Databricks Databricks-Certified-Data-Engineer-Associate Exam With Confidence Using Practice Dumps

In the Medallion Architecture promoted by Databricks, the Bronze layer represents the raw ingestion layer where data is stored in its original form with minimal transformation. The primary goal is to ensure data fidelity and traceability , not to apply business logic or heavy transformations. Typical operations in the Bronze layer include schema inference (or enforcement), ingestion of raw data, and the addition of metadata columns such as ingestion timestamp, load date/time, and process identifiers. These metadata fields help track lineage, enable debugging, and support incremental processing downstream. Extensive data cleansing (A), aggregations (B), and business transformations (C) are intentionally deferred to the Silver and Gold layers , where data is refined and structured for analytics. By keeping Bronze data close to the source and append-only, organizations maintain a reliable audit trail and can reprocess data if needed. Therefore, adding operational metadata like load timestamps and process IDs is the correct characteristic of Bronze layer transformations.

=========

Option B preserves the historical evidence needed to diagnose the regression. The Lakeflow Jobs UI provides run history, duration information, task details, links to the compute resource, and task metrics. Comparing several slow and normal runs shows whether the increase is consistent or intermittent. Cluster metrics such as CPU utilization, memory pressure, disk I/O, and network activity can then reveal resource spikes or bottlenecks that correlate with the longer runs. Automatically increasing cluster size before diagnosis could hide the cause and increase cost. Archiving old runs or resetting a baseline removes useful comparison data. Rerunning the job once does not make seven days of historical information irrelevant and cannot establish a trustworthy new baseline. Therefore, comparing run-history and cluster metrics, as stated in option B, is the appropriate troubleshooting action.

================

The COPY INTO statement is an idempotent operation, which means that it will skip any files that have already been loaded into the target table1. This ensures that the data is not duplicated or corrupted by multiple attempts to load the same file. Therefore, if the data engineer runs the same command every day without specifying the names of the files to be copied with the FILES keyword or a glob pattern with the PATTERN keyword, the statement will only copy the first file that matches the source location and ignore the rest. To avoid this problem, the data engineer should either use the FILES or PATTERN keywords to filter the files to be copied based on the date or some other criteria, or delete the files from the source location after they are copied into the table2. References: 1: COPY INTO | Databricks on AWS 2: Get started using COPY INTO to load data | Databricks on AWS