my question is trivial, but I'm new with RCPP and still have not mastered. I wanted to make a function that given a categorical vector and two integers returns the subset of vector content between the two integers. You know, ["A","B","C","D] and 1 and 2, returns me ["B","C"].
I did the next code but doesn't work.
// [[Rcpp::export]]
Rcpp::StringVector Segment(Rcpp::StringVector x, int start, int end) {
Rcpp::StringVector s(end-start+1);
for(int i=start; i <= end; i++){
s[i]=x[i];
}
return(s);
}
Try to use CharacterVector. The solution could be like this.
// [[Rcpp::export]]
CharacterVector Segment( CharacterVector x, int start, int end){
CharacterVector r(end-start+1);
int ind=0;
for( int i=0; i<=r.size(); i++){
if((i>=start) & (i<=end)){
r[ind]=x[i];
ind+=1;
}
}
return(r);
}
Related
I am using Rcpp to speed up some R code. However, I'm really struggling with types - since these are foreign in R. Here's a simplified version of what I'm trying to do:
#include <RcppArmadillo.h>
#include <algorithm>
//[[Rcpp::depends(RcppArmadillo)]]
using namespace Rcpp;
using namespace arma;
// [[Rcpp::export]]
NumericVector fun(SEXP Pk, int k, int i, const vec& a, const mat& D) {
// this is dummy version of my actual function - with actual arguments.;
// I'm guessing SEXP is going to need to be replaced with something else when it's called from C++ not R.;
return D.col(i);
}
// [[Rcpp::export]]
NumericVector f(const arma::vec& assignment, char k, int B, const mat& D) {
uvec k_ind = find(assignment == k);
NumericVector output(assignment.size()); // for dummy output.
uvec::iterator k_itr = k_ind.begin();
for(; k_itr != k_ind.end(); ++k_itr) {
// this is R code, as I don't know the best way to do this in C++;
k_rep = sample(c(assignment[assignment != k], -1), size = B, replace = TRUE);
output = fun(k_rep, k, *k_itr, assignment, D);
// do something with output;
}
// compile result, ultimately return a List (after I figure out how to do that. For right now, I'll cheat and return the last output);
return output;
}
The part I'm struggling with is the random sampling of assignment. I know that sample has been implemented in Rarmadillo. However, I can see two approaches to this, and I'm not sure which is more efficient/doable.
Approach 1:
Make a table of theassignment values. Replace assignment == k with -1 and set its "count" equal to 1.
sample from the table "names" with probability proportional to the count.
Approach 2:
Copy the relevant subset of the assignment vector into a new vector with an extra spot for -1.
Sample from the copied vector with equal probabilities.
I want to say that approach 1 would be more efficient, except that assignment is currently of type arma::vec, and I'm not sure how to make the table from that - or how much of a cost there is to converting it to a more-compatible format. I think I could implement Approach 2, but I'm hoping to avoid the expensive copy.
Thanks for any insights you can provide.
many variable declaration is not coherent with the assignment made by you, like assignment = k is impossible to compare as assignment has real value and k is a char. as the queston is bad written I feel free to change the variables type. this compile..
// [[Rcpp::depends(RcppArmadillo)]]
#include <RcppArmadillo.h>
#include <RcppArmadilloExtensions/sample.h>
// [[Rcpp::export]]
arma::vec fun(const Rcpp::NumericVector& Pk, int k, unsigned int i, const arma::ivec& a, const arma::mat& D)
{
return D.col(i);
}
// [[Rcpp::export]]
Rcpp::NumericMatrix f(const arma::ivec& assignment, int k, unsigned int B, const arma::mat& D)
{
arma::uvec k_ind = find(assignment == k);
arma::ivec KK = assignment(find(assignment != k));
//these 2 row are for KK = c(assignment[assignment != k], -1)
//I dont know what is this -1 is for, why -1 ? maybe you dont need it.
