6 date and time functions

6.0.1 Date and Time Base Functions in R

Date and time manipulation is a common task in data analysis. R provides several base functions to work with dates and times. Here are some of the most commonly used base R functions for handling date and time data:

# Get the current date
current_date <- Sys.Date()
print(current_date)

[1] "2025-12-17"

# Get the current date and time
current_datetime <- Sys.time()
print(current_datetime)

[1] "2025-12-17 15:42:18 UTC"

6.0.2 Creating Date and Time Objects

6.0.3 Year Formats

Code	Meaning	Example (`format(dt, ...)`)
`%Y`	4-digit year	`"2025"`
`%y`	2-digit year (00–99)	`"25"`
`%C`	Century (year / 100, 00–99)	`"20"` (for 2025; OS-dep.)

6.0.4 Month Formats

Code	Meaning	Example
`%m`	Month number (01–12)	`"11"`
`%b`	Abbrev. month name (locale)	`"Nov"`
`%B`	Full month name (locale)	`"November"`
`%h`	Same as `%b` (abbrev. month name)	`"Nov"`

6.0.5 Day Formats

Code	Meaning	Example
`%d`	Day of month (01–31)	`"14"`
`%e`	Day of month (1–31, padded with space)	`"14"` (or `" 7"` for 7th; OS-dep.)
`%j`	Day of year (001–366)	`"318"`

6.0.6 Weekday Formats

Code	Meaning	Example
`%a`	Abbrev. weekday name (locale)	`"Fri"`
`%A`	Full weekday name (locale)	`"Friday"`
`%w`	Weekday number (0–6, 0 = Sunday)	`"5"`
`%u`	ISO weekday number (1–7, 1 = Monday) (OS-dep.)	`"5"` (Friday)

6.0.7 Time Formats

Code	Meaning	Example
`%H`	Hour (00–23)	`"14"`
`%I`	Hour (01–12)	`"02"`
`%M`	Minute (00–59)	`"30"`
`%S`	Second (00–61, allows for leap)	`"05"`
`%p`	AM/PM indicator (locale)	`"PM"`

###Locale-dependent “full date/time” codes

###These depend on your system locale (language + region):

Code	Meaning	Example
`%c`	Date and time representation	`"Fri Nov 14 16:05:30 2025"`
`%x`	Date representation	`"11/14/25"` (varies by locale)
`%X`	Time representation	`"16:05:30"`

# Create a Date object
dt <- as.Date("2025-11-14")
# Create a POSIXct object (date-time)
dt <- as.POSIXct("2025-11-14 16:05:30")
# Format examples
format(dt, "%Y-%m-%d")  # e.g. "2025-11-14"

[1] "2025-11-14"

format(dt, "%m/%d/%y")  # e.g. "11/14/25"

[1] "11/14/25"

format(dt, "%B %d, %Y") # e.g. "November 14, 2025"

[1] "November 14, 2025"

format(dt, "%c")  # e.g. "Fri Nov 14 16:05:30 2025"

[1] "Fri Nov 14 16:05:30 2025"

format(dt, "%x")  # e.g. "11/14/25" (depends on locale)

[1] "11/14/25"

format(dt, "%X")  # e.g. "16:05:30"

[1] "16:05:30"

SAS informat	R parsing
`ddmmyy10.`	`as.Date(x, "%d/%m/%Y")`
`ddmmyyd10.`	`as.Date(x, "%d-%m-%Y")`
`ddmmyyp10.`	`as.Date(x, "%d.%m.%Y")`
`mmddyy10.`	`as.Date(x, "%m/%d/%Y")`
`date9.`	`as.Date(x, "%d%b%Y")`
`julian7.`	`as.Date(strptime(x, "%Y%j"))`
`time8.`	`as.POSIXct(x, "%H:%M:%S")`
`time10.`	`as.POSIXct(x, "%I:%M:%S%p")` (if am/pm)
`datetime18.`	`as.POSIXct(x, "%d%b%Y:%H:%M:%S")`
`datetime20.`	`as.POSIXct(x, "%d%b%Y:%I:%M:%S%p")`

