Feature request: adorn_n_formatting()

[Thyrisis]

I need to present numbers as well as percentages in many tables I create with tabyl() and since those numbers can be large, they need to be formattable - with either commas or dots for legibility.

I know this is possible by pre-formatting and use of the ns argument in adorn_ns() but I suggest making this simpler and more consistent with the existing adorn_pct_formatting() by adding a new function - adorn_n_formatting(). This would also avoid the need to create multiple output objects in multiple steps.

I suggest the arguments for this could include big.mark to be consistent with other packages existing syntax.

Proposed code example

set.seed(1)
raw_data <- data.frame(sex = rep(c("m", "f"), 3000),
											 age = round(runif(3000, 1, 102), 0))
raw_data$agegroup = cut(raw_data$age, quantile(raw_data$age, c(0, 1/3, 2/3, 1)))

comparison <- raw_data %>%
	tabyl(agegroup, sex, show_missing_levels = F) %>%
	adorn_totals(c("row", "col")) %>%
	adorn_percentages("col") %>%
	adorn_pct_formatting(digits = 1) %>% 
	adorn_ns(position = "rear") %>% 
	adorn_n_formatting(big.mark = ",")

Comparable current code example

set.seed(1)
raw_data <- data.frame(sex = rep(c("m", "f"), 3000),
											 age = round(runif(3000, 1, 102), 0))
raw_data$agegroup = cut(raw_data$age, quantile(raw_data$age, c(0, 1/3, 2/3, 1)))

comparison <- raw_data %>%
	tabyl(agegroup, sex, show_missing_levels = F) %>%
	adorn_totals(c("row", "col")) %>%
	adorn_percentages("col") %>%
	adorn_pct_formatting(digits = 1)

formatted_ns <- attr(comparison, "core") %>% 
	adorn_totals(c("row", "col")) %>% 
	mutate_if(is.numeric, format, big.mark = ",")

comparison %>% 
	adorn_ns(position = "rear", ns = formatted_ns)

[sfirke]

Thanks for this suggestion. I was on the fence about #378 and it sounds like this is a vote for implementing that.

I think the formatting would have to be part of adorn_ns() rather than a subsequent function, as once the Ns are merged into the values it would be much more difficult to manipulate them.

What do you think about a new argument to adorn_ns that would be a function that is applied to all numeric columns of the Ns? The most basic would be format_ns <- function(x) { format(x, big.mark = ",") }. You'd declare that function on a single line, once, and then could have calls like adorn_n_formatting(format_function = format_ns) EDIT: adorn_ns(format_function = format_ns).

[Thyrisis]

Apologies, I did try to avoid a duplicate request but didn't find #378 - yes that's very much the same requirement.

I understand your point about amending the formats after the adorn so yes, another approach would be to support the format() function as a ... pass through of adorn_ns(). This is how pandoc.table in the pander package does this and it works well.

I think I'm arguing for a way to do this all inline and without repetition rather than requiring another step - also for consistency with standard elements of the tidyverse as a consistent design philosophy. My problem with defining a function would be that it's another step - passing through the format() elements seems more elegant and consistent.

[Chanelle280]

Has this been implemented yet?

[sfirke]

Hi. Sorry my engagement is sporadic. I edited my comment above to show what it would look like in-line. I think we agree it should be part of adorn_ns, I don't know why I described it as a new function. I think this should not be too hard and I'll put it on the milestone list for v2.2, since there's clearly demand for it.

[sfirke]

Consolidating into one issue, the use case stated in #459 was:

I'm trying to add a comma to large number but I'm very confused as to how to do this.

adorn_totals(where = "row") %>%
adorn_percentages("col") %>%
adorn_pct_formatting(digits = 1) %>%
adorn_ns(position = "front", big.format = ",")

[davidbudzynski]

Just here to say adding the big.format argument would make my life much easier! Thanks for your work on this package.

[sfirke]

Alright I'm gonna try to sneak this in to 2.2 I think ...

[sfirke]

@Thyrisis @Chanelle280 @davidbudzynski @rasmusrhl this is now merged into the main branch! If you're interested in taking it for a spin, there's a short window in which I could act on any feedback before submitting the next version to CRAN.

Otherwise of course I'd still like your feedback on how I implemented it, at any time.

[davidbudzynski]

Looks good to me, thanks for your work on this!

[Chanelle280]

Looks great! Thanks for working on this. We use tabyl all the time and this will make our work a lot easier when we also include percentages.