KK.insert_rows(KK.n_rows, 1);
KK(KK.n_rows - 1) = -1;
arma::uvec k_ind_not = find(assignment != k);
Rcpp::NumericVector k_rep(B);
arma::mat output(D.n_rows,k_ind.n_rows); // for dummy output.
for(unsigned int i =0; i < k_ind.n_rows ; i++)
{
k_rep = Rcpp::RcppArmadillo::sample(KK, B, true);
output(arma::span::all, i) = fun(k_rep, k, i, assignment, D);
// do something with output;
}
// compile result, ultimately return a List (after I figure out how to do that. For right now, I'll cheat and return the last output);
return Rcpp::wrap(output);
}
this is not optimized (as the question is bogus), this is badly written, beccause as I think R would be sufficiently faster in searching index of a vector (so do this in R and implemement only fun in Rcpp)...is not useful to waste time here, there are other problems that need a solver implemented in Rcpp , not this searching stuff...
but this is not a useful question as you are asking more for an algorithm than for example signature of function
I have to convert individual elements of Rcpp::IntegerVector into their string form so I can add another string to them. My code looks like this:
#include <Rcpp.h>
using namespace Rcpp;
// [[Rcpp::export]]
Rcpp::String int_to_char_single_fun(int x){
// Obtain environment containing function
Rcpp::Environment base("package:base");
// Make function callable from C++
Rcpp::Function int_to_string = base["as.character"];
// Call the function and receive its list output
Rcpp::String res = int_to_string(Rcpp::_["x"] = x); // example of original param
// Return test object in list structure
return (res);
}
//[[Rcpp::export]]
Rcpp::CharacterVector add_chars_to_int(Rcpp::IntegerVector x){
int n = x.size();
Rcpp::CharacterVector BASEL_SEG(n);
for(int i = 0; i < n; i++){
BASEL_SEG[i] = "B0" + int_to_char_single_fun(x[i]);
}
return BASEL_SEG;
}
/*** R
int_vec <- as.integer(c(1,2,3,4,5))
BASEL_SEG_char <- add_chars_to_int(int_vec)
*/
I get the following error:
no match for 'operator+'(operand types are 'const char[3]' and 'Rcpp::String')
I cannot import any C++ libraries like Boost to do this and can only use Rcpp functionality to do this. How do I add string to integer here in Rcpp?
We basically covered this over at the Rcpp Gallery when we covered Boost in an example for lexical_cast (though that one went the other way). So rewriting it quickly yields this:
Code
// We can now use the BH package
// [[Rcpp::depends(BH)]]
#include <Rcpp.h>
#include <boost/lexical_cast.hpp>
using namespace Rcpp;
using boost::lexical_cast;
using boost::bad_lexical_cast;
// [[Rcpp::export]]
std::vector<std::string> lexicalCast(std::vector<int> v) {
std::vector<std::string> res(v.size());
for (unsigned int i=0; i<v.size(); i++) {
try {
res[i] = lexical_cast<std::string>(v[i]);
} catch(bad_lexical_cast &) {
res[i] = "(failed)";
}
}
return res;
}
/*** R
lexicalCast(c(42L, 101L))
*/
Output
R> Rcpp::sourceCpp("/tmp/lexcast.cpp")
R> lexicalCast(c(42L, 101L))
[1] "42" "101"
R>
Alternatives
Because converting numbers to strings is as old as computing itself you could also use:
itoa()
snprintf()
streams
and probably a few more I keep forgetting.
As others have pointed out, there are several ways to do this. Here are two very straightforward approaches.
1. std::to_string
Rcpp::CharacterVector add_chars_to_int1(Rcpp::IntegerVector x){
int n = x.size();
Rcpp::CharacterVector BASEL_SEG(n);
for(int i = 0; i < n; i++){
BASEL_SEG[i] = "B0" + std::to_string(x[i]);
}
return BASEL_SEG;
}
2. Creating a new Rcpp::CharacterVector
Rcpp::CharacterVector add_chars_to_int2(Rcpp::IntegerVector x){
int n = x.size();
Rcpp::CharacterVector BASEL_SEG(n);
Rcpp::CharacterVector myIntToStr(x.begin(), x.end());
for(int i = 0; i < n; i++){
BASEL_SEG[i] = "B0" + myIntToStr[i];
}
return BASEL_SEG;
}
Calling them:
add_chars_to_int1(int_vec) ## using std::to_string
[1] "B01" "B02" "B03" "B04" "B05"
add_chars_to_int2(int_vec) ## converting to CharacterVector
[1] "B01" "B02" "B03" "B04" "B05"
I have a dataframe 'tmp' on which I need to do perform calculation using the last row of another dataframe 'SpreadData'. I am using following code:
for(i in 1:ncol(tmp)){for(j in 1:nrow(tmp)){PNLData[j,i] = 10*tmp[j,i]*SpreadData[nrow(SpreadData),i]}}
Is there any faster method using mapply or something else so that I need not to use for loop.
