Arrange by a dynamically specified column

Question

I am trying use the arrange() function from dplyr but specify the column dynamically.

Here is a reprex:

dt <- tibble(
  foo = 1:10,
  bar = 101:110
)

sort_by <- 'bar'

dt %>% arrange(sort_by) # does not work

dt %>% arrange(!!quo_name(sort_by)) # does not work either

What is the correct way to pass a dynamic column name to arrange()?

Answer 1

Here are 4 approaches to pass string variables in arrange.

library(dplyr)
library(rlang)

1. arrange_at

dt %>% arrange_at(sort_by)

2. across - _at verbs are deprecated so we can use across from dplyr 1.0.0 instead.

dt %>% arrange(across(all_of(sort_by)))

3. Using non-standard evaluation with sym and !!.

dt %>% arrange(!!sym(sort_by))

4. Using .data pronoun

dt %>% arrange(.data[[sort_by]])

Answer 2

Try get

dt %>% arrange(get(sort_by))