lubridate days_in_month() in R

The lubridate days_in_month() function returns how many calendar days a date's month contains, a value from 28 to 31, with February corrected automatically for leap years.

days_in_month(ymd("2024-02-10"))         # 29 (leap-year Feb)
days_in_month(ymd("2023-02-10"))         # 28 (non-leap Feb)
days_in_month(Sys.Date())                # days in the current month
days_in_month(c(d1, d2, d3))             # vectorized over many dates
unname(days_in_month(d))                 # drop the month-name label
ceiling_date(d, "month") - days(1)       # last calendar day as a Date
mday(d) == days_in_month(d)              # TRUE when d is month-end

What days_in_month() does in one sentence

days_in_month() counts the calendar days in a date's month. You pass it a Date or date-time value and it returns how many days that month holds, anywhere from 28 to 31. The function reads only the month and year of the input, never the day component, so any date inside February 2024 returns the same answer.

The return value is a named integer vector. A single date gives one number labelled with the month's three-letter abbreviation, which is why printed output shows a header like Feb above the count. February is the only month whose length changes, and days_in_month() resolves that by checking whether the input's year is a leap year.

Syntax

days_in_month() takes exactly one argument. The signature is days_in_month(x), where x is the date object whose month you want measured. There are no extra options, no na.rm flag, and no format string to remember.

The argument x accepts several date classes:

• Date objects, the most common input, created by ymd() or as.Date().
• POSIXct and POSIXlt date-times, including values carrying a time zone.
• Period objects, where the month slot drives the result.

The output is an integer vector the same length as x. Each element is named with the month abbreviation of the matching input. An NA date propagates to an NA count rather than throwing an error.

Five common patterns

Most real uses of days_in_month() fall into a handful of shapes. Load lubridate first, then start with the leap-year contrast that makes the function worth calling at all.

Both dates sit in February, but 2024 is a leap year and 2023 is not, so the counts differ. You never hardcode 28 when days_in_month() already knows the rule.

The function is vectorized, so you can measure a whole sequence of months at once.

A frequent goal is the last calendar day of a month. Build it from the first of the month plus the day count, or snap forward with ceiling_date().

The day count also normalizes monthly figures into daily ones, which matters for prorating rent, salary, or subscription charges.

Finally, the count gives a clean test for whether a date is the last day of its month.

days_in_month() vs the alternatives

days_in_month() is the shortest leap-year-aware way to count a month's days. You can reach the same number with date arithmetic or base R, but those routes carry more code and more chances to forget February.

The decision rule is simple. If you only need the number, call days_in_month(). If you are already computing the month's first and last dates, the boundary difference reuses that work. Reach for base R only when adding lubridate as a dependency is not an option.

Common pitfalls

Three mistakes account for most days_in_month() confusion. Each has a quick fix.

First, feed days_in_month() a parsed date, not a raw string. A non-ISO string like "02/10/2024" cannot be coerced and returns NA, so always wrap text in ymd() first.

Second, the result is a named vector. When you assign it into a data frame column or compare it, the Feb style names can show up in unexpected places. Strip them with unname().

Third, remember the function ignores the day. days_in_month(ymd("2024-02-29")) and days_in_month(ymd("2024-02-01")) both return 29 because only the month and year are read.

Try it yourself

Related lubridate functions

days_in_month() works alongside the rest of lubridate's date toolkit. These functions cover the adjacent tasks:

• leap_year() tests whether a year has 366 days, the rule behind February's count.
• floor_date() snaps a date down to the start of its month.
• ceiling_date() snaps a date up to the next month boundary.
• rollback() returns the last day of the previous month.
• mday() extracts the day-of-month number for comparisons.

FAQ

Does days_in_month() account for leap years?

Yes. When the input date falls in February, days_in_month() checks whether that year is a leap year and returns 29 for leap years and 28 otherwise. It uses the full Gregorian rule, so century years like 1900 return 28 while 2000 returns 29. You never need to apply the leap-year formula yourself.

What does days_in_month() return for February?

It returns 28 or 29 depending on the year. For example, days_in_month(ymd("2024-02-15")) returns 29 because 2024 is a leap year, while days_in_month(ymd("2023-02-15")) returns 28. Every other month returns a fixed value of 30 or 31.

Can days_in_month() handle a vector of dates?

Yes, the function is fully vectorized. Pass a vector of Date or POSIXct values and it returns one integer per element, each named with that date's month abbreviation. This makes it convenient inside mutate() or any column-wise operation on a data frame of dates.

Why does the output have a name like "Feb"?

days_in_month() returns a named integer vector, and the names are the three-letter month abbreviations of the input dates. The names are purely informational. If they interfere with a comparison or a data frame column, wrap the call in unname() to get a plain integer vector.

How do I get the last day of the month with days_in_month()?

Add days_in_month(d) - 1 days to the first of the month: floor_date(d, "month") + days(days_in_month(d) - 1). An equivalent route is ceiling_date(d, "month") - days(1), which steps to the next month start and subtracts one day.