Microsoft DP-700 Practice Test 21

Question 200

An ETL job appends a revised historical order under the same order key, leaving both old and new amounts. What defect caused it?

The target load treats the correction as an unrelated insert instead of matching and updating the existing business key.
The source correction was assigned the same stable order key, so MERGE cannot distinguish it.
The historical order was intentionally preserved by Type 2 versioning with valid effective dates.
V-Order caused a duplicate SQL fact to appear after physical compaction.

Explanation

Correct Answer

A correction requires key-based reconciliation or intentional versioning; append-only handling duplicates a fact when historical rows are mutable.

Incorrect Answers

A stable key is exactly what enables a correction-aware update. The pipeline instead treats the incoming row as an unrelated append.

Version history could be a legitimate policy, but the scenario describes uncoordinated duplicate amounts under the same order key without the specified Type 2 contract.

V-Order reorganizes physical storage; it does not create new logical rows with the same order key. The append write path did.

Question 201

The ingestion code is updated to handle historical corrections. What validates it?

Submit a new order identifier instead of correcting an existing one.
Load an order, submit a corrected amount with the same business key, and verify the target reflects the specified update/version policy without unintended duplicates.
Check that two rows exist for the order without examining the chosen history policy.
Query only today’s orders and assume past-day corrections were included.

Explanation

Correct Answer

The controlled correction confirms a previously loaded fact is adjusted rather than merely adding a second row.

Incorrect Answers

An insert tests new-key ingestion and cannot show that historical facts are matched and updated correctly.

Duplicate active facts can overstate revenue. Validation needs the declared update/version semantics for one business key.

The target may still have the outdated historical amount. Verify the exact changed prior-period key after the load.

Question 202

A Spark Structured Streaming consumer writes events into Delta and must survive a notebook restart without rereading the entire source. Which mechanism should be configured?

Store the Structured Streaming checkpoint only in a temporary local driver folder.
Increase the trigger interval to one hour and assume the last offset persists automatically.
Use a durable, compatible streaming checkpoint location with an idempotent or transactional sink so the restarted query resumes committed progress.
Write directly into Delta using independent append jobs without identifying their last successful source offset.

Explanation

Correct Answer

Structured Streaming tracks offsets and state through its checkpoint; a durable location permits restart from saved progress rather than reprocessing the full history.

Incorrect Answers

A restarted Fabric notebook may lose that ephemeral folder and reread old source offsets.

Trigger cadence does not replace a durable source-offset and state checkpoint.

Append-only jobs risk replay duplicates after restart unless checkpoints and sink idempotence are configured.

Question 203

A Structured Streaming job restarts with a new empty checkpoint directory and loads old source messages again. Why?

The source messages were deleted by the Delta VACUUM operation before the query resumed.
Increasing executor memory changed the source’s original event IDs.
The new checkpoint retained the offsets from the original streaming query automatically.
The new checkpoint has no previous offsets or state, so the query treats the source as a fresh stream according to its configured starting position.

Explanation

Correct Answer

Changing or deleting a checkpoint discards the existing progress record and can cause replay or incorrect stateful aggregation.

Incorrect Answers

The scenario explicitly identifies a new empty checkpoint. VACUUM on a target table is not evidence that source offsets or stream state were restored.

Memory size does not define previously committed streaming offsets. The restarted query lacked the original durable checkpoint state.

Checkpoint state resides at its configured persistent location. A different empty path has no prior offsets to reuse and may replay the source.

Question 204

After configuring a persistent streaming checkpoint, which recovery test demonstrates proper behavior?

Process a controlled event batch, stop and restart the query with the same checkpoint, then send a second batch and reconcile unique target IDs and offsets.
Restart the stream with a new empty checkpoint and count every event as a new valid insert.
Observe that the Spark session starts successfully but never check the sink after another batch.
Compare only the checkpoint folder size before and after stopping the stream.

Explanation

Correct Answer

A restart fixture verifies state retention and sink behavior beyond a single uninterrupted streaming session.

Incorrect Answers

A new checkpoint does not test recovery of committed offsets. The same durable checkpoint path must be reused across restart.

Startup success cannot detect replay duplicates or missing records; the target business keys and offsets must be reconciled.

File size alone cannot show exact-once business outcomes or whether the next batch resumes at the correct source position.

Question 205

A source adds a nullable `promo_code` field while a Fabric incremental pipeline is already running. Which handling best preserves existing and incoming data?

Delete the destination table and reload all historical rows for any nullable field addition.
Define a supported schema-evolution policy and update mappings/nullable-field handling so old records remain valid and new records carry the optional value.
Silently reinterpret the new source column by physical position in a fixed schema.
Discard all incoming records that contain promo_code and retain only legacy rows.

Explanation

Correct Answer

Schema evolution must be explicit in the ingestion and sink contract; compatible nullable additions need not require destructive rebuilding of history.

Incorrect Answers

A compatible nullable column can be handled through explicit mapping or supported schema evolution without data loss.

Position-based mapping can misassign values; the new promo_code must be resolved by intended field name/type.

Dropping new-schema events violates the requirement to preserve both existing and incoming data.

Question 206

A newly added source column causes a sink mapping error in a strict pipeline. Select TWO likely corrective checks.

Compare the source schema with the fixed target schema and update the supported mapping or evolution settings.
Nullable promo_code values always cause a mapping error, even when the target explicitly permits NULL.
A new source field automatically maps by position into any existing target column of the same name.
The source update timestamp failed to advance, so the new column was rejected during the schema check.
Check the new column’s nullability and type for old rows, then test both prechange and postchange input records.

