r包新pca包暫時不用安裝rcpp_1 0 0rcppdocrcpp-quickref_h

1、Rcpp Quick Reference GuideDirk Eddelbuettela and Romain Franoisba; bhttps:/romain.rbind.io/This version was compiled on November 5, 2018This document provides short code snippets that are helpful for using the Rcpp (Eddelbuettel et al., 2018; Eddelbuettel and Franois, 2011

2、; Eddel-buettel, 2013).Rcpp | quickref | R | C+Important NotesUsing matricesCreate simple vectorsInline C+ Compile in RExtract and set singleelements/package=RcppRcpp Vignette | November 5, 2018 | 14/ extract single values double x0 = xx0; double x1 = xx(1);double y0 = yyfoo

3、; double y1 = yybar;# Note - this is R code.# cppFunction in Rcpp allows rapid testing.require(Rcpp)cppFunction(NumericVector exfun(NumericVector x, int i) x = x*i;return x;)SEXP x; std:vector y(10);/ from SEXPNumericVector xx(x);/ of a given size (filled with 0)NumericVector xx(10);/ . with a defau

4、lt for all valuesNumericVector xx(10, 2.0);/ range constructorNumericVector xx(y.begin(), y.end();/ using createNumericVector xx = NumericVector:create(1.0, 2.0, 3.0, 4.0);NumericVector yy = NumericVector:create(Namedfoo = 1.0,_bar= 2.0);/ _ short for Named/ Initializing from SEXP,/ dimensions handl

5、ed automaticallySEXP x;NumericMatrix xx(x);/ Matrix of 4 rows & 5 columns (filled with 0)NumericMatrix xx(4, 5);/ Fill with valueint xsize = xx.nrow() * xx.ncol();for (int i = 0; i xsize; i+) xxi = 7;/ Same as above, using STL fillstd:fill(xx.begin(), xx.end(), 8);/ Assign this value to single eleme

6、nt/ (1st row, 2nd col)xx(0,1) = 4;/ Reference the second column/ Changes propagate to xx (same applies for Row)NumericMatrix:Column zzcol = xx( _, 1); zzcol = zzcol * 2;/ Copy the second column into new objectNumericVector zz1 = xx( _, 1);/ Copy submatrix (top left 3x3) into new objectNumericMatrix

7、zz2 = xx( Range(0,2), Range(0,2);/ If you experience compiler errors, please check/ that you have an appropriate version of g+./ See Rcpp-FAQ for more information./ Many of the examples here imply the following: #include using namespace Rcpp;/ The cppFunction will automatically add this./ Or, prefix

8、 Rcpp objects with the Rcpp namespace/ as e.g. in:Rcpp:NumericVector xx(10);/ set single valuesxx0 = 2.1;xx(1) = 4.2;yyfoo = 3.0;/ grow the vectoryyfoobar = 10.0;Interface with RFirst step in R.STL interfaceAdditional C+.Rcpp AttributesIn C+.On the command-line.Back in R.2 | https:/cran.r-project.or

9、g/package=RcppEddelbuettel and Franois# From shell, above package directoryR CMD build myPackageR CMD check myPackage_1.0.tar.gz # Optional R CMDTALL myPackage_1.0.tar.gz/ Add code below into C+ file Rcpp_example.cpp#include using namespace Rcpp;/ Place the Rcpp:export tag/ right above function decl

10、aration./ Rcpp:exportdouble muRcpp(NumericVector x)int n = x.size(); / Size of vectordouble sum = 0;/ Sum value/ For loop, note cpp index shift to 0for(int i = 0; i x = Rx).NumericVector xx(x);/ clone is slower and uses extra memory./ Safe. No side effects.NumericVector yy(clone(y);xx0 = yy0 = -1.5;

11、int zz = xx0;/ use wrap() to return non-SEXP objects, e.g:/ return(wrap(zz);/ Build and return a listList ret; retx = xx;rety = yy;return(ret);/ sum a vector from beginning to enddouble s = std:accumulate(x.begin(),x.end(), 0.0);/ prod of elements from beginning to endint p = std:accumulate(vec.begi

12、n(),vec.end(), 1, std:multiplies();/ inner_product to compute sum of squaresdouble s2 = std:inner_product(res.begin(),res.end(), res.begin(), 0.0);# In R, create a package shell. For details, # see the Writing R Extensions manual and # the Rcpp-package vignette.Rcpp.package.skeleton(myPackage)# Add

13、R code to pkg R/ directory. Call C+ # function. Do type-checking in R.myfunR - function(Rx, Ry) ret = .Call(myCfun, Rx, Ry,package=myPackage)return(ret)require(myPackage)aa - 1.5bb - 1.5cc - myfunR(aa, bb) aa = bb# FALSE, C+ modifies aaaa - 1:2bb - 1:2cc - myfunR(aa, bb)identical(aa, bb)# TRUE, R/C+

14、 types dont match # so a copy was madeexfun(1:5, 3)# Use evalCpp to evaluate C+ expressionsevalCpp(std:numeric_limits:max()Random Number Generation functionsIn R:.Rcpp ExtensionsRcpp sugarEnvironmentEddelbuettel and FranoisRcpp Vignette| November 5, 2018 | 3/ Obtain an R environment Environment stat

15、s(package:stats); Environment env( 2 ); / by position/ Special environments Environment:Rcpp_namespace(); Environment:base_env(); Environment:base_namespace(); Environment:global_env(); Environment:empty_env();/ Extract function from specific environmentNumericVector x =NumericVector:create(-2.0,-1.

