Microsoft DP-700 Practice Test 41

Question 400

A notebook’s default-lakehouse relative path fails after deployment to a new workspace. The notebook was not attached to a lakehouse there. Why?

The relative storage path lacks the lakehouse attachment required to resolve its Files mount in the destination notebook session.
A notebook's relative `Files` path works without a default lakehouse in every workspace.
The referenced CSV is automatically converted to a warehouse SQL table when the notebook runs.
The notebook must create a Fabric domain before reading any file path.

Explanation

Correct Answer

The same text path is not portable unless the destination notebook has an equivalent default lakehouse reference.

Incorrect Answers

The `/lakehouse/default` mount depends on a valid attached/default lakehouse; deployment can leave it absent.

A relative file read needs storage attachment, not SQL table conversion.

Domain association doesn't define the default lakehouse mount used by Spark.

Question 401

A notebook file path is corrected. Which test confirms it now reads the intended source?

Check only that a CSV with the same basename exists in a different workspace.
In a fresh notebook session, read a known file by the corrected attachment or URI and compare schema and a known record to the actual source.
Open the source file through a workspace Admin GUI instead of the scheduled notebook identity.
Verify the notebook returns any rows without reconciling expected file contents and path.

Explanation

Correct Answer

Actual Spark data access verifies both path resolution and effective permissions rather than only that the lakehouse item is visible.

Incorrect Answers

A file elsewhere doesn't validate the notebook's resolved mount or intended data source.

Privileged portal access can succeed even when the runtime has no default lakehouse or lacks permissions.

Reading the wrong file with a similar name can still return rows and conceal a deployment path error.

Question 402

A Spark notebook fails with out-of-memory errors during a join on a highly skewed customer ID. What should the engineer investigate?

Disable the join entirely and report incomplete output as a performance success.
Increase a Power BI report font size to distribute Spark keys.
Inspect task/partition distribution and shuffle behavior, then choose an appropriate repartition, salting or join-strategy correction for the measured skew.
Set every row’s customer ID to the same constant.

Explanation

Correct Answer

Skew concentrates large amounts of join data in a small number of tasks, potentially exhausting executor memory despite sufficient aggregate cluster resources.

Incorrect Answers

Removing necessary matching logic violates the transformation result even if the notebook no longer OOMs. It does not implement the corrective control requested in this incident.

Presentation formatting cannot alter partitioning or executor memory. It does not implement the corrective control requested in this incident.

Collapsing keys worsens skew and destroys correct relational matching. It does not implement the corrective control requested in this incident.

Question 403

Most Spark join tasks finish quickly but one executor runs out of memory processing a dominant key. What explains the imbalance?

The failing join is slow only because the destination Delta table is not certified.
A single dominant customer ID can be fixed by globally raising driver memory only.
The job needs a new Dataflow Gen2 query to interpret skewed customer IDs.
The join keys are unevenly distributed and one partition receives a disproportionate share of rows or shuffle data.

Explanation

Correct Answer

One hot partition can OOM even when average task sizes and total data volume appear manageable.

Incorrect Answers

Endorsement doesn't change Spark key frequency, memory pressure, or task partition skew.

A skewed join stresses executor tasks/partitions; driver-only memory doesn't rebalance join work.

Power Query definitions don't repair skew in a Spark join's physical task distribution.

Question 404

An engineer changes partitioning to mitigate a skewed Spark join. What comparison validates the fix?

Run the same input and join, compare per-task input/shuffle sizes, peak executor memory and completed row cardinality.
Compare overall wall-clock time but ignore straggler task distribution and join result counts.
Check one small balanced input sample without the dominant customer key.
Verify only that executor OOM exceptions disappear, ignoring duplicate or dropped join records.

Explanation

Correct Answer

Balanced stage metrics with preserved output detect whether skew actually improved without dropping data.

Incorrect Answers

Faster timing might conceal missing rows; inspect skew metrics and preserve source-to-join cardinality.

A uniform fixture cannot reveal whether the new partition plan handles the observed heavy key.

Resource stability is necessary but insufficient if an unsafe skew fix changes join semantics.

Question 405

A source evolves from integer `Price` to decimal `Price` while older files retain the original schema. How should a Spark pipeline avoid inconsistent totals?

Drop all historical rows instead of supporting the existing schema.
Read with an explicit compatible schema evolution/cast policy, normalize numeric types and null handling, and test both historical and new partitions.
Rename the Price field without changing casts.
Tune only Spark driver UI settings and ignore numeric types.

Explanation

Correct Answer

Schema inference may differ by file batch; reproducible contracts prevent incompatible casts or unexpected rounding after evolution.

Incorrect Answers

Deleting valid history fails the requirement to preserve correct longitudinal totals. It does not implement the corrective control requested in this incident.

Field naming cannot reconcile decimal versus integer precision in computation. It does not implement the corrective control requested in this incident.

Interface settings do not determine the arithmetic semantics of cast expressions. It does not implement the corrective control requested in this incident.

Question 406

A Spark job starts producing different sums after a source type change. One partition casts decimals to integer while another keeps precision. Why?

