lubridate as_date() in R: Convert Inputs to Date Objects
The lubridate as_date() function coerces strings, POSIXct date-times, numeric values, and existing dates into a clean Date object. It is the safer, type-aware replacement for base R's as.Date() with sensible defaults across every input type.
as_date("2024-01-15") # parse ymd string
as_date(c("2024-01-15", "2024-02-20")) # parse vector
as_date("01/15/2024", format = "%m/%d/%Y") # explicit format
as_date(as.POSIXct("2024-01-15 23:30:00", tz="UTC")) # strip time
as_date(now(), tz = "America/New_York") # date in your tz
as_date(19367) # days since 1970-01-01
as_date(as.Date("2024-01-15")) # Date passthroughNeed explanation? Read on for examples and pitfalls.
What as_date() does in one sentence
as_date(x) converts x to a Date object using a method dispatched on the input class. Strings are parsed as year-month-day, POSIXct values are stripped to their calendar date, and numerics are interpreted as days since 1970-01-01.
This single function replaces three base R idioms: as.Date(x), as.Date(x, origin = "1970-01-01"), and as.Date(format(x, "%Y-%m-%d")). It is the workhorse coercion for any column you suspect should be a date.
Syntax
as_date(x, tz = NULL, format = NULL, origin = lubridate::origin). x is the input; tz controls which time zone defines the calendar date; format overrides ymd parsing for character input.
The default behavior reads strings as year first, month second, day third. Anything else needs the format argument or one of the explicit parsers like mdy() or dmy().
Convert different input types
1. Character string
as_date() is forgiving about separators. Hyphens, slashes, dots, and no-separator compact forms all parse as long as the order is year-month-day.
2. Vector of strings
The function is vectorized and returns a Date vector of the same length. Any unparseable element becomes NA with a warning.
3. POSIXct date-time
When the input is a date-time, as_date() extracts the calendar date. The tz argument controls which time zone defines that date. Without tz, the date is computed in UTC.
tz if your data was stored in a non-UTC zone, otherwise dates near midnight will silently shift.4. Numeric (days since 1970)
Numeric input is treated as days since the Unix epoch. This makes as_date() a one-step replacement for as.Date(x, origin = "1970-01-01").
5. Already-Date passthrough
Passing a Date returns the same Date. This makes as_date() safe inside data pipelines where some columns may already be the target type.
as_date() vs as.Date() vs ymd()
as_date() dispatches on input class, as.Date() requires explicit origin for numerics and format for non-ISO strings, and ymd() only accepts character or factor input. Use the comparison table below to pick correctly.| Function | Inputs accepted | String default | Numeric origin |
|---|---|---|---|
lubridate::as_date() |
character, POSIXt, Date, numeric, factor | ymd auto-detect | 1970-01-01 |
base::as.Date() |
character, POSIXct, factor | "%Y-%m-%d" |
none, must pass |
lubridate::ymd() |
character, factor | year-month-day | not applicable |
lubridate::as_datetime() |
character, Date, numeric | ymd_hms | seconds since 1970 |
Decision rule. If you have mixed input types or unknown class, use as_date(). If you know you have a string in a specific order, prefer ymd(), mdy(), or dmy() for clearer intent. If you need a date-time, use as_datetime().
Common pitfalls
Three traps catch most users. First, non-ymd strings return NA; pass format or use mdy()/dmy(). Second, POSIXct values near midnight shift dates if you forget tz. Third, factor input works but character is more predictable; convert factors with as.character() first if results look strange.
as_date() with dplyr::mutate() for type-safe imports. When reading CSVs that store dates as strings, do df |> mutate(date = as_date(date)) once at the top of the pipeline. Every downstream date function then works without surprise, because the column class is guaranteed.Try it yourself
Try it: Convert a vector of three string dates and a POSIXct value into a single Date vector. Save it to ex_dates.
Click to reveal solution
Explanation: as_date() returns a Date vector for both inputs, so c() concatenates them without coercion to character. Mixing classes inside c() would silently convert to character; coercing first guarantees a clean Date vector.
Related lubridate functions
ymd(),mdy(),dmy(): explicit string parsers when you know the date part order.ymd_hms(),mdy_hms(),dmy_hms(): same for date-time strings.as_datetime(): returns POSIXct instead of Date.today(): current date in your local time zone, equivalent toas_date(now()).parse_date_time(): handles vectors with multiple possible formats.
FAQ
What is the difference between as_date() and as.Date() in R?
as_date() is lubridate's coercion function and dispatches on input class. It accepts character, POSIXct, numeric, and Date inputs with sensible defaults: ymd parsing for strings, 1970-01-01 origin for numerics, time-zone-aware stripping for date-times. Base R's as.Date() requires an explicit origin for numerics and an explicit format for non-ISO strings, and it errors on common shapes that as_date() handles silently.
Does as_date() handle time zones correctly?
Yes, but you must pass the tz argument when the input is POSIXct. Without tz, the calendar date is computed in UTC, which can shift dates by one day for values near midnight in non-UTC zones. Always specify tz when your timestamps come from a known zone, for example as_date(ts, tz = "America/New_York").
Can as_date() parse non-ISO date strings?
Yes, with the format argument. By default it expects year-month-day order. For other orders, either pass format = "%m/%d/%Y" (or whatever strptime format applies) or use the explicit parsers mdy() and dmy(). The explicit parsers communicate intent more clearly to the next reader of your code.
What does as_date(0) return?
as_date(0) returns "1970-01-01", the Unix epoch. Numeric input is treated as days since 1970-01-01. This is useful when reading data exported from systems that store dates as integer day offsets, such as some database extracts or Excel sheets after numeric formatting.
Should I use as_date() or ymd() for parsing strings?
Use ymd() (or mdy()/dmy()) when you know the input is a string with a specific order. The function name documents intent. Use as_date() when the input could be a string OR another type, such as inside a function that accepts mixed arguments. Both produce the same Date result for valid ymd strings.