R devel - Feb 2010 - Error with cut.POSIXt and daylight savings time switchover dates

The following code:

cut(as.POSIXct("2009-11-01 04:00:00",
tz="America/Los_Angeles"), "1 day")

gives the error:

Error in seq.int(0, to - from, by) : 'to' must be finite

This is related to November 1st, 2009 being the switchover date from daylight
savings time to standard time in the America/Los_Angeles time zone.  In
particular, in cut.POSIXt, the starting time (start) is converted to a POSIXlt,
and the individual members are manipulated.  Because a spacing of "1
day" is requested, the hour, minute, and second are manually set to 0.  In
doing so, the represented time is now before the 2:00am PDT->PST change. 
This value is passed to seq.int (as the argument from), which dispatches to
seq.POSIXt.  seq.POSIXt eventually does from <- unclass(as.POSIXct(from))
which evaluates to NA because as.POSIXct(from) is NA.  The seq.int call in the
next line then passes NA as the "to" argument, causing the output
error (which comes from the C-code of seq.int).

Bringing it all together, the sequence of commands that causes the problems is:

tm <- as.POSIXlt("2009-11-01 04:00:00",
tz="America/Los_Angeles")
tm$hour <- 0
as.POSIXct(tm)
# [1] NA

Is there a timezone/daylight savings time safe way to get to the beginning of
the day in cut.POSIXt so that invalid dates are not sent to the other functions?
Alternatively, can as.POSIXct.POSIXlt be made to handle these manually
manipulated dates correctly?

--
Brian Diggs, Ph.D.
Senior Research Associate, Department of Surgery, Oregon Health & Science
University

On 2/1/2010 3:57 PM, Brian Diggs wrote:
> The following code:
> 
> cut(as.POSIXct("2009-11-01 04:00:00",
tz="America/Los_Angeles"), "1 day")
> 
> gives the error:
> 
> Error in seq.int(0, to - from, by) : 'to' must be finite
> 
> This is related to November 1st, 2009 being the switchover date from
> daylight savings time to standard time in the America/Los_Angeles
> time zone.  In particular, in cut.POSIXt, the starting time (start)
> is converted to a POSIXlt, and the individual members are
> manipulated.  Because a spacing of "1 day" is requested, the
hour,
> minute, and second are manually set to 0.  In doing so, the
> represented time is now before the 2:00am PDT->PST change.  This
> value is passed to seq.int (as the argument from), which dispatches
> to seq.POSIXt.  seq.POSIXt eventually does from <-
> unclass(as.POSIXct(from)) which evaluates to NA because
> as.POSIXct(from) is NA.  The seq.int call in the next line then
> passes NA as the "to" argument, causing the output error (which
comes
> from the C-code of seq.int).
> 
> Bringing it all together, the sequence of commands that causes the
> problems is:
> 
> tm <- as.POSIXlt("2009-11-01 04:00:00",
tz="America/Los_Angeles")
> tm$hour <- 0 
> as.POSIXct(tm)
> # [1] NA
> 
> Is there a timezone/daylight savings time safe way to get to the
> beginning of the day in cut.POSIXt so that invalid dates are not sent
> to the other functions?  Alternatively, can as.POSIXct.POSIXlt be
> made to handle these manually manipulated dates correctly?
I realized I forgot my session info:
> sessionInfo()
R version 2.10.1 (2009-12-14) 
i386-pc-mingw32 

locale:
[1] LC_COLLATE=English_United States.1252 
[2] LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

> -- Brian Diggs, Ph.D. Senior Research Associate, Department of
> Surgery, Oregon Health & Science University