Databricks Certified Data Engineer Associate Practice Test 55
Question 540
At Subscription video network, the ingestion schema rescue workflow handles viewer activity data. The problem is that the viewer_activity_data_south CSV feed has unrecognized optional columns and the ingest silently discards them. Evidence: raw samples show new attributes that do not appear in the bronze output. Constraint: unknown values must remain auditable even before mapping is approved. The team has not selected a remedy. It first needs evidence that distinguishes the actual dependency or data fault. Which investigation is the strongest first step?
- Roll out controlled schema rules and record each new rescued field
- Retain unexpected fields in a rescued-data column or quarantined bronze path
- Compare incoming headers with inferred schema and rescued fields for viewer_activity_data_south
- Overwrite source files after dropping all unrecognized fields at Subscription video network
Explanation
Correct Answer
C:Compare incoming headers with inferred schema and rescued fields for viewer_activity_data_south directly investigates the evidence (raw samples show new attributes that do not appear in the bronze output) and distinguishes possible causes of the viewer_activity_data_south CSV feed has unrecognized optional columns and the ingest silently discards them. It establishes what must be corrected without assuming a fix before verifying the stated business requirement: unknown values must remain auditable even before mapping is approved.
Incorrect Answers
A:Roll out controlled schema rules and record each new rescued field may introduce a correction, but the question does not ask for installation or migration. A rollout alone cannot replace the required pre-change investigation, especially given the evidence that raw samples show new attributes that do not appear in the bronze output.
B:Retain unexpected fields in a rescued-data column or quarantined bronze path is a plausible lasting design once chosen, but it does not perform the required pre-change investigation at this stage. In the viewer_activity_data_south scenario, the observed evidence is that raw samples show new attributes that do not appear in the bronze output; the requirement remains: unknown values must remain auditable even before mapping is approved.
D:Discarded raw fields cannot be reconstructed during audit. The proposed action (Overwrite source files after dropping all unrecognized fields at Subscription video network) does not meet the requirement that unknown values must remain auditable even before mapping is approved, nor does it constitute the required pre-change investigation.
Question 541
A data engineer at Consumer finance platform is responsible for loan repayment events. The scheduled job timezone issue is that loan_repayment_events_island overnight export fires one hour early after daylight-saving changes. The supporting evidence is that run history shows schedule configured in a different timezone than business cutoff. The engineer must ensure that the export must complete relative to the local reporting calendar. The team has not selected a remedy. It first needs evidence that distinguishes the actual dependency or data fault. What analysis is necessary before deciding on remediation?
- Assume all local timezones have the same UTC offset year-round at Consumer finance platform
- Change the job schedule in a test environment and review next fire times
- Inspect schedule timezone, cron expression and daylight-saving policy for loan_repayment_events_island
- Configure Lakeflow scheduled trigger using an explicit timezone and cutoff
Explanation
Correct Answer
C:Inspect schedule timezone, cron expression and daylight-saving policy for loan_repayment_events_island directly investigates the evidence (run history shows schedule configured in a different timezone than business cutoff) and distinguishes possible causes of loan_repayment_events_island overnight export fires one hour early after daylight-saving changes. It establishes what must be corrected without assuming a fix before verifying the stated business requirement: the export must complete relative to the local reporting calendar.
Incorrect Answers
A:Timezones with dst can shift relative to utc. The proposed action (Assume all local timezones have the same UTC offset year-round at Consumer finance platform) does not meet the requirement that the export must complete relative to the local reporting calendar, nor does it constitute the required pre-change investigation.
B:Change the job schedule in a test environment and review next fire times may introduce a correction, but the question does not ask for installation or migration. A rollout alone cannot replace the required pre-change investigation, especially given the evidence that run history shows schedule configured in a different timezone than business cutoff.
D:Configure Lakeflow scheduled trigger using an explicit timezone and cutoff is a plausible lasting design once chosen, but it does not perform the required pre-change investigation at this stage. In the loan_repayment_events_island scenario, the observed evidence is that run history shows schedule configured in a different timezone than business cutoff; the requirement remains: the export must complete relative to the local reporting calendar.
