Hi again!

Great question — when using Liquid Clustering, you do not need to run OPTIMIZE manually, even if you're truncating and reloading the entire dataset.

Once you’ve defined CLUSTER BY (x, y, z), Liquid Clustering automatically organizes the data in the background over time. It’s designed exactly for use cases like yours — large, frequent rewrites.

That said, right after a big reload, queries might still hit unclustered data initially. If that's a problem, you can trigger a manual OPTIMIZE WHERE on recent partitions, but in most cases, Liquid will handle it soon after.

Let me know if you're seeing delays even after switching to Liquid.