Change cells' values in RMarkdown PDF and linebreak

I have the following dataset:

structure(list(...1 = c("Catalunya", "PIB", "Demanda interna1", 
"Demanda externa1", "Espanya", "PIB", "Demanda interna1", "Demanda externa1", 
"Zona euro", "PIB", "Demanda interna1", "Demanda externa1", "PIB d'Alemanya", 
"PIB de França", "PIB d'Itàlia", "Internacional", "PIB de la UE-27", 
"PIB dels EUA", "PIB del Japó"), `2020` = c(NA, -11.9, -7.4, 
-4.5, NA, -11.2, -9, -2.2, NA, -6.1, -5.57, -0.53, -3.8, -7.5, 
-9, NA, -5.6, -2.2, -4.3), `2021` = c(NA, 6.2, 3.8, 2.4, NA, 
6.4, 5.9, 0.5, NA, 5.6, 4.24, 1.36, 3.2, 6.4, 8.3, NA, 5.7, 5.8, 
2.1), `2022` = c(NA, 5.5, 2.7, 2.8, NA, 5.5, 5.12183164381576, 
0.378168356184236, NA, 3.6, 3.76, -0.16, 1.8, 2.5, 3.7, NA, 3.5, 
1.9, 1.1), `1r 2023` = c(NA, 3.2, 1.1, 2.1, NA, 4.1, 1.3, 2.8, 
NA, 1.2, 0.63, 0.57, -0.2, 1, 2.1, NA, 1.1, 1.71772942152937, 
2), `2n 2023` = c(NA, 2.1, 1.2, 0.9, NA, 2, 2.2, -0.2, NA, 0.5, 
0.46, 0.04, 0.1, 1.1, 0.3, NA, 0.4, 2.38237207285343, 1.7), `3r 2023` = c(NA, 
2.1, NA, NA, NA, 1.8, 1.6, 0.2, NA, 0.1, NA, NA, -0.4, 0.7, 0, 
NA, 0.1, 3, 1.2)), row.names = c(NA, -19L), class = c("tbl_df", 
"tbl", "data.frame"))

And I am creating a table in Markdown (to PDF), using this code:

options(knitr.kable.NA = "-")

p3_1[3,1] <- p3_1[7,1] <- p3_1[11,1] <- "$\\text{Demanda interna}^1$"
p3_1[4,1] <- p3_1[8,1] <- p3_1[12,1] <- "$\\text{Demanda externa}^1$"


kable(p3_1,
      align = "lrrrrrr",
      col.names = c("", names(p3_1[-1])),
      booktabs=TRUE,
      digits=1,
      format.args = list(big.mark="."),
      linesep = "" ,
      caption="Producte interior brut (\\% de variació interanual, llevat dels casos indicats)", escape = FALSE) %>%
  add_header_above(c("", "Any" = 3, "Trimestre" = 3), bold=T) %>% 
  kable_styling(latex_options=c("HOLD_position","scale_down"), bootstrap_options = c("condensed"), position="center") %>% 
  column_spec(1, bold=T, width="4cm") %>%
  column_spec(2:7, width="2cm") %>%
  row_spec(0, bold = T) %>%
  row_spec(c(1, 5, 9, 16), background="lightgray") %>%
  row_spec(19, extra_css = "border-bottom: 1px solid;") %>%
  footnote(general="Idescat; Eurostat; INE; BEA; ESRI; Agència Estatal d'Administració Tributària.", general_title = "Font:", footnote_as_chunk = T,)  %>%
  footnote(general="$^{A}$Dada avançada. $^{1}$Contribució al creixement. $^{2}$Sèrie original corregida d'efectes de calendari. 
           $^{3}$En valors constants. Exclou estacions de servei.",  escape = F, general_title = "Nota:", threeparttable = TRUE, footnote_as_chunk = T,)

Which results in:

How can I delete the "-" in Catalunya, Espanya and Zona Eurorows? So "-" will only appear in those cells where the value is missing.

And also, how can I add a line break in the caption: "Producte interior brut \linebreakhere (% de variació interanual, llevat dels casos indicats)"

Solution

This approach uses kableExtra::packrows() option plus some tidyverse data wrangling to format and shape the data for the table.

Note I manually changed the c with a cidilla accent to c due to an encoding issue with my setup. And removed the html arguments which do nothing for a pdf table.

---
output: pdf_document
---

```{r packages, include=FALSE}

library(kableExtra)
library(dplyr)
library(tidyr)
library(stringr)
library(forcats)


```



```{r data, include=FALSE}

p3_1 <- structure(list(...1 = c("Catalunya", "PIB", "Demanda interna1", 
"Demanda externa1", "Espanya", "PIB", "Demanda interna1", "Demanda externa1", 
"Zona euro", "PIB", "Demanda interna1", "Demanda externa1", "PIB d'Alemanya", 
"PIB de Franca", "PIB d'Itàlia", "Internacional", "PIB de la UE-27", 
"PIB dels EUA", "PIB del Japó"), `2020` = c(NA, -11.9, -7.4, 
-4.5, NA, -11.2, -9, -2.2, NA, -6.1, -5.57, -0.53, -3.8, -7.5, 
-9, NA, -5.6, -2.2, -4.3), `2021` = c(NA, 6.2, 3.8, 2.4, NA, 
6.4, 5.9, 0.5, NA, 5.6, 4.24, 1.36, 3.2, 6.4, 8.3, NA, 5.7, 5.8, 
2.1), `2022` = c(NA, 5.5, 2.7, 2.8, NA, 5.5, 5.12183164381576, 
0.378168356184236, NA, 3.6, 3.76, -0.16, 1.8, 2.5, 3.7, NA, 3.5, 
1.9, 1.1), `1r 2023` = c(NA, 3.2, 1.1, 2.1, NA, 4.1, 1.3, 2.8, 
NA, 1.2, 0.63, 0.57, -0.2, 1, 2.1, NA, 1.1, 1.71772942152937, 
2), `2n 2023` = c(NA, 2.1, 1.2, 0.9, NA, 2, 2.2, -0.2, NA, 0.5, 
0.46, 0.04, 0.1, 1.1, 0.3, NA, 0.4, 2.38237207285343, 1.7), `3r 2023` = c(NA, 
2.1, NA, NA, NA, 1.8, 1.6, 0.2, NA, 0.1, NA, NA, -0.4, 0.7, 0, 
NA, 0.1, 3, 1.2)), row.names = c(NA, -19L), class = c("tbl_df", 
"tbl", "data.frame"))


```


```{r}

options(knitr.kable.NA = "-")

# data wrangling bit

# vector of grouping names
grp_nm <- c("Catalunya", "Espanya", "Zona euro", "Internacional")

p3_1 <- 
  p3_1 |> 
  mutate(grp = ifelse(...1 %in% grp_nm, ...1, NA),
         ...1 = case_when(str_detect(...1, "interna1") ~ "$\\text{Demanda interna}^1$",
                          str_detect(...1, "externa1") ~ "$\\text{Demanda externa}^1$",
                          TRUE ~ ...1))|>
  fill(grp) |> 
  filter(!...1 %in% grp_nm) |> 
  select(grp, everything())
  

kable(p3_1[, -1],
      align = "lrrrrrr",
      col.names = c("", names(p3_1[-(1:2)])),
      booktabs=TRUE,
      digits=1,
      format.args = list(big.mark="."),
      linesep = "" ,
      caption="Producte interior brut (\\% de variació interanual, llevat dels casos indicats)", escape = FALSE) |>
  add_header_above(c("", "Any" = 3, "Trimestre" = 3), bold=T) |>
  kable_styling(latex_options=c("HOLD_position","scale_down"), position="center") |> 
  column_spec(1, bold=T, width="4cm") |> 
  column_spec(2:7, width="2cm") |> 
  row_spec(0, bold = T) |> 
  pack_rows(index = table(p3_1$grp), background="lightgray") |>
  footnote(general="Idescat; Eurostat; INE; BEA; ESRI; Agència Estatal d'Administració Tributària.", general_title = "Font:", footnote_as_chunk = T,)  |>
  footnote(general="$^{A}$Dada avançada. $^{1}$Contribució al creixement. $^{2}$Sèrie original corregida d'efectes de calendari.
           $^{3}$En valors constants. Exclou estacions de servei.",  escape = F, general_title = "Nota:", threeparttable = TRUE, footnote_as_chunk = T,)

```

So you end up with: