[R]

From: Simon Wotherspoon (Simon.Wotherspoon@utas.edu.au)
Date: Mon 24 Mar 2003 - 15:17:41 EST

• Next message: Nirmala Ravishankar: "[R] Robust standard errors"
• Previous message: Jim Lemon: "[R] Scripting with an external editor"

Message-id: <JPEJIEHCLCCMMBFGMPDGGEJJCCAA.Simon.Wotherspoon@utas.edu.au>

Hi,
        In 1.6.2 at least, when constructing a sequence of dates with seq and the
by argument inherits from difftime, the "units" attribute seems to be
ignored. E.g.
try

str <- strptime("23/3/2003 06:00:00","%d/%m/%Y %H:%M:%S")
end <- strptime("23/3/2003 06:02:00","%d/%m/%Y %H:%M:%S")
step <- strptime("23/3/2003 06:01:00","%d/%m/%Y %H:%M:%S") -
  strptime("23/3/2003 06:00:00","%d/%m/%Y %H:%M:%S")
seq(str,end,step)

A possible fix is to replace the lines

  if (inherits(by, "difftime")) {
     by <- unclass(by)
  }

in seq.POSIXt with

  if (inherits(by, "difftime")) {
    by <- switch(attr(by, "units"), secs = by, mins = 60 * by, hours = 60 *
                 60 * by, days = 60 * 60 * 24 * by, weeks = 60 * 60 *
                 24 * 7 * by)
  }

as in "-.POSIXt", but this may not be the best solution.

I can't really see the point of representing difftime objects in terms of
different units, given POSIXct is always stored as seconds. Would it be
better to move the whole "unit" business into a print method, more or less
as it is for POSIXct?

Simon.

---
______________________________________________
R-help@stat.math.ethz.ch mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help

• Next message: Nirmala Ravishankar: "[R] Robust standard errors"
• Previous message: Jim Lemon: "[R] Scripting with an external editor"

This archive was generated by hypermail 2.1.3 : Tue 01 Jul 2003 - 09:11:26 EST