Ran the file through our deduper and only flagged 1.2% duplicates. That's the cleanest I've seen for United States business data.