16、0,0.0,1.0,2.0); IntegerVector y =IntegerVector:create(-2, -1, 0, 1, 2);NumericVector xx = abs( x ); IntegerVector yy = abs( y );bool b = all( x 2 ).is_true();NumericVector xx = ceil( x ); NumericVector xx = ceiling( x ); NumericVector yy = floor( y ); NumericVector yy = floor( y );NumericVector xx =

17、 exp( x ); NumericVector yy = exp( y );NumericVector xx = head( x, 2 ); IntegerVector yy = head( y, 2 );IntegerVector xx = seq_len( 10 ); IntegerVector yy = seq_along( y );/ Enable C+11/ Rcpp:plug(cpp11)/ Enable OpenMP (excludes macOS)/ Rcpp:plug(openmp)/ Use the RcppArmadillo package/ Requires diff

18、erent header file from Rcpp.h #include / Rcpp:depends(RcppArmadillo)Rcpp:sourceCpp(path/to/file/Rcpp_example.cpp) x - 1:5all.equal(muRcpp(x), mean(x) all.equal(var(x),varRcpp(x)/ Set seedRNGScope scope;/ For details see Section 6.7.1-Distribution/ functions of the Writing R Extensions manual./ome ca

19、ses (e.g. rnorm), dist-specific/ arguments can be omitted; when in doubt,/ specify all dist-specific arguments. The use/ of doublesrather than integers for dist-/ specific arguments is recommended. Unless/ explicitly specified, log=FALSE./ Equivalent to R calls NumericVector xx = runif(20); NumericV

20、ector xx1 = rnorm(20); NumericVector xx1 = rnorm(20, 0); NumericVector xx1 = rnorm(20, 0, 1);/ Example vector of quantiles NumericVector quants(5); for (int i = 0; i 5; i+) quantsi = (i-2);/ in R, dnorm(-2:2)NumericVector yy = dnorm(quants) ; NumericVector yy = dnorm(quants, 0.0, 1.0) ;/ in R, dnorm

21、(-2:2, mean=2, log=TRUE)NumericVector yy = dnorm(quants, 2.0, true) ;/ Note - cannot specify sd without mean/ in R, dnorm(-2:2, mean=0, sd=2, log=TRUE)NumericVector yy = dnorm(quants, 0.0, 2.0, true) ;/ To get original R api, use Rf_*double zz = Rf_rnorm(0, 2);NumericVector xx = rep( x, 3 ); Numeric

22、Vector xx = rep_len( x, 10 ); NumericVector xx = rep_each( x, 3 );IntegerVector yy = rev( y );double varRcpp(NumericVector x, bool bias = true)/ Calculate the mean using C+ functiondouble mean = muRcpp(x); double sum = 0;int n = x.size();for(int i = 0; i n; i+)sum += pow(xi - mean, 2.0); / Squareret

23、urn sum/(n-bias); / Return varianceCalling Functions in RModulesReferencesEddelbuettel D (2013). Seamless R and C+ Integration with Rcpp. Use R!Springer, New York. ISBN 978-1-4614-6867-7.Eddelbuettel D, Franois R (2011). “Rcpp: Seamless R and C+ Integration.”Journal of Statistical Software, 40(8), 1

24、18. URL /v40/i08/.Eddelbuettel D, Franois R, Allaire J, Ushey K, Kou Q, Russel N, Chambers J, Bates D (2018). Rcpp: Seamless R and C+ Integration. R package version 1.0.0, URL http:/CRAN.R-P/package=Rcpp.4 | /package=RcppEddelbuettel and Fran

25、ois/ Do NOT expect to have a performance gain/ when calling R functions from R!/ Retrieve functions from default loaded env.Function rnorm(rnorm);rnorm(100, _mean = 10.2, _sd = 3.2 );/ Passing in an R function and obtaining results/ Make sure function conforms with return type!NumericVector callFunc

26、tion(NumericVector x,Function f) NumericVector res = f(x);return res;/* R# The following is R code executed # by sourceCpp() as a convenience. x = 1:5callFunction(x, sum)*/ Warning - Module-based objects do not persist/ across quit(save=yes)/reload cycles. To be/ safe, save results to R objects and

27、remove/ module objects before exiting R./ To create a module-containing package from R:/ Rcpp.package.skeleton(mypackage, module=TRUE)class Bar public:Bar(double x_) : x(x_), nread(0), nwrite(0) double get_x( ) nread+;return x;void set_x( double x_) nwrite+; x = x_;IntegerVector stats() const returnInte