Comments and Ratings

Reply to Martin Heroux:
You can change your code in this way, and it will work.
%%
status = openR;

% Running a simple K-S test to determine whether two samples have the same distribution
a = rand (100,1); % creating a random signal
i = pi/100 :pi/100: pi;
b = (sin (i))' + a; % creating second signal

% Putting both signals into R
putRdata('a',a);
putRdata('b',b);

% Running ks.test in R and retrieving the output
evalR ('output <- ks.test(a, b)')
getRdata ('output')

closeR;

I want to launch R from MATLAB.
I have been using R2.7.0 and DCOM Server 2.5.0 successfully.

I want to upgrade R to 2.8.0 (and later 2.9.2) and I realize that I have to change the DCOM Server. I did the following:
1. Downloaded and installed R 2.8.0 from http://cran.r-project.org/bin/windows/base/old
2. Changed Window system path to point to R-2.8.0/bin
3. Downloaded and installed R_Scilab_DCOM3.0-1B5.exe from http://cran.r-project.org/bin/windows/base (took all options)
4. Installed R package rscproxy_1.3-1.zip from http://cran.r-project.org/web/packages/rscproxy/index.html

I rebooted the computer and run statement openR in MATLAB, a Visual C++ Runtime error box appears, with error message:

Also, I ran the provide (D)COM Server test (Server 1 – Basic Test) from the Windows Start button. In the popped up Window “StatConnector Test”, I click “Start R”, and I got the Visual C++ Run-time error message:
Program: C:\...\Statco~1.exe abnormal program termination.
When error dialog box is OKed, the error “Initializing R .. Function call failed” appeared in the StatConnector Test window.

As a result to this, MATLAB failed to run openR, resulting in a similar Visual C++ Runtime error.

Thanks for help.
Patrick Tai.

i had a similar problem to sophie. I downloaded the package rscproxy to my R installation and the problem went away.

I added R_HOME in the environment variables but still I have the following error:
??? Error: Object returned error code: 0x80040013

when the code is calling: R_lInK_hANdle.Init('R');

It works at XP but when I tried to use at Vista, it did not work. Do I need extra step to make it work? Or openR is not for vista?

Hi,
First of all i want to thank to all developer in such a tool. It seems very useful to me.

However, can you please help me about setting the R_HOME and PATH variables? I can not connect Matlab and R still.

Thanks.
Necati Sekkeli

Very handy and lightweight set of functions. Thank you!

To answer some questions:
1) actxserver is a MATLAB function found only in Windows
2) make sure you install the R-(D)COM interface, paying attention to the instructions about setting your R_HOME and PATH variables in the environment variables for your system. I received this same error, but once I restarted my machine and set the paths (and restarted MATLAB) it worked just fine.

A great idea! However did not work for me.
I tried it under unix and linux. The complain is that 'actxserver' is not found.
Am I blind for something obvious?
Is it only for Windows-users?
A comment will be appreciated.

R_lInK_hANdle.SetSymbol does not work for me.
Could someone help?

I have the same problem. I can't read out "vocano" and the error message is "Invoke Error, Dispatch Exception". What's wrong with it? Thanks.

Fantastic idea and potentially very useful.. Worked well initially, now having trouble (a) moving regression output (eg residual vector) back to Matlab correctly; (b) a simple correlation of 2 vectors in gives lots of 'ActiveX errors'...

 I can read the matrix "vocano" in R command window, but I can't read out "vocano" in using volcano = getRdata('volcano') in matlab; what's wrong with it ?
thanks

This is great to see. I like working with mass spec data, and I'm thrilled to see this work incorporated with Matlab.

Useful, instructive, and nicely presented.

It is a great tool.
However, I got the following error message when using the closeR function:
"Cannot close R session. Error: The remote procedure call failed."
I use Matlab(R2006a), R(2.2.1), and RSrv250. Seems like it is a bug?

easy to install, add extreme power to matlab, especially when using models (regression etc.) mixing both qualitative and quantitative variables.

Works great. Any way to get structures and data frames from R?

Easy to install!