Microsoft DP-700 Practice Test 28

Question 270

After correcting a Dataflow Gen2 destination decimal-to-integer mapping, what regression test should confirm no unintended rounding, overflow, or lost data across its three output queries?

Inspect only the first successful destination table and declare the entire Dataflow complete.
Compare transformations in Preview while ignoring the persisted destination row sets.
Inspect per-output refresh success and reconcile fractional, boundary and null input fixtures against stored data types and values in all three lakehouse target tables.
Check an unrelated notebook's most recent completion timestamp.

Explanation

Correct Answer

The fixture detects unintended truncation and overflow, while full destination inspection verifies that all three outputs still write accurately instead of merely completing refresh.

Incorrect Answers

A multi-output dataflow must validate all intended output mappings and target tables, not just one.

Preview doesn't prove per-query refresh and writes occurred in the lakehouse.

A notebook success is not evidence that any specific Dataflow Gen2 destination has refreshed.

Question 271

A fact table must join a Type 2 customer dimension. A customer has multiple historical rows, and facts need the version effective at their transaction date. Which join is correct?

Join facts to the dimension using customer business key alone, selecting all historical rows.
Join exclusively to the currently active dimension record for all historical facts.
Use the latest loaded surrogate key regardless of the fact transaction timestamp.
Join on the customer business key plus an effective-from/effective-to date condition or the already resolved version-specific surrogate key.

Explanation

Correct Answer

A Type 2 dimension intentionally has multiple records per business key, so date-bounded matching preserves one correct version per fact.

Incorrect Answers

That creates multiple matches for facts when a customer has more than one effective dimension version.

Older facts would be attributed to today's attributes rather than the customer version valid at transaction time.

Load order isn't the historical effective-date relation and can incorrectly assign old facts to newer versions.

Question 272

A notebook joins facts to a Type 2 dimension using only customer business key and fact totals double. What explains the multiplication?

Each fact matched multiple historical dimension versions because the join omitted the effective-date boundary or unique surrogate member.
The Type 2 dimension is missing a Spark `broadcast` hint so every business key matches twice.
The join needs to remove `customer_id` and keep only the current-state flag.
A warehouse query engine cannot join to tables with historical versions.

Explanation

Correct Answer

SCD Type 2 records share business keys across versions; a one-column key join can become many-to-many.

Incorrect Answers

A broadcast hint affects execution strategy, not effective-dated relationship cardinality.

Dropping the business key breaks customer matching and a current-only filter loses history needed for older facts.

Type 2 tables are joinable; the predicate must identify the single correct version for the fact date.

Question 273

A Type 2 dimension join is repaired. Which QA fixture most directly detects remaining duplicate joins?

Test only one customer who has never changed attributes or acquired a second dimension row.
Use a customer with two nonoverlapping dimension versions and facts before and after the change, then confirm each fact maps to exactly one version and totals stay constant.
Confirm the final report's row count equals the number of dimension rows.
Compare the earliest and latest surrogate keys without checking each fact's effective interval.

Explanation

Correct Answer

This checks temporal join correctness and fact-grain preservation, not only query compilation.

Incorrect Answers

A single-version customer never reveals the many-to-many historical match defect.

Report grain is fact rows, not the number of slowly changing dimension versions.

Key extremes don't prove correct mapping for facts occurring between dimension changes.

Question 274

A reporting model combines staged order headers and lines, but the grain of its proposed sales fact is ambiguous. What design decision should be made first?

Declare the fact grain to be one row per customer without checking order/header/line relationships.
Keep both header and line amounts in a single flattened fact without allocating header measures.
Define a single explicit fact grain, such as one row per order line, and align all measures and dimension keys with that grain.
Choose a grain after writing the final report queries and changing dimension keys ad hoc.

Explanation

Correct Answer

A declared grain determines uniqueness and additive measure semantics; otherwise header- and line-level values can be counted repeatedly.

Incorrect Answers

Customer grain loses order-line detail and can multiply or collapse header measures in unintended ways.

Header amounts repeated per line will overstate totals under line-grain aggregation.

A stable grain must precede fact key and measure design, otherwise joins and totals cannot be validated.

Question 275

A sales fact table combines one header amount with every line from the same order, and totals are inflated. What modeling defect explains it?

The header amount is already at order-line grain because it is repeated on each line.
The model must add more order-line records to preserve order header totals.
The target SQL view isn't endorsed as a certified reporting item.
Header-level measures were repeated at the lower line-level fact grain without allocation or a separate compatible fact.

Explanation

Correct Answer

Mixing measure grains duplicates header totals for multi-line orders even when each line record is otherwise correct.

Incorrect Answers

Repetition doesn't change a measure's true grain; it overcounts unless allocated or modeled separately.

Additional lines make double-counting worse; a consistent fact grain or separate header fact is required.

Endorsement status does not control mathematical grain or duplicate header amounts.

Question 276

After defining the sales fact grain as order line, what audit is useful?

Check that the fact key is unique per order line and reconcile line-level sums and any separately handled header totals against source orders.
Validate only total order count and disregard summed sales measures.
Group by customer but not OrderID/LineID to confirm line uniqueness.
Compare visual formatting of header and line totals in the report.

