[R-SIG-Finance] aggregate over POSIX timestamps not returning unique values

Gabor Grothendieck ggrothendieck at gmail.com
Thu Feb 28 21:29:59 CET 2008


Its a problem with daylight savings time.  If you use chron
or do this first:

Sys.setenv(TZ = "GMT")

to put your session into GMT time then daylight savings time
won't exist and the problem does not arise.

Suggest you read R News 4/1.

Also please make code with random numbers
reproducible by using set.seed, e.g.
set.seed(1)



On Thu, Feb 28, 2008 at 3:11 PM, Markus Loecher
<markus at insightfromdata.com> wrote:
>
> I am puzzled by the following observation. It seems that aggregating
> over POSIX timestamps sometimes returns duplicate entries in the output.
> For example:
>
> tmp <- zoo(cumsum(rnorm(35041)), order.by =
> seq(as.POSIXct("2007-01-01 10:00:00"), to = as.POSIXct("2008-01-01
> 10:00:00"), by = 15*60))
>  plot(tmp)
>  tmp.h <- aggregate(coredata(tmp), by =
> list(as.POSIXct(trunc(index(tmp),"hour"))), FUN = mean)
>  plot(zoo(tmp.h[,2], as.POSIXct(as.character(tmp.h[,1]))))
>
>  > Warning message:
>  > some methods for "zoo" objects do not work if the index entries
> in >'order.by' are not unique in: zoo(tmp.h[, 2], as.POS...
>
> When I use the more elegant aggregate.zoo() version of this, the
> problem is a non issue:
>
> tmp.h <- aggregate(tmp, by = as.POSIXct(trunc(index(tmp),"hour")), FUN = mean)
>
> I would still l like to understand this, any help would be greatly appreciated,
> Thanks,
> Markus
>
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
>



More information about the R-SIG-Finance mailing list