[R] [Rscript] difficulty passing named arguments from commandline

[William Dunlap]

The trailling arguments to Rscript, generally read by commandArgs(TRUE), come
into R as a vector of character strings.  Your script can interpret those character
strings in many ways.  The OP's script processed them all with
   eval(parse(text=arg[i]))
so all the arguments had to be valid R expressions: strings must be quoted, unquoted
things are treated as names of R objects, slash means division, "=" and "<-" mean
assignment, etc.

If that is a problem, don't use parse() to interpret the strings; use sub() or strsplit()
to extract substrings and do what you want with them.  (This is somewhat safer
than using eval(parse(text=)) because it can do less.)

E.g.,

% R --quiet --args ./file.txt logFile=./file.log var3=1/3 'stringVar <- paste("var3 is", var3)'
> # note last argument had to be quoted because otherwise the shell would
> # interpret the "<" as redirected input, spaces as breaks between arguments, etc.
> args <- commandArgs(TRUE)
> args
[1] "./file.txt"
[2] "logFile=./file.log"
[3] "var3=1/3"
[4] "stringVar <- paste(\"var3 =\", var3)"
> dataFile <- args[1]
> assign(sub("=.*$", "", args[2]), sub("^[^=]*=", "", args[2]))
> assign(sub("=.*$", "", args[3]), eval(parse(text=sub("^[^=]*=", "", args[3]))))
> eval(parse(text=args[4]))
> objects()
[1] "args"      "dataFile"  "logFile"   "stringVar" "var3"
> dataFile
[1] "./file.txt"
> logFile
[1] "./file.log"
> stringVar
[1] "var3 is 0.333333333333333"
> var3
[1] 0.3333333

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com


> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf
> Of Ben Tupper
> Sent: Thursday, September 06, 2012 7:39 PM
> To: R Help
> Cc: Tom Roche
> Subject: Re: [R] [Rscript] difficulty passing named arguments from commandline
> 
> Hi Tom,
> 
> Using the script below I had no trouble. Here's an example
> 
> Minke:R ben$ Rscript --no-init-file ./argCatcher.rs ./the=trouble .with=./args is=the.y
> argue
> ./the=trouble
> .with=./args
> is=the.y
> argue
> 
> 
> # script starts here
> # argumentCatcher.rs
> args = commandArgs(trailing = TRUE)
> cat(args, sep = "\n")
> q(runLast = FALSE, status = 0, save = "no")
> # script ends here
> 
> What does your R --version report?
> 
> Minke:R ben$ R --version
> R version 2.15.0 (2012-03-30)
> Copyright (C) 2012 The R Foundation for Statistical Computing
> ISBN 3-900051-07-0
> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
> 
> 
> Cheers,
> Ben
> 
> 
> 
> 
> 
> 
> On Sep 6, 2012, at 7:27 PM, Tom Roche wrote:
> 
> >
> > Wanting a commandline solution (for a problem detailed @
> >
> >
> http://mailman.unidata.ucar.edu/mailing_lists/archives/netcdfgroup/2012/msg00279.ht
> ml
> >
> > ) I turned to Rscript, and whacked out the q'n'd
> >
> > https://github.com/TomRoche/GEIA_to_netCDF/blob/master/netCDF.stats.to.stdout.r
> >
> > However it wasn't as quick as hoped, because I spent quite a bit of time
> > figuring out how to pass the arguments. This works (note the quoting):
> >
> > $ Rscript ./netCDF.stats.to.stdout.r 'netcdf.fp="./GEIA_N2O_oceanic.nc"'
> 'var.name="emi_n2o"'
> >> For ./GEIA_N2O_oceanic.nc var=emi_n2o
> >> 	cells=64800
> >> 	obs=36143
> >> 	min=5.96e-08
> >> 	max=1.17e+03
> >> 	mean=99.5
> >> 	med=67.7
> >
> > but this fails
> >
> > roche at amad1:/project/inf14w/roche/GEIA_to_netCDF $ Rscript
> ./netCDF.stats.to.stdout.r 'netcdf.fp=./GEIA_N2O_oceanic.nc' 'var.name=emi_n2o'
> >> Error in eval(expr, envir, enclos) : object '.' not found
> >> Calls: eval -> eval
> >> Execution halted
> >
> > and this fails
> >
> > roche at amad1:/project/inf14w/roche/GEIA_to_netCDF $ Rscript
> ./netCDF.stats.to.stdout.r netcdf.fp="GEIA_N2O_oceanic.nc" var.name="emi_n2o"
> >> Error in eval(expr, envir, enclos) :
> >>  object 'GEIA_N2O_oceanic.nc' not found
> >> Calls: eval -> eval
> >> Execution halted
> >
> > and this fails
> >
> > roche at amad1:/project/inf14w/roche/GEIA_to_netCDF $ Rscript
> ./netCDF.stats.to.stdout.r netcdf.fp=./GEIA_N2O_oceanic.nc var.name=emi_n2o
> >> Error in eval(expr, envir, enclos) : object '.' not found
> >> Calls: eval -> eval
> >> Execution halted
> >
> > Must the quoting be so strict/brittle, or am I missing something?
> >
> > Also, It Would Be Nice if there was more in the `help(Rscript)` examples
> > about argument passing. I for one found the current examples quite terse
> > and unhelpful.
> >
> > TIA, Tom Roche <Tom_Roche at pobox.com>
> >
> > ______________________________________________
> > R-help at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> > and provide commented, minimal, self-contained, reproducible code.
> 
> Ben Tupper
> Bigelow Laboratory for Ocean Sciences
> 180 McKown Point Rd. P.O. Box 475
> West Boothbay Harbor, Maine   04575-0475
> http://www.bigelow.org
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.