In the case that R code just isn’t fast enough, one available option is to convert your R code into C++. This does take some extra effort the first time around, and you have to know a bit about C++, but it can be a very powerful approach. Converting code to C++ is especially useful if you have to run a large number of loops in R. The more appropriate choice would be to vectorize the R code, but sometimes you have no other option.

The link between R and C++ is a package appropriately named RCpp. It makes your life easy by not having to worry about garbage collection, and there is even code completion for the C++ code!

Note that often only key time intensive functions need to be converted to C++.

Setting up RCpp

Here are the steps to set up Rcpp, and write a basic function in C++ that can be used in R.

Create a C++ file with a function that you want to use in R. We can do this right from RStudio. File -> New File -> C++ File will open one right up for us. Let’s create a simple function that will sort a vector of numbers that we pass it. If you are familiar with C++, there may be a few things in the code that seem strange. Ignore them for now, we will get to them later.

#include <Rcpp.h>
using namespace Rcpp;
using namespace std;

// [[Rcpp::export]]
void sortPrinter(NumericVector x) {
  Rcout << "[C++] Before: " << x << '\n'; 
  sort(x.begin(),x.end());
  Rcout << "[C++] After: " << x << '\n';  
}

Install the RCpp package.

Cpp + R = RCpp

Setting up RCpp