A company runs daily ETL jobs using AWS Glue to transform data from an S3 data lake into a data warehouse. They want to ensure that each job run processes only the new or updated data since the last successful run. Which AWS Glue feature should they use to achieve this.