CH works just fine for cleaning up rows: Delete with mutations sync=1, or use optimize with deduplicate by, or use aggregate trees and optimize final, or query aggregate tables with final=1.

Numerous ways to achieve removal of old/stale rows.