Utility function to round all numeric columns in an input-output style table. It is mainly intended for reproducibility and comparability with external sources that report rounded values. Non-numeric columns are left unchanged.
Special cases:
If digits = NULL (default), the function returns the input unchanged
(no rounding).
Values exactly equal to 1e-06 are preserved to avoid suppressing small
"epsilon" entries that occur in published IOTs and SUTs.
round_table(data_table, digits = NULL)A tibble (if input was a tibble) or data frame with numeric
columns rounded according to digits. Non-numeric columns are
unchanged. If digits = NULL or invalid, the table is returned
unchanged.
A symmetric input–output table, a use table, a supply table, or a margins/taxes table. Must be a data frame or tibble.
Integer scalar giving the number of digits for rounding.
If NULL, no rounding is performed. If not numeric, a warning is
issued and the table is returned unchanged.
Rounding conventions in published tables differ across sources:
Eurostat (2008, Manual of Supply, Use and Input-Output Tables), presents benchmark IOTs rounded to integers (millions of EUR).
UN (2018, Handbook on SUTs and IOTs), notes that examples may not sum exactly because of rounding (p. 15).
This function allows the user to replicate such rounded presentations
while keeping analytic pipelines consistent. Internally, rounding should
be used with care: repeated rounding in intermediate steps may accumulate
error. For modelling, keep digits = NULL and apply rounding only when
reproducing published sources.