Implementing Data Engineering Solutions Using Azure Databricks DP-750 Practice Test 70
Question 690
During a lakehouse implementation review, engineers observe: A late-arriving purchase refers to a customer identifier whose dimension record has not yet been ingested. What is the BEST next action?
- Choose and document a supported format and catalog integration with explicit read/write compatibility tests.
- Reevaluate data layout using representative event-time filters and a suitable clustering key.
- Evaluate dimensional normalization against query complexity, maintenance and expected cardinality.
- Define a deliberate unknown-member or late-dimension reconciliation pattern without silently dropping facts.
Explanation
Correct Answer
D:The most appropriate action is: Define a deliberate unknown-member or late-dimension reconciliation pattern without silently dropping facts. It addresses the stated condition: A late-arriving purchase refers to a customer identifier whose dimension record has not yet been ingested. The technical basis is: Late facts and missing dimension keys require explicit referential handling. Without this control, inner joining only currently known dimension rows can undercount legitimate transactions. A direct acceptance check is: The purchase remains traceable and later dimension arrival resolves the customer association without duplication.
Incorrect Answers
A:This alternative primarily addresses a different Azure Databricks problem: A cross-engine consumer expects a Delta table but source is plain Iceberg and engineers assume identical mutation semantics. The corresponding measure there would be: Choose and document a supported format and catalog integration with explicit read/write compatibility tests. The stated situation instead calls for: Define a deliberate unknown-member or late-dimension reconciliation pattern without silently dropping facts.
B:This alternative primarily addresses a different Azure Databricks problem: An event stream is partitioned by ingestion day but analysts constantly filter by event timestamp ranges crossing midnight. The corresponding measure there would be: Reevaluate data layout using representative event-time filters and a suitable clustering key. The stated situation instead calls for: Define a deliberate unknown-member or late-dimension reconciliation pattern without silently dropping facts.
C:This alternative primarily addresses a different Azure Databricks problem: A customer table includes both region hierarchy and customers; teams disagree whether to snowflake descriptive attributes. The corresponding measure there would be: Evaluate dimensional normalization against query complexity, maintenance and expected cardinality. The stated situation instead calls for: Define a deliberate unknown-member or late-dimension reconciliation pattern without silently dropping facts.
Question 691
The team chooses: Create an authorized connection and supported foreign catalog for the external source. Its observed issue is: A workspace needs to query a foreign transactional source without first loading every row into a Delta table. What is the strongest technical basis?
- Catalog separation and workspace access controls can help enforce organizational and geography boundaries.
- Lakehouse Federation lets governed queries access supported external systems through a connection abstraction.
- The object type should reflect whether data is computed on read or persisted and maintained.
- Abstraction views can separate a consumer-facing contract from internal physical table evolution.
Explanation
Correct Answer
B:The decisive mechanism is: Lakehouse Federation lets governed queries access supported external systems through a connection abstraction. That explains the selected action: Create an authorized connection and supported foreign catalog for the external source. In the original scenario: A workspace needs to query a foreign transactional source without first loading every row into a Delta table. The specific unmanaged risk is: Copying the whole remote source for a narrow lookup can create unnecessary latency, duplication and privacy exposure. Validate with: A restricted query through the foreign catalog returns current permitted rows and honors source connectivity limits.
Incorrect Answers
A:This principle concerns a separate data-engineering condition: A platform architect creates one catalog for subsidiaries with incompatible regional data-residency permissions. Its targeted response is: Split catalogs or governance boundaries along residency and ownership requirements and validate workspace bindings. The current case is governed more directly by: Lakehouse Federation lets governed queries access supported external systems through a connection abstraction.
C:This principle concerns a separate data-engineering condition: A data pipeline creates a physical table for a reporting transformation that would be better maintained as a query definition. Its targeted response is: Choose a view, materialized view or stored table based on freshness, cost and refresh responsibilities. The current case is governed more directly by: Lakehouse Federation lets governed queries access supported external systems through a connection abstraction.
D:This principle concerns a separate data-engineering condition: A reporting service needs to keep a stable presentation schema while underlying source columns are renamed by engineering. Its targeted response is: Define a stable contract through managed views and test compatibility before changing source objects. The current case is governed more directly by: Lakehouse Federation lets governed queries access supported external systems through a connection abstraction.
Question 692
A platform architect adopts: Apply a deterministic deduplication rule based on stable order identity and source version before aggregation. For this requirement: An incoming orders table contains repeated order IDs because a vendor retransmits the same daily extract. Which rationale MOST directly supports it?
- Set operations and joins differ in duplicate behavior and result cardinality.
