Skip to content

scan or source a text file into a list

5 messages · Andy Bunn, Adaikalavan Ramasamy, Deepayan Sarkar +2 more

Original
Andy Bunn
# 
I've ported somebody else's rather cumbersome Matlab model to R for
colleagues that want a free build of the model with the same type of I/O.

The Matlab model reads a text file with the initial parameters specified as:

C:\Data\Carluc\Rport>more Params.R
# Number of years to simulate
nYears = 50;
# Initial year for graphing purposes
year0 = 1970;
# NPP/GPP ratio (cpp0 unitless)
fnr = 0.30;
# Quantum efficency
qe  = 0.040;

That is, there are four input variables (for this run - there can be many
more) written in a way that R can understand them. In R, I can have the
model source the parameter text file easily enough and have the objects in
the workspace. The model function in R takes a list at runtime. How can I
have R read that file and put the contents into the list I need?

E.g.,
[1] "fnr"    "nYears" "qe"     "year0"
[1] 0.3
[1] 50
$fnr
[1] 0.3

$nYears
[1] 50


The model is then run with
I can't build inputParamList "by hand" as above because the number of
initial parameters changes with the model run and this runs in a wrapper.

Any thoughts? Some combination of paste with scan or parse?
-Andy
_
platform i386-pc-mingw32
arch     i386
os       mingw32
system   i386, mingw32
status
major    2
minor    0.0
year     2004
month    10
day      04
language R
Adaikalavan Ramasamy
# 
Not sure if I understand your question.

If you know the complete list of all possible variable names, then you
can grep for each occurrence and split on the "=". Otherwise, I hope you
have some sort of markers (like main()) to seperate the input and actual
commands. I think perl and bash are probably more suitable for this kind
of job.

Then you can try commandArgs() to feed into R.
On Thu, 2004-11-11 at 15:57, Andy Bunn wrote:
Deepayan Sarkar
#
On Thursday 11 November 2004 09:57, Andy Bunn wrote:
Approaching this slightly differently, how about 

read.pars <- function(file)
{
    foo <- read.table(file, as.is = TRUE)
    ans <- lapply(strsplit(foo[[3]], ";"), as.numeric)
    names(ans) <- foo[[1]]
    ans
}

Deepayan
Roger Bivand
#
On Thu, 11 Nov 2004, Andy Bunn wrote:

            

      
      
      
    

        
I think I would be tempted to try out read.dcf() for this, which returns a 
character matrix of field values in named columns, from which values cast 
to numeric:

            

      
      
      
    

        
nYears: 50
year0: 1970
fnr: 0.30
qe: 0.040

            

      
      
      
    

        
nYears year0  fnr    qe     
[1,] "50"   "1970" "0.30" "0.040"

            

      
      
      
    

        
chr [1, 1:4] "50" "1970" "0.30" "0.040"
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:4] "nYears" "year0" "fnr" "qe"


Since you will know the mode you need, you could read the same-mode fields 
together and cast them:

            

      
      
      
    

        
nYears year0
[1,]     50  1970

and make x a data frame if say x$nYears is needed, which gets to list:

            

      
      
      
    

        
$nYears
[1] 50

$year0
[1] 1970

and c() of the different mode lists to make inputParamList. I think 
treating everything as numeric may be enough, and the DCF format is 
flexible.

Roger

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Gabor Grothendieck
    

    

      
      #
    

  


  

      
Andy Bunn <abunn <at> whrc.org> writes:

: 
: I've ported somebody else's rather cumbersome Matlab model to R for
: colleagues that want a free build of the model with the same type of I/O.
: 
: The Matlab model reads a text file with the initial parameters specified as:
: 
: C:\Data\Carluc\Rport>more Params.R
: # Number of years to simulate
: nYears = 50;
: # Initial year for graphing purposes
: year0 = 1970;
: # NPP/GPP ratio (cpp0 unitless)
: fnr = 0.30;
: # Quantum efficency
: qe  = 0.040;
: 
: That is, there are four input variables (for this run - there can be many
: more) written in a way that R can understand them. In R, I can have the
: model source the parameter text file easily enough and have the objects in
: the workspace. The model function in R takes a list at runtime. How can I
: have R read that file and put the contents into the list I need?
: 
: E.g.,
: > rm(list = ls())
: > source("Params.R")
: > ls()
: [1] "fnr"    "nYears" "qe"     "year0"
: > fnr
: [1] 0.3
: > nYears
: [1] 50
: > foo.list <- list(fnr = fnr, nYears = nYears)
: >
: > foo.list
: $fnr
: [1] 0.3
: 
: $nYears
: [1] 50
: 
: 
: The model is then run with
: > CarlucR(inputParamList = foo.list, ...)
: 
: I can't build inputParamList "by hand" as above because the number of
: initial parameters changes with the model run and this runs in a wrapper.
: 


The following sources the file in the environment that
exists within function f returning a list of all variables in
that environment except those variables that begin with a dot. 
You could alternately replace the source(...) statement with 
eval(parse(.file)) .

R> f <- function(.file) { source(.file, local = TRUE); as.list(environment()) }
R> f("clipboard")
$qe
[1] 0.04

$fnr
[1] 0.3

$year0
[1] 1970

$nYears
[1] 50

: