Your company is integrating data from a legacy system using ODBC connectivity into a modern data lake on S3. The data includes various formats such as XML, JSON, and CSV. Which AWS service would you recommend for transforming this diverse data into a unified format like Apache Parquet?