Question 542
Consider the collection_telemetry_island workload at Municipal waste service: a streaming file reader for collection_telemetry_island reprocesses historical data after every job restart. Inspection found that the streaming query creates a new checkpoint directory on each run. The governed outcome must ensure that a restart must resume discovered files instead of replaying all existing input. The failure and requirements have been diagnosed; the team must select the lasting pipeline or governance approach. Which data-engineering approach resolves the documented weakness?
- Configure cloudFiles with a stable checkpoint path for the collection_telemetry_island stream
- Delete checkpoint state as part of each successful load at Municipal waste service
- Inspect checkpoint continuity, discovery progress and input file identifiers
- Monitor checkpoint accessibility and unexpected replay volume for collection_telemetry_island
Explanation
Correct Answer
A:Configure cloudFiles with a stable checkpoint path for the collection_telemetry_island stream is the intended durable approach for collection_telemetry_island: it addresses the verified problem (a streaming file reader for collection_telemetry_island reprocesses historical data after every job restart) while preserving the required behavior (a restart must resume discovered files instead of replaying all existing input). A diagnostic or monitoring step alone would not create this target-state behavior.
Incorrect Answers
B:Deleting checkpoint progress can cause full replay. The proposed action (Delete checkpoint state as part of each successful load at Municipal waste service) does not meet the requirement that a restart must resume discovered files instead of replaying all existing input, nor does it constitute the target approach.
C:Inspect checkpoint continuity, discovery progress and input file identifiers could clarify the original fault, but the available evidence already establishes its scope; it is not the target approach now requested. The collection_telemetry_island workflow still requires that a restart must resume discovered files instead of replaying all existing input.
D:Monitor checkpoint accessibility and unexpected replay volume for collection_telemetry_island would help detect future recurrence, but metrics alone do not deliver the target approach. The collection_telemetry_island requirement is that a restart must resume discovered files instead of replaying all existing input.
Question 543
At Air quality nonprofit, the privilege revocation workflow handles pollution measurements. The problem is that pollution_measurements_urban former intern still reads a sensitive table after leaving the team. Evidence: UC audit shows the intern belongs to a group with lingering SELECT privilege. Constraint: departed identities must lose data access promptly. The first acceptance run passed. The current task is to detect future regressions and changes over time. Which monitoring approach will detect newly emerging defects?
- Remove inappropriate group privileges or membership while keeping service access
- Monitor stale memberships and unauthorized access attempts to pollution_measurements_urban
- Attempt access as the removed principal and confirm denial
- Rename the sensitive table instead of revoking access at Air quality nonprofit
Explanation
Correct Answer
B:Monitor stale memberships and unauthorized access attempts to pollution_measurements_urban creates recurring visibility into the failure mode (pollution_measurements_urban former intern still reads a sensitive table after leaving the team). Unlike a single successful test, the measure can reveal drift, lag, or new failures after later source and deployment changes while departed identities must lose data access promptly.
Incorrect Answers
A:Remove inappropriate group privileges or membership while keeping service access is a plausible lasting design once chosen, but it does not perform continuing operational detection at this stage. In the pollution_measurements_urban scenario, the observed evidence is that UC audit shows the intern belongs to a group with lingering SELECT privilege; the requirement remains: departed identities must lose data access promptly.
C:Attempt access as the removed principal and confirm denial is a one-time acceptance check. It does not supply continuing operational detection for pollution_measurements_urban; a passing test cannot substitute for the specific task now required.
D:Renaming an object is not an access-control boundary. The proposed action (Rename the sensitive table instead of revoking access at Air quality nonprofit) does not meet the requirement that departed identities must lose data access promptly, nor does it constitute continuing operational detection.
Question 544
At Wholesale produce exchange, the trade settlement files workflow depends on trade_settlement_files_central. Engineers report that trade_settlement_files_central deployment works only when the developer logs in interactively; the evidence is that CI pipeline fails with an expired human authentication token. They require that unattended deployments require restricted non-personal credentials. The target approach is approved. The current task is to introduce it safely with bounded scope and rollback. Which controlled implementation step is appropriate now?