Thanks
You can use sweep():
PNLData <- sweep(10 * tmp, 2, SpreadData[nrow(SpreadData), ], "*")
PS1: you can replace SpreadData[nrow(SpreadData), ] by tail(SpreadData, 1).
PS2: I think this makes two copies of your data. If you have a large matrix, you'd better use Rcpp.
Edit: Rcpp solution: put that an .cpp file and source it.
#include <Rcpp.h>
using namespace Rcpp;
// [[Rcpp::export]]
NumericMatrix rcppFun(const NumericMatrix& x,
const NumericVector& lastCol) {
int n = x.nrow();
int m = x.ncol();
NumericMatrix res(n, m);
int i, j;
for (j = 0; j < m; j++) {
for (i = 0; i < n; i++) {
res(i, j) = 10 * x(i, j) * lastCol[j];
}
}
return res;
}
And do in R PNLData <- rcppFun(tmp, SpreadData[nrow(SpreadData), ]).
I created a cumsum function in an R package with rcpp which will cumulatively sum a vector until it hits the user defined ceiling or floor. However, if one wants the cumsum to be bounded above, the user must still specify a floor.
Example:
a = c(1, 1, 1, 1, 1, 1, 1)
If i wanted to cumsum a and have an upper bound of 3, I could cumsum_bounded(a, lower = 1, upper = 3). I would rather not have to specify the lower bound.
My code:
#include <Rcpp.h>
#include <float.h>
#include <cmath>
using namespace Rcpp;
// [[Rcpp::export]]
NumericVector cumsum_bounded(NumericVector x, int upper, int lower) {
NumericVector res(x.size());
double acc = 0;
for (int i=0; i < x.size(); ++i) {
acc += x[i];
if (acc < lower) acc = lower;
else if (acc > upper) acc = upper;
res[i] = acc;
}
return res;
}
What I would like:
#include <Rcpp.h>
#include <float.h>
#include <cmath>
#include <climits> //for LLONG_MIN and LLONG_MAX
using namespace Rcpp;
// [[Rcpp::export]]
NumericVector cumsum_bounded(NumericVector x, long long int upper = LLONG_MAX, long long int lower = LLONG_MIN) {
NumericVector res(x.size());
double acc = 0;
for (int i=0; i < x.size(); ++i) {
acc += x[i];
if (acc < lower) acc = lower;
else if (acc > upper) acc = upper;
res[i] = acc;
}
return res;
}
In short, yes its possible but it requires finesse that involves creating an intermediary function or embedding sorting logic within the main function.
In long, Rcpp attributes only supports a limit feature set of values. These values are listed in the Rcpp FAQ 3.12 entry
String literals delimited by quotes (e.g. "foo")
Integer and Decimal numeric values (e.g. 10 or 4.5)
Pre-defined constants including:
Booleans: true and false
Null Values: R_NilValue, NA_STRING, NA_INTEGER, NA_REAL, and NA_LOGICAL.
Selected vector types can be instantiated using the
empty form of the ::create static member function.
CharacterVector, IntegerVector, and NumericVector
Matrix types instantiated using the rows, cols constructor Rcpp::Matrix n(rows,cols)
CharacterMatrix, IntegerMatrix, and NumericMatrix)
If you were to specify numerical values for LLONG_MAX and LLONG_MIN this would meet the criteria to directly use Rcpp attributes on the function. However, these values are implementation specific. Thus, it would not be ideal to hardcode them. Thus, we have to seek an outside solution: the Rcpp::Nullable<T> class to enable the default NULL value. The reason why we have to wrap the parameter type with Rcpp::Nullable<T> is that NULL is a very special and can cause heartache if not careful.
