max(diet$dox), units = "weeks")

Epi::cal.yr(bdat)
diet2 <- Epi::cal.yr(diet)
str(diet2)

4 Time

4.1 Getting time

Universal functions: strptime() and format():

bd <- '1994-09-22 20:30:00'
class(bd)
bdtime <- strptime(x = bd, format = '%Y-%m-%d %H:%M:%S', tz = "Asia/Shanghai")

class(bdtime)
bdtime$wday
unclass(bdtime)

format(bdtime, format = '%d.%m.%Y')

Other functions:

date()
Sys.Date()
Sys.time()

4.2 Calculating time

bdtime + 1
bdtime2 <- strptime(
  '1995-09-01 7:30', format = '%Y-%m-%d %H:%M', 
  tz = 'Asia/Shanghai')
bdtime2 - bdtime
difftime(time1 = bdtime2, time2 = bdtime, units = 'secs')
mean(c(bdtime, bdtime2))

5 The “lubridate” package

6 The “calendR” package

7 Further readings

8 Exercises

  1. Today

    1. Which day of the year is it?

    2. Convert it as “Year, Month in English and Date, Weekday in English”, such as “2023, January 1, Sunday”.

    3. Suppose we get to know each other at 15:00, September 13, 2022. How many days have we known each other until now? How many hours? How many seconds?

    4. The anniversary is September 13, 2023. What day is it? How about in 2024? Plot a graph for September 13 of each year, with the weekday as x and the year as y.

  2. The airquality dataset

    1. Create a new column “Date”, showing the date in the format of “Year-month-day”, such as “2023-01-31”.

    2. Plot a graph for each atmospheric variable (i.e. ozone in ppb, solar radiation in W m-2, wind speed in m s-1, air temperature in °C) against the date.

    3. Create a new column “Weekday”, showing the day in the week (1 as Monday and 7 as Sunday).

    4. Calculate the mean values of each atmospheric variable for each weekday.

    5. Plot a graph for each atmospheric variable with the mean values against the weekdays.

  1. The Smart Flux data set The smartflux.zip data set was obtained from the flux observation at the Outdoor Research and Teaching Station (ORTS), XJTLU. It was supposed to be an automatic continuous measurement at a half-hourly base, saved in daily files. However, there were missing data in the records for some reason. How many data files are missing? How many records are missing?