Answer: Parquet

The correct answer is Parquet (D) because Snowflake documentation specifically states that floating-point numbers are not truncated when unloading to Parquet files. The community discussion with 100% consensus and upvoted comments references official Snowflake documentation that confirms Parquet preserves floating-point number precision during unloading. Other formats like CSV, JSON, and ORC may truncate or lose precision with floating-point numbers during the unloading process.

Author: LeetQuiz Editorial Team