Skip to content

Yet another batch mode question

3 messages · Mark E. Fenner, Brian Ripley

Mark E. Fenner
# 
Hello all and thanks in advance,

I know you can start R with either:

1.  R CMD BATCH myfile.R

or 

2. R < myfile.R

My question is this:  what if I want multiple input files and/or command
line arguments.  For example, under unix you can run a script command (in
a shell language, perl, or python) like:

python myScript.py

or better yet, if you have things set up right, you can place a
#!/usr/bin/python line at the beginning of the file and then just type (at
the shell command line):

./myScript.py

Now, if I wanted to specify a file to work with and a method to use, I
would just do:

./myScript.py file1 method1

How could I get this behavior in R?  With the commandArg() function I can
retrieve the arguments at the command line, but I'm puzzling over the best
way to implement this system at the command line so that it will be
consistent regardless of how someone chooses to run the script (be it
method 1 or 2 above).

I suppose I could create a script file in {python, perl, bash, etc} that
would simply take the arguments I want and provide them to R in a uniform
manner but a "standalone" R method would be preferable.

Regards,
Mark

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
Brian Ripley
#
On Mon, 8 Oct 2001, Mark E. Fenner wrote:

            

      
      
      
    

        
Um, you can't: try it to see the fatal error.

            

      
      
      
    

        
At present 1 does not pass along further arguments, but sends R method 2
(well, with enough flags to make it work).

If you give 2 further arguments, you will get a strident warning like

auk% R foo
ARGUMENT 'foo' __ignored__

R : Copyright 2001, The R Development Core Team
...

            

      
      
      
    

        
Your examples are entirely Unix, so I am not sure if you are interested in
anything else, but R needs to be.  Already on Windows R allows a file name
to be used as an extra argument: this is needed for file associations and
for drag-and-drop to work.  We have considered allowing that on Unix too.
so that

R /path/to/.Rdata

and

R /path/to/somescript.R

might work.  (Only the first is implemented on Windows at present.)

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Mark E. Fenner
    

    

      
      #
    

  


  

      
On Tue, 9 Oct 2001, Prof Brian D Ripley wrote:

        

            

      
      
      
    

        
I did forget the "--no-save" or "--vanilla" or "--save" option that is
required for #2 but otherwise:

[mfenner at seneca testForList]$ cat test.R
x <- array(trunc(3*runif(100)), c(10,10))
mean(x)

[mfenner at seneca testForList]$cat results
Script started on Tue Oct  9 09:41:52 2001

[mfenner at seneca testForList]$ R --no-save < test.R
R : Copyright 2001, The R Development Core Team
Version 1.3.1  (2001-08-31)

<snip>

            

      
      
      
    

        
[1] 0.94

            

      
      
      
    

        
[mfenner at seneca testForList]$ R CMD BATCH test.R
[mfenner at seneca testForList]$ more test.Rout

R : Copyright 2001, The R Development Core Team
Version 1.3.1  (2001-08-31)
 
<snip>

            

      
      
      
    

        
[1] 0.8

            

      
      
      
    

        
[1] 1.35 0.09 1.47 0.00 0.00

            

      
      
      
    

        
[mfenner at seneca testForList]$ exit
Script done on Tue Oct  9 09:42:43 2001
[mfenner at seneca testForList]$

In short, they both work for me.  You can supress the echoing of commands
by placing a options(echo = FALSE) type line in your script file.

            

      
      
      
    

        
So, do people not use R in this fashion under Unix (as a general
rule)?  That is, running R from the command line with a script and an
input file.

BTW, I am primarily interested in the Unix environment.  Scripting under
Windows raises my blood pressure :) (unless it is in cygwin).

Regards,
Mark

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._