Feature request: `shuffle_rows()` to shuffle rows in a tibble

[larsentom]

Feature request

Add a function shuffle_rows() to randomly reorder rows in the current tibble/slice:

dataset_tbl |>
  dplyr::shuffle_rows()

Motivation

Currently, shuffling all rows in a slice/tibble requires
dataset_tbl |> dplyr::slice_sample(prop = 1) or similar variations.

This approach has two main problems:

1. Unclear intent... because it is not obvious whether the goal is to shuffle rows or sample rows.
2. Fragile code... because small changes to the arguments passed can silently alter the result:
   • if prop is changed from 1, the output is no longer a complete shuffle;
   • if replace = TRUE is used (accidentally or via copy-paste), rows may repeat or disappear.

Introducing shuffle_rows() would make the goal clear and make code more robust.

Implementation sketch

shuffle_rows <- function(.data, ..., .by = NULL) {
  dplyr::slice(
    .data,
    .by = {{ .by }},
    base::local({
      number_of_rows <- dplyr::n()
      base::sample.int(n = number_of_rows, size = number_of_rows, replace = FALSE)
    })
  )
}

Example use

datasets::penguins |>
  shuffle_rows()

# # A tibble: 3 × 8
#   species   island bill_len bill_dep flipper_len body_mass sex     year
#   <fct>     <fct>     <dbl>    <dbl>       <int>     <int> <fct>  <int>
# 1 Adelie    Biscoe     37.6     19.1         194      3750 male    2008
# 2 Chinstrap Dream      46.9     16.6         192      2700 female  2008
# 3 Chinstrap Dream      42.4     17.3         181      3600 female  2007
# ...