POSIXlt$zone and $gmtoff questions

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

POSIXlt$zone and $gmtoff questions

R devel mailing list
I've been searching for patterns in why some POSIXlt objects have the zone
and gmtoff components and some don't and why gmtoff is sometimes NA when
the zone is known.  Is there a pattern or is it just that the additional
fields and workarounds were added in an ad hoc way?

E.g.,  as.POSIXlt adds the zone and gmtoff components for all strings and
logical NA inputs if the time zone is not GMT or UTC

f <- function (lt)  {
    stopifnot(inherits(lt, "POSIXlt"))
    cat(format(lt), ", $zone=", deparse(lt$zone), ", $gmtoff=",
        deparse(lt$gmtoff), "\n", sep = "")
}
f(as.POSIXlt("2018-03-08 16:31", tz="US/Pacific"))
# 2018-03-08 16:31:00, $zone="PST", $gmtoff=NA_integer_
f(as.POSIXlt(NA, tz="US/Pacific"))
# NA, $zone="", $gmtoff=NA_integer_
f(as.POSIXlt(NA_character_, tz="US/Pacific"))
# NA, $zone="", $gmtoff=NA_integer_


But in GMT or UTF it omits the zone and gmtoff components unless you give
it a single character NA

f(as.POSIXlt("2018-03-08 16:31", tz="GMT"))
# 2018-03-08 16:31:00, $zone=NULL, $gmtoff=NULL
f(as.POSIXlt(NA, tz="GMT"))
# NA, $zone=NULL, $gmtoff=NULL
f(as.POSIXlt(NA_character_, tz="GMT"))
# NA, $zone="", $gmtoff=NA_integer_


Another oddity is that as.POSIXlt(characterData, tz="not-GMT") fills the
gmtoff component with NAs even though the zone and isdst components give
the information required to figure out the gmtoff.  as.POSIXlt(POSIXctData)
does give proper values to gmtoff

f(as.POSIXlt("2019-03-08", tz="US/Pacific"))
# 2019-03-08, $zone="PST", $gmtoff=NA_integer_
f(as.POSIXlt(as.POSIXct("2019-03-08", tz="US/Pacific")))
# 2019-03-08, $zone="PST", $gmtoff=-28800L


Is this last an efficiency issue?

Bill Dunlap
TIBCO Software
wdunlap tibco.com

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel