[Rcpp-devel] Problems indexing matrics

[Hadley Wickham]

Hi all,

As I'm (slowly) getting more proficient with Rcpp, I'm attempting to
write a 2d convolution function. My latest attempt:

convolve2d <- cxxfunction(signature(sampleS = "numeric", kernelS =
"numeric"), plugin = "Rcpp", '
    Rcpp::NumericMatrix sample(sampleS), kernel(kernelS);
    int x_s = sample.nrow(), x_k = kernel.nrow();
    int y_s = sample.ncol(), y_k = kernel.ncol();

    Rcpp::NumericMatrix output(x_s + x_k - 1, y_s + y_k - 1, 0.0);
    for (int row = 0; row < x_s; row++) {
      for (int col = 0; col < y_s; col++) {
        for (int i = 0; i < x_k; i++) {
          for (int j = 0; j < y_k; j++) {
            output(row, col) += sample(row + i, col + j) * kernel(i, j);
          }
        }
      }
    }
    return output;
')

fails to compile with a not very informative error message:

Error in compileCode(f, code, language = language, verbose = verbose) :
  Compilation ERROR, function(s)/method(s) not created!
/usr/include/c++/4.2.1/bits/stl_iterator_base_types.h: In
instantiation of ‘std::iterator_traits<double>’:
/Users/hadley/R/Rcpp/include/Rcpp/internal/wrap.h:346:   instantiated
from ‘SEXPREC* Rcpp::internal::range_wrap(InputIterator,
InputIterator) [with InputIterator = double]’
/Users/hadley/R/Rcpp/include/Rcpp/internal/wrap.h:847:   instantiated
from ‘SEXPREC* Rcpp::wrap(InputIterator, InputIterator) [with
InputIterator = double]’
/Users/hadley/R/Rcpp/include/Rcpp/vector/Vector.h:418:   instantiated
from ‘void Rcpp::Vector<RTYPE>::assign(InputIterator, InputIterator)
[with InputIterator = double, int RTYPE = 14]’
/Users/hadley/R/Rcpp/include/Rcpp/vector/Vector.h:249:   instantiated
from ‘Rcpp::Vector<RTYPE>::Vector(InputIterator, InputIterator) [with
InputIterator = double, int RTYPE = 14]’
/Users/hadley/R/Rcpp/include/Rcpp/vector/Matrix.h:60:   instantiated
from ‘Rcpp::Matrix<RTYPE>::Matrix(const i

How can I use that error message to figure out what I'm doing wrong?

I'm also wondering if since I'm using () to index, should I be using 1
based indices?  I think that's the difference for () and [] for
vectors, but I might just be confused.

(Of course a separate question is whether this code actually
implements 2d convolution or not, but it should hopefully be close)

Hadley


-- 
Assistant Professor / Dobelman Family Junior Chair
Department of Statistics / Rice University
http://had.co.nz/