NA values in Rcpp conditional

Question

I am having trouble with conditionals in Rcpp. The best way to explain my problem is through an example.

z <- seq(from=1,to=10,by=0.1)
z[c(5,10,15,20,40,50,80)] <- NA
src <- '
 Rcpp::NumericVector vecz(z);
 for (int i=0;i<vecz.size();i++) {
   if (vecz[i] == NA_REAL) {
     std::cout << "Here is  a missing value" << std::endl;
   }
  }
'
func <- cxxfunction(signature(z="numeric"),src,plugin="Rcpp")
func(z)
# NULL

From my understanding, NA_REAL represents the NA values for the reals in Rcpp and NA_Integer represents the NA values for integers. I'm not sure why the above conditional never returns true given z.

Answer 1

You might want to use the R C level function R_IsNA.

require(Rcpp)
require(inline)
z <- seq(from=1, to=10, by=0.1)
z[c(5, 10, 15, 20, 40, 50, 80)] <- NA

src <- '
 Rcpp::NumericVector vecz(z);
 for (int i=0; i< vecz.size(); i++) {
   if (R_IsNA(vecz[i])) {
     Rcpp::Rcout << "missing value at position " << i + 1  << std::endl;
   }
  }
'

func <- cxxfunction(signature(z="numeric"), src, plugin="Rcpp")

## results
func(z)

missing value at position 5
missing value at position 10
missing value at position 15
missing value at position 20
missing value at position 40
missing value at position 50
missing value at position 80
NULL