I know everyone wants "just one more feature", but I've been looking for a tiny kv for a side project that allows you to query by key prefix. I haven't had time to build one yet, and have honestly been hoping to stumble across one. Only the huge kv's seem to offer this, despite the existence of off-the-shelf hashing algorithms that will preserve lexicographic ordering.
That's a really good point and honestly a very reasonable feature for a minimal store. TinyKV stores keys in a HashMap internally, so it doesn't currently support lexicographic ordering or prefix scanning efficiently. But you're absolutely right if we switched to a BTreeMap (or exposed one as an option), prefix queries could be both fast and natural. I'll add this to the roadmap. Would you mind opening a GitHub issue with your specific use case?