Select Two (Multi-Select) Questions

Azure SQL Database

Why wrong: Relational database, not a data lake.

Azure Synapse Analytics

Why wrong: Analytics service, not storage.

Azure Cosmos DB

Why wrong: NoSQL database, not a data lake.

Disable schema evolution to prevent accidental schema changes

Why wrong: Schema evolution is a feature, not a requirement for ACID; disabling it is not a best practice.

Run OPTIMIZE commands frequently to compact small files

Why wrong: OPTIMIZE improves performance, not ACID compliance.

Store data in CSV format to simplify schema enforcement

Why wrong: CSV does not support ACID; Delta Lake uses Parquet.

Use the largest resource class for all queries

Why wrong: Reduces concurrency.

Use round-robin distribution for large fact tables

Why wrong: Not optimal; leads to data movement.

Adjust the 'Streaming units' to handle higher throughput.

Why wrong: Streaming units scale performance but do not handle event ordering.

Enable 'Event hub capture' to store raw events for reprocessing.

Why wrong: Capture stores events but does not affect processing order within Stream Analytics.

It automatically evolves the schema without any configuration.

Why wrong: Schema evolution requires configuration options.

It provides sub-second latency for real-time streaming.

Why wrong: Auto Loader is optimized for near-real-time, not sub-second.

It provides built-in deduplication of records.

Why wrong: Deduplication must be implemented separately.

Azure Monitor

Why wrong: Monitor is for observability, not data transformation.

Azure Storage

Why wrong: Storage is for storing data, not transforming it.

Azure Event Hubs

Why wrong: Event Hubs is for data ingestion, not transformation.

Partitions are automatically aligned with distributions.

Why wrong: Partitions and distributions are separate concepts.

Partitioning is only supported on tables with clustered rowstore indexes.

Why wrong: Partitioning works with both columnstore and rowstore.

Scale up the SQL pool to a higher DWU.

Why wrong: Scaling up increases resources but is not a specific optimization technique.

Use heap indexes for fact tables.

Why wrong: Heaps are not recommended for data warehouse workloads.

Use round-robin distribution for all large fact tables.

Why wrong: Round-robin distributes evenly but does not optimize joins.

Enable Transparent Data Encryption (TDE) on the database.

Why wrong: TDE encrypts the entire database at rest, not specific columns.

Create a security policy with filter predicates to restrict access to specific columns.

Why wrong: Filter predicates are for row-level security, not column-level.

Implement Always Encrypted to encrypt columns at the application level.

Why wrong: Always Encrypted is not supported in dedicated SQL pools.

Blob Capacity

Why wrong: Storage metric, not pipeline performance.

SQL Pool DWU Used

Why wrong: Metric for Synapse SQL pool, not ADF.

Azure Data Factory

Why wrong: ADF can copy data to/from Delta Lake but does not implement lakehouse natively.

Azure Cosmos DB

Why wrong: Cosmos DB is a NoSQL database, not a data lake.

Azure SQL Database

Why wrong: Azure SQL Database does not support Delta Lake format.

Use Azure Data Lake Storage as the staging location for better throughput.

Why wrong: Azure SQL Database does not support Data Lake Storage as staging.

Use PolyBase to load directly from Blob Storage to Azure SQL Database.

Why wrong: PolyBase is supported for Synapse, not Azure SQL Database.

Always skip staging to reduce latency.

Why wrong: Skipping staging may cause timeouts for large datasets.

Use Azure RBAC to restrict access to the SQL pool.

Why wrong: RBAC controls access, not encryption in transit.

Disable SSL encryption to reduce latency.

Why wrong: Disabling SSL reduces security.

Use Azure ExpressRoute to connect to the SQL pool.

Why wrong: ExpressRoute provides private connectivity but does not encrypt data.

Sensitivity labels

Why wrong: Sensitivity labels are from Microsoft Information Protection and are applied to assets, but they are not a component of the Data Map itself.

Data flows

Why wrong: Data flows are part of Azure Data Factory, not Purview Data Map.

Configure a self-hosted integration runtime on a public cloud VM.

Why wrong: Self-hosted IR is for on-premises data sources.

Use the default Auto-resolve Integration Runtime for all data flows.

Why wrong: Auto-resolve may use public endpoints.

Allow all public IP addresses to access the Azure Synapse workspace.

Why wrong: Public access reduces security.

Azure Synapse Analytics pipeline

Why wrong: Not needed for this ingestion pipeline.

Power BI

Why wrong: Used for reporting, not data ingestion.

A dataset that references the stored procedure.

Why wrong: A dataset is not required. The stored procedure name and parameters can be defined directly within the stored procedure activity, eliminating the need for a separate dataset.

A self-hosted integration runtime.

Why wrong: Azure IR can connect to Azure SQL Database directly.

A staging blob storage account.

Why wrong: Staging is not needed for stored procedure execution.

Deploy a self-hosted integration runtime on a VM in Azure.

Why wrong: Self-hosted IR is for on-premises connectivity, not retry.

Use staged copy with Azure Data Lake as intermediate storage.

Why wrong: Staging helps with large data, not transient errors.

Increase the degree of copy parallelism.

Why wrong: Parallelism increases throughput, not reliability.

Enable soft delete to recover data

Why wrong: Soft delete increases storage cost due to retention.

Use geo-redundant storage (GRS) for durability

Why wrong: GRS increases cost.

Encrypt data with Azure Storage Service Encryption

Why wrong: Encryption does not affect cost.

Azure Policy to audit access

Why wrong: Auditing, not masking.

Azure SQL Database dynamic data masking

Why wrong: Does not apply to data lake directly.

Azure Storage blob-level access policies

Why wrong: Controls access, not masking.