Skip to contents

Changelog

Source: NEWS.md

datey 0.2.0

Breaking changes

  • Switched the interval convention for datey_interval from left-closed, right-open [start, end) to left-open, right-closed (start, end] for consistency with survival analysis and actuarial exposed-to-risk, for which a period ending in death should include the time of death.

    This means interval %includes% t is now TRUE when start < t <= end and so code that tests membership at an interval’s start or end will produce different results. In particular, datey(valid_years_start) is no longer included in all_of_time but datey(valid_years_end) is.

  • A datey_interval is now formatted and printed as (start, end].

datey 0.1.2

CRAN release: 2026-09-16

  • Added rep() methods for datey, durationy and datey_interval. Previously rep() dropped the class and returned bare integer or double values. As rep_len() and rep.int() fall back to rep() dispatch, they are fixed too. This also fixes pmax() and pmin(), which previously returned the wrong values for datey and durationy when arguments of different lengths were recycled.

datey 0.1.1

CRAN release: 2026-07-14

  • Fixed undefined behaviour (negation of NA_INTEGER) when parsing durationy from text in C++. Results are unaffected on conforming hardware.

datey 0.1.0

CRAN release: 2026-07-07

Initial CRAN release.

  • datey and durationy represent dates and durations on a discrete annual grid, with exact, associative arithmetic (+, -, comparisons, mean(), range(), etc.).
  • start_day(), mid_day(), end_day() and datey() map calendar year/month/day (and day-fraction), fractional years and Date, POSIXct, POSIXlt to datey.
  • to_ymdf() and $year, $month, $day and $day_fraction provide the breakdown of a datey into year, month, day and day fraction components.
  • is_start_day() and is_mid_day() provides checking of datey alignment.
  • as.double() and as.integer() provide convenient conversion to numeric types as years, with mixed number and datey/durationy arithmetic adhering to the same convention using implicit conversion.
  • NA_datey_, NA_durationy_ and NA_datey_interval_ provide type-correct NAs, with full is.na()/anyNA() support.
  • Text parsing and formatting for datey, durationy and datey_interval.
  • pillar_shaft() methods for readable printing of datey/durationy/ datey_interval columns in tibbles.
  • is_leap_year() for integer years and dateys.
  • datey_interval (constructed with datey_interval() or %to%) represents a half-open [start, end) date interval. Operations include:
    • == and != for equality comparison.
    • %includes% (or interval_includes()) to test whether an interval contains a datey or numeric value.
    • & for intersection: returns an empty collapsed interval for adjacent intervals, NA_datey_interval_ for non-intersecting non-adjacent intervals.
    • is_collapsed(), is_proper() and related predicates.
    • Constants all_of_time and NA_datey_interval_.
  • seq() is supported for datey and durationy via a by argument (note: length.out is not supported).