printing blanks instead of NA's when using formattable in R

Question

Consider the example data.frame

df <- data.frame(
  id = 1:4,
  name = c("Bob", "Ashley", "James", "David"), 
  age = c(48, NA, 40, 28),
  test1_score = c(18.9, 19.5, NA, 12.9),
  stringsAsFactors = FALSE)

I'm using the R package formattable to make a pretty table.

library(formattable)
formattable(df, list(
age = color_tile("white", "orange"),
test1_score = color_bar("pink", 'proportion', 0.2)
))

It used to be that NA's were automatically not printed, and a blank was printed instead. It seems like this is no longer the default, but I would still like to print a blank for the NA. Replacing NA like this works:

df[is.na(df)]=''
formattable(df, list(
  age = color_tile("white", "orange"),
  test1_score = color_bar("pink", 'proportion', 0.2)
))

However if I try to format one of the columns to force it to have 2 decimal places, the pesky NA's return:

df$age = digits(df$age, digits=2)
formattable(df, list(
age = color_tile("white", "orange"),
test1_score = color_bar("pink", 'proportion', 0.2)
))

If I remove the NA again, the NA goes away, but so do the decimal places

df[is.na(df)] = ''
formattable(df, list(
age = color_tile("white", "orange"),
test1_score = color_bar("pink", 'proportion', 0.2)
))

I believe the reason is that digits converts df$age to a formattable numeric object and creates the NA, and df[is.na(df)] = '' converts df$age to a formattable character object:

> df$age = digits(df$age, digits=2)
> df$age
[1] 48.00  NA   40.00 28.00
> class(df$age)
[1] "formattable" "numeric"    
> df[is.na(df)] = ''
> df$age
[1] "48" "  " "40" "28"
> class(df$age)
[1] "formattable" "character"

Any ideas on a solution?

Ultimately I'd also like to use this with a filtered data.frame, where I use the code from Filtering dataframes with formattable to ensure that the color scale stays the same when filtering the data.frame:

df$age = digits(df$age, digits=2)
  subset_df <- function(m) {
    formattable(df[m, ], list(
      age = x ~ color_tile("white", "orange")(df$age)[m],
      test1_score = x ~ color_bar("pink", 'proportion', 0.2)(df$test1_score)[m],
      test2_score = x ~ color_bar("pink", 'proportion', 0.2)(df$test2_score)[m]
    ))
  }

subset_df(1:3)

The problem doesn't seem to be with this code though.

user_6897086 · Answer

Another solution which worked for me was using str_remove_all(). Because color_bar() in formattable produces HTML output as a character, you can just remove the string "NA".

Note that this has the potential to mess up the HTML if you happened to have NA anywhere else. Also worth noting is that I wrap a percent function around your_var. This was the best way I could come up with to convert my numeric to percent and apply color_bar(). The code is below:

df %>%
    # First mutate w/color_bar()
    mutate(your_var= color_bar("green", na.rm=T)(percent(your_var, digits = 1))) %>% 
    # Second mutate
    mutate(your_var = str_remove_all(your_var, "NA"))

Output from First mutate

NA

Output from Second mutate

Also, in case anyone hasn't seen this yet: Awesome tables in HTML - Integration with formattable

printing blanks instead of NA's when using formattable in R

2 Answers

Output from First mutate

Output from Second mutate