An earlier post illustrated that R object attributes can be set at the C++ level. Naturally, we can also read them from an object. This proves particularly useful for xts objects which are, in essence, numerical matrices with added attributed that are used by a rich set of R operators and functions.

Here, we show how to access these attributes.

#include <Rcpp.h>

using namespace Rcpp;
 
// [[Rcpp::export]]
std::vector<std::string> xtsAttributes(NumericMatrix X) {
    std::vector<std::string> nm = RObject(X).attributeNames();
    return nm;
}

A first example simply creates a random xts object of twenty observations. We then examine the set of attributes and return it in a first program.

  suppressMessages(library(xts))
  set.seed(42)
  n <- 20
  Z <- xts(100+cumsum(rnorm(n)), order.by=ISOdatetime(2013,1,12,20,21,22) + 60*(1:n))
  xtsAttributes(Z)


[1] "dim"         "index"       "class"       ".indexCLASS" "tclass"     
[6] ".indexTZ"    "tzone"      

The same result is seen directly in R:

  names(attributes(Z))


[1] "dim"         "index"       "class"       ".indexCLASS" "tclass"     
[6] ".indexTZ"    "tzone"      

  all.equal(xtsAttributes(Z), names(attributes(Z)))


[1] TRUE

Now, given the attributes we can of course access some of these. The index() function xts objects returns the index. Here, we know we have a Datetime object so we can instantiate it at the C++ level. (Real production code would test types etc).

// [[Rcpp::export]]
DatetimeVector xtsIndex(NumericMatrix X) {
    DatetimeVector v(NumericVector(RObject(X).attr("index")));
    return v;
}

  xtsIndex(Z)


 [1] "2013-01-12 20:22:22 CST" "2013-01-12 20:23:22 CST"
 [3] "2013-01-12 20:24:22 CST" "2013-01-12 20:25:22 CST"
 [5] "2013-01-12 20:26:22 CST" "2013-01-12 20:27:22 CST"
 [7] "2013-01-12 20:28:22 CST" "2013-01-12 20:29:22 CST"
 [9] "2013-01-12 20:30:22 CST" "2013-01-12 20:31:22 CST"
[11] "2013-01-12 20:32:22 CST" "2013-01-12 20:33:22 CST"
[13] "2013-01-12 20:34:22 CST" "2013-01-12 20:35:22 CST"
[15] "2013-01-12 20:36:22 CST" "2013-01-12 20:37:22 CST"
[17] "2013-01-12 20:38:22 CST" "2013-01-12 20:39:22 CST"
[19] "2013-01-12 20:40:22 CST" "2013-01-12 20:41:22 CST"

Further operations such as subsetting based on the datetime vector or adjustments to time zones are left as an exercise.