Closed k00ni closed 5 years ago
It's the same approach I would take for in-memory sparqling. I've actually started work on a new (property) graph store that will use SQLite quite a bit. It's a a very powerful DB engine. It needs a slightly different approach for partitioning larger datasets, but if I ever get to ARC3, it would also be based on SQLite instead of MySQL.
Could you provide some more information about your work? That would be great.
Also, larger datasets and SQLite seems very contrary. Do you think of a "group" of SQLite files or some kind of in-memory+offline database?
yes, very very many sqlite files, dynamically combined via SQLite's ATTACH option.
@semsol: I was thinking about the idea to "copy" your SPARQL => SQL engine to SQLite and emulate an In-Memory store, because SQLite can run in memory and is basically always available.
Do you have any major concerns or might this worth a shot?