Collapse rows

[ratnanil]

A feature that I'm missing in this great package is the ability to "span cells" between rows. This would be a feature similar as gt currently behaves when passing it a grouped dataframe. However, instead of adding the grouped column as an additional row, it would be cool if there was a way to merge cells with the same value, similar to the function kableExtra::collapse_rows (see way below)

Current behaviour

library(gt)
library(dplyr)

iris %>%
  group_by(Species) %>%
  sample_n(3) %>%
  gt()

Desired output as demonstrated with kableExtra:

library(kableExtra)

iris %>%
  group_by(Species) %>%
  sample_n(3) %>%
  kable() %>%
  collapse_rows(5)

PS: Issue #108 (comment) contains a link to this very feature from kableExtra, but that might be a mistake

[Nova-Scotia]

This is exciting! I'm using the dev version of gt and I can get this to work using the example in #2008

 library(gt)

df = dplyr::tibble(
    vals = 1:8,
    rowname_left  = c("A",  "A",  "A",  "A",  "B",  "B",  "B",  "B"),
    rowname_mid   = c("AA", "AA", "AZ", "AZ", "BA", "BA", "BZ", "BZ"),
    rowname_right = c("H", "I", "J", "K", "L", "M", "N", "O")
)

gt(
    df,
    rowname_col = c("rowname_left", "rowname_mid", "rowname_right")
) |> tab_stubhead(c("left", "middle", "right"))

However, if I just wish to group by one column, it doesn't work since the hierarchical stub layout only happens with multiple values in rowname_col. So when I try

gt(
    df,
    rowname_col = "rowname_left") |> 
  tab_stubhead("left")

I don't get my observations grouped by 'left':

Are there any workarounds for this?

[rich-iannone]

The single column case is a bit disconnected from the new implementation but can be done with this (though it's not obvious coming from the new way of using rowname_col=):

library(gt)

df = dplyr::tibble(
  vals = 1:8,
  rowname_left  = c("A",  "A",  "A",  "A",  "B",  "B",  "B",  "B"),
  rowname_mid   = c("AA", "AA", "AZ", "AZ", "BA", "BA", "BZ", "BZ"),
  rowname_right = c("H", "I", "J", "K", "L", "M", "N", "O")
)

gt(
  df,
  groupname_col = "rowname_left",
  row_group_as_column = TRUE
  ) |> 
  tab_stubhead("left")

I've been torn on whether to make your last example just collapse/group values, but then there's no way to have them ungrouped (without another argument). So I guess I might just improve the gt() documentation article?

[Nova-Scotia]

Improving the documentation is always good! I tried to do this in an interactive table with sorting and got pretty frustrated so had to leave it for the time being.