Materializing Datasets built on Datasets

When materializing a dataset based on a materialized dataset, does Sigma take into account when the first datasets has finished materializing?
What’s the best way to stagger schedules to ensure the freshest data?

Looking into this more, it looks like when materializing a Dataset build on a materialized Dataset, we don’t hit the materialized table. We instead re-run the entire query and materialize the results.
To get the most efficient query, I ended up rebuilding from base sources instead of off of other datasets, since I was hitting some Snowflake memory errors. Letting Sigma optimize the full query instead of breaking it into part helped fix the issue.