alt Hacker NewsThat second point is huge – Rkyv does have limited support for in-place mutation, but it is quite limited!
If you added support for running jq natively, that would be very cool. Lite³ brings the B-trees, jq brings the query parser and bytecode, combined, you get SQLite :P
Yes, in fact the name Lite³ was chosen because it is lighter than SQLite.
I thought about implementing something like jq or JSON query, and this is very possible. It is like sending a mini-database that can be queried at speeds thousands of times faster than any JSON library is able to parse.
One interesting effect of being a zero-copy format is that the 'parsing speed' can exceed the memory bandwidth of the CPU, since to fulfill a query you do not actually need to parse the entire dataset. You only walk the branches of the tree that are actually required.
I've talked to some other people that have also shown interest in this idea. There doesn't really seem to exist a good schemaless single-file format that supports advanced queries. There is only SQLite and maybe HDF5.