- Set difference helps identify migration gaps without retrieving matched records.
- Deduplication should use meaningful business keys, not arbitrary full-row equality alone.
- Time zone conversion and calendar boundary choice affect the grouping of events.
Explanation
Correct Answer
C:The decisive mechanism is: Deduplication should use meaningful business keys, not arbitrary full-row equality alone. That explains the selected action: Apply a deterministic deduplication rule based on stable order identity and source version before aggregation. In the original scenario: An incoming orders table contains repeated order IDs because a vendor retransmits the same daily extract. The specific unmanaged risk is: Duplicate orders inflate sales totals and can trigger incorrect downstream payouts. Validate with: A seeded repeated-ID sample yields one chosen current order version and reconciled totals.
Incorrect Answers
A:This principle concerns a separate data-engineering condition: A team wants only customer IDs present in both systems but uses an inner join that multiplies repeated rows. Its targeted response is: Use an appropriate intersect or semi-join operation based on set and duplicate requirements. The current case is governed more directly by: Deduplication should use meaningful business keys, not arbitrary full-row equality alone.
B:This principle concerns a separate data-engineering condition: An analyst must find records present in the source system but absent in a target migration. Its targeted response is: Use a left anti-join or EXCEPT operation with aligned schemas and explicit null semantics. The current case is governed more directly by: Deduplication should use meaningful business keys, not arbitrary full-row equality alone.
D:This principle concerns a separate data-engineering condition: A date field mixes UTC timestamps and local store times but a pipeline groups all values by the raw date string. Its targeted response is: Normalize time zone semantics before extracting business-local reporting dates. The current case is governed more directly by: Deduplication should use meaningful business keys, not arbitrary full-row equality alone.
Question 693
An implementation skips: Evaluate compatible instance pools or supported serverless startup behavior rather than keeping unused all-purpose clusters awake. Engineers are observing: A scheduled workflow launches dozens of short-lived classic clusters and loses several minutes each time waiting for virtual machines. What consequence is MOST likely?
- An upgrade or runaway task on the shared cluster can compromise finance pipeline availability.
- A stream may stop unexpectedly and leave event consumption behind its permitted freshness objective.
- Ignoring startup overhead can make small jobs miss a narrow delivery window even when their Spark tasks are fast.
- An unverified runtime switch may break a previously working production ingestion notebook.
Explanation
Correct Answer
C:The direct residual exposure is: Ignoring startup overhead can make small jobs miss a narrow delivery window even when their Spark tasks are fast. The triggering condition is: A scheduled workflow launches dozens of short-lived classic clusters and loses several minutes each time waiting for virtual machines. The omitted safeguard is: Evaluate compatible instance pools or supported serverless startup behavior rather than keeping unused all-purpose clusters awake. This matters because: Pool capacity can shorten classic cluster startup but reserved idle capacity must still be budgeted. A corrective test would show: Job run traces show lower startup latency and the combined pool or serverless cost stays within budget.
Incorrect Answers
A:This failure mode belongs mainly to another situation: A notebook pipeline uses a shared cluster with unstable third-party dependencies while a critical finance job needs controlled isolation. Its preventative action would be: Run the finance job on dedicated supported job compute with its own pinned libraries and permissions. The most direct exposure in the current scenario is: Ignoring startup overhead can make small jobs miss a narrow delivery window even when their Spark tasks are fast.
B:This failure mode belongs mainly to another situation: A long-running streaming pipeline is configured with aggressive auto-termination during brief periods without arriving events. Its preventative action would be: Select stream-appropriate compute and lifecycle settings so monitoring and availability targets are met. The most direct exposure in the current scenario is: Ignoring startup overhead can make small jobs miss a narrow delivery window even when their Spark tasks are fast.
D:This failure mode belongs mainly to another situation: A developer selects a new Databricks Runtime version for production despite an untested JDBC driver dependency. Its preventative action would be: Pin a supported Databricks Runtime version and validate application libraries against it before deployment. The most direct exposure in the current scenario is: Ignoring startup overhead can make small jobs miss a narrow delivery window even when their Spark tasks are fast.
Question 694
A data product owner adopts: Inspect Unity Catalog data lineage and table history in Catalog Explorer, checking actual supported capture coverage. The reason is: An analyst wants to know which upstream pipeline produced a metric displayed in a board dashboard. What observed result BEST proves effectiveness?
- A newly classified test table receives the prescribed mask and an authorized exception is documented.
- A seeded transformation appears with its expected upstream source and downstream target in lineage.
- A test GRANT or REVOKE event appears with actor, timestamp and target in the audit sink.
