I would have offloaded the work to the database but the integers in question are from an already complex query involving a three-way join, one of which is a self-join. I'm just barely familiar enough with SQL that I can write the query and do have to think a bit each time I re-examine it. However, if those results can be cached in a temporary table while still preserving the results of that original query, then I'll explore that option just in case. For right now, I like the readability of the Perl-side processing.