running r scripts or commands with interpretor in unix for unix-layman

Question

I am layman to unix and sofar I using R in windows. For example I type following in my R session (in R gui).

# this is a my funny example script 
X <- 1:         


        

Answer 1

Assuming you save your script in a simple text file with the name so.R, you can run it under Linux/Unix by typing R at the prompt. Once in R enter

  source('so.R')

to execute the script inside the R environment (this assumes the so.R file is in the same directory as you are when you issue this command).

To run the script from the Linux/Unix command line use the following command:

  R CMD BATCH so.R

Note that I got the plot to show when I ran the script inside of R, but from the Linux command line it doesn't show. I suspect it gets quickly displayed and then goes away, so there will be a R command that you have to look up to make it pause after it displays the plot.

Answer 2

If your program is going to work on a single dataset, then simple-r might be the solution:

http://code.google.com/p/simple-r/

It is especially designed for simple statistical analysis as a part of Linux command line. For example, if one wants to plot some data, 'r -p data.txt' will do the job; for getting correlation coefficient: 'r cor data.txt' will suffice.

Answer 3

I'm guessing from the way you worded your question that you maybe SSH'ed into a linux machine? Or that you installed Ubuntu, for example, on your usual laptop/PC.

Assuming it's the second case: open a terminal and type sudo apt-get install r-base. Then type R. Then type

X <- 1:10
Y <- 21:30
plot(X, Y)
myfun <- function (x){
              x1 <- x^0.2
              return (x1)
             }
myfun(X)

Since your question is about unix versus linux rather than R, you might also try http://unix.stackexchange.com. There is a lot to be said about the differences between linux and unix, but all you probably need to know is: download Ubuntu, burn it onto a disc, then restart your computer with the disc in your CD drive.

Hope this helps.

Answer 4

The examples below show two ways to run R code in a shell script. Both examples will also define functions without executing them, if the scripts are loaded to an interactive R session via the source() function.

The first example allows you to give arguments as you would to any other shell script, but will not pass additional R-options to R (because Rscript gives "--args" to R as one of the arguments).

The second example allows you to give additional R-options, but generates (harmless) warning messages unless you give "--args" as one of the script arguments. This version is best avoided unless you have special requirements.

prototype-Rscript.r

#!/usr/bin/env Rscript
# Prototype R script for use at command line in Linux, Mac OS X, UNIX

# References:
#   Manual "A Introduction to R", available via help.start() from the R Console
#   Appendix "B.1 Invoking R from the command line" in "A Inroduction to R",

showArguments <- function(argv)  {
    print(argv)
    0
}

if ( ! interactive() )  {
    # set some error return codes
    SCRIPT_ERROR <- 10                      # see documentation for quit()
    SCRIPT_ARG_ERROR <- SCRIPT_ERROR + 1

    # Define ARGV as script path concatenated to script arguments
    ARGV <- commandArgs(FALSE)          # start with all the arguments given to R
    scriptPath <- sub("^--file=", "", grep("^--file=", ARGV, value=TRUE)) [[1]]
    ARGV <- c(scriptPath, commandArgs(TRUE))

    if (length(ARGV) < 2)   {
        cat(file=stderr(), sep="",
            "Usage: ", ARGV[[1]], " [ options ] item ...\n",
            "       Do something with item\n",
            "       See script for details\n")
        quit(save="no", status=SCRIPT_ARG_ERROR)
    }
    quit(save="no", status=showArguments(ARGV))
}

prototype-shellscript.r

#!/usr/bin/env R --slave --vanilla --quiet -f
# Prototype R script for use at command line in Linux, Mac OS X, UNIX

# References:
#   Manual "A Introduction to R", available via help.start() from the R Console
#   Appendix "B.1 Invoking R from the command line" in "A Inroduction to R",

showArguments <- function(argv)  {
    print(argv)
    0
}

if ( ! interactive() )  {
    # set some error return codes
    SCRIPT_ERROR <- 10                      # see documentation for quit()
    SCRIPT_ARG_ERROR <- SCRIPT_ERROR + 1

    # Define ARGV as the arguments given to this script (after argument “-f”)
    ARGV <- commandArgs(FALSE)          # start with all the arguments given to R
    ARGV <- ARGV[(grep("-f", ARGV) [[1]] + 1):length(ARGV)]
    if ( any(grepl("--args", ARGV) ))   {   # remove arguments intended only for R
        ARGV <- c(ARGV[[1]], commandArgs(TRUE))
    }

    if (length(ARGV) < 2)   {
        cat(file=stderr(), sep="",
            "Usage: ", ARGV[[1]], " [ R_options ] --args [ options ] item ...\n",
            "       Do something with item\n",
            "       See script for details\n")
        quit(save="no", status=SCRIPT_ARG_ERROR)
    }
    quit(save="no", status=showArguments(ARGV))
}