You save $34.99
A shared Fabric backlog includes two mandatory controls. First: A warehouse dashboard refreshes slowly and users blame source ingestion without checking actual SQL workload. Second: Daily notebook outputs should generate an alert when row counts drop outside an accepted band. Which two-part implementation satisfies them together?
Step 1 resolves requirement A: Query history exposes SQL execution trends distinct from copy/load duration. Step 2 resolves requirement B: Data quality anomalies may not appear as compute failures unless explicitly measured. Both acceptance tests are required: the slow query shape and its historical duration are identified; and a deliberately truncated batch triggers an anomaly alert while the prior good batch does not.
Data quality anomalies may not appear as compute failures unless explicitly measured. The other proposed action (Check the semantic model refresh history and dependencies after the upstream load) does not implement the first control. Query history exposes SQL execution trends distinct from copy/load duration.
Query history exposes SQL execution trends distinct from copy/load duration. Data quality anomalies may not appear as compute failures unless explicitly measured. The proposed steps (Check the semantic model refresh history and dependencies after the upstream load; compare multiple refresh-history durations and per-query execution behavior) address other capabilities instead of this required pair.
Query history exposes SQL execution trends distinct from copy/load duration. The other proposed action (compare multiple refresh-history durations and per-query execution behavior) addresses a different operation. Data quality anomalies may not appear as compute failures unless explicitly measured.
The project has two blockers. Blocker A: A shortcut was created but users assume it permanently copies data even if source permissions are revoked. Blocker B: An application writes updates into a mirrored source while engineers propose editing the replicated Fabric target table directly. What pair of actions resolves both blockers?
Step 1 resolves requirement A: A shortcut is a reference to data, not an independently secured copy with unlimited access. Step 2 resolves requirement B: Mirroring is a replication pattern, not a general bidirectional transaction editor. Both acceptance tests are required: an unauthorized reader is denied at the shortcut/source boundary; and the source update propagates and no downstream process attempts unsupported target writes.
A shortcut is a reference to data, not an independently secured copy with unlimited access. The other proposed action (use a partitioned window ordered by quality rank and modified time, not arbitrary dropDuplicates) addresses a different operation. Mirroring is a replication pattern, not a general bidirectional transaction editor.
Mirroring is a replication pattern, not a general bidirectional transaction editor. The other proposed action (Define a type-safe source-to-target schema mapping and test conversion behavior) does not implement the first control. A shortcut is a reference to data, not an independently secured copy with unlimited access.
A shortcut is a reference to data, not an independently secured copy with unlimited access. Mirroring is a replication pattern, not a general bidirectional transaction editor. The proposed steps (Define a type-safe source-to-target schema mapping and test conversion behavior; use a partitioned window ordered by quality rank and modified time, not arbitrary dropDuplicates) address other capabilities instead of this required pair.
The data platform team needs a narrowly scoped fix: An analyst is a member of both a broad OneLake data role and a restricted role and unexpectedly sees extra tables. Which approach satisfies the documented test that the analyst's effective permissions exclude private tables after the excess grant is removed?
Multiple OneLake grant roles combine granted access; a restrictive grant does not revoke access granted elsewhere. Verify that the analyst's effective permissions exclude private tables after the excess grant is removed.
Supported Fabric audit events expose administrative or item activities without being the same as pipeline data logs. However, this alone does not show that the analyst's effective permissions exclude private tables after the excess grant is removed.
Sensitivity labels mark data classification and inform downstream governance; they do not automatically grant read access. However, this alone does not show that the analyst's effective permissions exclude private tables after the excess grant is removed.
Item audit events do not alone identify each source row or transformation step. However, this alone does not show that the analyst's effective permissions exclude private tables after the excess grant is removed.
A Fabric engineer has implemented the following change: Map source fields by validated names and schema, not solely ordinal position. The original requirement was: A new raw CSV export changes column order without changing column names, and positional loads swap values. Which observation best validates that the change works?
Schema-aware mapping prevents silent column misplacement after a source reorder.
Streaming architectures ingest and query events with low delay compared with scheduled batch loads. It does not show that a swapped-column test either loads correctly or fails validation without corrupting values.
Unknown dimension handling preserves fact integrity without fabricating dimension matches. It does not show that a swapped-column test either loads correctly or fails validation without corrupting values.
Transport retry can redeliver messages even when the consumer recovered correctly. It does not show that a swapped-column test either loads correctly or fails validation without corrupting values.
A data engineer must validate the completed action: Review shuffle stages and partition counts and tune aggregation strategy for the workload. It targets this failure: A Spark transformation spends most runtime in a massive shuffle after grouping raw events with very high cardinality. Which test outcome should be accepted?
Excessive shuffle overhead can dominate distributed grouping cost.
