During November 2024, Zhou Zemin enhanced data ingestion efficiency for the Shopify/tidb repository by refactoring the Parquet sampling logic in the Lightning import process. He redesigned the system to sample average row size once per table rather than per file, reducing computational overhead and accelerating data size estimation for restore operations. This approach enabled more accurate planning and faster execution of large-scale data loads by calculating total data size from the sampled average row size and total row count. Zhou applied his expertise in Go, data import, and performance optimization, delivering a focused, well-scoped feature that improved throughput and operational accuracy.