Handling missing values in a col_merge

[erhoppe]

Thanks for the great package! It's super useful. I wondered if it would be possible to add an option for dealing with missing data in cols_merge so that you could either override sub_missing or somehow provide an alternative pattern for cells where you only have missing data.

I'm encountering a situation where I have a subgroup of data for a particular column that is missing data, so my summary with the median, min, and max are all missing. I would like to just have a single "n.p" in that cell rather, but currently, it doesn't appear as if this is possible.

library(tidyverse)
library(gt)

tibble(
  category = c("a","b","c","d"),
  median = c(5,10,8, NA_real_),
  min = c(1,6,4,NA_real_),
  max = c(8,25,12,NA_real_),
) %>% 
  gt(rowname_col = "category") %>% 
  cols_merge(columns = c(median, min, max),
             pattern = "{1}<br>({2},{3})") %>%
  tab_spanner(
    label = "Measurement",
    columns = c(median)
  ) %>% 
  cols_label(
    median = html("Median<br>(min,max)")
  ) %>% 
  sub_missing(columns =
                c(median),
              rows = everything(), missing_text = "n.p.") %>%
  sub_missing(columns =
                c(min,max),
              rows = everything(), missing_text = "") %>% 
  tab_source_note(
    source_note = "n.p. = annotation not provided by authors"
  )

Would something like this be possible?

[rich-iannone]

Thanks for filing this issue and providing a great example! This has come up a few times and I think we need some extra syntax in the pattern to enclose a run of text. Where any NAs are encountered, the entire span would be omitted.

[erhoppe]

That sounds like a great solution for it. I look forward to seeing what you come up with. :) Thanks again!