AWS Certified Data Engineer Associate DEA-C01 Practice Question

AWS Glue job bookmarks maintain state about previously processed data, allowing a job to skip files it has already loaded. Enabling bookmarks turns the load into an incremental batch ingestion and sharply reduces the amount of data that must be read and written each night. Merely increasing DPUs raises cost but still processes all historical files, S3 Transfer Acceleration affects network transfer rather than eliminating redundant reads, and rewriting the job as a Python shell would remove Spark parallelism and require code changes without addressing the root cause.