Check for `Inf` or `-Inf` in the `.Call` R Interface

October 3, 2023

In a R package using compiled code via .Call the macro R_FINITE
can be used to check if a double takes a special value. R_NegInf and
R_PosInf can be used to set it to the negative or the positive
infinite value. But how can we test that a double is equal to the
negative or the positive infinite value?

This is useful for instance when implementing a probability distribution function which should take the values zero and one at the infinite values.

>Solution :

R (and C) have you covered here, you can simply compare for equality:

> Rcpp::cppFunction("bool isPosInf(double x) { return x == R_PosInf; }")
> sapply(c(-Inf, Inf, 0), isPosInf)
[1] FALSE  TRUE FALSE
> 
> Rcpp::cppFunction("bool isNegInf(double x) { return x == R_NegInf; }")
> sapply(c(-Inf, Inf, 0), isNegInf)
[1]  TRUE FALSE FALSE
>

Note that while I use Rcpp for convenience there is nothing specific to Rcpp here as the two symbols compared-to are from the R headers.