t1flex() still causes errors when more than one header column has the same N

[Koehlibert]

Sorry for the potentially duplicated issue, but the error in #103 seems to have resurfaced.
When I have multiple levels in my table1 headers, applying t1flex throws an error. Here's a reprex:

library(table1)
library(flextable)
library(dplyr)
df <- data.frame(
  l1 = rep(c("A", "B"), each = 200),
  l2 = rep(rep(c("x", "y"), each = 100), 2),
  x = rnorm(400)
)
table1(~x | l1/l2, data = df) %>% t1flex() %>% flextable::autofit()

R version 4.2.2, all packages up to date.

[fthielen]

I think that this package is no longer actively maintained as the issue you raise (including the previous one) were never fixed in the CRAN version of the package.

What helped for me was to install the developed version of this package:

devtools::install_github("benjaminrich/table1") together with the developer version of flextable devtools::install_github("davidgohel/flextable")

[benjaminrich]

The package is still actively maintained, I've just been busy with other priorities. I am planning a new CRAN release, which will hopefully happen soon. Thanks for your patience.

[Koehlibert]

Thanks for the replies. In the meantime, for others having this issue, I got a solution from stackexchange:

I wrote this little helper function that make the tables usable for t1flex:

t1flex_helper <- function(t1) {
  obj <- attr(t1, "obj")
  pad <- function(x) paste0(x, strrep(" ", seq_along(x) - 1))
  split(obj$headings[1,], obj$headings[1,]) <-
    split(obj$headings[1,], obj$headings[1,]) |> lapply(pad)
  obj$headings
  
  split(obj$labels$strata, obj$labels$strata) <-
    split(obj$labels$strata, obj$labels$strata) |> lapply(pad)
  obj$labels$strata
  attr(t1, "obj") <- obj
  t1
}```