Loop Over Rows & Columns of Rcpp Matrix in R (2 Examples)
In this post, we show you how to make a loop over the rows or columns of an Rcpp matrix in R. Rcpp is an R package which allows you to integrate C++ code directly into R, we recommend you to take a look at our post about it when you never worked with Rcpp before.
This post has the following structure.
So let’s start right away!
Example 1: With Standard Rcpp
We load the Rcpp package.
if (!require('Rcpp', quietly = TRUE)) { install.packages('Rcpp') } library('Rcpp') # Load package 'Rcpp'
Define a function called “fun_cpp” which takes a matrix as input and uses a loop over the matrix rows, adding the row index to each row. Important for users of R: Indices start with 0 in C++!
cppFunction(' Rcpp::NumericMatrix fun_cpp( Rcpp::NumericMatrix Mat ) { // You can get the number of rows and columns of matrix Mat as follows Rcout << "Mat.nrow(): " << Mat.nrow() << std::endl; Rcout << "Mat.ncol(): " << Mat.ncol() << std::endl; int nrow_Mat = Mat.nrow(); // Conduct loop over the rows of matrix Mat // For each row: add the index of the row to the matrix values for ( int i=0; i<nrow_Mat; i++ ) { Mat(i, _) = Mat(i, _) + i; } return Mat; } ')
Equivalently, we can write the same function looping over the columns of the matrix instead, called “fun_cpp2”.
cppFunction(' Rcpp::NumericMatrix fun_cpp2( Rcpp::NumericMatrix Mat ) { // You can get the number of rows and columns of matrix Mat as follows Rcout << "Mat.nrow(): " << Mat.nrow() << std::endl; Rcout << "Mat.ncol(): " << Mat.ncol() << std::endl; int ncol_Mat = Mat.ncol(); // Conduct loop over the columns of matrix Mat // For each column: add the number of the column to the matrix values for ( int i=0; i<ncol_Mat; i++ ) { Mat(_, i) = Mat(_, i) + i; } return Mat ; } ')
Try out both functions for an example matrix.
Mat <- matrix(1:(2*3), ncol = 2, nrow = 3) Mat # [,1] [,2] # [1,] 1 4 # [2,] 2 5 # [3,] 3 6 fun_cpp(Mat) # Mat.nrow(): 3 # Mat.ncol(): 2 # [,1] [,2] # [1,] 1 4 # [2,] 3 6 # [3,] 5 8 fun_cpp2(Mat) # Mat.nrow(): 3 # Mat.ncol(): 2 # [,1] [,2] # [1,] 1 5 # [2,] 2 6 # [3,] 3 7
Example 2: With RcppArmadillo
As a data scientist, a very useful additional package is RcppArmadillo. The package brings the rich linear algebra library Armadillo from C++ into R. We have written a blog post about it here. A documentation of the Armadillo functions and classes can be found here.
The equivalent function of fun_cpp() (looping over the rows) with RcppArmadillo is written as
cppFunction(depends = "RcppArmadillo", ' arma::mat fun_cpp_arma(arma::mat Mat) { // You can get the number of rows and columns of matrix Mat as follows Rcout << "Mat.n_rows: " << Mat.n_rows << std::endl; Rcout << "Mat.n_cols: " << Mat.n_cols << std::endl; int nrow_Mat = Mat.n_rows; // Conduct loop over the rows of matrix Mat // For each row: add the number of the row to the matrix values for ( int i=0; i<nrow_Mat; i++ ) { Mat.row(i) = Mat.row(i) + i; } return Mat ; } ')
and the equivalent function of fun_cpp2() (looping over the columns) with RcppArmadillo is written as
cppFunction(depends = "RcppArmadillo", ' arma::mat fun_cpp_arma2(arma::mat Mat) { // You can get the number of rows and columns of matrix Mat as follows Rcout << "Mat.n_rows: " << Mat.n_rows << std::endl; Rcout << "Mat.n_cols: " << Mat.n_cols << std::endl; int ncol_Mat = Mat.n_cols; // Conduct loop over the columns of matrix Mat // For each column: add the number of the column to the matrix values for ( int i=0; i<ncol_Mat; i++ ) { Mat.col(i) = Mat.col(i) + i; } return Mat ; } ')
You can see that the functions return the same output as fun_cpp() and fun_cpp2():
Mat <- matrix(1:(2*3), ncol = 2, nrow = 3) Mat # [,1] [,2] # [1,] 1 4 # [2,] 2 5 # [3,] 3 6 fun_cpp_arma(Mat) # Mat.n_rows: 3 # Mat.n_cols: 2 # [,1] [,2] # [1,] 1 4 # [2,] 3 6 # [3,] 5 8 fun_cpp_arma2(Mat) # Mat.n_rows: 3 # Mat.n_cols: 2 # [,1] [,2] # [1,] 1 5 # [2,] 2 6 # [3,] 3 7
Video & Further Resources
You might ask yourself: As an R user why would you bother to know how to loop through the rows of a matrix in Rcpp? Take a look at our post about the Rcpp package here. There, we show you that the integration of C++ code into R with Rcpp can significantly speed up your code.
For more information on C++, you can take a look at the following YouTube video of Klaus Iglbergers presentation about the Blaze library at the CppCon 2016, uploaded by the CppCon channel.
We have more posts on https://statisticsglobe.com/ which you might want to like into:
- sugar Functions in Rcpp Package in R (Example)
- Construct 3D Array in Rcpp Package in R (2 Examples)
- Call User-Defined R Function from C++ Using Rcpp Package (Example)
- R programming Overview
We showed you some examples of how to define Rcpp functions which loop over the rows and columns of a matrix. For comments or questions, use the section below.
This page was created in collaboration with Anna-Lena Wölwer. Have a look at Anna-Lena’s author page to get more information about her academic background and the other articles she has written for Statistics Globe.
I’m Joachim Schork. On this website, I provide statistics tutorials as well as code in Python and R programming.
