You have a dataset stored in Amazon S3 in CSV format. Your goal is to extract data for a specific month and to retrieve only 3 out of the 10 available columns. Given the need to minimize both compute and network costs, which AWS service or feature should you use to achieve this?