time_seq {timeplyr}

Time based version of base::seq()

Description

Time based version of base::seq()

Usage

time_seq(
  from,
  to,
  time_by,
  length.out = NULL,
  time_type = getOption("timeplyr.time_type", "auto"),
  week_start = getOption("lubridate.week.start", 1),
  time_floor = FALSE,
  roll_month = getOption("timeplyr.roll_month", "preday"),
  roll_dst = getOption("timeplyr.roll_dst", "NA")
)

time_seq_sizes(
  from,
  to,
  time_by,
  time_type = getOption("timeplyr.time_type", "auto")
)

time_seq_v(
  from,
  to,
  time_by,
  time_type = getOption("timeplyr.time_type", "auto"),
  roll_month = getOption("timeplyr.roll_month", "preday"),
  roll_dst = getOption("timeplyr.roll_dst", "NA"),
  time_floor = FALSE,
  week_start = getOption("lubridate.week.start", 1)
)

time_seq_v2(
  sizes,
  from,
  time_by,
  time_type = getOption("timeplyr.time_type", "auto"),
  time_floor = FALSE,
  week_start = getOption("lubridate.week.start", 1),
  roll_month = getOption("timeplyr.roll_month", "preday"),
  roll_dst = getOption("timeplyr.roll_dst", "NA")
)

Arguments

from

Start date/datetime of sequence.

to

End date/datetime of sequence.

time_by

Time unit increment.
Must be one of the three:

• string, specifying either the unit or the number and unit, e.g time_by = "days" or time_by = "2 weeks"

• named list of length one, the unit being the name, and the number the value of the list, e.g. list("days" = 7). For the vectorized time functions, you can supply multiple values, e.g. list("days" = 1:10).

• Numeric vector. If time_by is a numeric vector and x is not a date/datetime, then arithmetic is used, e.g time_by = 1.

length.out

Length of the sequence.

time_type

If "auto", periods are used for the time expansion when days, weeks, months or years are specified, and durations are used otherwise. If durations are used the output is always of class POSIXt.

week_start

day on which week starts following ISO conventions - 1 means Monday (default), 7 means Sunday. This is only used when time_floor = TRUE.

time_floor

Should from be floored to the nearest unit specified through the time_by argument? This is particularly useful for starting sequences at the beginning of a week or month for example.

roll_month

Control how impossible dates are handled when month or year arithmetic is involved. Options are "preday", "boundary", "postday", "full" and "NA". See ?timechange::time_add for more details.

roll_dst

See ?timechange::time_add for the full list of details.

sizes

Time sequence sizes.

Details

This works like seq(), but using timechange for the period calculations and base::seq.POSIXT() for the duration calculations. In many ways it is improved over seq as dates and/or datetimes can be supplied with no errors to the start and end points. Examples like,
time_seq(now(), length.out = 10, by = "0.5 days", seq_type = "dur") and
time_seq(today(), length.out = 10, by = "0.5 days", seq_type = "dur")
produce more expected results compared to
seq(now(), length.out = 10, by = "0.5 days") or
seq(today(), length.out = 10, by = "0.5 days").

For a vectorized implementation with multiple start/end times, use time_seq_v()/time_seq_v2()

time_seq_sizes() is a convenience function to calculate time sequence lengths, given start/end times.

Value

time_seq returns a time sequence.
time_seq_sizes returns an integer vector of sequence sizes.
time_seq_v returns time sequences.
time_seq_v2 also returns time sequences.

See Also

seq_id time_seq_id

Examples

library(timeplyr)
library(lubridate)

# Dates
today <- today()
now <- now()

time_seq(today, today + years(1), time_by = "day")
time_seq(today, length.out = 10, time_by = "day")
time_seq(today, length.out = 10, time_by = "hour")

time_seq(today, today + years(1), time_by = list("days" = 1)) # Alternative
time_seq(today, today + years(1), time_by = "week")
time_seq(today, today + years(1), time_by = "fortnight")
time_seq(today, today + years(1), time_by = "year")
time_seq(today, today + years(10), time_by = "year")
time_seq(today, today + years(100), time_by = "decade")

# Datetimes
time_seq(now, now + years(1), time_by = "12 hours")
time_seq(now, now + years(1), time_by = "day")
time_seq(now, now + years(1), time_by = "week")
time_seq(now, now + years(1), time_by = "fortnight")
time_seq(now, now + years(1), time_by = "year")
time_seq(now, now + years(10), time_by = "year")
time_seq(now, today + years(100), time_by = "decade")

# You can seamlessly mix dates and datetimes with no errors.
time_seq(now, today + days(3), time_by = "day")
time_seq(now, today + days(3), time_by = "hour")
time_seq(today, now + days(3), time_by = "day")
time_seq(today, now + days(3), time_by = "hour")

# Choose between durations or periods

start <- dmy(31012020)
# If time_type is left as is,
# periods are used for days, weeks, months and years.
time_seq(start, time_by = "month", length.out = 12,
         time_type = "period")
time_seq(start, time_by = "month", length.out = 12,
         time_type = "duration")
# Notice how strange base R version is.
seq(start, by = "month", length.out = 12)

# Roll forward or backward impossible dates

leap <- dmy(29022020) # Leap day
end <- dmy(01032021)
# 3 different options
time_seq(leap, to = end, time_by = "year",
         roll_month = "NA")
time_seq(leap, to = end, time_by = "year",
         roll_month = "postday")
time_seq(leap, to = end, time_by = "year",
         roll_month = getOption("timeplyr.roll_month", "preday"))