The job's differences are caused by Decimal arithmetic requiring Fabric sensitivity labels.
The old integer and new decimal schemas are guaranteed to sum identically without conversion.
Inconsistent numeric conversion across schema versions truncates or changes values before aggregation.
The source Delta table must be moved to a different Fabric domain to preserve cents.

Explanation

Correct Answer

A mixed-schema dataset requires a common numerical type and deliberate conversion semantics to preserve totals.

Incorrect Answers

Labels don't dictate integer truncation; inconsistent explicit casts across source partitions cause numeric loss.

Decimal fractional values can be truncated when cast to integer; a common target type must be applied.

Domain classification cannot correct loss of numeric precision in Spark data types.

Question 407

A Spark schema-evolution fix is deployed. What regression test is most useful?

Test only whole-dollar prices and accept matching totals.
Compare only number of rows without validating `Price` value scale and arithmetic.
Read a single recent partition, ignoring older integer-schema files.
Aggregate a fixture containing old integer and new decimal records and reconcile expected precision, nulls and totals across both partitions.

Explanation

Correct Answer

Mixed-version data catches silent truncation or schema mismatch hidden by tests using only new files.

Incorrect Answers

Integer-only cases cannot detect fractional amounts lost under an incorrect cast.

Row counts can match while sums differ because decimals were truncated or coerced.

Evolution compatibility requires both old and new files to produce correct normalized decimal amounts.

Question 408

An Eventstream source is configured but produces authentication errors immediately after a source key rotation. What should operators repair?

Update the supported Eventstream source connection credentials and verify the source identity and endpoint can receive events, then republish or restart the source route as required.
Increase Eventhouse query concurrency as the first source authentication fix.
Rename the Eventstream item to the source tenant ID.
Disable all source-error monitoring when authorization fails.

Explanation

Correct Answer

A source authorization error occurs before events reach the processing operators, requiring connector credentials rather than sink tuning.

Incorrect Answers

Query capacity cannot make an upstream connector secret or access grant valid. It does not implement the corrective control requested in this incident.

A label change cannot update the credentials used by the source protocol. It does not implement the corrective control requested in this incident.

Suppressing errors hides the problem and leaves the ingest path broken. It does not implement the corrective control requested in this incident.

Question 409

Eventstream Live view receives no records, and connector logs report unauthorized from the upstream source. The Eventhouse destination is healthy. What is the likely issue?

The destination Eventhouse table must be renamed when the upstream key rotates.
The source connector cannot authenticate or lacks rights to receive upstream events, independent of the healthy downstream sink.
The Live view is empty because the Eventhouse KQL query acceleration policy is disabled.
The sensor source was assigned to a different workspace domain.

Explanation

Correct Answer

Successful destination health does not prove that the Eventstream is receiving data from its configured producer.

Incorrect Answers

A source connector's authorization is independent of the destination table's display name.

A source unauthorized error occurs before sink querying and cannot be fixed by external-table caching.

Domain organization doesn't invalidate a rotated upstream authentication secret by itself.

Question 410

Source authentication has been repaired for Eventstream. What test verifies recovery?

Verify that the Eventhouse sink can accept an unrelated manual event.
Check only that Eventstream's connection form saved the new key, not that it receives events.
Generate a known upstream event and confirm its source arrival and matching downstream destination record under the repaired source identity.
Run a source event through a different connector identity than the production service.

Explanation

Correct Answer

End-to-end event tracing demonstrates that both authorization and routing are operational.

Incorrect Answers

Sink health alone cannot prove that the upstream connector authenticates and receives live source records.

Configuration persistence isn't proof of successful source authentication or resumed delivery.

An alternate principal can hide the actual rotated-secret failure; test the configured route.

Implementing Data Engineering Solutions Using Microsoft Fabric (DP-700) Practice Tests

Microsoft DP-700 Practice Test 39 - 10 Questions (Last Reviewed 26-Sep-2026)
Microsoft DP-700 Practice Test 40 - 10 Questions (Last Reviewed 25-Sep-2026)
Microsoft DP-700 Practice Test 41 - 10 Questions (Last Reviewed 24-Sep-2026)
Microsoft DP-700 Practice Test 42 - 10 Questions (Last Reviewed 23-Sep-2026)
Microsoft DP-700 Practice Test 43 - 10 Questions (Last Reviewed 22-Sep-2026)

Pass Your Microsoft DP-700 Exam Easy!

Question 400

Explanation

Correct Answer

Incorrect Answers

Question 401

Explanation

Correct Answer

Incorrect Answers

Question 402

Explanation

Correct Answer

Incorrect Answers

Question 403

Explanation

Correct Answer

Incorrect Answers

Question 404

Explanation

Correct Answer

Incorrect Answers

Question 405

Explanation

Correct Answer

Incorrect Answers

Question 406

Explanation

Correct Answer

Incorrect Answers

Question 407

Explanation

Correct Answer

Incorrect Answers

Question 408

Explanation

Correct Answer

Incorrect Answers

Question 409

Explanation

Correct Answer

Incorrect Answers

Question 410

Explanation

Correct Answer

Incorrect Answers

Implementing Data Engineering Solutions Using Microsoft Fabric (DP-700) Practice Tests