Move summary_rows() to top of group

[alex-lauer]

Hi {gt} community,

I am trying to build a demographic table for clinical reporting and am currently struggling with the position of the summary row. Is there a way to force summary_rows() to move the summary column to the top of the group?

Thank you so much in advance!

library(formatters)
library(tidyr)
library(dplyr)
library(gt)

df <- formatters::ex_adsl |> 
  dplyr::group_by(ARM) |> 
  dplyr::mutate(N = n()) |> 
  dplyr::ungroup() |> 
  dplyr::group_by(ARM, SEX) |> 
  dplyr::summarize(n = n(),
                   pct = n()/mean(N)) |> 
  dplyr::ungroup() |> 
  tidyr::pivot_wider(id_cols = SEX, names_from = ARM, values_from = c(n, pct)) |> 
  dplyr::mutate(across(where(is.numeric), ~ifelse(is.na(.), 0, .)),
                category = paste0(attr(SEX, "label"), ", n (%)")) |> 
  dplyr::select(category, everything()) |> 
  dplyr::rename(label = SEX)

  
gt(df, 
   rowname_col = "label",
   groupname_col = "category") |> 
  tab_header(
    title = "x.x: Study Subject Data",
    subtitle = html("x.x.x: Demographic Characteristics<br>Table x.x.x.x: Demographic Characteristics - Full Analysis Set"),
    preheader = c("Protocol: XXXXX", "Cutoff date: DDMMYYYY")
  ) |> 
  summary_rows(groups = c("Sex, n (%)"),
               columns = starts_with("n_"),
               fns = list(n = ~sum(.)),
               missing_text = "",
               formatter = fmt_integer) |> 
  fmt_percent(columns = starts_with(c("pct_")), decimals = 1) |> 
  cols_merge_n_pct(col_n = "n_A: Drug X", col_pct = "pct_A: Drug X") |> 
  cols_merge_n_pct(col_n = "n_B: Placebo", col_pct = "pct_B: Placebo") |> 
  cols_merge_n_pct(col_n = "n_C: Combination", col_pct = "pct_C: Combination") 

[rich-iannone]

Hi @alex-lauer , unfortunately this is impossible but the idea for this has been around for awhile (#177).

The only feasible way to do this now is to include the summary data in the input data table and style the resulting table so it has the proper line styles.

[alex-lauer]

Thanks for getting back to me so quickly @rich-iannone! Would love to see this implemented in the future, a simple match.arg(., c("top", "bottom")) would be totally sufficient for me.