Closed cehbrecht closed 3 years ago
We might decide to optimize this but haven't had any performance issue up to now. We replicated CMIP5 and CORDEX.
sqlite limits seem hard to reach. https://sqlite.org/limits.html
We have this implementation in our TODO list (probably available in 2022)
I just started to use synda ... have no experience here ... just guessing :)
synda uses the sqlite database. I have noticed that it protocols each downloaded file in the sqlite database. I would guess that if you keep only one instance of synda to replicate cmip6 with millions of files it might slow down ... just a guess, not sure about it.
You don't use SQLAlchemy? If you would do so, one would have the option to use a different database like postgres.
Or maybe using some kind of key-value store?