[Assuming you mean duplicate HN submissions]

I normalize the URL and use HN's Algolia search API to find an existing story with a matching URL. I don't currently handle multiple HN submissions of the same article; I just use the matching story I find.

That being said, I could see creating a historical snapshot of all the times an article has been submitted and building a blended comments section in the future.