Data movement bottlenecks can be independent of Spark compute. It does not show that the corrected job produces identical aggregates with lower shuffle/runtime costs.
Queued ingestion batching combines small chunks to improve throughput, but adds a bounded freshness delay. It does not show that the corrected job produces identical aggregates with lower shuffle/runtime costs.
Reducing intermediate row counts can lower join and shuffle cost. It does not show that the corrected job produces identical aggregates with lower shuffle/runtime costs.
During a Microsoft Fabric deployment review, the following problem is identified: A Dataflow Gen2 writes staged data to a Lakehouse slowly because the optimized transfer path is disabled. The reviewer requires the following result: refresh duration and CU consumption improve for the selected staged workload. What should the engineer implement?
Optimized staged-to-Lakehouse copy can avoid extra serialization and transfer overhead. Verify that refresh duration and CU consumption improve for the selected staged workload.
VACUUM removes files no longer needed beyond the retention threshold; aggressive cleanup can break readers or history. However, this alone does not show that refresh duration and CU consumption improve for the selected staged workload.
Excessive partition cardinality creates metadata and file overhead that can dominate scans. However, this alone does not show that refresh duration and CU consumption improve for the selected staged workload.
Early stream filtering reduces downstream ingestion and query overhead. However, this alone does not show that refresh duration and CU consumption improve for the selected staged workload.
A Fabric release must resolve two independent technical requirements. Requirement A: A streaming query needs a nonoverlapping five-minute aggregate that resets at each fixed boundary. Requirement B: Spark Structured Streaming must resume a Delta write after a compute restart without rereading every event as new. Select TWO actions that directly satisfy both requirements.
Tumbling windows partition time into fixed, nonoverlapping intervals. Confirm that events at two adjacent intervals appear in distinct five-minute aggregates.
Checkpoints persist streaming offsets, progress and state for recoverable jobs. Confirm that a restarted query continues from committed progress without duplicating Delta output.
Arg_max selects the row with the greatest time value in each group.
Make-series generates regular time bins, including configured defaults for empty periods.
After a production change, the runbook records: Include source-system identity and immutable batch IDs in the landing path and record metadata. The previous problem was: Different divisions send files with the same filename; the ingest process overwrites one division's data. What test result confirms effective remediation?
Namespace and batch isolation prevent filename collisions across producers.
Type 1 updates replace the current attribute without retaining prior versions. It does not show that two files with the same basename land as separately traceable batches.
Business-key upserts prevent replayed snapshots from multiplying logical records. It does not show that two files with the same basename land as separately traceable batches.
Completeness and validity gates detect incorrect loads that still return a success status. It does not show that two files with the same basename land as separately traceable batches.
Two data engineering defects have separate acceptance criteria. First: A Dataflow Gen2 refresh fails with a data destination validation error after a source column was renamed. Second: An Eventstream destination table rejects events after a new field type was introduced in source JSON. Which TWO changes are necessary to resolve both?
Dataflow source/query schema and destination mapping must agree. Confirm that the refresh succeeds and fields land in the correct typed columns.
Schema drift can break a streaming destination even when the source remains connected. Confirm that both old and new valid event shapes map correctly or are quarantined by policy.
Type overflow creates a deterministic mapping failure, not a transient network fault.
Rate limiting is generally transient and worsens with uncontrolled immediate retries.
A team is working on two related Fabric work items. Work item A: A business analyst needs Power Query merge and column cleanup without writing PySpark. Work item B: A Fabric pipeline must ingest from an on-premises SQL Server accessible only through a corporate gateway. Select TWO specific implementation steps.
Managed gateway connectivity enables approved on-premises source access for supported connectors. Confirm that the Copy activity reads the authorized source without opening the server publicly.
Power Query-based dataflow transformations suit low-code batch shaping. Confirm that the dataflow outputs a correctly typed joined table after refresh.
Shortcuts can expose external data without per-refresh duplication.
Mirroring replicates supported source changes into Fabric with managed synchronization.
A production analytics workflow has this requirement: A PySpark transformation runs out of driver memory after calling collect() on a huge DataFrame. What is the most appropriate change if success must be shown by this test: a large input completes with bounded driver memory consumption?
Collect moves data to driver memory and can exhaust it irrespective of executor capacity. Verify that a large input completes with bounded driver memory consumption.
Stream routing can discard records during transformations before they reach a sink. However, this alone does not show that a large input completes with bounded driver memory consumption.
Many-to-many join expansion can inflate measures even when SQL executes without errors. However, this alone does not show that a large input completes with bounded driver memory consumption.
KQL table name resolution and authorization depend on the selected database context. However, this alone does not show that a large input completes with bounded driver memory consumption.
Please check your mailbox for a message from support@examlabs.com and follow the directions.
