gmane.comp.lang.r.rosuda.devel

Re: Possible bug in REXPDouble.NA?

Simon Urbanek — 2012-05-06T16:27:38

On May 6, 2012, at 11:09 AM, Víctor Pascual Cid wrote:


That is correct, R encodes NA in doubles as a very specific NaN value. There is no NaN concept in the other so R can only use "regular" values there.

Cheers,
Simon

Possible bug in REXPDouble.NA?

Víctor Pascual Cid — 2012-05-06T15:09:42

Hi there,

The value of REXPDouble.NA seems to be NaN, while in REXPInteger and REXPLogical has a specific value. Is that correct, or it can be considered as a bug?

System.out.println("Double NA: " + REXPDouble.NA);
System.out.println("Integer NA: " + REXPInteger.NA);
System.out.println("Logical NA: " + REXPLogical.NA);

Double NA: NaN
Integer NA: -2147483648
Logical NA: -128

Cheers,

Víctor

Re: Unable to run Rserve -dbg version

Simon Urbanek — 2012-04-24T15:06:14

On Apr 24, 2012, at 11:55 AM, Suresh Krishna wrote:


You can do that by simply connecting to the Rserve instance with R. There are more options with Rserve 1.x but that is still work in progress.

Note, however, that Rserve supports isolated, parallel connections (something you can't do on Windows), so at the moment you connect the instance is that same as the master process, but anything you do from that point on is private to you (= only the client can see that).



Sure, you're welcome. Suggestions for improvement are always welcome as well.

Cheers,
Simon

Re: Unable to run Rserve -dbg version

Suresh Krishna — 2012-04-24T15:55:58


Yes, I was missing the point. I was expecting something more like an R  
session (if called from within R) where I could access the same workspace  
as the one being accessed by my program (like rcom() does for Windows).

Thanks for the helpful correction and all your work !

Very best, Suresh

Re: Unable to run Rserve -dbg version

Simon Urbanek — 2012-04-24T14:53:45

On Apr 24, 2012, at 11:44 AM, Suresh Krishna wrote:


I think you're missing the point - the debug version doesn't daemonize - it runs in your tty so you can look at the output, abort it etc. It doesn't fork, either. That is the point of the debug version, so you can easily debug it, look at the protocol output, errors, warnings etc.

Cheers,
Simon

Unable to run Rserve -dbg version

Suresh Krishna — 2012-04-24T15:44:35

Hello,

I am having some trouble trying to run Rserve-dbg.so on Mac OS 10.6.

I am using:

alias Rservedbg="R CMD  
/Library/Frameworks/R.framework/Versions/Current/Resources/library/Rserve/libs/x86_64/Rserve-dbg.so"

Now:

Suresh-iMac:~ suresh$ Rservedbg
WARNING: ignoring environment value of R_HOME
Rserve 0.6-8 (338) (C)Copyright 2002-2011 Simon Urbanek
$Id: Rserv.c 338 2012-02-11 20:52:25Z urbanek $

Loading config file /etc/Rserv.conf
Failed to find config file /etc/Rserv.conf

R version 2.15.0 (2012-03-30)
Copyright (C) 2012 The R Foundation for Statistical Computing
ISBN 3-900051-07-0
Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

   Natural language support but running in an English locale

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

Rserve: Ok, ready to answer queries.

<... and here the Terminal hangs, I do not get the prompt back ...>

Hope there is an easy solution ?

Thanks, Suresh

Re: problem installing Deducer on Windows 764-bit

Ian Fellows — 2012-04-18T23:55:35

Hum,

I'm moving this discussion over to the stats-rosuda-devel mailing
list, as it pertains to iplots not Deducer.


My windows VM is Vista 32-bit, and does not display the behavior that
you report. Can you load the JGR console (and iplots) using exe
launcher?

Best,
Ian


On Wed, Apr 18, 2012 at 2:45 PM, Phillip Holcomb <holcomb.phil< at >gmail.com> wrote:

Re: iplots : Moving points

Athula — 2012-04-05T23:40:51


Mohit,

Perhaps you might also want to look at cranvas (package) and its friends ?

