across does not handle lists of functions stored as an external variable

[Xitian9]

across is normally able to accept a list of functions which will all be applied to the selected columns. This works in dtplyr if the list is supplied directly in across, but not if it's stored as a variable and substituted in.

library(dplyr)

df <- data.table::data.table(
  x = seq(1, 10)
)

# Store list of summary functions as a variable
summaries <- list(
  "Mean" = mean,
  "SD" = sd
)

# Throws an error that summaries is not a function
df |>
  dtplyr::lazy_dt() |>
  summarise(across("x", summaries)) |>
  data.table::as.data.table()

# Works fine if not a lazy data table
df |>
  summarise(across("x", summaries)) |>
  data.table::as.data.table()

# Also works fine if the list is written within the call
df |>
  dtplyr::lazy_dt() |>
  summarise(across("x", list("Mean" = mean, "SD" = sd))) |>
  data.table::as.data.table()