- Configure CI with an approved service principal and scoped deployment grants
- Rotate the pipeline to noninteractive credentials and test target validation
- Run unattended CI from a clean agent and verify only intended resources change
- Store the developer's personal token unencrypted in source control at Wholesale produce exchange
Explanation
Correct Answer
B:Rotate the pipeline to noninteractive credentials and test target validation provides a bounded way to introduce the approved solution into trade_settlement_files_central and detect an early mismatch. The existing signal is CI pipeline fails with an expired human authentication token; a staged change protects the service requirement (unattended deployments require restricted non-personal credentials) and makes rollback practical.
Incorrect Answers
A:Configure CI with an approved service principal and scoped deployment grants is a plausible lasting design once chosen, but it does not perform the controlled implementation step at this stage. In the trade_settlement_files_central scenario, the observed evidence is that CI pipeline fails with an expired human authentication token; the requirement remains: unattended deployments require restricted non-personal credentials.
C:Run unattended CI from a clean agent and verify only intended resources change is a one-time acceptance check. It does not supply the controlled implementation step for trade_settlement_files_central; a passing test cannot substitute for the specific task now required.
D:Source-controlled long-lived personal secrets create exposure and ownership risk. The proposed action (Store the developer's personal token unencrypted in source control at Wholesale produce exchange) does not meet the requirement that unattended deployments require restricted non-personal credentials, nor does it constitute the controlled implementation step.
Question 545
At Retail pharmacy network, the inventory adjustments workflow depends on inventory_adjustments_highland. Engineers report that nested arrays in inventory_adjustments_highland events are written as uninterpreted strings; the evidence is that sample JSON contains an array of line items under each order. They require that analytics requires one record per line item while retaining the parent ID. The new approach is deployed. The current task is to prove correct output and authorized behavior, not just successful installation. Which post-change acceptance check is strongest?
- Land a raw bronze copy then curate line items into a governed silver table
- Verify expected row multiplication and preserved parent/child relationships
- Inspect schema, array cardinality and nested field types for inventory_adjustments_highland
- Split raw JSON text on commas to infer nested objects at Retail pharmacy network
Explanation
Correct Answer
B:Verify expected row multiplication and preserved parent/child relationships directly tests the behavior that should change after the correction. With the observed evidence (sample JSON contains an array of line items under each order), the acceptance result must demonstrate actual data or permission correctness, while retaining the requirement that analytics requires one record per line item while retaining the parent ID.
Incorrect Answers
A:Land a raw bronze copy then curate line items into a governed silver table may introduce a correction, but the question does not ask for installation or migration. A rollout alone cannot replace the post-change acceptance test, especially given the evidence that sample JSON contains an array of line items under each order.
C:Inspect schema, array cardinality and nested field types for inventory_adjustments_highland could clarify the original fault, but the available evidence already establishes its scope; it is not the post-change acceptance test now requested. The inventory_adjustments_highland workflow still requires that analytics requires one record per line item while retaining the parent ID.
D:String splitting is not valid nested json parsing. The proposed action (Split raw JSON text on commas to infer nested objects at Retail pharmacy network) does not meet the requirement that analytics requires one record per line item while retaining the parent ID, nor does it constitute the post-change acceptance test.
Question 546
The match_session_events_central pipeline at Cloud gaming cooperative has this failure: match_session_events_central contains many tables with a sensitive tag but policy enforcement differs by owner. The team observed that audit shows only some similarly tagged tables have manual row filters. Business requirement: one central rule must apply automatically to matching governed objects. The failure and requirements have been diagnosed; the team must select the lasting pipeline or governance approach. What lasting solution should be selected for the verified conditions?
- Use ungoverned free-text tags and assume they enforce security at Cloud gaming cooperative
- Inventory governed tag assignments and effective policy scope for match_session_events_central
- Define a Unity Catalog ABAC policy using governed tags at an appropriate parent scope
- Monitor tag changes, policy conditions and exemption grants for match_session_events_central
Explanation
Correct Answer
C:Define a Unity Catalog ABAC policy using governed tags at an appropriate parent scope is the intended durable approach for match_session_events_central: it addresses the verified problem (match_session_events_central contains many tables with a sensitive tag but policy enforcement differs by owner) while preserving the required behavior (one central rule must apply automatically to matching governed objects). A diagnostic or monitoring step alone would not create this target-state behavior.
