Merge pull request #1 from BerriJ/develop

Release 0.1.0

Author: BerriJ

.github/workflows/R-CMD-check-Dev.yaml

@@ -25,7 +25,7 @@ jobs:
       R_KEEP_PKG_SOURCE: yes
 
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-pandoc@v2
 

.github/workflows/R-CMD-check.yaml

@@ -26,7 +26,7 @@ jobs:
       R_KEEP_PKG_SOURCE: yes
 
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-pandoc@v2
 

.github/workflows/pkgdown.yaml

@@ -11,7 +11,7 @@ jobs:
     env:
       GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-r@v2
 

CRAN-SUBMISSION

@@ -1,3 +1,3 @@
-Version: 0.0.2
-Date: 2022-03-01 09:49:30 UTC
-SHA: 11f3b178b03cc42cf402292a360254958a1ea13f
+Version: 0.1.0
+Date: 2023-09-27 08:05:00 UTC
+SHA: bec2f04098b3d23048a71441174f229ff33bd459

DESCRIPTION

@@ -1,8 +1,8 @@
 Package: dccpp
 Type: Package
 Title: Fast Computation of Distance Correlations
-Version: 0.0.2
-Date: 2022-03-01
+Version: 0.1.0
+Date: 2023-09-27
 Authors@R: c(
     person(given = "Jonathan",
              family = "Berrisch",
@@ -12,13 +12,12 @@ Authors@R: c(
              )
 Description: Fast computation of the distance covariance 'dcov' and distance correlation 'dcor'.  The computation cost is only O(n log(n)) for the distance correlation (see Chaudhuri, Hu (2019) <arXiv:1810.11332> <doi:10.1016/j.csda.2019.01.016>). The functions are written entirely in C++ to speed up the computation.
 License: GPL (>= 3)
-SystemRequirements: C++11
 URL: https://dccpp.berrisch.biz/, https://github.com/BerriJ/dccpp
 BugReports: https://github.com/BerriJ/dccpp/issues
 Encoding: UTF-8
 Imports: Rcpp (>= 1.0.8)
 LinkingTo: Rcpp, RcppArmadillo
-RoxygenNote: 7.1.2
+RoxygenNote: 7.2.3
 Suggests: 
     testthat (>= 3.0.0)
 Config/testthat/edition: 3

NEWS.md

@@ -1,3 +1,10 @@
+dccpp 0.1.0 (Release date: 2023-09-27)
+==============
+
+## Fixes
+
+Fix overflow in `dcov` and `dcor` for large sample sizes (> 46340).
+
 dccpp 0.0.2 (Release date: 2022-01-03)
 ==============
 

README.md

@@ -3,9 +3,9 @@ The dccpp R Package
 ======================
 
 <!-- badges: start -->
-[![GitHub Workflow Status](https://img.shields.io/github/workflow/status/berrij/dccpp/R-CMD-check?label=Build&style=for-the-badge)](https://github.com/BerriJ/dccpp/actions/workflows/R-CMD-check.yaml)
-[![GitHub Workflow Status](https://img.shields.io/github/workflow/status/berrij/dccpp/pkgdown?label=Documentation&style=for-the-badge)](https://dccpp.berrisch.biz/)
-[![Lifecycle: stable](https://img.shields.io/badge/Lifecycle-stable-green?style=for-the-badge)](https://lifecycle.r-lib.org/articles/stages.html#stable)
+[![R-CMD-check](https://img.shields.io/github/actions/workflow/status/berrij/dccpp/R-CMD-check.yaml?&style=for-the-badge)](https://github.com/BerriJ/dccpp/actions/workflows/R-CMD-check.yaml)
+[![GitHub Workflow Status (branch)](https://img.shields.io/github/actions/workflow/status/berrij/dccpp/pkgdown.yaml?label=Documentation&style=for-the-badge)](https://dccpp.berrisch.biz/)
+[![Lifecycle: stable](https://img.shields.io/badge/Lifecycle-stable-brightgreen?style=for-the-badge)](https://lifecycle.r-lib.org/articles/stages.html#stable)
 <!-- badges: end -->
 
 Fast computation of the distance covariance `dcov` and distance correlation `dcor`. The computation cost is only O(n log(n)) for the distance correlation (see Chaudhuri, Hu, 2019, [arXiv](https://arxiv.org/abs/1810.11332), [elsevier](https://doi.org/10.1016/j.csda.2019.01.016)). The functions are written entirely in C++ to speed up the computation.

src/Makevars

@@ -9,8 +9,6 @@
 ## And with R 3.4.0, and RcppArmadillo 0.7.960.*, we turn C++11 on as OpenMP
 ## support within Armadillo prefers / requires it
 
-CXX_STD = CXX11
-
 INC = -I../inst/include
 
 PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS) $(INC)

src/Makevars.win

@@ -9,8 +9,6 @@
 ## And with R 3.4.0, and RcppArmadillo 0.7.960.*, we turn C++11 on as OpenMP
 ## support within Armadillo prefers / requires it
 
-CXX_STD = CXX11
-
 INC = -I../inst/include
 
 PKG_CXXFLAGS = $(SHLIB_OPENMP_CXXFLAGS) $(INC)

src/dc.cpp

@@ -31,19 +31,19 @@ using namespace arma;
 //[[Rcpp::export]]
 double dcov(const arma::colvec &x, const arma::colvec &y)
 {
-    int n = x.n_elem;
+    long long int n = x.n_elem;
 
     // Sort elements by x order
     vec ys = y.elem(sort_index(x));
-    vec xs = arma::sort(x);
+    vec xs = sort(x);
 
     vec si = cumsum(xs);
 
     // Element wise multiplication
     vec ax = regspace(-(n - 2), 2, n) % xs + (si(si.n_elem - 1) - 2 * si);
     mat v = join_horiz(xs, ys, xs % ys);
 
-    arma::umat idx(n, 2, fill::zeros);
+    umat idx(n, 2, fill::zeros);
     idx.col(0) = regspace<uvec>(1, n);
 
     mat iv1(n, 1, fill::zeros);
@@ -54,10 +54,10 @@ double dcov(const arma::colvec &x, const arma::colvec &y)
     int i = 1;
     int r = 1;
     int s = 2;
-
-    int gap, k, st1, e1, st2, e2, kf;
+    int gap, k, e1, e2, kf;
+    long long int st1, st2;
     double idx1, idx2;
-    arma::uvec idxr(idx.n_rows);
+    uvec idxr(idx.n_rows);
     mat csumv(v.n_rows + 1, v.n_cols);
 
     idxr = idx.col(r - 1);