- Authorized privacy reviewers see full test values while ordinary report roles see only permitted redactions.
Explanation
Correct Answer
B:The strongest acceptance proof is: A seeded transformation appears with its expected upstream source and downstream target in lineage. It specifically verifies: Inspect Unity Catalog data lineage and table history in Catalog Explorer, checking actual supported capture coverage. The original issue was: An analyst wants to know which upstream pipeline produced a metric displayed in a board dashboard. The underlying mechanism is: Data lineage links producers, transformations and consumers where observed activity is supported. Without it, without lineage the team may repair the wrong pipeline after a metric regression.
Incorrect Answers
A:This result would validate a different measure: Create compatible central ABAC policies based on governed classifications and verify inherited coverage. That is relevant to the distinct problem: A central security team manually adds row masking rules to every department table and new datasets repeatedly lack protection. Here the acceptance criterion is: A seeded transformation appears with its expected upstream source and downstream target in lineage.
C:This result would validate a different measure: Configure supported Databricks audit logging and route relevant events to a secured retention destination. That is relevant to the distinct problem: A compliance auditor requests a record of who changed Unity Catalog grants but audit logs are not collected. Here the acceptance criterion is: A seeded transformation appears with its expected upstream source and downstream target in lineage.
D:This result would validate a different measure: Develop an approved role-aware column mask and test its outputs across supported query clients. That is relevant to the distinct problem: A data product stores PII in a shared Delta table but several reports require different masking for each audience. Here the acceptance criterion is: A seeded transformation appears with its expected upstream source and downstream target in lineage.
Question 695
A team omits the recommended measure: Define meaningful latency, error and backlog thresholds with alert routing based on the business SLO. Its environment shows: An alert triggers on every harmless microbatch delay while real hour-long ingestion stalls go unnoticed. What is the MOST direct residual risk?
- Restarting a healthy compute resource cannot repair a denied storage credential or table grant.
- Over-maintenance can spend more resources rewriting files than it saves in queries.
- Ignoring metadata and file-layout changes can increase latency even though row counts decline.
- Excessive false alerts can obscure serious freshness incidents.
Explanation
Correct Answer
D:The direct residual exposure is: Excessive false alerts can obscure serious freshness incidents. The triggering condition is: An alert triggers on every harmless microbatch delay while real hour-long ingestion stalls go unnoticed. The omitted safeguard is: Define meaningful latency, error and backlog thresholds with alert routing based on the business SLO. This matters because: Alert quality depends on actionable thresholds and distinguishing transient variation from sustained failure. A corrective test would show: A short benign delay does not page staff while a sustained backlog triggers the documented severity.
Incorrect Answers
A:This failure mode belongs mainly to another situation: An operations team sees Dataflow-like ingestion failures but cannot identify whether the problem is a Unity Catalog grant or a Spark stage. Its preventative action would be: Correlate error messages, job run IDs, catalog authorization logs and Spark execution evidence before remediation. The most direct exposure in the current scenario is: Excessive false alerts can obscure serious freshness incidents.
B:This failure mode belongs mainly to another situation: A new Delta table maintenance job combines repeated OPTIMIZE and VACUUM at every microbatch even though changes are tiny. Its preventative action would be: Set a measured maintenance cadence or evaluate supported predictive optimization for eligible managed tables. The most direct exposure in the current scenario is: Excessive false alerts can obscure serious freshness incidents.
C:This failure mode belongs mainly to another situation: A query on a Delta table slows after a small percentage of rows are deleted but data files are not optimized. Its preventative action would be: Review deletion-vector behavior, table layout and suitable optimization for the supported runtime. The most direct exposure in the current scenario is: Excessive false alerts can obscure serious freshness incidents.
Question 696
The following control has NOT been implemented: Profile aggregation key distribution, shuffle partitions and executor memory before choosing a targeted change. In this case: A Spark workload repeatedly fails with executor out-of-memory errors on a high-cardinality groupBy. Which exposure remains?
- Overlapping backlogs can amplify compute consumption and create conflicting writes.
- Blindly restarting the cluster may reproduce the same failure at the same stage.
- A recurring failure may never be diagnosed if its only logs are destroyed repeatedly.
- A small number of hot tasks can dominate the join while extra workers remain idle.
Explanation
Correct Answer
B:The direct residual exposure is: Blindly restarting the cluster may reproduce the same failure at the same stage. The triggering condition is: A Spark workload repeatedly fails with executor out-of-memory errors on a high-cardinality groupBy. The omitted safeguard is: Profile aggregation key distribution, shuffle partitions and executor memory before choosing a targeted change. This matters because: Wide aggregations redistribute data and can create memory-intensive hot partitions. A corrective test would show: A controlled representative load finishes without OOM and output group counts reconcile.
