[Rd] [R] Passing character strings from C code to R

Prof Brian Ripley ripley at stats.ox.ac.uk
Fri Jul 15 17:10:32 CEST 2005


Reading the posting guide would have shown you that R-help was quite 
inappropriate, so I have moved this to R-devel.

allocString does not allocate a character vector, which is what you 
want to set as an attribute (not a 'string', whatever that is).  (It 
creates a CHARSXP, a type that should not be user-visible directly and for 
which STRING_ELT is inappropriate.)

You want something like (PROTECTS omitted):

Str = allocVector(STRSXP, 1);
SET_STRING_ELT(Str, 0, mkChar(comment));

There are many similar examples in the R sources for you to browse.  There 
is even a shortcut that manages the PROTECTS,

Str = mkString(comment);

(That allocString and mkString return different types indicate why I wrote 
`whatever that is': the term is too loose to be useful.)

So you could just have

if(comment && strlen(comment)
    setAttrib(Ret, install("com"), mkString(comment));
if(comment) Free(comment);


On Fri, 15 Jul 2005, Tuszynski, Jaroslaw W. wrote:

> Hi,
>
> I have a C code which produces array of integers and potentially a string,
> and I have problems passing the string out. Here is the relevant part of the
> code:
>
> 1   PROTECT(Ret = allocVector(INTSXP, n));
> 2   ret     = (int*) INTEGER(Ret);  /* get pointer to R's Ret */
> 3   for(i=0; i<n; i++) ret[i] = data[i];
> 4   Free(data);
> 5   i=1;
> 6   if (comment) { // comment was found
> 7     n = strlen(comment);
> 8     if(n>0) {    // and it actually have some length
> 9        Rprintf(" '%s' %i\n", comment, n);
> 10       PROTECT(Str = allocString(n+1));
> 11       str = CHAR(STRING_ELT(Str, 0));
> 12       strcpy(str, comment);
> 13       Rprintf(" '%s' %i\n", str, n);
> 14       setAttrib(Ret, install("comm"), Str);
> 15       i=2;
> 16     }
> 17     Free(comment);
> 18   }
> 20   UNPROTECT(i);
>
> Print statement in line 9 gives right results, but program crashes before
> print statement in line 13.
> Any ideas to what am I doing wrong?

-- 
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 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595



More information about the R-devel mailing list