Skip to content

Save R Objects in C Code

4 messages · Brian Ripley, Ziqiang Zhao, Simon Urbanek

Original
Brian Ripley
#
On 26/06/2013 09:17, Ziqiang Zhao wrote:
Call that R code (using eval).

If you look at the R code for save() you will see that a substantial 
amount of it is interpreted code.
Please see the posting guide and what it says about HTML mail.
3 days later
Ziqiang Zhao
# 
Thanks for your help.
Since I don't know how to reference the current environment, I create a new
one to do so.
It seems that it works.
Not sure if it is helpful to anyone else out there, but here is my solution:
SEXP Save(SEXP X,SEXP File)
{
	char Name='X';

	SEXP EnvList;
	PROTECT(EnvList=allocList(1));
	SET_TAG(EnvList,install(&Name));
	SETCAR(EnvList,X);

	SEXP Env;
	PROTECT(Env=allocSExp(ENVSXP));
	SET_FRAME(Env,EnvList);
	SET_ENCLOS(Env,R_BaseEnv);
	SET_HASHTAB(Env,R_NilValue);
	SET_ATTRIB(Env,R_NilValue);

	int L=length(STRING_ELT(File,0));

	char *CMD=(char *)malloc((L+18)*sizeof(double));
	CMD[0]='s';CMD[1]='a';
CMD[2]='v';CMD[3]='e';CMD[4]='(';CMD[5]='X';CMD[6]=',';CMD[7]='f';CMD[8]='i'
;CMD[9]='l';CMD[10]='e';CMD[11]='=';CMD[12]='"';
	for(int k=0;k<L;k++)
	{
		CMD[k+13]=CHAR(STRING_ELT(File,0))[k];
	}
	CMD[L+13]='"';	CMD[L+14]=')';	CMD[L+15]=';';
CMD[L+16]='\n';CMD[L+17]='\0';

	SEXP RCMD,REXP;
	PROTECT(RCMD=allocVector(STRSXP,1));
	SET_STRING_ELT(RCMD,0,mkChar(CMD));
	ParseStatus Status;
	REXP=PROTECT(R_ParseVector(RCMD,-1,&Status,R_NilValue));
	eval(VECTOR_ELT(REXP,0),Env);
	UNPROTECT(4);
	return R_NilValue;
}

--------------------
Ziqiang Zhao
2013-06-29

-----Original Message-----
From: Prof Brian Ripley [mailto:ripley at stats.ox.ac.uk] 
Sent: Wednesday, June 26, 2013 4:32 PM
To: johnzhao2835 at gmail.com
Cc: r-devel at r-project.org
Subject: Re: [Rd] Save R Objects in C Code
On 26/06/2013 09:17, Ziqiang Zhao wrote:
Call that R code (using eval).

If you look at the R code for save() you will see that a substantial amount
of it is interpreted code.
Please see the posting guide and what it says about HTML mail.
Simon Urbanek
#
On Jun 29, 2013, at 11:14 AM, Ziqiang Zhao wrote:

            

      
      
      
    

        
There are tons of issues with this (first I meant to point them all out but then realized it would make the e-mail 10-times longer than the original code). So to other readers, do not use that code! Not only is it ugly, but it has overflows, leaks and it uses API calls that do something entirely different than what the author thought they do. In addition, you really don't want to be parsing anything - you can construct LANGSXP directly with the values - there are examples in R-exts how to do that properly. As for the environment - if you don't know what you're doing, just use R_GlobalEnv - it's the closes to what most people need (if you need something else then you know how to get that env ;)).

Cheers,
Simon