[R] #!/usr/bin/env Rscript --vanilla ??

Berend Hasselman bhh at xs4all.nl
Fri Apr 13 21:42:33 CEST 2012


On 13-04-2012, at 10:32, Martin Maechler wrote:

> I think that's my first true question (rather than answer)
> to R-help.
> 
> As R has, for a long time, become my primary scripting and
> programming language, I'm prefering at times to write  Rscript
> files instead of shell scripts, notably when R has nice ways to
> do some of the things.
> On a standard standalone platform with standard R, 
> I would start such a script with
> ---------------------------------------
> #! /usr/bin/Rscript --vanilla
> ---------------------------------------
> (yes, the "--vanilla" is important to me, in this case)
> 
> However; as, at work, my scripts have to work correctly on quite a
> few different (unixy : several flavors of Linux, Solaris, MacOS X) platforms,
> *and* as an R developer, I have many different versions of R
> installed simultaneously, using /usr/bin/Rscript  is not an
> option.
> Rather, I'd use the /usr/bin/env trick :
> 
> ---------------------------------------
> #! /usr/bin/env Rscript
> ---------------------------------------
> 
> which finds Rscript in "the correct" place, according to the
> current PATH.  All fine till now.
> 
> PROBLEM:  It does not work with '--vanilla' or any other argument:
>     If I start my script with	  
> #! /usr/bin/env Rscript --vanilla
>     the error message simply is
> /usr/bin/env: Rscript --vanilla: No such file or directory
> 
> I have tried a few variations on the theme, using quotes in
> different places, but have not succeeded till now.
> Any suggestions?

I had similar problems running R scripts from BBEdit on Mac OS X.
The problem could only be solved by making a shell script that uses the file extension to determine what to run.

Searching internet yielded these pages which may be helpful:

http://www.in-ulm.de/~mascheck/various/shebang/

http://stackoverflow.com/questions/4303128/how-to-use-multiple-arguments-with-a-shebang-i-e

Berend



More information about the R-help mailing list