BUG: Scaling fails when the scaling column is the only ID column

[spencerpease]

Describe the bug
When scaling data where the only ID column is the column you want to scale over, the function errors. I believe there is some internal assumption that there will be more ID columns and scale() is trying to do some grouping by those non-existent columns.

To Reproduce

library(data.table)

dt <- data.table(
  id = c("a", "a1", "a2", "a3"),
  value = c(10, 2, 3, 4)
)

mapping <- data.table(
  parent = "a",
  child = c("a1", "a2", "a3")
)

dt_scaled <- hierarchyUtils::scale(
  dt = dt,
  id_cols = "id",
  value_cols = "value",
  col_stem = "id",
  col_type = "categorical",
  mapping = mapping
)

This produces the error and warning:

Scaling id
Scaling 1 of 1: a
Error in merge.data.table(unique(children_dt[, .SD, .SDcols = by_id_cols]),  : 
  A non-empty vector of column names for `by` is required.
In addition: Warning message:
In merge.data.table(unique(children_dt[, .SD, .SDcols = by_id_cols]),  :
  You are trying to join data.tables where 'x' argument is 0 columns data.table.

Expected behavior
If the scaling column is the only ID column, no internal grouping should be applied.

Desktop (please complete the following information):

• OS: Linux
• R: 4.1.3