[R-SIG-Mac] R script in batch mode on Mac

[didier peeters]

Thank you very much Berend !

I haven't tried them all but your third proposition suits me perfectly, it's much simpler that way.  

Didier


Le 7 nov. 2012 à 13:11, Berend Hasselman <bhh at xs4all.nl> a écrit :

> 
> On 07-11-2012, at 12:08, didier peeters wrote:
> 
>> Hi list,
>> 
>> I would like the run a R script in a batch mode.  So i use a command line like this :
>> 
>> for i in *.xlsx; do cat $i | echo 'argv <- $i; source("path/to/script.R")' | R --vanilla --slave ; done
>> 
>> which gives the error " Erreur : '$' inattendu(e) dans "argv <- $" "
>> 
>> i've also tried  :
>> 
>> for i in *.xlsx; do cat $i | R CMD BATCH --args $i "path/to/script.R" ; done
>> 
>> which gives me nothing.  
>> 
>> The idea is to read dozens of similar excel files in a folder with gdata package and to write their content in another single file.  
>> My script works fine with a read.xls when i specify a file name, but and here I get the name of the file with :
>> argv <- commandArgs(TRUE)
>> 
>> This is all I've been able to find from various docs and I'm not familiar with the command line.  
>> What am I doing wrong ?
>> Could anyone help me ?
> 
> 
> Try Rscript like this
> 
> for k in *.R; do
>   Rscript -e "source('path/to/script.R')" $k
> done
> 
> and get the name of the file to process from commandArgs()
> 
> or
> 
> Rscript -e "source('path/to/script.R')" *.xlsx
> 
> and loop over the elements of commandArgs() in the script.
> 
> or
> 
> use list.files() -- something like list.files(pattern="^.*\\.xlsx") -- in the script then you don't need to loop in the shell script.
> 
> Berend
> 
>