Re: Actually in many cases
Even such single query data sets can benefit from being in a database, providing you set it up correctly which would presumably require a degree of normalisation.
For your single use, keeping things in a flat file might suffice. Though your "simply sort" notion might not always be that simple in practice. Performance is also directly related to the sort order.
I regularly import large data sets into a db and have to take the necessary measures to do this fast enough (drop and recreate indices). But seeing as I'm running random queries, using a db is the only sane way to do this.