MikkelFJ
MikkelFJ
random thoughts: Overall, I like many of the suggestions, but there are too many optional and variable parameters in the proposal which would make things slow. There is a lot...
We also need a proper NULL type for representing database data.
> it would be awesome to have more scalar types available: This can be done with fixed length arrays in the current format. However, a typedef in the schema would...
That's what a typedef will do for you. Meanwhile it can be wrapped in a struct.
> can fixed length array types be used the same as built-in scalars (say unit64)? would be great, because if so, yes, a built-in typedef would fully solve this .....
@adsharma @aardappel I'm not reading all comments closely ATM, but regarding multiple keys, then FlatCC (for C) already supports multiple keys. FlatCC does not automatically sort on a key while...
> If I were to use flatbuffers instead of SQL DDL to represent a database, we would have one flatbuffer for the primary key + data and one flatbuffer for...
@aardappel on slow: I am referring to branching. Whenever the reader has to make a choice, things slow down. Currently Flatbuffers is a half the speed of optimally reading a...
@aardappel "drop nested tables" .. what does that mean?" The support to store a flatbuffer table inside another table in a uint8 array has taken at least 40% of the...
It's mostly because correct padding introduces an endless number of special cases to ensure correct alignment within the parent container and also because vtables must be contained within the local...