Question 140 A partner drops files into a Unity Catalog volume only after a manual approval, and the ingest must start when files are available. Which combination is suitable? An untracked notebook copy-paste operation A fixed hourly full reset of the destination A file-arrival trigger with a checkpointed idempotent ingestion task A Git commit hook as the only evidence of source arrival Explanation Correct Answer C:An event-driven run can respond to arrival while ingestion state protects retries. Incorrect Answers A:Manual extraction lacks reliable retries and source inventory tracking. B:Resetting target data is unnecessary and risks losing history. D:Code commits do not directly signal when a governed volume receives new objects.
Question 141 An external database is accessible through JDBC, but its rows include duplicate primary keys caused by a source defect. What should the landing workflow do? Preserve raw records and establish an explicit duplicate-key quarantine or reconciliation rule Change the Databricks workspace URL Delete random duplicate rows without recording policy Assume JDBC automatically enforces source business uniqueness Explanation Correct Answer A:Separating ingestion from curated deduplication retains evidence and prevents silently arbitrary silver results. Incorrect Answers B:Workspace naming does not resolve duplicate source keys. C:Arbitrary row drops can hide important conflicting source changes. D:JDBC is a connection protocol and does not fix source-data quality defects.
Question 142 A source supplies a JSON payload too large for a simple one-page API response and includes continuation tokens. What key extraction behavior is required? Stop after the first successful HTTP 200 response without inspecting the next-page token Increase an API page-size parameter and assume that one page now contains every available record Iterate through supported pagination and reconcile each committed page Request pages until a fixed hard-coded page count is reached without using the returned continuation tokens Explanation Correct Answer C:All continuation pages must be fetched and durably recorded before completeness is asserted. Incorrect Answers A:HTTP success confirms a page, not the full paginated extraction; later pages would be missed. B:Larger response pages can reduce requests but do not remove the need to follow and reconcile continuation tokens for datasets exceeding one page. D:An assumed fixed page count can miss late or extra pages; the source-provided continuation sequence must determine extraction completion.
Question 143 A new JSON field arrives during Auto Loader ingestion and the stream resumes after schema evolution. Which downstream step is still necessary before treating the field as business-trusted? Validate its meaning, type and quality before silver/gold use Assume the field is certified just because the stream restarted Disable all record-quality rules permanently Grant every user MODIFY on the new table Explanation Correct Answer A:Automatic technical schema discovery does not establish the semantic correctness of a new business attribute. Incorrect Answers B:A successful restart proves processing but not business interpretation. C:Removing validation would expose untrusted source values to analysts. D:Wider permissions do not validate the newly introduced attribute.
Question 144 A batch process needs to ingest only files whose names begin with region- and end in .parquet. What should its SQL file load use? Filter the destination dashboard result after all source files have been ingested Only ALTER TABLE RENAME COLUMN A supported COPY INTO PATTERN for the desired source filenames Load the entire directory and apply a target table filter afterward Explanation Correct Answer C:A file-name pattern restricts candidate objects while the format option declares Parquet interpretation. Incorrect Answers A:Filtering report rows does not prevent ingestion of unwanted source files or implement the file-selection constraint. B:Column renaming is unrelated to source object selection. D:Post-load filtering can import unwanted files and does not implement the source filename-selection requirement of PATTERN.
Question 145 A company is deciding between ingestion directly from a supported database connector and periodically exporting the database to files. Which requirement matters most? Whether database table names start with a vowel Choose the method solely from the number of source database tables Supported source connector behavior, data freshness and operations burden Optimize only the downstream report refresh interval without checking the source connector behavior Explanation Correct Answer C:Connector support and change-capture requirements determine whether managed ingestion reduces custom export steps. Incorrect Answers A:Naming letters do not determine a connector's capabilities or operational fit. B:Table count alone does not establish connector support, acceptable freshness, governance, or operational cost. D:Faster report refreshes cannot compensate for a connector that does not capture required database changes or meet ingestion freshness needs.
Question 146 A stream processes events from files that can arrive late relative to business event time. Which field should be used for event-time analytics after ingestion? The object-storage list order A preserved source event timestamp with an explicit late-data policy Use the notebook execution start time instead of the source record update time for incremental selection Use cloud-object arrival time as event time for every source record Explanation Correct Answer B:Event-time metrics must not automatically substitute ingestion time for the actual business occurrence. Incorrect Answers A:List order is not guaranteed to match business event timestamps. C:Processing time does not describe source row revisions; late or retried changes can be incorrectly ordered or skipped. D:Late-arriving events can differ greatly from arrival time; source event timestamps and lateness policy are needed for event-time analytics.
Question 147 A source supplier issues a corrected extract of a previously processed file under the exact same name. What should engineers not rely on? A versioned source-file naming policy An explicit corrected-file delivery and reconciliation process Default COPY INTO detecting and reprocessing that file solely because contents changed A business-key merge in the curated layer Explanation Correct Answer C:COPY INTO skips files already tracked as loaded by default, including modified files with known identity. Incorrect Answers A:Distinct versioned identities can make corrections auditable when coupled with idempotent target logic. B:A controlled correction process is appropriate for previously ingested source information. D:A merge can reconcile replacement values when the corrected records are actually loaded.
Question 148 A team processes source events containing a region code that drives later access restrictions. When should that field be retained? Only after publishing a global unrestricted gold report Never, because data ingestion cannot include metadata columns In the raw landing stage with provenance so downstream policy can classify and validate it Retain the region code only after aggregation in gold Explanation Correct Answer C:Preserving region labels from source supports correct later governance and anomaly investigation. Incorrect Answers A:Delaying classification until after publication exposes sensitive cross-region information. B:An ingestion pipeline can retain source metadata and business attributes. D:If source-level region is discarded earlier, later validation and authorized regional filtering may be impossible to reconstruct.
Question 149 A Python job calls a REST endpoint and receives JSON with partial results and an HTTP 206 response. What should the author verify? Revoke SELECT on the final table Assume any 2xx response means all expected records were returned Increase the target table ZORDER columns The endpoint's documented partial-content and continuation semantics Explanation Correct Answer D:A noncomplete response cannot be treated as a full logical dataset without understanding the protocol. Incorrect Answers A:Access changes do not change HTTP response completeness. B:Successful transport status does not prove dataset completeness. C:Storage layout does not request missing API content.
Question 150 A source-object ingest job has the correct target schema but fails because the cloud storage credential is expired. Which fix is most direct? Grant a reporting user dashboard edit permission Renew or repair the approved source-storage authentication and test access Rewrite the downstream gold aggregation SQL Increase the number of ingestion retries without repairing source credentials Explanation Correct Answer B:A source authentication failure must be resolved at the storage access boundary before loading can succeed. Incorrect Answers A:Dashboard permissions cannot renew storage identity credentials. C:Aggregation logic does not repair expired source credentials. D:Repeating a request with expired authentication cannot restore authorization; the credential must first be renewed or repaired.