Using #include <vector> fails checking R package

The most recently added vignette entitled Thirteen Simple Steps for Creating an R Package with an External Library covers the topic you are asking for.

Very broadly, there can be three types of packages with an external library:

• header-only, which is easy as you only need to take care of the -I... flag
• embedded, which can be easy as small libraries can be included in the package
• external for presumably larger libraries which is the hardest variant

External libraries are hardest because the very portable nature of R makes you then worry about how to build with that library on Windows, macOS, and different Linux flavours.

All that said, here your problem must be more fundamental. We cannot say much as you did not provide an MCVE but it is easy to show that Rcpp does of course know where the STL vector class is (initial code is one line, broken here for exposition only)

R> Rcpp::cppFunction("std::vector<int> doubleMe(std::vector<int> x) { \
      std::vector<int> y(x.size()); for (size_t i=0; i<x.size(); i++) \
           y[i] = x[i]+x[i]; return y; }")
R> doubleMe(1:3)
[1] 2 4 6
R> 

which clearly shows that Rcpp knows where to find the header---so either you installed compilers and libraries, or R, in some very unusual way on your computer, or called R the wrong way. Please see A Brief Introduction to Rcpp for more details on Rcpp.

PS With Rcpp types it is just

R> Rcpp::cppFunction("IntegerVector doubleMe(IntegerVector x) { return x+x; }")
R> doubleMe(2:4)
[1] 4 6 8
R> 

PPS Of course you can also do all of this without Rcpp and how to do so is described in the (compulsory reading in this case) Writing R Extensions manual. On my talks page I have links to several (older) tutorials from a decade+ ago that show examples. As this is in fact cumbersome I would recommend Rcpp.

I guess you are using the Rcpp package. Could be a good idea to take a look at the Rcpp vignette. Also the official Writing R Extensions manual is useful here. These contain a lot of the information you need.

There is a difference, if you only need to use vector specifically or if you want to include just any non standard library.

In case you just need to use a vector you could just have LinkingTo: Rcpp in the DESCRIPTION file and after doing all the other required steps for adding Rcpp, a c++ file could look like this example from the vignette:

#include <Rcpp.h>
using namespace Rcpp;

// [[Rcpp::export]]
List rcpp_hello_world() {
  CharacterVector x = CharacterVector::create("foo", "bar");
  NumericVector y = NumericVector::create( 0.0, 1.0 ) ;
  List z = List::create( x, y ) ;

  return z ;
}

Thus you would use e.g. Rcpp::NumericVector provided by the Rcpp namespace.

This of course wouldn't work, if you need different libraries e.g. boost library. There are different ways to go / archive this, but there is also plenty of information out there, also SO questions with useful answers e.g. How to build a R package which use Rcpp with external c++ libraries?

You can also take a look at other R packages, that use Rcpp and additional c++ libraries (e.g. boost). DescTools is a nice example (Git DescTools).

I would suggest you try to follow the guide from the Rcpp vignette to set up Rcpp correctly and if it still fails you follow up here with more detailed information, which library exactly you are trying to add and where it fails.