Random sample

base::sample() treats a single numeric x with x >= 1 as 1:x and samples from that sequence. This surprises callers when x may have varying length, such as in sample(x). resample() always samples from the elements of x itself. See base::sample() for more details.

Usage

resample(x, size, replace = FALSE, prob = NULL)

Arguments

x

(Atomic vector)
The atomic vector to sample from.

size

(Scalar integer: [1, Inf))
The number of elements to sample.

replace

(Scalar logical)
Whether or not sampling should be done with replacement.

prob

(numeric: NULL)
A non-negative numeric vector of length length(x) whose elements sum to 1. Element prob[i] is the probability of sampling x[i]. NA values are not allowed.

Value

A vector of length size with elements sampled from x.

Details

When replace = FALSE, size must not exceed length(x).

See also

base::sample()

Examples

#----------------------------------------------------------------------------
# resample() examples
#----------------------------------------------------------------------------
library(bkbase)

set.seed(2)

# Sample without replacement
resample(x = 1:10, size = 5)
#> [1]  5  6  9  1 10

# Sample with replacement
resample(x = c("a", "b", "c"), size = 5, replace = TRUE)
#> [1] "a" "a" "a" "b" "c"

# Weighted sampling
resample(
  x = c("heads", "tails"),
  size = 3, replace = TRUE,
  prob = c(0.7, 0.3)
)
#> [1] "heads" "tails" "tails"

# A length-1 vector is sampled as-is, unlike base::sample()
resample(x = 5L, size = 1L)
#> [1] 5
sample(x = 5L, size = 1L) # samples from 1:5 instead
#> [1] 2