SAS format	Meaning	R equivalent
`worddate18.`	Month date, year (e.g., February 23, 2003)	`format(d, "%B %d, %Y")`
`weekdate24.`	Weekday, month date, year	`format(d, "%A, %B %d, %Y")`
`year.`	numeric year	`as.integer(format(d, "%Y"))`
`month.`	month number (1–12)	`as.integer(format(d, "%m"))`
`day.`	day of month	`as.integer(format(d, "%d"))`
`weekday.`	day of week (1–7)	`as.integer(format(d, "%u"))` (Mon = 1)

# Convert a character string to a Date object
date_string <- "2023-10-01"
date_object <- as.Date(date_string)
print(date_object)

[1] "2023-10-01"

# Convert a character string to a POSIXct object (date-time)
datetime_string <- "2023-10-01 12:34:56"
datetime_object <- as.POSIXct(datetime_string)
print(datetime_object)

[1] "2023-10-01 12:34:56 UTC"

# Extract components from a Date object
year <- format(date_object, "%Y")
month <- format(date_object, "%m")
day <- format(date_object, "%d")
print(paste("Year:", year, "Month:", month, "Day:", day))

[1] "Year: 2023 Month: 10 Day: 01"

# Extract components from a POSIXct object
hour <- format(datetime_object, "%H")
minute <- format(datetime_object, "%M")
second <- format(datetime_object, "%S")
print(paste("Hour:", hour, "Minute:", minute, "Second:", second))

[1] "Hour: 12 Minute: 34 Second: 56"

# Perform date arithmetic
tomorrow <- current_date + 1
yesterday <- current_date - 1
print(paste("Tomorrow:", tomorrow, "Yesterday:", yesterday))

[1] "Tomorrow: 2025-12-18 Yesterday: 2025-12-16"

# Calculate the difference between two dates
date_diff <- as.numeric(difftime(current_date, date_object, units = "days"))
print(paste("Difference in days:", date_diff))

[1] "Difference in days: 808"

# Format a Date object to a specific string format
formatted_date <- format(current_date, "%B %d, %Y")
print(formatted_date)

[1] "December 17, 2025"

# Format a POSIXct object to a specific string format
formatted_datetime <- format(current_datetime, "%Y-%m-%d %H:%M:%S")
print(formatted_datetime)

[1] "2025-12-17 15:42:18"

These functions allow you to create, manipulate, and format date and time data in R. For more advanced date and time handling, you might consider using the lubridate package, which provides a more user-friendly interface for working with dates and times.

A practical, CDISC‑friendly guide to working with dates/times and implementing partial date imputation in R using base R and lubridate.

6.1 1. Date & Time Basics in R

6.1.1 1.1 Core classes

Date: calendar dates without time.
POSIXct: seconds since 1970‑01‑01 (compact, good for storage/arithmetic).
POSIXlt: list of components (year, mon, mday, hour, min, sec) — convenient to extract parts.

6.1.2 1.2 Converting numerics to Date (origins matter)

Different systems use different origins for numeric dates: - Excel: 1900‑01‑01 (beware 1900 leap‑year bug in older Excels) - SAS: 1960‑01‑01 - R: 1970‑01‑01

num_dates <- c(0, 159, 464, 15674)

as.Date(num_dates, origin = "1960-01-01")  # SAS origin

[1] "1960-01-01" "1960-06-08" "1961-04-09" "2002-11-30"

as.Date(num_dates, origin = "1970-01-01")  # R origin

[1] "1970-01-01" "1970-06-09" "1971-04-10" "2012-11-30"

as.Date(num_dates, origin = "1900-01-01")  # Excel origin

[1] "1900-01-01" "1900-06-09" "1901-04-10" "1942-12-01"

