Free DP-700: Implementing Data Engineering Solutions Using Microsoft Fabric (beta) Exam Questions and Answers

You have a Fabric workspace named Workspace1 that contains an Apache Airflow job named Job1. The DAGs in Job1 process large files and fail with out-of-memory errors on the default compute. You need to increase the node size that Workspace1 uses for Apache Airflow jobs. What should you do?

• AA. In the workspace settings, edit the starter pool for Data workflows and select a larger node size
• BB. In the Admin portal capacity settings, select a larger node size for Apache Airflow jobs
• CC. In the Spark settings of the workspace, edit the default pool and select a larger node size
• DD. In the workspace settings, create a custom pool for Data workflows and select a larger node size

You have a Fabric deployment pipeline named ReleasePipeline that has Development, Test and Production stages. Each stage is assigned to a separate workspace. The Development workspace contains an Import mode semantic model named Model1 that uses an Azure SQL database named SalesDev as its data source. The Production stage must use an Azure SQL database named SalesProd. You need to ensure that Model1 connects to SalesProd after it is deployed to the Production stage. What should you configure?

• AA. A selective deployment of Model1 from the Test stage to the Production stage
• BB. A data source rule for Model1 on the Development stage
• CC. A deployment rule for the Azure SQL database item on the Production stage
• DD. A data source rule for Model1 on the Production stage

You have a Fabric workspace that contains the following items: - A warehouse named Warehouse1 that contains sales tables - A lakehouse named Lakehouse1 that contains 2 TB of JSON files - An eventhouse named Eventhouse1 that receives device telemetry You need to choose a transformation tool for each requirement. The solution must minimize development effort and use the item where the data already resides. For each of the following statements, answer Yes if the tool is appropriate for the requirement. Otherwise, answer No. Statement 1: Use T-SQL in Warehouse1 to rank the rows of existing sales tables by using window functions and store the result in a new table. Statement 2: Use Dataflow Gen2 to parse the 2 TB of JSON files in Lakehouse1 by calling a custom Python library for fuzzy matching. Statement 3: Use a KQL update policy to reshape the telemetry into a second table in Eventhouse1 as it is ingested.

Answer each statement, then reveal.

1. Use T-SQL in Warehouse1 to rank the rows of existing sales tables by using window functions and store the result in a new table.YesNo
2. Use Dataflow Gen2 to parse the 2 TB of JSON files in Lakehouse1 by calling a custom Python library for fuzzy matching.YesNo
3. Use a KQL update policy to reshape the telemetry into a second table in Eventhouse1 as it is ingested.YesNo

You have a Fabric workspace named Workspace1 that uses a custom Spark pool named Pool1 as the default pool. Pool1 has Medium nodes and autoscale set to a maximum of 6 nodes. Developers use the %%configure magic command in notebooks to request executors with more cores and memory than Pool1 defines, which causes capacity throttling during business hours. You need to prevent notebooks and Spark job definitions in Workspace1 from overriding the compute configuration of Pool1. What should you do in the Spark settings of Workspace1?

• AA. Turn off the high concurrency setting for notebooks
• BB. Turn off the Customize compute configurations for items setting on the Pool tab
• CC. Change the runtime version of Workspace1 to an earlier Fabric runtime release
• DD. Reduce the autoscale maximum of Pool1 from 6 nodes to 3 nodes

You are designing a Fabric solution for Contoso. Data engineers need a data store for relational data that supports T-SQL INSERT, UPDATE, and DELETE statements that run in multi-table transactions. Which data store should you use?

• AA. Mirrored database
• BB. Lakehouse
• CC. Warehouse
• DD. Eventhouse

You have a Fabric workspace named Workspace1 that contains a data pipeline named LoadSales. LoadSales contains a Copy data activity that copies 40 GB from an Azure SQL Managed Instance database to Lakehouse1. Recent runs take about three times longer than normal. You need to determine whether the extra time is spent in queuing, reading from the source, or writing to the destination. What should you review?

• AA. The Microsoft Fabric Capacity Metrics app report for the capacity that hosts Workspace1
• BB. The refresh history of the default semantic model of Lakehouse1
• CC. The Spark application details for Lakehouse1 in the monitoring hub
• DD. The run details of the Copy data activity in the LoadSales pipeline run

You have a Fabric workspace named Workspace1 that is connected to the main branch of an Azure DevOps Git repository. Workspace1 contains a notebook named Notebook1. You are a developer who must change Notebook1 without affecting other users of Workspace1. For each of the following selections, choose the correct option from the list. To work in isolation, use [Box 1: Branch out to new workspace / Switch branch / Disconnect workspace] in the Source control pane. After you edit Notebook1 in your isolated workspace, save the change to your branch by selecting [Box 2: Commit / Update all / Undo]. After a pull request is completed in Azure DevOps, bring the merged changes into Workspace1 by selecting [Box 3: Commit / Update all / Disconnect].

Fill each box, then reveal.

1. Box 1
2. Box 2
3. Box 3

You have a Fabric workspace that contains a warehouse named Warehouse1 and a lakehouse named Lakehouse1. Warehouse1 contains a dimension table named DimCustomer that has a surrogate key column named CustomerKey and a business key column named CustomerCode. DimCustomer contains a row that has CustomerKey -1 and the name Unknown. You are loading a fact table named FactSales from a staging table in Lakehouse1. Some staged rows have a CustomerCode that does not exist in DimCustomer. All rows must be loaded, and a Power BI report must show the sales of these rows under Unknown. What should you do when you transform the staged data?

• AA. Join the staging table to DimCustomer by using a left join on CustomerCode and leave CustomerKey as NULL for the unmatched rows
• BB. Join the staging table to DimCustomer by using an inner join on CustomerCode and write the unmatched rows to a rejects table
• CC. Store CustomerCode in FactSales as a degenerate dimension and omit CustomerKey
• DD. Join the staging table to DimCustomer by using a left join on CustomerCode and use -1 as CustomerKey for the unmatched rows

You plan to use domains in Fabric to organize workspaces by business area. Match each role to the task it is the least privileged role to perform. Each role may be used once, more than once, or not at all. Roles: Fabric admin Domain admin Domain contributor Workspace member Tasks: Task 1: Assign a workspace that you administer to a domain by using the workspace settings. Task 2: Specify the domain contributors for an existing domain. Task 3: Create a new domain.

Drag or tap each item onto its target, then reveal.

A notebook named Transform1 in a Fabric workspace runs much longer than usual. You suspect that one Spark stage is slow because a few tasks process far more data than the others (data skew). Put the following actions in the correct order to confirm the suspicion. Actions: Open the Spark UI from the run details. In the monitoring hub, open the run of Transform1 that was slow. Compare the maximum task duration with the median task duration in the stage summary metrics. In the Jobs tab, open the stage of the job that has the longest duration.

Put the items in order, then reveal.