Writing large formated data frames is slow

[JanMarvin]

Similar to #356. Writing a large data frame with dates is slow. This is caused by the for loop in add_numfmt().

library(openxlsx2)

beg <- "1900-1-1"
end <- "2000-11-18"

dat <- seq(
  from = as.POSIXct(beg, tz="UTC"),
  to   = as.POSIXct(end, tz="UTC"),
  by   = "day"
)

out <- data.frame(
  date = dat,
  chr  = as.character(dat),
  num  = seq_along(dat)-1
)

wb <- wb_workbook() %>% 
  wb_add_worksheet() %>% wb_add_data(x = out) %>% 
  wb_open()

[JanMarvin]

It's definitely unwanted and unexpected behavior. Yet atm I have no desire to touch this part of the code again. If this nags you, feel free to submit a PR.

[JanMarvin]

Likely it would speed things up, if we collect all unique previous styles, instead of asking every cell for its style. Maybe call table on the styles, modify the possible styles and replace their old id with the new id. This way we'd only have to loop through a couple of cell styles instead of every single cell.