See:

https://github.com/ggobi/cranvas/wiki


Cranvas not iplots, but it probably has the functionality that you are 
looking for (caveat: it currently works well on Macs, 
and I have been told it also works under Linux).

Athula.

 
Mohit Dayal <ken.sensei <at> gmail.com> writes:

iplots : Moving points

Mohit Dayal — 2012-03-31T11:19:09

Dear Rosuda group,

I am trying to develop an interactive clustering tool for which I require a
graphics system that allows one to move points interactively. (That is, the
user can brush one or more points, and then when the mouse is dragged, the
points move along). GGobi has a nice implementation of this, but that
system does not really provide me with all that I need (especially
close Rintegration). Therefore I would like to work with
iplots and add this feature.

I've gone through the source of iplots (at least the R side of things) and
I see the function ivar.update. This would do quite nicely for me, but I've
not been able to work out how to make iplots recognize a mouse drag event.
(I'm guessing this needs the java side?) I would appreciate any and all
help in this regard, and am ready to code the required parts myself.

Best,
Mohit Dayal
Researcher
Applied Statistics and Computing Lab
Indian School of Business
Hyderabad, India
www.isb.edu

Re: Rserve server problems

Simon Urbanek — 2012-03-28T12:50:42

Michael,

On Mar 28, 2012, at 8:40 AM, Simon Urbanek wrote:


Actually, sorry, I was skimming through this too quickly and missed the part that you are connecting to localhost and you get some bytes back. (Whether a connection get rejected or not depends on the IP of the established connection's peer so it may still apply). To make sure, you can always use proper networking and allow connection from the host IP (I don't know what networking mode you're using) - I would try that to make sure you don't have issues with the localhost forwarding.

First, try telnet localhost 6311 on the the *host* - there is likely a problem in the forwarding since you get garbage on the connection. Also make sure you have fw disabled for 6311 on the Windows machine.

Cheers,
Simon

Re: Rserve server problems

Simon Urbanek — 2012-03-28T12:40:55

On Mar 28, 2012, at 7:42 AM, Michael Bach wrote:


Yes, likely, you did not enable remote connections in the configuration (they are disabled by default for security reasons) so Rserve is refusing non-local connections. Add "remote enable" to the configuration file.

Cheers,
Simon

Rserve server problems

Michael Bach — 2012-03-28T11:42:35

Hi!

We are trying to use Rserve at work. I am using Rserve 0.6.1 on a Debian
VirtualBox virtual machine as guest and a Windows 7 as host. I want to
access Rserve through a little Java program as a test whether the port
forwarding from localhost port 6311 to debian guest OS works (so Rserve can
be accessed from the network). I checked whether the server is accepting
connections via `telnet localhost 6311' from the debian guest - and it
does. I then tried to connect from the Windows OS via this small Java
program:

import org.rosuda.REngine.*;
import org.rosuda.REngine.Rserve.*;

public class Rserve_test {
    public static void main(String[] args) {
        try {
            System.out.println("test");
            RConnection c = new RConnection("localhost", 6311);
            REXP x = c.eval("R.version.string");
            System.out.println(x.asString());
        } catch (Exception ex) {
        ex.printStackTrace();
        }
    }
}

This fails with the backtrace:

org.rosuda.REngine.Rserve.RserveException: Handshake failed: expected 32
bytes header, got -1
    at org.rosuda.REngine.Rserve.RConnection.<init>(RConnection.java:107)
    at org.rosuda.REngine.Rserve.RConnection.<init>(RConnection.java:60)
    at
at.ac.ait.fabianleimgruber.javar.Rserve_test.main(Rserve_test.java:10)

I found a thread reporting similar behaviour here
http://stackoverflow.com/questions/8925000/rserve-exception-handshake-failedthat

I turned on debugging Rserve via Rserve(TRUE) on the debian machine and get
the following messages:

Rserve: Ok, ready to answer queries.
connection accepted.
sending ID string.
malformed packet (n=5). closing socket to prevent garbage.
DUMP [5]: ff f4 ff fd 06
OUT.sendResp(void data)
DUMP [16]: 02 00 01 42 00 00 00 00 00 00 00 00 00 00 00 00
done.

