How use correctly Rcpp::pt( )

Question

I am reading the chapter 4 of "Seamless R and C++ Integration with Rcpp" and I had a little problem.

In the "listing 4.13" the book given a example about how to use a function of R. I tried use other functions (different of the example) and I had success. My code is here:

#include <Rcpp.h>

// [[Rcpp::export]]
Rcpp::DataFrame myrandom(Rcpp::NumericVector x) {
  int n = x.size();
  Rcpp::NumericVector y1(n), y2(n), y3(n);

  y1 = Rcpp::pexp(x,1.0,1,0);
  y2 = Rcpp::pnorm(x,0.0,1.0,1,0);
  y3 = Rcpp::ppois(x,3.0,1,0);

  return Rcpp::DataFrame::create(Rcpp::Named("Exp") = y1,Rcpp::Named("Norm") = y2, Rcpp::Named("Pois") = y3);
}


sourceCpp("random.cpp")
myrandom(c(0.5,1))

In this case is OK, but when I try use Rcpp::pt I don't have success. My code is here.

#include <Rcpp.h>

// [[Rcpp::export]]
Rcpp::DataFrame myrandom2(Rcpp::NumericVector x) {
  int n = x.size();
  Rcpp::NumericVector y1(n);

  y1 = Rcpp::pt(x,3.0,0,1,0);

  return Rcpp::DataFrame::create(Rcpp::Named("T") = y1);
}

sourceCpp("random2.cpp")
myrandom2(c(0.5,1))

/Library/Frameworks/R.framework/Versions/3.0/Resources/library/Rcpp/include/Rcpp/stats/nt.h: In function ‘Rcpp::stats::P2<RTYPE, NA, T> Rcpp::pt(const Rcpp::VectorBase<RTYPE, NA, VECTOR>&, double, double, bool, bool) [with int RTYPE = 14, bool NA = true, T = Rcpp::Vector<14>]’:
random2.cpp:8:   instantiated from here
/Library/Frameworks/R.framework/Versions/3.0/Resources/library/Rcpp/include/Rcpp/stats/nt.h:25: error: invalid conversion from ‘double (*)(double, double, int, int)’ to ‘double (*)(double, double, double, int, int)’
/Library/Frameworks/R.framework/Versions/3.0/Resources/library/Rcpp/include/Rcpp/stats/nt.h:25: error:   initializing argument 1 of ‘Rcpp::stats::P2<RTYPE, NA, T>::P2(double (*)(double, double, double, int, int), const Rcpp::VectorBase<RTYPE, NA, VECTOR>&, double, double, bool, bool) [with int RTYPE = 14, bool NA = true, T = Rcpp::Vector<14>]’
make: *** [random2.o] Error 1
llvm-g++-4.2 -arch x86_64 -I/Library/Frameworks/R.framework/Resources/include -DNDEBUG  -I/usr/local/include  -I"/Library/Frameworks/R.framework/Versions/3.0/Resources/library/Rcpp/include"    -fPIC  -mtune=core2 -g -O2  -c random2.cpp -o random2.o 

When I use Rcpp::pt(x,3) I want have control over de parameters od the function

pt(q, df, ncp, lower.tail = TRUE, log.p = FALSE)

I think that I'm not using correctly de function but I don't know what.

Answer 1

Aren't C++ compiler messages lovely? ;-)

I think you just helped squash a really old bug. For both t and F, R actually has two functions group (r|p|q|d)t and (r|p|q|d)nt (ditto for f and nf) where the second form allows for the non-centrality parameter. And I think out header for nt.h was wrong.

If you make this change in file include/Rcpp/stats/nt.h

@@ -22,7 +22,7 @@
 #ifndef Rcpp__stats__nt_h
 #define Rcpp__stats__nt_h

-RCPP_DPQ_2(t,::Rf_dt,::Rf_pt,::Rf_qt)
+RCPP_DPQ_2(nt,::Rf_dnt,::Rf_pnt,::Rf_qnt)

 #endif

(which you can just edit in the header, no reinstallation needed) then things seem to work:

#include <Rcpp.h>

// [[Rcpp::export]]
Rcpp::DataFrame myrandom(Rcpp::NumericVector x) {
  Rcpp::NumericVector y1, y2, y3, y4, y5, y6, y7, y8;

  y1 = Rcpp::pexp(x,1.0,1,0);
  y2 = Rcpp::pnorm(x,0.0,1.0,1,0);
  y3 = Rcpp::ppois(x,3.0,1,0);
  y4 = Rcpp::pt(x,3.0,1,0);
  y5 = Rcpp::pnt(x,3.0,2,TRUE,FALSE);
  y6 = Rcpp::pnt(x,3.0,2,TRUE,TRUE);
  y7 = Rcpp::pnt(x,3.0,2,FALSE,TRUE);
  y8 = Rcpp::pnt(x,3.0,2,FALSE,FALSE);

  return Rcpp::DataFrame::create(Rcpp::Named("Exp") = y1,
                                 Rcpp::Named("Norm") = y2, 
                                 Rcpp::Named("Pois") = y3,
                                 Rcpp::Named("t") = y4,
                                 Rcpp::Named("nt1") = y5,
                                 Rcpp::Named("nt2") = y6,
                                 Rcpp::Named("nt3") = y7,
                                 Rcpp::Named("nt4") = y8);
}

I still need to check the numerics against R, but at least it builds now. (And TRUE/FALSE versus 1/0 does not matter; that gets cast correctly).

Edit: And here is some output:

> sourceCpp("/tmp/so1.cpp")  
R> myrandom(seq(0.2, 0.5, by=0.1))  
       Exp     Norm      Pois        t       nt1      nt2        nt3      nt4   
1 0.181269 0.579260 0.0497871 0.572865 0.0351335 -3.34860 -0.0357655 0.964866   
2 0.259182 0.617911 0.0497871 0.608118 0.0434710 -3.13566 -0.0444441 0.956529  
3 0.329680 0.655422 0.0497871 0.642032 0.0535219 -2.92767 -0.0550074 0.946478  
4 0.393469 0.691462 0.0497871 0.674276 0.0654790 -2.72603 -0.0677212 0.934521   
R>   

Answer 2

It seems that the Student-t distribution doesn't have a non-centrality parameter ncp; from Rmath.h:

/* Student t Distibution */
inline double dt(double x, double n, int lg) { return ::Rf_dt(x, n, lg); }
inline double pt(double x, double n, int lt, int lg) { return ::Rf_pt(x, n, lt, lg); }
inline double qt(double p, double n, int lt, int lg) { return ::Rf_qt(p, n, lt, lg); }
inline double rt(double n) { return ::Rf_rt(n); }

I can compile and execute your code with the following change:

#include <Rcpp.h>

// [[Rcpp::export]]
Rcpp::DataFrame myrandom2(Rcpp::NumericVector x) {
  int n = x.size();
  Rcpp::NumericVector y1(n);

  y1 = Rcpp::pt(x, 3.0, false, true);

  return Rcpp::DataFrame::create(Rcpp::Named("T") = y1);
}