dplyr (version 0.5.0)

sample: Sample n rows from a table.

Description

This is a wrapper around sample.int to make it easy to select random rows from a table. It currently only works for local tbls.

Usage

sample_n(tbl, size, replace = FALSE, weight = NULL, .env = parent.frame())
sample_frac(tbl, size = 1, replace = FALSE, weight = NULL,
  .env = parent.frame())

Arguments

tbl

tbl of data.

size

For sample_n, the number of rows to select. For sample_frac, the fraction of rows to select. If tbl is grouped, size applies to each group.

replace

Sample with or without replacement?

weight

Sampling weights. This expression is evaluated in the context of the data frame. It must return a vector of non-negative numbers the same length as the input. Weights are automatically standardised to sum to 1.

.env

Environment in which to look for non-data names used in weight. Non-default settings for experts only.

Examples

by_cyl <- mtcars %>% group_by(cyl)

# Sample fixed number per group
sample_n(mtcars, 10)
sample_n(mtcars, 50, replace = TRUE)
sample_n(mtcars, 10, weight = mpg)

sample_n(by_cyl, 3)
sample_n(by_cyl, 10, replace = TRUE)
sample_n(by_cyl, 3, weight = mpg / mean(mpg))

# Sample fixed fraction per group
# Default is to sample all data = randomly resample rows
sample_frac(mtcars)

sample_frac(mtcars, 0.1)
sample_frac(mtcars, 1.5, replace = TRUE)
sample_frac(mtcars, 0.1, weight = 1 / mpg)

sample_frac(by_cyl, 0.2)
sample_frac(by_cyl, 1, replace = TRUE)