alt Hacker NewsLong article about why not to use UUIDv4 as Primary Keys, but.. Who is doing so? And why are they doing that? How would you solve their requirements? Just throwing out "you can use UUIDv7" doesn't help with, e.g., the size they take up.
Aren't people using (big)ints are primary keys, and using UUIDs as logical keys for import/export, solving portability across different machines?
Replies
Sayrus • yesterday at 11:24 AM
UUIDs are usually the go-to solution to enumeration problems. The space is large enough that an attacker cannot guess how many X you have (invoices, users, accounts, organizations, ...). When people replace the ints by UUIDv4, they keep them as primary keys.
➕ show 1 reply
This was written based on working on several Postgres databases at different companies of “medium” size as a consultant, that had excessive IO and latency and used UUID v4 PKs/FKs. They’re definitely out there. We could transform the schema for some key tables as a demonstration with big int equivalents and show the IO latency reduction. With that said, the real world PK data type migration is costly but becomes a business decision of whether to do or not.