I have a table containing multiple statistics per group (in this case: site). I have a couple of sites that have a really long name that forces the first column to expand to contain all the characters. This makes the table larger than necessary.

Hence, for the rows with the ‘row_type` = label I want all the cells in the columns after the label to merged into one cell. I think this will prevent the first column to expand in such way that it has a lot of white space.

Below is an example with a really long name (I left the HTML output out of the results as I am not sure if HTML is supported).

library(tibble)
library(dplyr)
library(gtsummary)


example_data <- tibble(species = c(replicate(10, "species_1"),
                                   replicate(10, "species_2"),
                                   replicate(10, "species_3")),
                       site_1 = rnorm(30),
                       site_2_realy_long_name_that_goes_on_and_on = rnorm(30),
                       site_3 = rnorm(30)) %>% 
  mutate(species = factor(species, levels = c("species_1", "species_2", "species_3")))

table_1 <- example_data %>% 
  tbl_summary(
    include = starts_with("site"),
    by = species, # split table by group
    missing = "no", # don't list missing data separately
    type = all_continuous() ~ "continuous2",
    statistic = list(all_continuous() ~ c("{N_miss}",
                                          "{mean} ({sd})",
                                          "{median} ({p25}, {p75})",
                                          "{min}, {max}")
    )
  )

table_1$table_body |>
  dplyr::filter(row_type == "label")
#> # A tibble: 3 × 8
#>   variable                var_type var_label row_type label stat_1 stat_2 stat_3
#>   <chr>                   <chr>    <chr>     <chr>    <chr> <chr>  <chr>  <chr> 
#> 1 site_1                  continu… site_1    label    site… <NA>   <NA>   <NA>  
#> 2 site_2_realy_long_name… continu… site_2_r… label    site… <NA>   <NA>   <NA>  
#> 3 site_3                  continu… site_3    label    site… <NA>   <NA>   <NA>

^{Created on 2024-08-07 with reprex v2.1.1}

I have tried to transform the table_1 to a gt object (table_1 |> as_gt()) and see if I can modify it from there. I have found functions to merge columns, but this will affect all rows. Other questions on stackoverflow deal with merging rows in a single column.

The desired output is a merged label rows with the label column and the stat_xx columns that contains the label.