Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

One of the issues this is trying to resolve is that v4 UUIDs wholly random nature makes them less than brilliant in the b-trees that SQL databases use for indices (I am actually not sure how much that really matters in practice for most projects).

However, Postgres has an alternate index implementation that uses a hash function. It is crash-safe now (previously wasn't) but you still can't use it to enforce a unique constraint. If you can use that, I suspect it might resolve the problem and would be easier than switching UUID generation algorithm.



It’s a shame v6 uuids didn’t become a thing[1]. It’s the entire reason they were invented.

1: https://datatracker.ietf.org/doc/draft-peabody-dispatch-new-...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: