XTS with unique time stamps?
On Mon, Jan 31, 2011 at 10:09 AM, Dirk Eddelbuettel <edd at debian.org> wrote:
On 31 January 2011 at 09:37, Jeffrey Ryan wrote:
| Brian, Worik
|
| w.r.t the new functionality in xts.
|
| It is so bleeding edge that Brian gave you the wrong name ;-) think
| "make [the] index unique". ?It probably will also be extended to do
| the former removal of subsequent non-unique observations/times as
| well.
|
| HTH,
| Jeff
|
|
| ?make.index.unique
|
| make.index.unique ? ? ? ? ? ? package:xts ? ? ? ? ? ? ?R Documentation
|
| Force Time Values To Be Unique
|
| Description:
|
| ? ? ?A generic function to force sorted time vectors to be unique.
| ? ? ?Useful for high-frequency time-series where original time-stamps
| ? ? ?may have identical values. For the case of xts objects, the
| ? ? ?default ?eps? is set to one-hundred microseconds. In practice this
| ? ? ?advances each subsequent identical time by ?eps? over the previous
| ? ? ?(possibly also advanced) value.
|
| Usage:
|
| ? ? ?make.index.unique(x, eps = 1e-05, ...)
|
| ? ? ?make.time.unique(x, eps = 1e-05, ...)
Why eps=1e-05? ?I wrote variants of this in-house and use
? ?incrementTimestamps <- function(times, incr=1.0e-6, ...) {
? ? ? ...
? ?}
Why not. ;-) On the topic of semantics, why 1.0e-6: 1.0 is redundant, given that a negative exponent will assure you a mode of "double". All kidding aside, Brian told me to. And it is user settable of course. Cuique suum Jeff
Dirk
|
| Arguments:
|
| ? ? ? ?x: An xts object, or POSIXct vector.
|
| ? ? ?eps: value to add to force uniqueness.
|
| ? ? ?...: unused
|
| Details:
|
| ? ? ?The returned time-series object will have new time-stamps so that
| ? ? ??isOrdered( .index(x) )? evaluates to TRUE.
|
| Value:
|
| ? ? ?A modified version of x.
|
| Note:
|
| ? ? ?Incoming values must be pre-sorted, and no check is done to make
| ? ? ?sure that this is the case. ?If the index values are of
| ? ? ?storage.mode ?integer?, they will be coerced to ?double?.
|
| Author(s):
|
| ? ? ?Jeffrey A. Ryan
|
| See Also:
|
| ? ? ??align.time?
|
| Examples:
|
| ? ? ?ds <- options(digits.secs=6) # so we can see the change
|
| ? ? ?x <- xts(1:10, as.POSIXct("2011-01-21") + c(1,1,1,2:8)/1e3)
| ? ? ?x
| ? ? ?make.index.unique(x)
|
| ? ? ?options(ds)
|
|
|
| On Mon, Jan 31, 2011 at 6:05 AM, Brian G. Peterson <brian at braverock.com> wrote:
| > On Monday, January 31, 2011 12:55:03 am Worik wrote:
| >> I am having trouble with non-unique time stamps in an xts.
| >>
| >> My underlying data has some repeated rows (in a csv file).
| >>
| >> How can I easily get rid of the duplicates?
| >>
| >> I feel I must be missing something simple. ?If not I can concoct an
| >> example to illustrate my problem.
| >
| > Worik,
| >
| > It depends on what you need.
| >
| > If you can remove the rows with duplicated indices, then a construction such
| > as:
| >
| > myxts<-myxts[!duplicated(index(myxts))]
| >
| > should work.
| >
| > If you need all of the observations, and need to artificially make them unique
| > (as is a common problem with tick data), then you will see discussion in the
| > list archives here and other places regarding adding artificial indices to high
| > frequency data while preserving order. You will need the latest xts from R-
| > Forge and use a construction like this:
| >
| > myxts<-make.unique.index(myxts)
| >
| > which will (by default) add .00001 sec to each non-unique index after the
| > first, preserving order, and providing every observation with a unique index.
| > Note that this presumes that the original order of the observations was
| > correct in the first place, no provision has been made if you have different
| > circumstances.
| >
| > Thanks to Jeff Ryan for (very) recently adding this second method.
| >
| > Regards,
| >
| > ?- Brian
| >
| > --
| > Brian G. Peterson
| > http://braverock.com/brian/
| > Ph: 773-459-4973
| > IM: bgpbraverock
| >
| > _______________________________________________
| > R-SIG-Finance at r-project.org mailing list
| > https://stat.ethz.ch/mailman/listinfo/r-sig-finance
| > -- Subscriber-posting only. If you want to post, subscribe first.
| > -- Also note that this is not the r-help list where general R questions should go.
| >
|
|
|
| --
| Jeffrey Ryan
| jeffrey.ryan at lemnica.com
|
| www.lemnica.com
|
| _______________________________________________
| R-SIG-Finance at r-project.org mailing list
| https://stat.ethz.ch/mailman/listinfo/r-sig-finance
| -- Subscriber-posting only. If you want to post, subscribe first.
| -- Also note that this is not the r-help list where general R questions should go.
--
Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com
_______________________________________________ R-SIG-Finance at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-finance -- Subscriber-posting only. If you want to post, subscribe first. -- Also note that this is not the r-help list where general R questions should go.
Jeffrey Ryan jeffrey.ryan at lemnica.com www.lemnica.com