How to pass a ts object to R via RSRuby

I've had absolutely no luck working with the HoltWinters function via RSRuby and R. How exactly do you 1) create a time series object through RSRuby and 2) successfully pass that to HoltWinters to get output?

Example:

@r = RSRuby.instance
=> #<RSRuby:0x106bfe6c0 @proc_table={}, @class_table={}, @default_mode=-1, @cache={"get"=>#<RObj:0x106bfe580>, "helpfun"=>#<RObj:0x106bfd3d8>, "help"=>#<RObj:0x106bfd3d8>, "NaN"=>NaN, "FALSE"=>false, "TRUE"=>true, "F"=>false, "NA"=>-2147483648, "eval"=>#<RObj:0x106bfdf18>, "T"=>true, "parse"=>#<RObj:0x106bfe0d0>}, @caching=true>
@r.assign('mytime',@r.ts(:data => [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34], :frequency => 12, :start => [1993,3], :end => [1995,3]))
=> [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25]
@r.HoltWinters(@r.mytime)
    RException: Error in decompose(ts(x[1L:wind], start = start(x), frequency = f), seasonal) : 
    time series has no or less than 2 periods

rsruby (0.5.1.1)

R version 2.12.2 (2011-02-25)

Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

:edit: a similar example inside R only... if I could get any output from HoltWinters via RSRuby (other 开发者_运维知识库than an error) I'd be very happy

> z <- ts(1:34, frequency = 12, start = c(1993,3), end = c(1995,3))
> z
     Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
1993           1   2   3   4   5   6   7   8   9  10
1994  11  12  13  14  15  16  17  18  19  20  21  22
1995  23  24  25                                    
> HoltWinters(z)
Holt-Winters exponential smoothing with trend and additive seasonal component.

Call:
 HoltWinters(x = z) 

Smoothing parameters:
 alpha:  1 
 beta :  0 
 gamma:  0 

Coefficients:
             [,1]
a    2.500000e+01
b    1.000000e+00
s1  -8.141636e-16
s2  -8.141636e-16
s3   9.621933e-16
s4   2.738550e-15
s5  -8.141636e-16
s6  -8.141636e-16
s7   7.401487e-17
s8  -8.141636e-16
s9   9.621933e-16
s10 -8.141636e-16
s11 -8.141636e-16
s12  9.621933e-16

---

If I understand your question correctly, you are really looking for guidance on using a specific interface. Up front, let me tell you that I don't use RSRuby, but I do use a different utility for R and Ruby integration.

Upon further examination, I see you really want a time series option... Which I believe... Yes, that can easily be produced from a dataframe by using the as.ts(data_frame) fxn.
http://stat.ethz.ch/R-manual/R-patched/library/stats/html/ts.html

What I've used is simple enough. If this is a major issue, maybe this can be of use to you.

require 'rserve/simpler'
r_object = Rserve::Simpler.new 

What I do next is to take a Hash where the keys correspond to dataframe columns, and the values are arrays, and I run the Rserve::Simpler fxn Hash.to_dataframe on them so that they are ready to be converted.

data = Hash.new()#insert data here
datafr = data.to_dataframe
r_object.converse(df: datafr) do 
  %Q{df$time <- strptime(as.character(df$time), "%Y-%m-%d %X")
     df$name <- factor(df$name)
  }
end

This uses straight R code in the converse block and handles everything nicely. I haven't tried this with your particular issue, but I know that this worked for me to import time data (Ruby Datetime class) from an array into R so that I could graph it. Good luck!