Second call to re.eval("source()) fails with JRI

Question

I have installed JRI to run with NetBeans 7.4 using 32-bit R 3.0.2 and Java jdk1.7.0_45, on Windows 7.

I am using the following Java function.

import org.rosuda.JRI.Rengine;
import org.rosuda.JRI.REXP;
import org.rosuda.JRI.RList;

void testJRI(){

    // Start JRI engine. 
    String[] Rargs = {"--vanilla"};
    Rengine re = new Rengine(Rargs, false, null);

    if (!re.waitForR()) {
        System.out.println("Cannot load R");
        return null;
    }

    REXP load=re.eval("source('C:\\\\searchPath\\\\nonparametricAnova.r')");
    re.end();

    return;
}

The first time the function is called, everything works fine. But the second time, the variable, load, is null - indicating failure.

I replaced the nonparametricAnova.r R function with the following simple script

simple<-function(){
  a=1
  a
}

which I named simple.r and called with

    REXP load=re.eval("source('C:\\\\searchPath\\\\simple.r')");

Now, the second time through, it hangs on

Rengine re = new Rengine(Rargs, false, null);

I have noticed that

load=re.eval("refClusterMeasurements<-read.csv(\"C:/SearchPath/fileName.csv\", header=TRUE)");

also fails the second time through, even if "name.csv" is a minimally sized file.

I don't know what is causing the issue but note that detaching a script doesn't make any sense in R. — Dason, Apr 07 '14 at 22:29
Have you tried sourcing a different (perhaps extremely simple) script in place of nonparametricAnova to see if you have the same issue? — Dason, Apr 07 '14 at 22:37
@Dason. I tried that and now it hangs on Rengine re = new Rengine(Rargs, false, null); I have edited my question to reflect that. Thanks, — OtagoHarbour, Apr 07 '14 at 22:53
@Yehoshaphat. I had edited my question to add the imports. Thanks, — OtagoHarbour, Apr 08 '14 at 13:21
Your quotes on your latest edits in the read.csv part are messed up. Is that what you had in your code? — Dason, Apr 08 '14 at 15:08
can you add a sample data set using `dput(data)` , ill run it in my machine. — Yehoshaphat Schellekens, Apr 08 '14 at 18:37
@Yehoshaphat. If I run the script more than once, it doesn't work on any data. I tried Rsession. It has the same basic usage as JRI but does not have this problem. Thanks, — OtagoHarbour, Apr 08 '14 at 18:55
could you post the answers (with the right `csv` url), so other users could enjoy is? — Yehoshaphat Schellekens, Apr 10 '14 at 05:24
@Yehoshaphat. I cannot post the data. It is priority. I am now using Rsession instead and it works fine. Thanks, — OtagoHarbour, Apr 10 '14 at 22:17

score 1 · Answer 1 · answered Apr 20 '14 at 06:08

This version works:

package stackoverflow;

import org.rosuda.JRI.REXP;
import org.rosuda.JRI.Rengine;

/**
 *
 * @author yschellekens
 */
public class StackOverflow {

/**
 * @param args the command line arguments
 */
public static void main(String[] args) {




 // Start JRI engine. 
    String[] Rargs = {"--vanilla"};
    Rengine re = new Rengine(Rargs, false, null);

      re.eval("source('C:/Users/yschellekens.INTRANET/Desktop/java projects/simple.R')");
       REXP value =re.eval("as.integer(a<-simple())");
   int a  = value.asInt();
        System.out.println(a);


    }
}

using the following R file:

simple<-function(){
  a=1
  return(a)
}

Java output:

run:
1

Notice several differences:

Don't use REXP load= before the re.eval("source...
Save your r file with .R
Change your dashes to : /

This code work's fine many times, please let me know if that solved the issue.

It was still hanging after using the above code. Then I added re.end(); just before return and it solved the hanging issue. — Rana, Oct 13 '16 at 08:40
when I am using it as a web service, the first call made returns response but for the second time onwards it got hanged eating memory and CPU. — Rana, Oct 13 '16 at 11:37
I got the solution at [here](http://stackoverflow.com/questions/39994141/calling-r-script-function-from-java-using-rjava/40104285#40104285) — Rana, Oct 18 '16 at 09:22

Second call to re.eval("source()) fails with JRI

1 Answers1