ggplot2 and scales package: the parameter negative_parens = TRUE does not work

I was trying to use the scales package to produce formatting for a complex table. The following helper applies the label_percent and label_number_si functions from the scales package.

For some reason, the negative_parens = TRUE is not producing the correct output:

prettify_numbers_as_percent <- function(x){   
lapply(as.list(as.numeric(x)),label_percent(accuracy = 1,  suffix = "%", negative_parens = TRUE, sep = " ")) %>%

    unlist() %>%
    return() }



prettify_numbers_as_si <- function(x){
  lapply(as.list(as.numeric(x)), label_number_si(accuracy = 1,  negative_parens = TRUE, sep = " ")) %>%
    unlist() %>%
    return()
}

When I run

prettify_numbers_as_si(50000) 
prettify_numbers_as_percent(0.05)

I get the expected output:

"50K"
"5%"

When I run

prettify_numbers_as_si(-50000) 
prettify_numbers_as_percent(-0.05)

I get the incorrect output, despite the fact that negative_parens = TRUE is set:

"-50K"
"-5%"

Does anyone know why parameter setting is failing?

The issue is that while other functions in the scales package have negative_parens= as arguments, label_percent and label_number_si do not. Consequently, you have to write in that logic to your functions:

new_percent <- function(x){
  perc <- lapply(
    as.list(as.numeric(x)),
    label_percent(
      accuracy=1, suffix = '%', sep=' '
    )
  ) %>%
  unlist()

  for(i in 1:length(perc)){
    if(substring(perc[i],1,1)=='-'){
      perc[i] <- paste0('(',substring(perc[i],2),')')
    }
  }
  return(perc)
}

new_numbers <- function(x){
  nums <- lapply(
    as.list(as.numeric(x)),
    label_number_si(
      accuracy = 1,  sep = " "
    )
  ) %>%
  unlist()
  for(i in 1:length(nums)){
    if (substring(nums[i],1,1)=='-'){
      nums[i] <- paste0('(',substring(nums[i],2),')')
    }
  }
  return(nums)
}

Since you know each value in your return that needs to be in parentheses will start with a "-", I'm using a for loop and substring() to iterate through each item and convert those that start with "-" to start and end with parentheses. Works pretty well:

test_perc <- c(-0.5, -0.05, 1.2, .23, -0.13)
test_nums <- c(6000,-60000, 74000, -56000000)

> new_percent(test_perc)
[1] "(50%)" "(5%)"  "120%"  "23%"   "(13%)"
> new_numbers(test_nums)
[1] "6K"    "(60K)" "74K"   "(56M)"