R help - Jan 2012 - plotOHLC(alpha3): Error in plotOHLC(alpha3) : x is not a open/high/low/close time series

R version 2.12.0, 64 bit on Windows.

 

Here is a short script that illustrates the problem:

 

library(tseries)

library(xts)

setwd('C:\\cygwin\\home\\Ted\\New.Task\\NKs-01-08-12\\NKs\\tests')

x = read.table("quotes_h.2.dat", header = FALSE, sep="\t",
skip=0)

str(x)

y <- data.frame(as.POSIXlt(paste(x$V2,substr(x$V4,4,8),sep="
"),format='%Y-%m-%d %H:%M'),x$V5)

colnames(y) <- c("tickdate","price")

str(y)

plot(y)

z <- as.irts(y)

str(z)

plot(z)

str(alpha3)

List of 2

$ time : POSIXt[1:98865], format: "2010-06-30 15:47:00"
"2010-06-30
15:53:00" "2010-06-30 17:36:00" ...

$ value: num [1:98865, 1:4] 9215 9220 9205 9195 9195 ...

  ..- attr(*, "dimnames")=List of 2

  .. ..$ : NULL

  .. ..$ : chr [1:4] "z.Open" "z.High" "z.Low"
"z.Close"

- attr(*, "class")= chr "ts"

- attr(*, "tsp")= num [1:3] 1 2 1

alpha3 <- as.xts(to.minutes3(z,OHLC = TRUE))

plotOHLC(alpha3)

Error in plotOHLC(alpha3) : x is not a open/high/low/close time series

 

The file quotes_h.2.dat contains real time tick data for futures contracts,
so the above manipulation is my attempt to just get a time series with one
column being a date/time and the other being tick price.  I believe I have
to use read.table to make a data frame, and then the manipulations to
combine the date and time fields from that feed, along with the price.

 

My first attempt at using to.minutes3 (and I am interested in the other
'to.period' functions too), is to get a regular time series to which I
can
apply rollapply, along with a function in which I use various autoregression
methods, along with forecasting for as long as the 95% confidence intervals
is reasonably close - I want to know how far into the future the forecast
contains useful information.  And then, I want to create a plot in which I
do the autoregression, and then plot the actual and forecast prices (along
with the confidence interval), as a function of time, embed that in a
function, which rollappply works with, so I can have a plot comprised of all
those individual plots (plotting only the comparison of actual and forecast
values).

 

It seems everything works adequately until I try the plotOHLC function
itself, which gives me the error in the subject line.

 

I would ask for two things: 

 

1) what the fix is to get rid of that error plotOHLC gives me

2) some tips on the 'walk-forward' method I am looking at using.

 

Thanks

 

Ted


	[[alternative HTML version deleted]]

Hi Ted,

On Tue, Jan 10, 2012 at 1:59 PM, Ted Byers <r.ted.byers at gmail.com>
wrote:
> R version 2.12.0, 64 bit on Windows.
>
> Here is a short script that illustrates the problem:
> library(tseries)
> library(xts)
> setwd('C:\\cygwin\\home\\Ted\\New.Task\\NKs-01-08-12\\NKs\\tests')
> x = read.table("quotes_h.2.dat", header = FALSE,
sep="\t", skip=0)
> str(x)
> y <- data.frame(as.POSIXlt(paste(x$V2,substr(x$V4,4,8),sep="
> "),format='%Y-%m-%d %H:%M'),x$V5)
> colnames(y) <- c("tickdate","price")
> str(y)
> plot(y)
> z <- as.irts(y)
> str(z)
> plot(z)
> str(alpha3)
> List of 2
> $ time : POSIXt[1:98865], format: "2010-06-30 15:47:00"
"2010-06-30
> 15:53:00" "2010-06-30 17:36:00" ...
> $ value: num [1:98865, 1:4] 9215 9220 9205 9195 9195 ...
> ?..- attr(*, "dimnames")=List of 2
> ?.. ..$ : NULL
> ?.. ..$ : chr [1:4] "z.Open" "z.High" "z.Low"
"z.Close"
> - attr(*, "class")= chr "ts"
> - attr(*, "tsp")= num [1:3] 1 2 1
>
This is a big clue.  Your alpha3 object is a list with two elements 1)
the datetime, and 2) the OHLC values as a ts object.  There's no
as.xts() method for this type of object.
> alpha3 <- as.xts(to.minutes3(z,OHLC = TRUE))
>
Look at str(z).  Why did you convert z to an irts object instead of
directly to an xts object?  to.minutes() expects an xts object.  Try
something like this instead:
z <- xts(y[,2], y[,1])
alpha3 <- to.minutes3(z, OHLC=TRUE)
> plotOHLC(alpha3)
> Error in plotOHLC(alpha3) : x is not a open/high/low/close time series
>
> The file quotes_h.2.dat contains real time tick data for futures contracts,
> so the above manipulation is my attempt to just get a time series with one
> column being a date/time and the other being tick price. ?I believe I have
> to use read.table to make a data frame, and then the manipulations to
> combine the date and time fields from that feed, along with the price.
>
> My first attempt at using to.minutes3 (and I am interested in the other
> 'to.period' functions too), is to get a regular time series to
which I can
> apply rollapply, along with a function in which I use various
autoregression
> methods, along with forecasting for as long as the 95% confidence intervals
> is reasonably close - I want to know how far into the future the forecast
> contains useful information. ?And then, I want to create a plot in which I
> do the autoregression, and then plot the actual and forecast prices (along
> with the confidence interval), as a function of time, embed that in a
> function, which rollappply works with, so I can have a plot comprised of
all
> those individual plots (plotting only the comparison of actual and forecast
> values).
>
> It seems everything works adequately until I try the plotOHLC function
> itself, which gives me the error in the subject line.
>
> I would ask for two things:
> 1) what the fix is to get rid of that error plotOHLC gives me
> 2) some tips on the 'walk-forward' method I am looking at using.
>
> Thanks
> Ted
>
>
> ? ? ? ?[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

--
Joshua Ulrich  |  FOSS Trading: www.fosstrading.com