NA processing with Quarto doesn't work

[cbgoodman]

When asking tinytable to process NA's with quarto = TRUE, there's no processing with TeX.

---
title: "testing"
format: pdf
---

```{r}
#| include: false
library(tidyverse)
library(tinytable)

assignments = tribble(
  ~Points,  ~Assignment,
  4 * 10  , "Analytical Assignments",
  0       , "Team Project Part 0 - County Selection",
  10      , "Team Project Part 1 - Summary Background",
  10      , "Team Project Part 2 - SWOT Analysis",
  10      , "Team Project Part 3 - Strategic Direction/Action Plan",
  10      , "Team Project Part 4 - Evaluation Framework",
  10      , "Team Project Part 5 - Plan Presentation",
  20      , "Class participation"
) |>
  mutate(Percent = Points / sum(Points),
         Percent = scales::label_percent(accuracy = 1)(Percent))

assignments_total = tibble(Assignment = "Total",
                            Points = sum(assignments$Points),
                            Percent = NA)

bind_rows(assignments, assignments_total) |>
  na_if(NA) |> 
  select(Assignment, Points, Percent) |> 
  tt(width = c(.5, .125, .125)) |> 
  format_tt(replace = "-", quarto = TRUE) |>
  style_tt(j = 1:3, align = "lcc")

Works fine with HTML (and Quarto processing turned off)

[vincentarelbundock]

Thanks a lot for the report @cbgoodman . The primary issue should be fixed on Github now.

Unfortunately, investigating this brought to my attention two more issues in dealing with Quarto.

Issue 1: Special characters

When compiling to LaTeX, Quarto will sometimes interpret some Base64 encoded strings as Markdown code. For example, if you call format_tt(replace="-", quarto=TRUE), Quarto will believe that - is the start of a bullet point list, and will insert a \begin{itemize} inside your table. This breaks compilation.

I recommend you use a "non-markdown" character or an empty space when replacing NAs.

See below for an example.

Issue 2: Quarto data processing in LaTeX

I believe that tinytable behaves in the way that Quarto developers suggest, in terms of inserting \QuartoMarkdownBase64{} macros around cell content that should be data processed by Quarto.

This works in HTML but unfortunately not in LaTeX. @andrewheiss once linked to this thread when explaining that the strategy only worked for HTML.

Andrew, do you know if a new workaround was proposed? I see that the original Quarto issue has been closed, but I can't quite figure out what the recommeded strategy is.

Example

---
title: "Citations in tables"
format: 
  pdf:
    keep-tex: true

references:
- type: article-journal
  id: Lovelace1842
  author:
  - family: Lovelace
    given: Augusta Ada
  issued:
    date-parts:
    - - 1842
  title: >-
    Sketch of the analytical engine invented by Charles Babbage, by LF Menabrea, 
    officer of the military engineers, with notes upon the memoir by the translator
  title-short: Molecular structure of nucleic acids
  container-title: Taylor’s Scientific Memoirs
  volume: 3
  page: 666-731
  language: en-GB
---


```{r}
#| message: false
library(tinytable)
x <- data.frame(Thing = 1:2, Citation = c("@Lovelace1842", NA))

tt(x) |> format_tt(replace = " ", quarto = TRUE)
```


```{r}
#| eval: false
tt(x) |> format_tt(replace = "+", quarto = TRUE)
```

[cbgoodman]

Thanks, @vincentarelbundock! I wish my bug reports had not uncovered fundamental issues with Quarto processing, but this isn't the first time it's happened. 🤷‍♂️

[cbgoodman]

Oh, right, context. I've been pestering @andrewheiss with Quarto bugs that are really not implemented features (that are also poorly documented) for ages.

[cbgoodman]

Can confirm the special characters issue is solved. The secondary issue still remains, but I expected that.

[vincentarelbundock]

Thanks for confirming. Closing because there is probably nothing I can do to fix the rest here.