← SASnRAdd an Interval to a Date

Introduction

Clinical protocols often define time points relative to a starting date — visit windows, cutoff dates, follow-up milestones.
Adding days is straightforward arithmetic, but adding months or years requires calendar awareness because months vary in length.
SAS uses the INTNX function with an alignment option that controls where the result lands within the target interval.
R tidyverse uses lubridate arithmetic: days(), months(), and the rollback-safe %m+% operator.
This lesson adds day, month, and year offsets to protocol-defined dates and compares how each language handles end-of-month edge cases.

Add an Interval to a Date

SAS (Base SAS)+

*=============================================================================;
* Add an interval to a date using INTNX
*=============================================================================;

*-----------------------------------------------------------------------------;
* Create sample visit schedule data
*-----------------------------------------------------------------------------;
data visits;
    input usubjid $ visit_dt :date9.;
    format visit_dt date9.;
    datalines;
S001 15JAN2025
S002 28FEB2025
S003 31MAR2025
S004 30APR2025
;
run;

*-----------------------------------------------------------------------------;
* Add various intervals using INTNX
*-----------------------------------------------------------------------------;
data visit_windows;
    set visits;

    * Add 30 days for next visit window;
    next_30d = intnx('DAY', visit_dt, 30);

    * Add 3 months - beginning of that month;
    month3_b = intnx('MONTH', visit_dt, 3, 'B');

    * Add 3 months - end of that month;
    month3_e = intnx('MONTH', visit_dt, 3, 'E');

    * Add 3 months - same day of month;
    month3_s = intnx('MONTH', visit_dt, 3, 'S');

    * Add 1 year - same day;
    year1_s = intnx('YEAR', visit_dt, 1, 'S');

    format next_30d month3_b month3_e month3_s year1_s date9.;
run;

proc print data=visit_windows noobs;
run;

We create four subjects with visit dates chosen to show edge cases: mid-month, end of February, end of March (31st), and end of April (30th).
intnx('DAY', visit_dt, 30) simply adds 30 calendar days — alignment does not matter for days.
The fourth argument of INTNX controls alignment within the target month: 'B' = beginning (1st), 'E' = end (last day), 'S' = same day, 'M' = middle (16th).
intnx('MONTH', visit_dt, 3, 'B') moves 3 months forward, then snaps to the 1st of that month — useful for reporting-period starts.
intnx('MONTH', visit_dt, 3, 'E') snaps to the last day of the target month — useful for reporting-period ends.
intnx('MONTH', visit_dt, 3, 'S') tries to keep the same day of month; if that day does not exist (e.g., 31st in a 30-day month), SAS rolls back to the last valid day.
intnx('YEAR', visit_dt, 1, 'S') adds one year keeping the same day; Feb 28 in a non-leap year stays Feb 28.
We run PROC PRINT and check subject S002 (28FEB2025): adding 3 months with 'S' alignment should give 28MAY2025.

R (tidyverse)+

# =============================================================================
# Add an interval to a date using lubridate
# =============================================================================
library(tidyverse)
library(lubridate)

# -----------------------------------------------------------------------------
# Create sample visit schedule data
# -----------------------------------------------------------------------------
visits <- tribble(
  ~usubjid, ~visit_dt,
  "S001",   "2025-01-15",
  "S002",   "2025-02-28",
  "S003",   "2025-03-31",
  "S004",   "2025-04-30"
) %>%
  mutate(visit_dt = ymd(visit_dt))

# -----------------------------------------------------------------------------
# Add various intervals using lubridate
# -----------------------------------------------------------------------------
visit_windows <- visits %>%
  mutate(
    # Add 30 days for next visit window
    next_30d = visit_dt + days(30),

    # Add 3 months - beginning of that month
    month3_b = floor_date(visit_dt %m+% months(3), unit = "month"),

    # Add 3 months - end of that month
    month3_e = ceiling_date(visit_dt %m+% months(3), unit = "month") - days(1),

    # Add 3 months - same day (rollback-safe)
    month3_s = visit_dt %m+% months(3),

    # Add 1 year - same day
    year1_s = visit_dt %m+% years(1)
  )

print(visit_windows)

We parse character dates into Date objects with ymd() and then use lubridate period arithmetic.
visit_dt + days(30) adds exactly 30 calendar days — identical to the SAS INTNX DAY result.
The %m+% operator adds months while handling end-of-month rollover safely; if the target day does not exist, it rolls back to the last valid day of that month.
To get the beginning of the target month (matching SAS 'B' alignment), we apply floor_date(..., unit = "month") after the month addition.
To get the end of the target month (matching SAS 'E' alignment), we use ceiling_date() to get the 1st of the next month, then subtract one day.
visit_dt %m+% months(3) on its own matches SAS 'S' alignment — same day of month, rolled back if needed.
visit_dt %m+% years(1) adds one year with the same rollback-safe behavior for leap-year edge cases.
We print the result and check subject S003 (March 31): adding 3 months with %m+% should give June 30, since June has only 30 days.

Happy learning!

Search lessons

Introduction

Add an Interval to a Date