Lazy evaluation may lead to unexpected results

[BerriJ]

Hi, I recently discovered an interesting (in hindsight) problem. Consider you want to make a simple 3x3 table containing NAs. Now you want to add 1,2,3 under the 3 respective diagonal elements of the table.

What I tried:

library(dplyr)
library(gt)

df <- as.data.frame(matrix(NA, nrow = 3, ncol = 3))
tab <- gt(df)
for (i in 1:3) {
    tab <- tab %>%
        text_transform(locations = cells_body(
            columns = all_of(i),
            rows = all_of(i)
        ), fn = function(x) {
            paste0(x, "<br>", i)
        })
}
tab See result here: [tab](https://i.stack.imgur.com/0c1fd.png)

However, add "3" to all diagonal elements of the table. This SO Thread suggests points out that R's lazy evaluation is messing things up here. The SO Thread also contains a solution > using lapply.

So there are some problems I see:

• People that are aware of the problem have to use lapply or apply or something. But translating loops into lapply calls is not always easy, especially if you deal with nested loops.
• People that are not aware won't know why they get unexpected results

I think looping through (part of) a table is a relatively common scenario. Is there something we can do about this? So for example can we force non-lazy evaluation inside of for-loops?

Let's Discuss! :)

[rich-iannone]

I've seen a similar looping issue before and I totally agree that this is not an uncommon thing to want to do. I'll have a closer look at this and I suspect it's a gt problem. Will create an issue from this.