Hashids.java VS meow_hash

You basically want Hashids but sequential? Why not simple convert a base 10 (0-9) number to hex? (0-F)

I found hashids [1] to be a great compromise between integer ids in the database and copyable non-enumerable strings on the client.

[1] https://hashids.org/

I wonder how this might compare to just storing regular autoincrementing ints in the database, and converting to/from hashids (https://hashids.org/) at the edge. It eliminates the collision concern and stores more compactly at the cost of a tiny amount of encode/decode when processing requests. You’d want to push it down as close to the database layer as possible to avoid inadvertent int ID leaks; I added native hashids support to clickhouse but I’m not sure what other database support might entail.

Since hashids are not really hashes and are not secure at all this is not even achieved. Hashids can be easily decoded without the salt by a simple brute-force attack described by the authors of hashid themselves right on their website: https://hashids.org/

You don't necessarily need to use UUIDs. You could use something like Hashids to generate random strings from your sequential IDs in a reversible way, so that users can't predict what their values will be, but you can decode them as needed.

You should not use UUIDv4 as a primary key. You can use normal int values and then use hashids to make them safe for URL. UUIDv7 might be good to use as well once they are more widely supported as well.

I'm late on the reply but I was using xxhash for something similar, but found that Meow hash was faster for me. Feel free to benchmark, since I'm sure it could vary depending on CPU architecture.

for everyone who doesn't think highly of Casey Muratori (or at least the way he conducts himself online), the author of Meow Hash, took the criticism quite graciously:

https://twitter.com/cmuratori/status/1417546500083568641

https://github.com/cmuratori/meow_hash/issues/80

Is a standardized hashing algorithm required? SHA256 takes several CPU cycle per byte, but there are several hashing algorithm utilizing AES-NI that gives them 15-16 bytes per cycles, which translates to tens of gigabytes per second on modern CPU. At least one of them has been ported to .NET Core, and from what I see properly calls AES primitives.

MeowHash is the fastest non-cryptographic hash available online, and if it were used in CPython, which performs a hash for every fundamental operation in the language, then Python would be much faster.

Meow hash claims 3-4x faster hashing over this, still passes smhasher, and is a few years old. https://mollyrocket.com/meowhash

Have you seen the meow hash implementation? Would that be suitable for this ?