Use substrings of column names in across

I have the following tibble:

df <- tibble(
  x = c(5, 5),
  Column_1 = c(0.5, 0.5),
  Column_2 = c(0.75, 0.75))

I would like to create two new columns, which are the product of column x and Column_1 and x and Column_2 respectively. These columns should be named ABC_1 and ABC_2.

I tried the following code:

df <- mutate(df,
  across(starts_with("Column_"),
              function(y) y * x,
              .names = paste0("ABC", gsub("Column_", "", "{.col}"))
        ))

However, this resulted in the following table:

x	Column_1	Column_2	ABCColumn_1	ABCColumn_2
5	0.5	0.75	2.5	3.75
5	0.5	0.75	2.5	3.75

The resulting columns are now named ABCColumn_1 and ABCColumn_2. Somehow creating a substring with gsub did not work. My desired output is:

x	Column_1	Column_2	ABC_1	ABC_2
5	0.5	0.75	2.5	3.75
5	0.5	0.75	2.5	3.75

Any ideas how I can elegantly solve this problem? In reality, the number of columns is variable, so I cannot hardcode 1 and 2 into my script.

EDIT: Sorry, my generic columns names were not well chosen. I chose a different one to show that the old and the new column names should not be related in any way except for the number.

Solution

The glue strings used in the .names argument of across() evaluate calls so you could do:

library(dplyr)

df %>%
  mutate(across(starts_with("Column"), ~ x * .x, .names = "ABC_{seq_along(.col)}"))

# A tibble: 2 × 5
      x Column_1 Column_2 ABC_1 ABC_2
  <dbl>    <dbl>    <dbl> <dbl> <dbl>
1     5      0.5     0.75   2.5  3.75
2     5      0.5     0.75   2.5  3.75

Or use sub() in the glue string:

df %>%
  mutate(across(starts_with("Column"), ~ x * .x, .names = "{sub('Column', 'ABC', .col)}"))