Incorrect Answers
A:Untested labels are not enforced access policies. The proposed action (Use ungoverned free-text tags and assume they enforce security at Cloud gaming cooperative) does not meet the requirement that one central rule must apply automatically to matching governed objects, nor does it constitute the target approach.
B:Inventory governed tag assignments and effective policy scope for match_session_events_central could clarify the original fault, but the available evidence already establishes its scope; it is not the target approach now requested. The match_session_events_central workflow still requires that one central rule must apply automatically to matching governed objects.
D:Monitor tag changes, policy conditions and exemption grants for match_session_events_central would help detect future recurrence, but metrics alone do not deliver the target approach. The match_session_events_central requirement is that one central rule must apply automatically to matching governed objects.
Question 547
The match_session_events_north pipeline at Cloud gaming cooperative has this failure: a raw Parquet copy of match_session_events_north is mistaken for a complete Delta table. The team observed that the copied directory lacks _delta_log while Delta history queries fail. Business requirement: restoration needs the committed schema and transaction history. The first acceptance run passed. The current task is to detect future regressions and changes over time. What should be tracked as an operational early warning?
- Copy only a selection of Parquet part files and declare the transaction log unnecessary at Cloud gaming cooperative
- Check backup completeness, retention and restoration failures
- Query historical versions and a current snapshot after restoring match_session_events_north
- Preserve the Delta table log together with supported table data and metadata
Explanation
Correct Answer
B:Check backup completeness, retention and restoration failures creates recurring visibility into the failure mode (a raw Parquet copy of match_session_events_north is mistaken for a complete Delta table). Unlike a single successful test, the measure can reveal drift, lag, or new failures after later source and deployment changes while restoration needs the committed schema and transaction history.
Incorrect Answers
A:Data files alone do not preserve delta transactional history. The proposed action (Copy only a selection of Parquet part files and declare the transaction log unnecessary at Cloud gaming cooperative) does not meet the requirement that restoration needs the committed schema and transaction history, nor does it constitute continuing operational detection.
C:Query historical versions and a current snapshot after restoring match_session_events_north is a one-time acceptance check. It does not supply continuing operational detection for match_session_events_north; a passing test cannot substitute for the specific task now required.
D:Preserve the Delta table log together with supported table data and metadata is a plausible lasting design once chosen, but it does not perform continuing operational detection at this stage. In the match_session_events_north scenario, the observed evidence is that the copied directory lacks _delta_log while Delta history queries fail; the requirement remains: restoration needs the committed schema and transaction history.
Question 548
At Environmental inspection office, the CI service principal workflow handles sampling field notes. The problem is that sampling_field_notes_north deployment works only when the developer logs in interactively. Evidence: CI pipeline fails with an expired human authentication token. Constraint: unattended deployments require restricted non-personal credentials. The failure and requirements have been diagnosed; the team must select the lasting pipeline or governance approach. What governed target implementation should the engineer design?
- Inspect authentication identity, workspace permissions and token expiry for sampling_field_notes_north
- Monitor deployment identity use and privilege drift for sampling_field_notes_north
- Store the developer's personal token unencrypted in source control at Environmental inspection office
- Configure CI with an approved service principal and scoped deployment grants
Explanation
Correct Answer
D:Configure CI with an approved service principal and scoped deployment grants is the intended durable approach for sampling_field_notes_north: it addresses the verified problem (sampling_field_notes_north deployment works only when the developer logs in interactively) while preserving the required behavior (unattended deployments require restricted non-personal credentials). A diagnostic or monitoring step alone would not create this target-state behavior.
Incorrect Answers
A:Inspect authentication identity, workspace permissions and token expiry for sampling_field_notes_north could clarify the original fault, but the available evidence already establishes its scope; it is not the target approach now requested. The sampling_field_notes_north workflow still requires that unattended deployments require restricted non-personal credentials.
B:Monitor deployment identity use and privilege drift for sampling_field_notes_north would help detect future recurrence, but metrics alone do not deliver the target approach. The sampling_field_notes_north requirement is that unattended deployments require restricted non-personal credentials.
