Weighted Random Sampling without Replacement

A collection of implementations of classical and novel algorithms for weighted sampling without replacement.

wrswoR

The goal of wrswoR is to provide faster implementations of weighted random sampling without replacement in R.

Installation

You can install the released version of wrswoR from CRAN with:

install.packages("wrswoR")

And the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("krlmlr/wrswoR")

Example

The functions in this package are a drop-in replacement to sample.int(n, size, replace = FALSE, prob = prob). With large n, sample.int() becomes too slow to be practical, unlike the functions in this package.

library(wrswoR)
set.seed(20200726)

sample_int_crank(20, 10, 1:20)
#>  [1]  8 18 14 17 11 15 10  4 13  5