add type=width argument to nchar function

[anajens]

Issue originally reported by lil128 on Jul 24, 2020

this is not an issue, but a feature enhancement.

The root concern is related to China filing, where the health authority is requiring translation on everything. I just wonder if it could make sense that our reporting system supports these characters.

an example using refactored rtables can be

rtable(
  header = c("测试\n测试2", "asda\nasdsa"),
  rrow("行行行行行行行行行行", 1,2)
)

you can see that the layout is not correct. though using nchar(type="width") does not always solve the issue(depending on the font you choose: these characters does not always occupy twice the width of the alphabets)

[anajens]

Here's another example from lil128

library(rtables)

df <- data.frame(
  a = sample(c('打算\ndasd', '的撒\ndasda\nasdas', '大苏打啊实打实的\n的撒大阿三大苏打'), replace = TRUE, size = 1000),
  b = rnorm(1000)
)

basic_table() %>%
  split_cols_by('a') %>%
  analyze('b') %>%
  build_table(df)

Result:

        的撒                    
       dasda   大苏打啊实打实的    打算 
       asdas   的撒大阿三大苏打   dasd
------------------------------
Mean   -0.01     0.05     -0.1

[gmbecker]

So changing the type to width in what I believe are all the places it would need to be results in a different but still obviously wrong rendering of the table.

It is difficult to tell, but it looks like some of the chinese characters are taking up 2 positions while others are not:

(this is with experimental local changes to formatable):

> cat(toString(m, nch_type= "width"))
                          的撒        
       大苏打啊实打实的   dasda   打算
       的撒大阿三大苏打   asdas   dasd
——————————————————————————————————————
Mean         0.02         -0.01   0.05



> cat(toString(m, nch_type= "char"))
                   的撒         
       大苏打啊实打实的   dasda    打算 
       的撒大阿三大苏打   asdas   dasd
——————————————————————————————
Mean     0.02     -0.01   0.05

Putting (keeping) a pin in this for now, but we'll leave it in the backlog