R help - Feb 2000 - Reading data from file made by C fwrite

Hello!

First, I must also congratulate the R core team with their accomplishment!
I have gotten to like R a lot, and I have recommended it for inclusion in
an "Astronomy for Linux"-distribution which is in use by many
professional
astronomers and observatories. 

I'm currently working on importing data from files created by a C program
(that I have not written myself, I have the source code, but I don't know
if the author wants to publish it yet). 

I really don't know much about C, but I can understand that the program
writes to a file using 
fwrite(MAP,sizeof(float),MAP_i*MAP_j,f)
where MAP most probably is a pointer to a two-dimensional array of single
precision floats. This is what I want to read into a matrix in R.

When I look at the files written by this program, it is clear that they
are not on a simple form, less says it is a binary file. However, it
appears to be kind of common, as I have seen other packages (such as IDL)
reading files like this without much ado... :-) So, can I read the data of
these files into R? (If not, I may sit down and try to hack it up myself,
some hints would be greatly appreciated).

BTW, something unrelated, if anybody has a couple of examples of
generating random numbers from distribution functions with a couple of
jumps, they would like to share, I'd be happy (I'm an astronomer with
very
little training in statistics, so I'm uncertain whether I've got my code
right (while it returns the expected results, I have a bad feeling)).

Best,

Kjetil
-- 
Kjetil Kjernsmo
Graduate astronomy-student                    Problems worthy of attack
University of Oslo, Norway            Prove their worth by hitting back
E-mail: kjetikj at astro.uio.no                                - Piet Hein
Homepage <URL:http://www.astro.uio.no/~kjetikj/>
Webmaster at skepsis.no 

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

On Tue, 29 Feb 2000, Kjetil Kjernsmo wrote:
> I'm currently working on importing data from files created by a C
program
> (that I have not written myself, I have the source code, but I don't
know
> if the author wants to publish it yet). 
> 
> I really don't know much about C, but I can understand that the program
> writes to a file using 
> fwrite(MAP,sizeof(float),MAP_i*MAP_j,f)
> where MAP most probably is a pointer to a two-dimensional array of single
> precision floats. This is what I want to read into a matrix in R.
Just link a bit of C in to read the format with fread. You will need to do
this anyway: R does not use the float type.

Here's a bit of code we use for similar purposes. This reads 64x64 slices
of 16 bit values from fMRI machine files.

/* Read a slice from a .img file */
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include "R.h"

void readimg(unsigned int *ans, int *nslice, char **name)
{
  int fd, i;
  unsigned int t1, t2;
  char *p, buf[64*64*2];
  long nread;
  
  fd = open(name[0], O_RDONLY);
  if(fd < 0) {
    PROBLEM "can't open file" RECOVER(NULL_ENTRY);
  }
  lseek(fd, 2*64*64*(*nslice), SEEK_SET);
  
  nread = read(fd, buf, 64 * 64*2);
  p = buf;
  for(i = 0; i < 64*64; i++) {
    t1 = (*p++)&0xFF;
    t2 = (*p++)&0xFF;
    *ans++ = 256 * t1 + t2;
  }
  close(fd);
}

called by

getslice <- function (nslice, name = "MRI.img") 
.C("readimg", ans = integer(64 * 64), as.integer(nslice), name)$ans

> When I look at the files written by this program, it is clear that they
> are not on a simple form, less says it is a binary file. However, it
> appears to be kind of common, as I have seen other packages (such as IDL)
> reading files like this without much ado... :-) So, can I read the data of
> these files into R? (If not, I may sit down and try to hack it up myself,
> some hints would be greatly appreciated).
Um. However do they know what it is (float, double, integer,
dimensions)....  Yes, they read it, often incorrectly!

Alternatively, netpbm has lots of filters to dump such formats to ASCII.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

On Tue, 29 Feb 2000, Kjetil Kjernsmo wrote:
> 
> I'm currently working on importing data from files created by a C
program
> (that I have not written myself, I have the source code, but I don't
know
> if the author wants to publish it yet). 
> 
> I really don't know much about C, but I can understand that the program
> writes to a file using 
> fwrite(MAP,sizeof(float),MAP_i*MAP_j,f)
> where MAP most probably is a pointer to a two-dimensional array of single
> precision floats. This is what I want to read into a matrix in R.
> 
We don't currently have any support for this in R. It's not that hard to
do, but it doesn't generalise that well to eg multiple variables.

What you want to do is to write a C function that accepts a filename and
a pointer to a vector of doubles. It should open the file, read each
float, cast it to double, and put it in the vector of doubles, then close
the file.
	
	void readTheFile(char **filename, double *workspace)

Calling this function from R would read the numbers into the supplied
vector, and you could then fix up the dimensions in R (I'm assuming you
know the size of the file here)

	a<-.C("readTheFile",filename,numeric(size.of.file))
	m<-a[[2]]
	dim(m)<-c(100,200) #for example


The manual on writing R Extensions should help.


	-thomas

Thomas Lumley
Assistant Professor, Biostatistics
University of Washington, Seattle

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._