C:Source-controlled long-lived personal secrets create exposure and ownership risk. The proposed action (Store the developer's personal token unencrypted in source control at Environmental inspection office) does not meet the requirement that unattended deployments require restricted non-personal credentials, nor does it constitute the target approach.
Question 549
At Wholesale produce exchange, the union versus union all workflow handles trade settlement files. The problem is that trade_settlement_files_urban combines regional extracts but normal UNION removes rows that are legitimately identical across different source batches. Evidence: sample data includes repeated values where batch lineage distinguishes observations. Constraint: the target must preserve all source rows for audit. The first acceptance run passed. The current task is to detect future regressions and changes over time. Which monitoring approach will detect newly emerging defects?
- Use UNION ALL when duplicates must be retained and add a batch identifier
- Monitor row-count reconciliation and unintended dedup for trade_settlement_files_urban
- Verify repeated identical source values remain in distinct batch records
- Use UNION DISTINCT and assume it preserves every source row at Wholesale produce exchange
Explanation
Correct Answer
B:Monitor row-count reconciliation and unintended dedup for trade_settlement_files_urban creates recurring visibility into the failure mode (trade_settlement_files_urban combines regional extracts but normal UNION removes rows that are legitimately identical across different source batches). Unlike a single successful test, the measure can reveal drift, lag, or new failures after later source and deployment changes while the target must preserve all source rows for audit.
Incorrect Answers
A:Use UNION ALL when duplicates must be retained and add a batch identifier is a plausible lasting design once chosen, but it does not perform continuing operational detection at this stage. In the trade_settlement_files_urban scenario, the observed evidence is that sample data includes repeated values where batch lineage distinguishes observations; the requirement remains: the target must preserve all source rows for audit.
C:Verify repeated identical source values remain in distinct batch records is a one-time acceptance check. It does not supply continuing operational detection for trade_settlement_files_urban; a passing test cannot substitute for the specific task now required.
D:Set union removes duplicates and can lose genuine repeated observations. The proposed action (Use UNION DISTINCT and assume it preserves every source row at Wholesale produce exchange) does not meet the requirement that the target must preserve all source rows for audit, nor does it constitute continuing operational detection.
Question 550
At Subscription video network, the Lakeflow Connect standard connector workflow handles viewer activity data. The problem is that a viewer_activity_data_coastal source has a compatible connector but existing custom scripts have divergent retries. Evidence: logs show different authentication and pagination paths across team-owned scripts. Constraint: ingestion must be standardized under governed destinations. The first acceptance run passed. The current task is to detect future regressions and changes over time. Which monitoring approach will detect newly emerging defects?
- Monitor source offset, failed pages and connector task errors for viewer_activity_data_coastal
- Compare loaded rows and source bookmarks before switching production
- Let every notebook invent a new unscheduled direct source sync at Subscription video network
- Use the supported Lakeflow Connect standard connector with scoped source access
Explanation
Correct Answer
A:Monitor source offset, failed pages and connector task errors for viewer_activity_data_coastal creates recurring visibility into the failure mode (a viewer_activity_data_coastal source has a compatible connector but existing custom scripts have divergent retries). Unlike a single successful test, the measure can reveal drift, lag, or new failures after later source and deployment changes while ingestion must be standardized under governed destinations.
Incorrect Answers
B:Compare loaded rows and source bookmarks before switching production is a one-time acceptance check. It does not supply continuing operational detection for viewer_activity_data_coastal; a passing test cannot substitute for the specific task now required.
C:Uncoordinated custom pulls are less governed and harder to recover. The proposed action (Let every notebook invent a new unscheduled direct source sync at Subscription video network) does not meet the requirement that ingestion must be standardized under governed destinations, nor does it constitute continuing operational detection.
D:Use the supported Lakeflow Connect standard connector with scoped source access is a plausible lasting design once chosen, but it does not perform continuing operational detection at this stage. In the viewer_activity_data_coastal scenario, the observed evidence is that logs show different authentication and pagination paths across team-owned scripts; the requirement remains: ingestion must be standardized under governed destinations.