Partial dates

Impute year-only or year-month to a full date

SAS: partial --DTC / character dates; impute with INTNX beginning/end or a fixed mid day

Clinical data often arrive as incomplete ISO-like strings: year only (2025) or year-month (2024-04). Derivation rules usually push these to the first, middle, or last day of the known period. Month bounds matter most for year-month; year-only is the same idea at year scale.

Always follow the study imputation convention and record how far you imputed (for example CDISC --DTF) when required. Examples below use generic STD-001-* subjects and one explicit mid rule.

Quick reference

library(dplyr)
library(lubridate)

# year-only "2025" -> first / mid / last day of year
ymd(paste0(year_txt, "-01-01"))
ymd(paste0(year_txt, "-06-30"))   # common mid-year convention
ymd(paste0(year_txt, "-12-31"))

# year-month "2024-04" -> first / mid / last day of month
ymd(paste0(ym_txt, "-01"))
ymd(paste0(ym_txt, "-15"))        # common mid-month convention
(ceiling_date(ymd(paste0(ym_txt, "-01")), "month") - days(1)) |> as.Date()

R (after parsing the known part)	SAS idea
first day of month/year	impute to period start (`INTNX` beginning)
mid day (often 15 / 30 Jun)	sponsor mid-point rule
last day of month/year	impute to period end (`INTNX` end)

Traps: document the mid rule (15 vs true midpoint). ceiling_date(..., "month") alone is next month start, not month end. Missing completely stays missing. Keep an imputation flag (--DTF) when the standard requires it.

Worked examples below.

Year-month to first, mid, last

Three partial year-month values, including February in a leap year.

library(dplyr)


Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

library(lubridate)


Attaching package: 'lubridate'

The following objects are masked from 'package:base':

    date, intersect, setdiff, union

partial_ym <- data.frame(
  USUBJID = c("STD-001-0001", "STD-001-0001", "STD-001-0002"),
  DTC_YM  = c("2024-02", "2024-03", "2025-12")
)

partial_ym |>
  mutate(
    START = ymd(paste0(DTC_YM, "-01")),
    MID   = ymd(paste0(DTC_YM, "-15")),
    END   = (ceiling_date(ymd(paste0(DTC_YM, "-01")), "month") - days(1)) |>
      as.Date()
  )

       USUBJID  DTC_YM      START        MID        END
1 STD-001-0001 2024-02 2024-02-01 2024-02-15 2024-02-29
2 STD-001-0001 2024-03 2024-03-01 2024-03-15 2024-03-31
3 STD-001-0002 2025-12 2025-12-01 2025-12-15 2025-12-31

`DTC_YM`	first	mid (day 15)	last
2024-02	2024-02-01	2024-02-15	2024-02-29
2024-03	2024-03-01	2024-03-15	2024-03-31
2025-12	2025-12-01	2025-12-15	2025-12-31

First day matches SAS INTNX('month', ..., 0, 'beginning') on a real date in that month. Last day matches INTNX(..., 'end'). Day 15 is a common mid-month convention; some protocols use other mid rules. Pick one and keep it consistent.

Month end is not bare ceiling_date

ceiling_date(start, "month") is the first day of the next month. Last day of the current month is that value minus one day.

Year-only to first, mid, last

Year-only strings are easier: fixed calendar anchors.

partial_y <- data.frame(
  USUBJID = c("STD-001-0001", "STD-001-0002"),
  DTC_Y   = c("2024", "2025")
)

partial_y |>
  mutate(
    START = ymd(paste0(DTC_Y, "-01-01")),
    MID   = ymd(paste0(DTC_Y, "-06-30")),
    END   = ymd(paste0(DTC_Y, "-12-31"))
  )

       USUBJID DTC_Y      START        MID        END
1 STD-001-0001  2024 2024-01-01 2024-06-30 2024-12-31
2 STD-001-0002  2025 2025-01-01 2025-06-30 2025-12-31

`DTC_Y`	first	mid (30 Jun)	last
2024	2024-01-01	2024-06-30	2024-12-31
2025	2025-01-01	2025-06-30	2025-12-31

30 June is a common mid-year convention (half-year boundary). Some conventions use 1 July or day 183. Document the choice the same way you document day 15 for months.

SAS analogs: INTNX('year', ..., 0, 'beginning') / 'end', or a fixed mid date constructed from the year.

Choosing first vs last vs mid

Situation	Typical choice
Conservative "earliest possible" onset	first of period
Conservative "latest possible"	last of period
Neutral / midpoint rule in SAP	mid (day 15 / 30 Jun or protocol rule)

Do not invent a mid rule in production code. Read the SAP or derivation specification. When CDISC date imputation flags apply, set --DTF to show which parts were imputed (for example D when day was imputed, M when month and day were imputed). Exact flag codes follow the standard and the study metadata.

Missing or unusable partial strings stay missing after failed parses. Do not silently coerce blank --DTC to period start.