Native grouping API

[ggrothendieck]

I have been using huxtable recently and will report a few points where the code might have been shorter if the feature existed.

Currently, inserting a row shifts the index of every row below it, forcing us to use cumulative sums and index math.

# We have to calculate the 'pos' for every single insertion.
df <- data.frame(Asset = c("A","B","C", "D","E","F"), Val = 1:6)
ht <- as_hux(df)

# We must calculate offsets manually
group_sizes <- c(3, 3)
insert_at <- c(1, 4) # Original indices

for (i in seq_along(insert_at)) {
  # (i-1) accounts for rows already added earlier in the loop
  current_pos <- insert_at[i] + (i - 1) 
  ht <- insert_row(ht, c("Group Header", ""), after = current_pos)
  ht <- set_colspan(ht, current_pos + 1, 1, 2)
}

What would be nice would be to shorten this to one line

# One command handles the splitting and header insertion automatically.
ht %>% add_group_headers(by = "Sector", style = list(bold = TRUE))

[hughjonesd]

I think the solution is narrower than the issue. The issue is that there's no way to insert multiple rows (or columns) into a huxtable at different places at once. (Of course, this is also true for standard data frames...)

The solution addresses the specific purpose of adding header rows. For that specific case, have you looked at column_to_header()?

df <- data.frame(Asset = c("A","B","C", "D","E","F"), Val = 1:6, Sector = c(1,1,1,2,2,2))
ht <- as_hux(df)
ht |> column_to_header("Sector", glue = "Sector: {value}", bold = TRUE)
                               Asset         Val
                               Sector: 1
                               A               1
                               B               2
                               C               3
                               Sector: 2
                               D               4
                               E               5
                               F               6

[ggrothendieck]

I definitely missed that. This was my first time using huxtable.