Am I doing something wrong/missing something? Help is greatly appreciated.

Thanks and Best Regards,
Michael Bach

Re: How to use Rengine

Simon Urbanek — 2012-03-23T17:01:46

On Mar 23, 2012, at 10:20 AM, I.Andoh wrote:


Please read the documentation - R can only be started once, there is not way to re-initialize R so you can't use new Rengine() more than once in one Java process.

Cheers,
Simon

How to use Rengine

I.Andoh — 2012-03-23T14:20:11

Dear sir;
Although the re operate with the following sample programs
the re2 does not operate but it becomes fatal error.
Please let me know whether what I should do.


import org.rosuda.JRI.REXP;
import org.rosuda.JRI.Rengine;
public class Rsample
{
    public static void main(String[] args)
    {

          Rengine re = new Rengine (new String[]{"--no-save"}, false, null);
        re.assign("a", new int[]{36});
        REXP result = re.eval("sqrt(a)");
        System.out.println(result.asDouble());
        re.end();

                  Rengine re2 = new Rengine (new String[]{"--no-save"}, 
false, null);
        re2.assign("b", new int[]{36});
        REXP result2 = re2.eval("sqrt(b)");
        System.out.println(result2.asDouble());
        re2.end();

}
}

error message
run:
6.0
#
# A fatal error has been detected by the Java Runtime Environment:
#
#  Internal Error (0xc0000029), pid=6332, tid=11692
#
# JRE version: 7.0_02-b13
# Java VM: Java HotSpot(TM) Client VM (22.0-b10 mixed mode, sharing 
windows-x86 )
# Problematic frame:
# C  [ntdll.dll+0x90572]
#
# Failed to write core dump. Minidumps are not enabled by default on client 
versions of Windows
#
# An error report file with more information is saved as:
# C:\NetBeans\Rsample\hs_err_pid6332.log
#
# If you would like to submit a bug report, please visit:
#   http://bugreport.sun.com/bugreport/crash.jsp
# The crash happened outside the Java Virtual Machine in native code.
# See problematic frame for where to report the bug.
#
Java Result: 1

Best Regards,

I.Andoh

Re: Problem with Rengine constructor, VM,and System Path

Steven Thorner — 2012-03-21T10:39:49

Hi Regie, 

I expect this is too late to be of use to you, but might be useful to others
with the same problem. 

I had a similar problem with exit after Rengine constructor call, with a message
including <terminated, exit value: 10>. 
I am running Eclipse Indigo (3.7) on
Windows 7, with R version 2.13.2. 

By experimenting with various Run Configuration settings, I eventually got a
meaningful error message.
This indicated that either R has to be properly installed,
resulting in a registry entry, or the R_HOME environmental variable has to be
set. 
I subsequently found two ways to fix the problem:

Either:
Set (or add) your R_HOME system (not User) environmental variable to point to
the folder that R was installed to. For me, this is :
C:\R\R-2.13.2. 
Note, this is not the complete path to your rterm.exe executable, only to the R
root folder. 
Note, also, that I find I always need to restart after changes to
paths, for them to have any effect.

Or:
Re-install R, specifying, at the appropriate stage, that you do not want default
install settings. 
Then, when it asks about registry entries, tick both boxes. 

I also had to import the JRI.jar, JRIEngine.jar and REngine.jar files in to my
Eclipse project from this path under the R home folder: 
/library/rJava/jri.
For each jar, I set the Native Library to this path under the R home folder:
/library/rJava/jri/x64
which is where my jri.dll file resides. (This is for 64-bit R, you may need the
path ending i386 if you're using 32-bit R). 

I found that, after doing this, I do not need to set any Run Configuration
settings for the project. 
I can run the rtest example (from the jri
installation) from within Eclipse, and it works fine. 

Hope this helps,
Steve

Re: Why JGR looks so different?

