Any methods to dynamically generate chunk label names based on R expressions in quarto? (!expr won't work)

[fcmeyer]

RStudio version: Version 2023.03.1+446 (2023.03.1+446)
Quarto version: 1.3.353
knitr version: 1.42
OS: macOS Ventura 13.3.1

I am familiar with the use of !expr '...' in the fig-cap chunk options, but it does not seem to work for the label option. Is there another way to dynamically generate chunk labels based on R code? Examples:

This works:

---
title: example
format: pdf
validate-yaml: false
---

Hello!

```{r}
x <- "mpg"
```

```{r}
#| fig-align: center
#| fig-dpi: 700
#| label: fig-mpg
#| fig-cap: !expr 'glue::glue("Histogram of {x}")'

library(ggplot2)
ggplot(mtcars) +
  aes(x = mpg) +
  geom_histogram(bins = 15L) +
  theme_light()
```

But this won't work:

I am familiar with the use of !expr '...' in the fig-cap chunk options, but it does not seem to work for the label option. Is there another way to dynamically generate chunk labels based on R code? Examples:

This works:

---
title: example
format: pdf
validate-yaml: false
---

Hello!

```{r}
x <- "mpg"
```

```{r}
#| fig-align: center
#| fig-dpi: 700
#| label: !expr 'glue::glue("fig-{x}")'
#| fig-cap: !expr 'glue::glue("Histogram of {x}")'

library(ggplot2)
ggplot(mtcars) +
  aes(x = mpg) +
  geom_histogram(bins = 15L) +
  theme_light()
```

This second example yields error:

processing file: test.qmd
                                                                                                            Quitting from lines 14-24 (test.qmd) 
Error in defaults[[name]] : invalid subscript type 'list'
Calls: .main ... process_group.block -> call_block -> get_code -> unlist -> <Anonymous>

Execution halted

Thank you!!!

[baptiste]

The knitr workaround for this is to generate chunks dynamically, e.g. with knit_expand() or similar templating system (brew, glue, etc.):

---
title: "Untitled"
engine: knitr
---


# Demos of `knit_expand()`

Build regression models based on a template; loop through all variables in `mtcars`:

```{r results='hide'}
library(knitr)
src = lapply(names(mtcars)[-1], function(i) {
knit_expand(text=c("# Regression on {{i}}\n", '```{r lm-{{i}}}', 'lm(mpg~{{i}}, data=mtcars)', '```'))
})
# knit the source
res = knit_child(text = unlist(src))
```

`r res`