6.1.3 1.3 Custom date formats

x1 <- c("01/05/1965", "08/16/1975")
as.Date(x1, "%m/%d/%Y")

x2 <- c("05-01-1965", "16-08-1975")
as.Date(x2, "%d-%m-%Y")

x3 <- c("01MAY1965", "16AUG1975")
as.Date(x3, "%d%b%Y")

# Unknown format? Try multiple patterns
x4 <- c("01MAY1965MORNING", "16AUGUST1975")
as.Date(x4, tryFormats = c("%d%b%Y", "%d%B%Y"))

6.1.4 1.4 Datetime strings (ISO 8601 & time zones)

x <- c("2015-10-19T10:15", "2018-12-19T23:05")
# Keep the literal 'T' in the format string
as.POSIXct(x, format = "%Y-%m-%dT%H:%M", tz = "UTC")

# Extract parts
xt <- as.POSIXlt(x, format = "%Y-%m-%dT%H:%M")
format(xt, "%Y-%m-%d")  # date part
format(xt, "%H:%M")     # time part

6.2 2. Lubridate: Fast, Friendly Parsing & Arithmetic

Install & load: install.packages("lubridate"); library(lubridate)

6.2.1 2.1 Intuitive parsers

ymd(), mdy(), dmy() for dates
ymd_hms(), ymd_hm(), ymd_h() for date‑times

ymd("20210529")
dmy("29/05/2021")
ymd_hm("2019/01/19T17:15")  # delimiter agnostic

6.2.2 2.2 Accessors and rounding

x <- ymd_hms("2009-08-03 12:01:59")
year(x); month(x); mday(x); wday(x); hour(x); minute(x); second(x)

floor_date(x, "month")   # 2009-08-01
ceiling_date(x, "month") # 2009-09-01
round_date(x, "month")    # 2009-08-01
rollback(x)               # last day of previous month

6.2.3 2.3 Timespans

Durations: exact seconds, e.g., ddays(2)
Periods: human units, e.g., months(1)
Intervals: start–end pair, e.g., interval(start, end)

start <- ymd("2015-10-19")
end   <- ymd("2018-12-19")
int   <- interval(start, end)
as.duration(int)       # exact seconds
as.period(int)         # years/months/days

6.2.4 3 ADaM‑Style Partial Date Imputation (Start & End)

Imputation rules vary by study; always follow the SAP. A common pattern:

6.2.5 Start date (`AESTDTC`)

YYYY-MM-DD → use as is
YYYY-MM → impute day = 01 (first of month)
YYYY → impute month = 01, day = 01 (01-Jan)

6.2.6 End date (`AEENDTC`)

YYYY-MM-DD → use as is
YYYY-MM → impute last day of that month
YYYY → impute 31-Dec

6.2.7 Censoring by death / last alive date

Additionally, bound AEENDT by death / last alive date (when present):

CENSOR = coalesce(DTHDT, LASTALVDT)
AEENDT = min(imputed_end, CENSOR)

6.2.8 Imputation flags

Add flags:

AESTDTF, AEENDTF:
- "D" → day imputed
- "M,D" → month & day imputed
Add "C" if we had to censor the end date at DTHDT / LASTALVDT.

library(dplyr, warn.conflicts = FALSE)
library(stringr, warn.conflicts = FALSE)
library(lubridate, warn.conflicts = FALSE)

df <- data.frame(
  USUBJID    = sprintf("01-001-%03d", 1:6),
  AESTDTC    = c("2015-10-19", "2015-10", "2015", "2019-02", "2020-02", "2018-12-31"),
  AEENDTC    = c("2018-12-19", "2018-10", "2017-05-29", "2019-02", "2020-02", "2020"),
  RANDDT     = ymd(c("2015-10-01", "2015-10-01", "2015-01-10", "2019-01-15", "2020-02-10", "2018-12-01")),
  DTHDT      = ymd(c(NA, NA, NA, NA, "2020-08-03", NA)),
  LASTALVDT  = ymd(c(NA, NA, NA, NA, NA, "2020-12-15"))
)

