Yeah, the sqlite is *much* slower on initial scan. Not sure why, exactly. I’m doing everything one is supposed to do — pragma synchronous, wrapping all the inserts in a transaction…
Dunno. Guess that’s the price for a relational db.
In any event, to make it seem faster, I think I am going to change it so that initial scans are handled like a background scan — the db count will just gradually grow while it’s indexing for the first time. That way it shows up immediately, and you can connect and see it working.
Have to get incremental db updates working for that, though.