A daily AWS Glue ETL job processes XML data from an S3 bucket, which receives new data daily. The job reprocesses all data each time it runs. How can the solutions architect prevent reprocessing of old data?