Calling an R function

R is an open source programming language commonly used for data analysis. This page describes how to call an existing R function from a model running in the Analytica Enterprise edition.

** This is currently just a place to take notes -- I do not yet have a functioning example **

Using the RDCOMServer package

One approach of many is to use the RDCOMServer package in R.

Installation of R

1. Install R
2. Install the RDCOMServer package for R
   1. Download the Windows Binary ZIP file
   2. Unzip it into %R_HOME%\library. For me this was "C:\Program Files\R\R-4.3.0\library"
3. Install the SWinRegistry package.
   1. Download the Windows Binary ZIP file
   2. Unzip it into %R_HOME%\library. For me this was "C:\Program Files\R\R-4.3.0\library"
   3. Run: Rcmd install RWinRegistry
4. Install the R package Ruuid.
   1. Start R
   2. install.packages("BiocManager")

Install packages

1. Launch R
2. Run:

   install.packages("devtools")

   devtools::install_github("omegahat/RDCOMServer")

-- This results in an error. RDCOMServer not yet successfully installed. This part is being researched.

Server-side code

• Note: This part is untested, since I haven't yet been able to successful install the required R packages. Work in progress.

Here's example R-code that creates a simple COM server with a single method named multiply:

library(RDCOMServer) 
 
serverDesc <- COMCreate("R.Server")
serverDesc$setClass("R.Server")
 
# Create the `multiply` method
serverDesc$addMethodDef(
  functionName = "multiply",
  args = c("x" = "numeric", "y" = "numeric"),
  return = "numeric",
  def = function(self, x, y) x * y
)
 
# Register the server
serverDesc$register()
 
# Start the server
COMServer()

Analytica-side call

1. Create the variable:

   Variable RServer ::= COMCreateObject("R.Server")

   Note: Launch the R-side server before evaluating RServer

2. Call the method within a different variable:

   Variable Y ::= RServer->multiply(5,7)