Incorrect Answers
A:This failure mode belongs mainly to another situation: A job scheduled each hour starts running for 80 minutes and overlaps the next run. Its preventative action would be: Investigate stage runtime growth and enforce appropriate concurrency policy before raising frequency. The most direct exposure in the current scenario is: Blindly restarting the cluster may reproduce the same failure at the same stage.
C:This failure mode belongs mainly to another situation: An operator restarts a failed Databricks cluster before collecting the Spark executor and driver logs. Its preventative action would be: Preserve diagnostic events and Spark UI stage evidence before restart where possible. The most direct exposure in the current scenario is: Blindly restarting the cluster may reproduce the same failure at the same stage.
D:This failure mode belongs mainly to another situation: A DataFrame join spills heavily because a single join key occurs in most rows. Its preventative action would be: Use query-profile evidence and a supported skew mitigation such as adaptive execution or key-aware repartitioning. The most direct exposure in the current scenario is: Blindly restarting the cluster may reproduce the same failure at the same stage.
Question 697
A platform architect adopts: Use an appropriate intersect or semi-join operation based on set and duplicate requirements. For this requirement: A team wants only customer IDs present in both systems but uses an inner join that multiplies repeated rows. Which rationale MOST directly supports it?
- Spark can reduce shuffle costs by broadcasting a sufficiently small dimension where appropriate.
- SQL NULL indicates absent or unknown values while numeric zero is a valid value.
- Set operations and joins differ in duplicate behavior and result cardinality.
- Aggregation alone cannot distinguish valid repeat enrollments from accidental duplicates.
Explanation
Correct Answer
C:The decisive mechanism is: Set operations and joins differ in duplicate behavior and result cardinality. That explains the selected action: Use an appropriate intersect or semi-join operation based on set and duplicate requirements. In the original scenario: A team wants only customer IDs present in both systems but uses an inner join that multiplies repeated rows. The specific unmanaged risk is: An unrestricted join can multiply keys and exaggerate customer counts. Validate with: A test with repeated IDs produces the expected unique or multiset intersection by policy.
Incorrect Answers
A:This principle concerns a separate data-engineering condition: A pipeline joins a huge fact table with a tiny dimension and a normal shuffle join creates excessive network traffic. Its targeted response is: Evaluate a suitable broadcast join based on measured dimension size and Spark execution plans. The current case is governed more directly by: Set operations and joins differ in duplicate behavior and result cardinality.
B:This principle concerns a separate data-engineering condition: A team filters out null quantities but needs to preserve rows with zero as valid inventory states. Its targeted response is: Implement explicit null predicates rather than treating zero and null as equivalent. The current case is governed more directly by: Set operations and joins differ in duplicate behavior and result cardinality.
D:This principle concerns a separate data-engineering condition: A student dataset uses SQL GROUP BY with unhandled duplicate enrollment rows before reporting course completion. Its targeted response is: Validate business keys and choose a deterministic latest-enrollment record before aggregation. The current case is governed more directly by: Set operations and joins differ in duplicate behavior and result cardinality.
Question 698
A production data platform experiences this condition: A team filters out null quantities but needs to preserve rows with zero as valid inventory states. Which targeted measure should be selected?
- Use an incremental append pattern with suitable deduplication when source events cannot change.
- Use a compatible unpivot or stack transformation with explicit month and amount columns.
- Implement explicit null predicates rather than treating zero and null as equivalent.
- Use a Delta MERGE keyed by SKU with appropriate matched-update and not-matched-insert logic.
Explanation
Correct Answer
C:The most appropriate action is: Implement explicit null predicates rather than treating zero and null as equivalent. It addresses the stated condition: A team filters out null quantities but needs to preserve rows with zero as valid inventory states. The technical basis is: SQL NULL indicates absent or unknown values while numeric zero is a valid value. Without this control, conflating zero with missing data hides legitimate stockouts and distorts service metrics. A direct acceptance check is: A seeded dataset retains zero-quantity items while rows with null quantities follow the documented rule.
Incorrect Answers
A:This alternative primarily addresses a different Azure Databricks problem: An append-only telemetry feed includes newly observed events but engineers use MERGE with expensive full-table scans on every insert. The corresponding measure there would be: Use an incremental append pattern with suitable deduplication when source events cannot change. The stated situation instead calls for: Implement explicit null predicates rather than treating zero and null as equivalent.
