[R-SIG-Finance] Convertion of a zoo object to a ts object

[Ulrich Staudinger]

Hi Sylvain,

i know this is a R mailing list, but let me add some information on how 
i solve the very same problem.

I usually do the whole data obtaining and preparation in ccapi2, i then 
export this data from ccapi2 as a CSV file which i load into R.

Data preparation includes for example aligning multiple timeseries, 
dealing with extreme outliers due to unreasonable trades.
For aligning multiple time series i use the java method alignTimeSeries 
viewable at 
http://activestocks.de:8081/viewrep/ccapi3/src/core/ccapi/core/util/TimeSeriesUtil.java?r=1.7#l44 


Would be good if RMetrics would integrate such a method. Yahoo does drop 
a few days in long timeseries, usually.

Hope this helps,
Best regards,
Ulrich Staudinger


Sylvain BARTHELEMY schrieb:
> Dear All,
>
> I use get.hist.quote to read historical prices on various quotes from
> Finance Yahoo. It works fine on daily prices and most of the time on monthly
> prices but it is sometimes difficult to synchronize monthly data without
> duplicating months when extracting more than one quote. 
>
> Here is an example where many months are duplicated (see at the end of the
> time series that 2006-05 is duplicated because CAC40 is quoted on 2006-05-02
> and AXA on 2006-05-01):
> x <- get.hist.quote("^FCHI", compression = "m", start = as.Date(0), quote =
> "Close")
> x <- cbind( xx, get.hist.quote("CS.PA", compression = "m", start =
> as.Date(0), quote = "Close") )
>
> For the moment, to avoid these problems, I use something like:
> readFromYahooM <- function(nam,quote) {
>  xx.0 <- get.hist.quote(nam, compression = "m", start = as.Date(0), quote =
> quote)
>  xx <- ts( coredata(xx.0), frequency = 12, start =
> as.numeric(as.yearmon(as.Date(start(xx.0)[1]))) )
>  xx
> }
> x <- readFromYahooM("^FCHI","Close")
> x <- cbind(x, readFromYahooM("CS.PA","Close"))
> colnames(x) <- c("FCHI","CS.PA")
>
> It works, but there is a strong assumption that all the months are present
> in the two time series and the fact is that xx is a zoo object and zoo
> objects can contain irregular time series ! I am sure that there should be a
> more elegant way to do a conversion from a zoo object to a ts object or to
> replace the "start" parameter by the exact vector of zoo dates and or
> yearmon but I don't find any way do to that.
>
> I anyone can help me, I would appreciate.
> Thanks in advance.
>
> ---
> Sylvain Barthélémy
> Head of Quantitative Research, TAC
> www.tac-financial.com | www.sylbarth.com
>
>
>
>
> 15:59
>
> _______________________________________________
> 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.
>