Identifying the OS from R

comments 8

Sometimes a bit of R code needs to know what operating system it’s running on. Here’s a short account of where you can find this information and a little function to wrap the answer up neatly.

Operating systems are a platform issue, so let’s start with the constants in the list .Platform. For Windows the OS.type is just “windows”, but confusingly it’s “unix” for Unix, Linux, and my Mac OSX laptop. To be fair that’s because OSX is built on a base of tweaked BSD Unix. But it does seem that .Platform won’t distinguish OSX from a more traditional Unix or Linux machine.

Nor, by the way, will the value of GUI. When we’re running inside RStudio this is just “RStudio” and from the that’s bundled with the R distribution it’s “AQUA” on OSX and “Rgui” on Windows. Worse, from my command line (Terminal) GUI is “X11”, even though it uses Aqua for the graphics instead. Presumably it’s this same value on Unix and Linux where we really would be using X11.

We can also ask about the R that we’re running by looking at the R.version list. On my machine this has os as “darwin13.4.0”, which is much the same information that presents more neatly.

Amusingly, the help page for R.version can’t quite decide whether or not we should use os to determine which operating system we’re running on: The ‘Note’ tells us not to use it, and the second piece of example code uses it anyway with a comment about it being a “good way to detect OSX”.

Another source of information is the function. On my machine it says that sysname is “Darwin”, which for various not very interesting reasons confirms that I’m running OSX. More straightforwardly, on Windows it’s “Windows” and on Linux it’s “Linux”. That all looks like what we want, except for the cryptic note in the Help pages suggesting that this function is not always implemented.

I’ve never used a machine where returned NULL. Have you?

To sum up then. The easiest way to find out what we want to know is to check["sysname"], remembering that “Darwin” means it’s OSX.

If we’re more cautious or we’re on a mystery machine where really isn’t implemented, then we can first check whether .Platform$OS.type is “windows” or “unix” and if it’s “unix” look carefully at the R.version$os string. This backup route is a bit more involved, so I’ve wrapped everything up in a function that tries to return the lower case name of the operating system we’re running on.

get_os <- function(){
  sysinf <-
  if (!is.null(sysinf)){
    os <- sysinf['sysname']
    if (os == 'Darwin')
      os <- "osx"
  } else { ## mystery machine
    os <- .Platform$OS.type
    if (grepl("^darwin", R.version$os))
      os <- "osx"
    if (grepl("linux-gnu", R.version$os))
      os <- "linux"

Corrections or improvements are welcome.


  1. Will says

    Right. When is available one can basically just use that. I still have no idea when it isn’t. On OpenVMS perhaps. Symbian? ZX Spectrum?

  2. Tom Mitch says

    It is unclear that the OS is needed. What may be needed is the list
    of features and functions in the OS as well as external process methods.

    In the linux world many programs including R work with system differences
    with a configure script.
    ~/src/R/R-3.2.0 $ file configure
    configure: POSIX shell script, ASCII text executable

    The results of running this script and friends are the “specific” answers most are looking for when they
    want to know what the OS is. The number of Fortran compilers this script looks for opens
    other questions.

    It is true that the OS does identify a set of specifics that covers most bases correctly.
    The only thing I would add to your answer is to comment the call to as to
    why. i.e. // so I can correctly emit ksh scripting commands //

  3. Muddassir says this piece of code O.S version and distinguish between OSX and Linux . But where are you printing the information if the O.S is windows ?

    • Will says

      Hint: the if statements selectively override a variable. What value does that take if they aren’t triggered?

  4. Pingback: Distilled News | Data Analytics & R

Leave a Reply to Will Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.