B:This alternative primarily addresses a different Azure Databricks problem: A large orders dataset has one column per month but a dashboard expects month-value rows for filtering and grouping. The corresponding measure there would be: Use a compatible unpivot or stack transformation with explicit month and amount columns. The stated situation instead calls for: Implement explicit null predicates rather than treating zero and null as equivalent.
D:This alternative primarily addresses a different Azure Databricks problem: A product catalog receives corrections for existing SKUs plus entirely new items on each incremental load. The corresponding measure there would be: Use a Delta MERGE keyed by SKU with appropriate matched-update and not-matched-insert logic. The stated situation instead calls for: Implement explicit null predicates rather than treating zero and null as equivalent.
Question 699
A platform architect adopts: Include performance and end-to-end tests using representative scale and data distributions. For this requirement: An engineering test suite passes on miniature in-memory data yet fails in production on skewed billion-row joins. Which rationale MOST directly supports it?
- Reproducible deployments require versioned execution dependencies as well as application code.
- Deployment isolation depends on correct values for each environment's object references.
- Immutable release versions enable controlled recovery of pipeline configuration and code.
- Correctness under small inputs does not establish workload performance at production cardinality.
Explanation
Correct Answer
D:The decisive mechanism is: Correctness under small inputs does not establish workload performance at production cardinality. That explains the selected action: Include performance and end-to-end tests using representative scale and data distributions. In the original scenario: An engineering test suite passes on miniature in-memory data yet fails in production on skewed billion-row joins. The specific unmanaged risk is: Skew and memory limits may remain invisible until a full-scale run misses its SLA. Validate with: A load test includes known hot keys and verifies latency, spills and output totals within targets.
Incorrect Answers
A:This principle concerns a separate data-engineering condition: A team cannot reproduce a prior release because package versions are resolved dynamically during deployment. Its targeted response is: Pin package and runtime versions in the governed release artifact and record its dependency manifest. The current case is governed more directly by: Correctness under small inputs does not establish workload performance at production cardinality.
B:This principle concerns a separate data-engineering condition: A multi-environment setup uses the same catalog for test and production because bundle target variables were omitted. Its targeted response is: Define and validate environment-specific catalog, workspace and identity targets in deployment configuration. The current case is governed more directly by: Correctness under small inputs does not establish workload performance at production cardinality.
C:This principle concerns a separate data-engineering condition: An operations group wants to roll back a broken Lakeflow workflow but keeps no previous package version. Its targeted response is: Tag and retain approved bundle releases and document a verified redeploy or rollback procedure. The current case is governed more directly by: Correctness under small inputs does not establish workload performance at production cardinality.
Question 700
A production data team leaves out: Validate timezone-aware timestamp types and quarantine ambiguous or unparseable event values. Its scenario is: A regulated dataset requires all timestamps to parse as UTC but several rows have ambiguous local values. Which failure mode is MOST relevant?
- Numerically valid but impossible values can contaminate customer metrics and machine-learning features.
- Unclear expectation handling may allow financial errors to reach gold tables or stop all low-risk events unnecessarily.
- Overly strict validation can block legitimate records while loose validation may permit missing keys.
- Ambiguous timestamp parsing can move financial events across reporting windows.
Explanation
Correct Answer
D:The direct residual exposure is: Ambiguous timestamp parsing can move financial events across reporting windows. The triggering condition is: A regulated dataset requires all timestamps to parse as UTC but several rows have ambiguous local values. The omitted safeguard is: Validate timezone-aware timestamp types and quarantine ambiguous or unparseable event values. This matters because: Type checks and temporal semantics must be aligned to prevent incorrect time-series results. A corrective test would show: A sample of valid UTC timestamps passes and known ambiguous strings produce the expected exception record.
Incorrect Answers
A:This failure mode belongs mainly to another situation: A bronze stream sometimes includes ages below zero or above 120 yet the transformation casts every value to integer. Its preventative action would be: Define a range expectation with metrics and the approved drop, fail or quarantine disposition. The most direct exposure in the current scenario is: Ambiguous timestamp parsing can move financial events across reporting windows.
B:This failure mode belongs mainly to another situation: A Lakeflow pipeline has a quality rule but nobody has decided whether invalid transactions should stop publishing. Its preventative action would be: Configure expectations with an explicit fail, drop or retention behavior aligned to business severity. The most direct exposure in the current scenario is: Ambiguous timestamp parsing can move financial events across reporting windows.
C:This failure mode belongs mainly to another situation: A new customer feed is missing optional contact information for 8% of rows and the team treats every null as pipeline failure. Its preventative action would be: Profile source missingness and define field-specific required versus optional expectations. The most direct exposure in the current scenario is: Ambiguous timestamp parsing can move financial events across reporting windows.