datetimeoffset (version 1.0.0)

datetimeoffset-invalid: Invalid datetimeoffset datetimes

Description

invalid_detect() detects invalid datetimes. invalid_any() returns TRUE if any datetimes are invalid. invalid_count() returns number of invalid datetimes. invalid_remove() removes invalid datetimes. invalid_resolve() resolves invalid datetimes.

Usage

# S3 method for datetimeoffset
invalid_detect(x)
# S3 method for datetimeoffset
invalid_resolve(x, ..., invalid = "NA", nonexistent = "NA")
# S3 method for datetimeoffset
invalid_any(x)
# S3 method for datetimeoffset
invalid_count(x)
# S3 method for datetimeoffset
invalid_remove(x)

Value

invalid_detect(), invalid_remove(), and invalid_resolve() return datetimeoffset() vectors. invalid_count() returns an integer and invalid_any() returns a logical value.

Arguments

x

A datetimeoffset() object.

...

Ignored.

invalid

Invalid date resolution strategy. See clock::invalid_resolve().

nonexistent

Nonexistent (because of DST spring forward) time resolution strategy. See clock::as_zoned_time.clock_naive_time().

Details

datetimeoffset() datetimes can be considered invalid for three main reasons:

1. An invalid "calendar date" such as "2020-02-30" (there are less than 30 days in February).

2. A "nonexistent" datetime due to a Daylight Savings Time "spring forward" such as "2020-03-08T02:59:59[America/Los_Angeles]"

3. Incorrect UTC offsets such as "2020-03-08T01:59:59-08[America/New_York]" (that particular Eastern time has a UTC offset of -05)

Examples

# invalid date because April only has 30 days
dts <- c("2019-04-30T03:30:00", "2019-04-31T02:30:00")
dts <- as_datetimeoffset(dts)
clock::invalid_detect(dts)
clock::invalid_any(dts)
clock::invalid_count(dts)
clock::invalid_remove(dts)
clock::invalid_resolve(dts)
clock::invalid_resolve(dts, invalid = "previous")
clock::invalid_resolve(dts, invalid = "previous-day")

# non-existent time because of DST "spring forward"
if ("America/Los_Angeles" %in% OlsonNames()) {
  dt <- as_datetimeoffset("2020-03-08T02:59:59[America/Los_Angeles]")
  print(clock::invalid_detect(dt))
  clock::invalid_resolve(dt, nonexistent = "roll-forward")
}

# incorrect UTC offsets
if ("America/New_York" %in% OlsonNames()) {
  dt <- as_datetimeoffset("2020-03-08T01:59:59-08[America/New_York]")
  print(clock::invalid_detect(dt))
  clock::invalid_resolve(dt)
}