Using Sugar Function cumsum()

Ross Bennett — written Dec 22, 2012 — source

The traditional way to compute the cumulative sum of a vector is with a for loop. This is demonstrated with the function cumsum1().

#include <Rcpp.h>
#include <numeric>   	// for std::partial_sum
using namespace Rcpp;
 
// [[Rcpp::export]]
NumericVector cumsum1(NumericVector x){
    // initialize an accumulator variable
    double acc = 0;
        
    // initialize the result vector
    NumericVector res(x.size());
        
    for(int i = 0; i < x.size(); i++){
         acc += x[i];
         res[i] = acc;
    }
    return res;
}

The C++ standard template library (STL) has the partial_sum() function that computes the cumulative sum of a vector. This is demonstrated with the function cumsum2().

// [[Rcpp::export]]
NumericVector cumsum2(NumericVector x){
    // initialize the result vector
    NumericVector res(x.size());
    std::partial_sum(x.begin(), x.end(), res.begin());
    return res;
 }

With Rcpp sugar, there is a cumsum() function which makes writing this function in C++ very similar to using the cumsum function in R.

// [[Rcpp::export]]
NumericVector cumsum_sug(NumericVector x){
    return cumsum(x);    // compute the result vector and return it
}

And we can of course compare the versions discussed here with the base R variant.

x <- 1:10
all.equal(cumsum1(x), cumsum2(x), cumsum_sug(x), cumsum(x))

[1] TRUE

tags: sugar

Extending R with C++ and Fortran — Dirk Eddelbuettel and JBrandon Duck-Mayr
Benchmarking Rcpp code with RcppClock — Zach DeBruine
Simulation Smoother using RcppArmadillo — Tomasz Woźniak

Using Sugar Function cumsum()

Related Articles