It can be useful for improving ingestion pipeline: put your pdf collection in a temp table and then extract with pure SQL the information you want.