[R] zoo, zooreg, & ISOdatetime

[Gabor Grothendieck]

On 8/26/05, David James <djames at frontierassoc.com> wrote:
> I create a zooreg object that runs from Jan-1-2002 0:00 to Jun-1-2005
> 0:00...
> 
> regts.start = ISOdatetime(2002, 1, 1, hour=0, min=0, sec=0, tz="")
> regts.end = ISOdatetime(2005, 6, 1, hour=0, min=0, sec=0, tz="")
> regts.zoo <- zooreg( NA, regts.start, regts.end, deltat=3600 )
> 
> Upon inspection:
>  > regts.zoo[1:3]
> 2002-01-01 00:00:00 2002-01-01 01:00:00 2002-01-01 02:00:00
>                  NA                  NA                  NA
> 
>  > regts.zoo[29926:29928]
> 2005-05-31 22:00:00 2005-05-31 23:00:00 2005-06-01 00:00:00
>                  NA                  NA                  NA
> 
> However:
>  > summary(regts.zoo)
> Error in "row.names<-.data.frame"(`*tmp*`, value = c("2002-01-01
> 00:00:00",  :
>     duplicate 'row.names' are not allowed
> 
> I don't understand why it claims that there are duplicate row.names.

Because you are using tz = "" and when the time falls back in fall 2am is 1am 
so you wind up having two 1am's which is not allowed.   The times
themselves are
ok but the names are duplicated which is not allowed either.  Thanks for
finding this problem.

> Any advice?

1. Use tz = "GMT" rather than tz = "" since GMT has no daylight savings time 
and won't result in duplicate names.

> regts.start = ISOdatetime(2002, 1, 1, hour=0, min=0, sec=0, tz="GMT")
> regts.end = ISOdatetime(2005, 6, 1, hour=0, min=0, sec=0, tz="GMT")
> regts.zoo <- zooreg( NA, regts.start, regts.end, deltat=3600 )

2. As advised last time use chron if you don't need time zones.

3. Read the R News 4/1 article I previously recommended, which has the
same advice as #2 plus a lot more info on dates and times.

4. If you are using time zones and do need POSIXct and tz = "" then let me know 
privately and I will send you the devel version of zoo where I just fixed it.

> 
> I probably could use the aggregate function to clean this up, but I
> don't see why it should be needed (provided that I do things properly
> in the first place).

Yes.  It should not be needed.

> 
> Thanks,
> David
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
>