Parsing a string in R

Question

I am writing a R script which I will be running as a script from the command line. I am passing command line arguments to the script using keywords as follows:

myscript.R --arg1=100 --arg2='hello' --arg3=3.14159

I want to write an R function that will return the command line values into a dictionary like object (i.e. a list of lists in R), filling unsupplied arguments with defaults.

for e.g.

parseArguments <- function() {
    options <- commandArgs(TRUE)

    # options now contains "--arg1=100 --arg2='hello' --arg3=3.14159"
    # parse string held in variable named options and stuff into a list 
    # .... Do some string manip ....

    args <- list()
    args['arg1'] <- 100
    args['arg2'] <- 'hello'
    args['arg3'] <- 3.14159
    args['arg4'] <- 123  # Not found in parsed line so we use a hard coded default

    return (args)
}

Can someone help fill in the blanks?

IRTFM · Accepted Answer · 2012-03-11 22:48:23Z

3

> parseArguments <- function() {
+     text1 = "--arg1=100 --arg2='hello' --arg3=3.14159" 
+     eval(parse( text= gsub("\\s", ";", gsub("--","", text1))))
+     args <- list()
+     args['arg1'] <- arg1
+     args['arg2'] <- arg2
+     args['arg3'] <- arg3
+     args['arg4'] <- 123  # Not found in parsed line so we use a hard coded default
+ 
+     return (args)
+ }
> argres <- parseArguments()
> argres
$arg1
[1] 100

$arg2
[1] "hello"

$arg3
[1] 3.14159

$arg4
[1] 123

To address the unkown number of arguments modification to the problem:

 parseArguments <- function() {
     text1 = "--arg1=100 --arg2='hello' --arg3=3.14159" 
     eval(parse(text=gsub("\\s", ";", gsub("--","", text1))))
     args <- list()
     for( ar in ls()[! ls() %in% c("text1", "args")] ) {args[ar] <- get(ar)}
     return (args)
 }
 argres <- parseArguments()
 argres
 #---------
$arg1
[1] 100

$arg2
[1] "hello"

$arg3
[1] 3.14159

edited Mar 11, 2012 at 22:48

answered Mar 11, 2012 at 22:29

IRTFM

264k22 gold badges381 silver badges503 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

Homunculus Reticulli Over a year ago

Ah, I'm glad I asked in here - it would have taken me forever to come up with that solution. One quick question - regarding default values for arguments not specified. The example I gave was a bit contrived. It is not clear from the snippet you provided, how to determine which arguments were not provided at the command line, so that default values could be substituted. Could you please clarify that?. Thanks!

IRTFM Over a year ago

Inside a function you can use ls() to find out what objects exist locally. You could loop over all of the local objects as names to the arg entries and use get() to supply the values. I will add code.

Homunculus Reticulli Over a year ago

Thanks. Lots of interesting new stuff for me. Not sure I understand everything (lots of magic going on), but I hope to learn from the code.

Homunculus Reticulli Over a year ago

DWIn: Could you please explain the meaning of this cryptic expression? for( ar in ls()[! ls() %in% c("text1", "args")] ) - its the list comprehension part that I don't get. Thanks!

IRTFM Over a year ago

I was just removing the other local named objects, text1 and args from the items added to the list, args. ls() returns a character vector of all the names of objects in the local environment, unless you point it at a different environment that is.

|

Vincent Zoonekynd · Accepted Answer · 2012-03-11 22:59:05Z

You can split each argument into the name and value parts, using strsplit or a regular expression. The following does not try to check the type of the arguments: everything will be returned as a string.

parseArgs <- function(...) {
  o <- commandArgs(TRUE)
  # The defaults arguments should be named
  defaults <- list(...)
  stopifnot( length(defaults) == length(names(defaults)) )
  stopifnot( all( names(defaults) != "" ) )
  # All the arguments should be of the form "--foo=bar"
  re <- "^--(.*?)=(.*)"
  stopifnot( all(grepl(re, o)) )
  # Extract the values and names
  r <- gsub(re, "\\2", o)
  names(r) <- gsub("^--(.*?)=(.*)", "\\1", o)
  r <- as.list(r)
  # Add the default values
  missing <- setdiff( names(defaults), names(r) )
  append(r, defaults[missing])
}
print( parseArgs() )
print( parseArgs(foo=1, bar=2) ) # With default values

Collectives™ on Stack Overflow

Parsing a string in R

2 Answers 2

6 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

6 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related