Explanation

Correct Answer

Fact uniqueness and additive-measure checks verify that the dimensional model preserves the chosen transaction granularity.

Incorrect Answers

Order counts can be correct even when header amounts are multiplied across all lines.

Customer aggregates can hide duplicate order lines; test at the declared composite grain.

Display styling cannot establish that measures reconcile to their natural grains.

Question 277

A source adds a new `customer_segment` column that a strict Fabric sink does not yet recognize. How should the load be revised safely?

Discard every record with customer_segment until all consumers have been rewritten.
Assess compatibility, explicitly update the source-to-target mapping or supported evolution settings, and test required downstream consumers.
Allow automatic positional column mapping without checking target field names and types.
Overwrite the full target with only the newest JSON schema and drop earlier records.

Explanation

Correct Answer

Schema changes affect target contracts and transformations; explicitly managed evolution prevents silent dropping or misalignment of fields.

Incorrect Answers

A compatible added nullable attribute need not require losing incoming records during a managed evolution.

New columns can shift positions; an explicit compatible mapping or supported schema evolution is safer.

A controlled evolution should preserve historical data and handle previously absent nullable values.

Question 278

An ingestion mapping begins failing after a new field appears in source JSON. The destination requires an exact schema contract. What caused it?

A source field addition requires deleting every existing target Delta row before parsing.
The job fails because its domain label does not contain the new field name.
The incoming payload no longer matches the expected source/target mapping and the pipeline has no approved evolution policy for the added field.
Source JSON can never be changed without retiring the Fabric workspace.

Explanation

Correct Answer

A strict sink can reject unexpected properties or fail casts until mappings are intentionally revised.

Incorrect Answers

A new nullable column is a schema mapping/evolution issue; destructive deletion is not an inherent requirement.

Domain labels are organizational; they don't change strict source-to-target schema validation.

A managed schema-evolution or approved mapping update can accommodate compatible added fields.

Question 279

A Dataflow Gen2 destination uses explicit column mapping. Its contract now permits a nullable loyalty_tier column, but older source payloads omit the field. After updating the mapping, which end-to-end test verifies that the destination actually writes the newly mapped attribute without breaking historical records?

Test only newly shaped records and ignore historical rows without the added column.
Compare file names before and after the schema edit without querying destination columns.
Verify an editor preview but omit the production pipeline run and destination readback.
Run a real Dataflow Gen2 refresh using both old and new payloads, then query the target to confirm the mapped new loyalty_tier values persist and older rows read as null according to contract.

Explanation

Correct Answer

This tests the Dataflow Gen2 mapping and actual target readback; historical and newly mapped records must both work. A successful schema declaration alone cannot prove destination values were written correctly.

Incorrect Answers

New records can pass while old rows fail or receive incorrect default values.

File labels don't prove the expected column types, nullability and values in target rows.

Preview cannot establish that the strict destination accepted both old and new payload versions.

Question 280

A daily sales metric calculated in KQL disagrees with the same metric in SQL. Which normalization approach should the team use?

Define a common business date, grouping keys, null policy, filters and rounding semantics, then implement equivalent aggregates in both engines.
Make both KQL and SQL queries display the same metric name and assume sums are equal.
Force both engines to round each input record independently before aggregation without business approval.
Compare only overall daily totals and omit region/product grouping keys.

Explanation

Correct Answer

KQL and SQL can differ in null handling and expression typing; a shared aggregation contract is necessary for comparable results.

Incorrect Answers

Display labels do not align source data, date boundaries, numeric types or filter logic.

Premature rounding can change sums; reconcile numeric precision and rounding at the defined aggregation stage.

Totals can match across engines while their group-level allocations remain inconsistent.

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

Microsoft DP-700 Practice Test 26 - 10 Questions (Last Reviewed 09-Oct-2026)
Microsoft DP-700 Practice Test 27 - 10 Questions (Last Reviewed 08-Oct-2026)
Microsoft DP-700 Practice Test 28 - 10 Questions (Last Reviewed 07-Oct-2026)
Microsoft DP-700 Practice Test 29 - 10 Questions (Last Reviewed 06-Oct-2026)
Microsoft DP-700 Practice Test 30 - 10 Questions (Last Reviewed 05-Oct-2026)

Pass Your Microsoft DP-700 Exam Easy!

Question 270

Explanation

Correct Answer

Incorrect Answers

Question 271

Explanation

Correct Answer

Incorrect Answers

Question 272

Explanation

Correct Answer

Incorrect Answers

Question 273

Explanation

Correct Answer

Incorrect Answers

Question 274

Explanation

Correct Answer

Incorrect Answers

Question 275

Explanation

Correct Answer

Incorrect Answers

Question 276

Explanation

Correct Answer

Incorrect Answers

Question 277

Explanation

Correct Answer

Incorrect Answers

Question 278

Explanation

Correct Answer

Incorrect Answers

Question 279

Explanation

Correct Answer

Incorrect Answers

Question 280

Explanation

Correct Answer

Incorrect Answers

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