df_imp <- df |> 
  mutate(
    # Length of partial date strings
    AESTDTC_LEN = str_length(AESTDTC),
    AEENDTC_LEN = str_length(AEENDTC),
    ## --- AE start date imputation (minimum: first possible date) ---
    AESTDT = case_when(
      is.na(AESTDTC) ~ as.Date(NA),
      AESTDTC_LEN == 10 ~ ymd(AESTDTC, quiet = TRUE),
      AESTDTC_LEN == 7  ~ ymd(str_c(AESTDTC, "-01"), quiet = TRUE),
      AESTDTC_LEN == 4  ~ ymd(str_c(AESTDTC, "-01-01"), quiet = TRUE),
      TRUE ~ as.Date(NA)
    ),
    AESTDTF = case_when(
      is.na(AESTDTC) ~ NA_character_,
      AESTDTC_LEN == 10 ~ "",
      AESTDTC_LEN == 7  ~ "D",
      AESTDTC_LEN == 4  ~ "M,D",
      TRUE ~ "UNK"
    ),
    ## --- AE end date imputation (maximum: latest possible date) ---
    AEENDT_RAW = case_when(
      is.na(AEENDTC) ~ as.Date(NA),
      AEENDTC_LEN == 10 ~ ymd(AEENDTC, quiet = TRUE),
      AEENDTC_LEN == 7  ~ ymd(str_c(AEENDTC, "-01"), quiet = TRUE),
      AEENDTC_LEN == 4  ~ ymd(str_c(AEENDTC, "-12-31"), quiet = TRUE),
      TRUE ~ as.Date(NA)
    ),
    AEENDTF_RAW = case_when(
      is.na(AEENDTC) ~ NA_character_,
      AEENDTC_LEN == 10 ~ "",
      AEENDTC_LEN == 7  ~ "D",
      AEENDTC_LEN == 4  ~ "M,D",
      TRUE ~ "UNK"
    )
  ) %>%
  mutate(
    AEENDT_RAW = if_else(
      AEENDTC_LEN == 7 & !is.na(AEENDT_RAW),
      ceiling_date(AEENDT_RAW, "month") - days(1),
      AEENDT_RAW
    ),
    CENSOR = coalesce(DTHDT, LASTALVDT),
    AEENDT = case_when(
      !is.na(CENSOR) & !is.na(AEENDT_RAW) & AEENDT_RAW > CENSOR ~ CENSOR,
      TRUE ~ AEENDT_RAW
    ),
    AEENDTF = case_when(
      is.na(AEENDT) ~ AEENDTF_RAW,
      !is.na(CENSOR) & !is.na(AEENDT_RAW) & AEENDT_RAW > CENSOR & AEENDTF_RAW == ""  ~ "C",
      !is.na(CENSOR) & !is.na(AEENDT_RAW) & AEENDT_RAW > CENSOR & AEENDTF_RAW != "" ~ str_c(AEENDTF_RAW, ",C"),
      TRUE ~ AEENDTF_RAW
    )
  )

6.3 4. Durations & Analysis Variables

6.3.1 4.1 AE duration (AEDUR) in days

7 Using difftime (base) or as.numeric on Date differences

AEDUR <- as.integer(out$AEENDT - out$ASTDT)

7.0.1 6.2 Handy lubridate parsers (subset)

Dates: ymd(), mdy(), dmy(), yq()
Datetimes: ymd_hms(), ymd_hm(), ymd_h()
Times: hms(), hm(), ms()
Accessors: year(), month(), mday(), wday(), hour(), minute(), second(), date()
Rounding: floor_date(), ceiling_date(), round_date(), rollback()