The NULL value, unlike others on the real number line, will not be used to bound your values in this case. As a result, it is the perfect candidate to use on the function call. There are two choices you then have to make: use Rcpp::Nullable<T> as the parameters on the main function or create a "logic" helper function that has the correct parameters and can be used elsewhere within your application without worry. I've opted for the later below.
#include <Rcpp.h>
#include <float.h>
#include <cmath>
#include <climits> //for LLONG_MIN and LLONG_MAX
using namespace Rcpp;
NumericVector cumsum_bounded_logic(NumericVector x,
long long int upper = LLONG_MAX,
long long int lower = LLONG_MIN) {
NumericVector res(x.size());
double acc = 0;
for (int i=0; i < x.size(); ++i) {
acc += x[i];
if (acc < lower) acc = lower;
else if (acc > upper) acc = upper;
res[i] = acc;
}
return res;
}
// [[Rcpp::export]]
NumericVector cumsum_bounded(NumericVector x,
Rcpp::Nullable<long long int> upper = R_NilValue,
Rcpp::Nullable<long long int> lower = R_NilValue) {
if(upper.isNotNull() && lower.isNotNull()){
return cumsum_bounded_logic(x, Rcpp::as< long long int >(upper), Rcpp::as< long long int >(lower));
} else if(upper.isNull() && lower.isNotNull()){
return cumsum_bounded_logic(x, LLONG_MAX, Rcpp::as< long long int >(lower));
} else if(upper.isNotNull() && lower.isNull()) {
return cumsum_bounded_logic(x, Rcpp::as< long long int >(upper), LLONG_MIN);
} else {
return cumsum_bounded_logic(x, LLONG_MAX, LLONG_MIN);
}
// Required to quiet compiler
return x;
}
Test Output
cumsum_bounded(a, 5)
## [1] 1 2 3 4 5 5 5
cumsum_bounded(a, 5, 2)
## [1] 2 3 4 5 5 5 5
Deep inside an MCMC algorithm I need to multiply a user-provided list of matrices with a vector, i.e., the following piece of Rcpp and RcppArmadillo code is called multiple times per MCMC iteration:
List mat_vec1 (const List& Mats, const vec& y) {
int n_list = Mats.size();
Rcpp::List out(n_list);
for (int i = 0; i < n_list; ++i) {
out[i] = as<mat>(Mats[i]) * y;
}
return(out);
}
The user-provided list Mats remains fixed during the MCMC, vector y changes in each iteration. Efficiency is paramount and I'm trying to see if I can speed up the code by not having to convert the elements of Mats to arma::mat that many times (it needs to be done only once). I tried the following approach
List arma_Mats (const List& Mats) {
int n_list = Mats.size();
Rcpp::List res(n_list);
for (int i = 0; i < n_list; ++i) {
res[i] = as<mat>(Mats[i]);
}
return(res);
}
and then
List mat_vec2 (const List& Mats, const vec& y) {
int n_list = Mats.size();
Rcpp::List aMats = arma_Mats(Mats);
Rcpp::List out(n_list);
for (int i = 0; i < n_list; ++i) {
out[i] = aMats[i] * y;
}
return(out);
}
but this does not seem to work. Any pointers of alternative/better solutions are much welcome.
Ok, I basically wrote the answer in the comment but it then occurred to me that we already provide a working example in the stub created by RcppArmadillo.package.skeleton():
// [[Rcpp::export]]
Rcpp::List rcpparma_bothproducts(const arma::colvec & x) {
arma::mat op = x * x.t();
double ip = arma::as_scalar(x.t() * x);
return Rcpp::List::create(Rcpp::Named("outer")=op,
Rcpp::Named("inner")=ip);
}
This returns a list the outer product (a matrix) and the inner product (a scalar) of the given vector.
As for what is fast and what is not: I recommend to not conjecture but rather profile and measure as much as you can. My inclination would be to do more (standalone) C++ code in Armadillo and only return at the very end minimizing conversions.