Explanation

Correct Answers

A strict mapping can fail when a source field is added and the destination does not accept it.

Historical source rows can lack values for the added optional field; the schema policy must handle these rows without invalid coercion.

Incorrect Answers

A nullable column can be valid if the supported schema and mappings accept it. The failure arises from rigid schema or type expectations, not nullability alone.

A strict mapping definition is not automatically expanded just because source schema evolves. The new field may need explicit target compatibility and mapping changes.

An unchanged watermark could omit a row, but the specified error is a sink schema/mapping mismatch caused by a newly added field.

Question 207

After supporting a new nullable source column, what acceptance test establishes compatibility?

Validate only rows emitted after the new column appeared.
Inspect the target schema declaration without executing either version of the source payload.
Verify a green job status without comparing promo_code values for new and legacy rows.
Re-run a representative old-schema row and a new-schema row, verifying both load successfully and populate or null-fill the optional field according to contract.

Explanation

Correct Answer

Testing both versions demonstrates backward-compatible schema handling and prevents a change from silently breaking historical records.

Incorrect Answers

The target must also accept old-schema records with promo_code absent or null. New-only samples cannot test backward compatibility.

A schema can appear compatible yet the actual Copy/merge mapping still fails or misplaces optional field values.

Success alone does not prove correct field population and null handling. The result must be reconciled per schema version.

Question 208

A team needs to query a Delta table already stored in a different Fabric lakehouse, without creating another physical copy. How should it expose the table?

Create a supported OneLake table shortcut targeting that Delta table and grant the consumer the required source and shortcut permissions.
Create a shortcut in Files and assume the SQL endpoint automatically registers its Delta table.
Create a scheduled full copy of the existing source table into every consuming workspace.
Grant full workspace Contributor to every consumer instead of configuring source permissions.

Explanation

Correct Answer

A Tables shortcut makes a reference to external or cross-workspace Delta data and can be recognized as a table without duplicating its files.

Incorrect Answers

Delta data exposed for T-SQL should be addressed via a supported Tables shortcut; Files alone is not automatically registered.

Copying duplicates data and violates the specific no-extra-physical-copy requirement.

Broad write privileges are neither needed nor a safe substitute for target/source read authorization.

Question 209

A new Tables shortcut appears in the lakehouse but its analyst cannot query the target Delta table. The source lakehouse denies access. What explains the failure?

The Tables shortcut was created without duplicating all data into the consuming workspace.
The shortcut is a reference, not an unrestricted copy; the effective identity or shortcut connection must have the required source authorization.
The analyst did not turn on scheduled nightly synchronization for the target Delta folder.
The analyst is not an administrator of every workspace that hosts a shortcut.

Explanation

Correct Answer

Creating a shortcut does not erase the producer’s data-security boundary or give all consumers blanket source privileges.

Incorrect Answers

A shortcut is an intentionally nonduplicating reference; denied source authorization, not missing copies, explains the access failure.

Internal OneLake shortcuts resolve the live target; lack of a copy schedule cannot explain a permission error.

Workspace-wide Admin is not required for authorized reading and would violate least privilege.

Question 210

A team replaces a copied Delta dataset with a Tables shortcut. What demonstrates that it points to the intended live data?

Reconcile the sizes of two physically copied tables after a nightly batch process.
Check whether raw Parquet files under Files have a matching display name.
Update a controlled record in the original Delta source and confirm a suitably authorized query against the shortcut reflects it without rerunning a physical copy job.
Promote the analyst to workspace Contributor and confirm an unrestricted query succeeds.

Explanation

Correct Answer

A source-to-shortcut read test verifies the logical reference and effective access, distinguishing it from a stale materialized duplicate.

Incorrect Answers

Copy counts can agree while a shortcut points to a wrong or stale source; this checks copying, not live reference.

A Files naming match does not establish a Tables shortcut's Delta target or effective reader access.

A privileged query can hide source-identity denials, failing to validate the intended least-privilege consumer.

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

Microsoft DP-700 Practice Test 19 - 10 Questions (Last Reviewed 16-Oct-2026)
Microsoft DP-700 Practice Test 20 - 10 Questions (Last Reviewed 15-Oct-2026)
Microsoft DP-700 Practice Test 21 - 10 Questions (Last Reviewed 14-Oct-2026)
Microsoft DP-700 Practice Test 22 - 10 Questions (Last Reviewed 13-Oct-2026)
Microsoft DP-700 Practice Test 23 - 10 Questions (Last Reviewed 12-Oct-2026)

Pass Your Microsoft DP-700 Exam Easy!

Question 200

Explanation

Correct Answer

Incorrect Answers

Question 201

Explanation

Correct Answer

Incorrect Answers

Question 202

Explanation

Correct Answer

Incorrect Answers

Question 203

Explanation

Correct Answer

Incorrect Answers

Question 204

Explanation

Correct Answer

Incorrect Answers

Question 205

Explanation

Correct Answer

Incorrect Answers

Question 206

Explanation

Correct Answers

Incorrect Answers

Question 207

Explanation

Correct Answer

Incorrect Answers

Question 208

Explanation

Correct Answer

Incorrect Answers

Question 209

Explanation

Correct Answer

Incorrect Answers

Question 210

Explanation

Correct Answer

Incorrect Answers

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