Markus Helbig — 2012-03-20T18:10:19

hi

i guess this is your window manager configuration which is different. jgr does nit do anything at look and feel which dependa on the user. so currently all java applications will habe this different looks at your setup.

cheers markis

On 20.03.2012, at 16:14, Iurie Malai <iurie.malai< at >gmail.com> wrote:

Re: Why JGR looks so different?

Iurie Malai — 2012-03-20T15:14:53

Sorry! These are the corrected links:
JGR started as a normal user:
http://farm7.staticflickr.com/6233/7000076003_8384b0662c_m.jpg
JGR started  with 'sudo' (as root):
http://farm8.staticflickr.com/7069/6853951548_c8964a08d5_m.jpg

Why JGR looks so different?

Iurie Malai — 2012-03-20T15:00:56

When I start JGR as a normal user, it looks like here:
http://farm7.staticflickr.com/6233/7000076003_8384b0662c_m.jpg.
When I start JGR with 'sudo' (as root), it looks like here:
http://farm8.staticflickr.com/7069/6853951548_c8964a08d5_m.jpg.
Why? From my point of view the second option is more clear (and
pleasant), but this is not a reason to run JGR everytime as a root.
Can I change this?

Regards,
Iurie Malai

How to install GNU R, JGR and Deducer in Ubuntu
https://sites.google.com/site/iuriemalai/en/how-to-install-gnu-r-jgr-and-deducer-in-ubuntu

_______________________________________________
stats-rosuda-devel mailing list
stats-rosuda-devel< at >listserv.uni-augsburg.de
http://mailman.rz.uni-augsburg.de/mailman/listinfo/stats-rosuda-devel

Re: How to kill/close JGR window from Rconcole?

Iurie Malai — 2012-03-20T14:48:35

Run in terminal:

    sudo killall java


În data de 7 martie 2012, 20:38, Iurie Malai a scris:

Rengine kill

I.Andoh — 2012-03-18T23:11:24

Hello

I don’t kill Rengine.
I run next proguram but can’t  execute  “re.end();”.

How to kill the progran.

package jrtest;

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

public class  JRtest {
public static void main(String[] args) {
      if (!Rengine.versionCheck()) {
        System.err.println("** Version mismatch - Java files don't match 
library version.");
        System.exit(1);
    }
    //start
     Rengine re = new Rengine (new String[]{"--no-save"}, false, null);

   double fluct[]={
         -1.510061151, -1.483533210 ,-2.454857885 
,-1.582381336, -2.189866550
, -1.259217396 , 0.122006351 ,-0.680441673 , 0.148050467 ,-0.424794143
, -0.606838795, -0.799486016 ,-0.184970425 , 0.666686479 ,-0.377533334
, -1.320383326, -0.756885509, -0.438778336 ,-1.031146806 , 0.191056368
, -1.154535333 ,-0.132690188 , 0.513218124  ,0.988845215 , 1.649814052
,  0.927037224 , 0.627193855 , 1.149564715 , 0.284609312 ,-0.054383389
, -0.645675322, -0.512513869 ,-0.443688139 , 0.148951284, -0.480104506
, 0.401391522 ,-0.007915557,  0.723504732 , 0.191544878, -0.544863121
,-0.055942539,  0.278057151,  1.922562703,  0.909566946 , 1.774394361
,  1.937790900,  0.921274700 , 1.278086288 , 1.063823536 , 1.014685088
,  0.384750044 , 0.898532960 , 0.436981014, -0.022530751 , 0.453789274
, -0.583885884 ,-1.115714711, -1.475849203 , 0.612645212,  1.711053651 };
     re.eval("library(\"fractal\")");
re.assign("fluct",fluct);
REXP result3 = re.eval("RoverS(fluct)}
System.out.println("Ros="+result3);
re.end();


Best Regard

I.Andoh

How to kill/close JGR window from R concole?

Iurie Malai — 2012-03-07T18:38:15

Is this possible to kill JGR Console from R (in terminal)?
(I know how to close JGR from his own menu).

Regards,
Iurie