1

I'm trying to remove string from at gt() table, but without success. In my table I want multiple column names/titles covering two columns named "Day 0", "Day 4", "Day 10", "Day 17" and sub columns named "mean" and "se". Probably easier to understand when looking at the data.

What I want right now is to remove the "day 0" (or e.g. "day 4") from the sub column. Any suggestions?

My code:

library(gt)

gt(mean_se)%>%
  tab_spanner(label = "Day 0", columns = matches("0")) %>%
  tab_spanner(label = "Day 4", columns = matches("4")) %>%
  tab_spanner(label = "Day 10", columns = matches("10")) %>%
  tab_spanner(label = "Day 17", columns = matches("17")) %>%
  tab_header(
    title = md("Biogenic volatile organic compound emissions"))
  
  # str_remove(mean_se, "day 0")
  # rename("mean"=="mean day 0")

Here's the data:

mean_se <- structure(list(mass = c("i.x33.031", "i.x35.034", "i.x36.017", 
"i.x39.959", "i.x40.023"), `mean day 0` = c("241.82", "0.36", 
"1.78", "0.2", "1.82"), `se day 0` = c("241.82", "0.36", "1.78", 
"0.2", "1.82"), `mean day 4` = c("32.94", "0.14", "0", "0", "1.74"
), `se day 4` = c("32.94", "0.14", "0", "0", "1.74"), `mean day 10` = c("266.28", 
"0.6", "0", "0", "1.58"), `se day 10` = c("266.28", "0.6", "0", 
"0", "1.58"), `mean day 17` = c("451.4", "0.48", "0", "0", "2.94"
), `se day 17` = c("451.4", "0.48", "0", "0", "2.94")), row.names = c(NA, 
-5L), class = "data.frame")
Nate
  • 10,361
  • 3
  • 33
  • 40
Tiptop
  • 533
  • 5
  • 19
  • 1
    this seems tedious but you could re-label the columns one by one like this: `gt(mean_se) %>% cols_label("mean day 0" = "mean")` – Nate Dec 08 '20 at 14:01

1 Answers1

2

One approach is with cols_label. This function accepts individuals arguments to rename a specific column, for example "mean day 0" = "mean". But it's a lot of work to type all those out. Instead, we can use .list = and provide a named list of all the columns we want to rename.

We can just repeat "mean" and "se" the appropriate number of times and name that character vector with the column names:

setNames(rep(c("mean","se"),(ncol(mean_se)-1)/2),names(mean_se)[-1])
# mean day 0    se day 0  mean day 4    se day 4 mean day 10   se day 10 mean day 17   se day 17 
#     "mean"        "se"      "mean"        "se"      "mean"        "se"      "mean"        "se"

To keep the order correct, we can use cols_move.

gt(mean_se)%>%
  tab_spanner(label = "Day 0", columns = matches("0")) %>%
  tab_spanner(label = "Day 4", columns = matches("4")) %>%
  tab_spanner(label = "Day 10", columns = matches("10")) %>%
  tab_spanner(label = "Day 17", columns = matches("17")) %>%
  cols_label(.list = setNames(rep(c("mean","se"),(ncol(mean_se)-1)/2),
                              names(mean_se)[-1])) %>%
  cols_move(columns = names(mean_se)[-1], after = names(mean_se)[1]) %>%
  tab_header(
    title = md("Biogenic volatile organic compound emissions"))

enter image description here

Ian Campbell
  • 23,484
  • 14
  • 36
  • 57
  • Thanks @Ian Campbell, it works quite well! However now Day 10 is suddenly before Day 4? I don't exactly understand how the setNames list works, but it saves me some time for sure. – Tiptop Dec 08 '20 at 14:19
  • Nice catch, I have edited my answer to include a call to `cols_move` which forces the order. – Ian Campbell Dec 08 '20 at 14:24
  • Looking for this exact solution but if you use 'select' and change the number of columns 'names()' doesn't work. Looks like gt doesn't allow 'names(.)' to function inside it. I could add a variable outside the pipe but was wondering if there is a solution inside the pipe. Thanks – seansteele Jan 19 '22 at 00:27
  • @seansteele The typical solution to this problem is to pipe the data into `{}` and then assign `.` to a temporary variable that exists within `{}` but not the global environment. See this Q&A for an example: https://stackoverflow.com/questions/40369832/assign-intermediate-output-to-temp-variable-as-part-of-dplyr-pipeline – Ian Campbell Jan 19 '22 at 00:35
  • Thanks! I had tried some of those solutions from other thread but they did pass the temp variable to gt. The `pipeR` solution worked though! – seansteele Jan 19 '22 at 02:03