Receiving an error of an unexpect symbol when running a script in R.

[krwinters11]

This is the error I am receiving:

command="r", R exited with code 1: Error: unexpected symbol in: "input <- read.csv( "c:\windows\temp\tmp2d19tj", check.names=FALSE) source('predictR.r') result" Execution halted

I am never actually calling read.csv in my script. Any suggestions?

[krwinters11 · Accepted solution]

Our problem came from the way we were taking in the variable. They were read in as characters, so in order to fix the problem, we needed to typecast our values back to numeric for R to read them correctly.
We added a line at the beginning of our function to "unlist" and cast to numeric:
y<-as.numeric(unlist(y))

[Patient]

Hi krwinters,

I don't know if this will help you, if I have understood very well you need to
Upload a csv file in your script. If yes I wil like to let you know that The command requires one parameters which is either a actual R language script:

1. | r "output = data.frame(Name=c('A','B','C'),Value=c(1,2,3))"
2. or just the name of a R script file that is uploaded to the app: | r myscript.r
3. Now, Input comes in asinput and you need you direct your results to output to get them back into Splunk. For example: | r "exp(mean(log( read.csv(input)))) -> output"
4. For more information, follow this link: http://stackoverflow.com/questions/26607592/splunk-query-language-for-r

[krwinters11]

this is my script:
predictR <- function(y){
o <- (sum(y>(100+3*sd(y)) ) + sum(y<(100-3*sd(y)) ))/2+1

fq <- length(y)/o

fit <- stl(ts(y, freq=fq), "per", robust=TRUE)
forecasted.data <- fit$time.series[1:102,1] + mean(y)

fitted.data <- fit$time.series[,2]+fit$time.series[,1]

forecast <- c(fitted.data, forecasted.data)
return(forecast)
}

I never call read...which is why I am confused.