[R] correct computation of timezones called 'strange' (was time difference bug?)

[Prof Brian Ripley]

You haven't given us the 'at a minimum' information that the posting guide 
requested, so we don't know your timezone.  But in America/Toronto on my 
F8 machine

d <- "2007-11-04 01:30:00"
dd <- as.POSIXct(d)
c(dd, dd+3600)

[1] "2007-11-04 01:30:00 EDT" "2007-11-04 01:30:00 EST"

Note the change of timezone.  Wikipedia says that is the right day for 
Canada ('from 2:00 a.m. LDT to 1:00 LST on the first Sunday in November 
since 2007').  Note that 01:34 is ambiguous that day (see ?as.POSIXct, 
surely part of the required homework).

The N American late change in rules in 2007 means that DST changes can be 
unreliable in OSes, but this one seems to be correct.

I do believe that you have failed to credit the careful work in getting 
this right in R.  Perhaps these are not times in Ontario and you forgot to 
tell R that?

On Thu, 14 Aug 2008, Bill Szkotnicki wrote:

> Hi,
> I am computing some time differences.
> Using the linux version of R 2.7.1
> And I am getting a strange result ( see below )
> I need the difference in minutes.
> Actually looking for where it is NOT 15 minutes.
> Would anyone know why this could be  happening?
> Or should I do this another way?
>
> Bill
>
> The script "k0.R"
> ===============
> a=read.table("buzwah005.txt")
> d=strptime(paste(a[,1]," ",a[,2]),"%m/%d/%Y %H:%M")
> tm=-24*60*as.numeric(as.POSIXlt("2007-1-1")-d) # time since midnight Jan 1
> print(tm[2]-tm[1])
> x=a[,3]
> diffs=round(tm[-1]-tm[-length(tm)])
> not15=(1:length(diffs))[diffs!=15]
> print(not15)
> ===============
>
> The R session....
>>
>> source("k0.R")
> [1] 15
> [1]  2113  5686 15712
>> class(d)
> [1] "POSIXt"  "POSIXlt"
>> d[15711:15714]
> [1] "2007-11-04 01:34:00" "2007-11-04 01:49:00" "2007-11-04 02:04:00"
> [4] "2007-11-04 02:19:00"
>> d[15712]-d[15711]
> Time difference of 15 mins
>> d[15713]-d[15712]
> Time difference of 1.25 hours
>> d[15714]-d[15713]
> Time difference of 15 mins
>>
>>
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595