DP-600 Prepare Data Practice Test 2 — 15 Questions

Question 1

You are optimizing a Spark job in Fabric that processes large Parquet files. You notice high memory consumption and frequent spills to disk. Which TWO actions should you take to improve performance?

Accepted Answer

Increase the number of Spark partitions. Optimizing memory usage involves managing how data is partitioned and read. Increasing the number of partitions allows for better parallel processing, preventing individual executors from overloading. Additionally, choosing appropriate file formats or using predicate pushdown ensures that only necessary data is loaded into memory. These steps minimize memory pressure, reduce disk spilling, and significantly decrease total execution time for large-scale data transformation workflows in Spark.

Answer

Decrease the number of Spark partitions

Answer

Increase the executor memory size

Answer

Use the default Spark configuration without changes

Question 2

Refer to the exhibit. You are using Data Wrangler in Fabric to clean a dataset. You apply a transformation that matches the JSON configuration shown. What is the effect of this operation on your dataset?

Accepted Answer

All rows with missing customer_id values are removed.. The JSON configuration indicates a data cleansing rule designed to enforce referential integrity or data quality by removing incomplete records. By dropping rows where the 'customer_id' is null, the transformation ensures that every record in the final dataset contains a valid identifier. This is a standard preprocessing step that prevents downstream errors in joins or aggregations where a missing primary key would otherwise cause data discrepancies or runtime failures.

Answer

The column customer_id is deleted from the table.

Answer

The null values in customer_id are replaced with 0.

Answer

The pipeline execution stops immediately.

Question 3

You need to store raw, unstructured data in a Microsoft Fabric Lakehouse. Which folder structure is the recommended best practice for organizing data in the 'Files' section?

Accepted Answer

Use a Bronze, Silver, Gold folder hierarchy.. The recommended approach for organizing data in a Lakehouse follows a medallion architecture: Bronze (raw), Silver (cleansed), and Gold (curated). For the 'Files' section specifically, placing raw data into a 'Bronze' folder maintains a clear separation of concerns. This organization ensures that data pipelines have a predictable path to read from and write to, which facilitates easier maintenance and governance as the data matures through the lifecycle.

Answer

Store all files directly in the root directory.

Answer

Create a unique folder for every individual user.

Answer

Rename all files to include the ingestion date.

Question 4

You are creating a shortcut in a Microsoft Fabric Lakehouse to access data stored in an Azure Data Lake Storage (ADLS) Gen2 account. What is the primary benefit of using a shortcut instead of copying the data?

Accepted Answer

Shortcuts avoid data duplication and synchronization overhead.. Shortcuts provide a virtualized link to the data without the need to move it. The primary benefit is the elimination of data duplication, which reduces storage costs and maintenance overhead. Because the data remains in the source location, any updates made to the source files are instantly reflected in Fabric. This approach ensures a 'single source of truth' and avoids the latency and complexity associated with periodic ingestion and synchronization jobs.

Answer

Shortcuts increase the query speed by caching data.

Answer

Shortcuts automatically convert files to Delta format.

Answer

Shortcuts provide read/write access to all file types.

Question 5

You have a dataset with a highly skewed distribution in a join key. When performing a join in a Spark notebook, you observe performance degradation. Which technique should you use to handle this skew?

Accepted Answer

Salting. Data skew occurs when one key is over-represented, causing a single executor to handle significantly more work than others, leading to bottlenecks. To resolve this, you should perform 'salting'. By appending a random integer to the skewed key in both datasets before joining, you distribute the data more evenly across the cluster partitions. After the join, you remove the salted integer to restore the original key values and complete the operation.

Answer

Broadcast join

Answer

Increasing the number of executor cores

Answer

Reducing the number of partitions

Question 6

Which THREE file formats are optimized for analytical workloads in Microsoft Fabric?

Accepted Answer

Parquet. Analytical workloads benefit from columnar storage formats that allow for efficient data compression and selective column retrieval. Parquet, Delta, and Avro (in specific streaming scenarios) or optimized Delta tables are standard in Fabric. These formats enable predicate pushdown, projection pushdown, and vectorized query execution, which are critical for high-performance analytics. Specifically, Delta is the foundational format for Fabric, providing ACID compliance, time travel, and advanced indexing that drastically improve the performance and reliability of data lakes.

Answer

CSV

Answer

JSON

Question 7

Refer to the exhibit. You are setting up a notebook in Fabric to ingest data. The source is provided in the exhibit. Which command should you use to load this data efficiently?

Accepted Answer

df = spark.read.option('header', 'true').csv('s3://raw-data'). The exhibit specifies a CSV source with GZIP compression. In Spark, the standard method for loading CSV files with specific parameters is using the 'spark.read.csv' function. Specifying options like 'header=True' and identifying the compression codec is vital for the spark engine to correctly decompress and parse the rows. Since the data is GZIP compressed, Spark automatically handles it, provided the format and options are correctly defined in the notebook session.

Answer

df = spark.read.json('s3://raw-data')

Answer

df = spark.read.parquet('s3://raw-data')

Answer

df = spark.read.text('s3://raw-data')

Question 8

You are cleaning data in Fabric. You have a column containing currency values as strings with symbols, such as '$1,200.50'. You need to convert this into a numeric type. What is the most efficient way to achieve this?

Accepted Answer

Use regexp_replace and cast in Spark SQL.. To convert currency strings to numeric values, you must remove the non-numeric characters (currency symbol and comma) before casting to a float or decimal type. Using Spark SQL's 'regexp_replace' function to strip these characters, followed by a 'cast' to a numeric type, is the most performant and scalable method, as it allows the transformation to be executed in parallel across the entire distributed dataset in the Spark cluster.

Answer

Use a Python loop to iterate through every row.

Answer

Export to Excel, clean it, and re-import.

Answer

Use the 'split' function on every character.

Question 9

You are monitoring a pipeline execution in Fabric. You notice that the 'Copy' activity is taking longer than expected. Which metric should you investigate to identify if the bottleneck is related to data movement throughput?

Accepted Answer

Data read and Data written metrics.. Investigating throughput metrics is crucial for identifying bottlenecks in data movement. The 'Data written' and 'Data read' metrics, combined with the duration, allow you to calculate the effective throughput. If the throughput is significantly lower than the network capacity or storage capability, it suggests issues like slow source response, network congestion, or excessive serialization/deserialization overhead during the copy process. Monitoring these allows you to tune parallelism settings to optimize the transfer performance.

Answer

CPU utilization on the driver node.

Answer

Memory usage of the Lakehouse.

Answer

Number of active users in the workspace.

Question 10

You are designing a data pipeline that must handle late-arriving data. Which strategy ensures that your reports are consistent even when data arrives after the scheduled batch window?

Accepted Answer

Use event-time partitioning and idempotent writes.. To handle late-arriving data, you should implement idempotent pipelines that support re-running or 'backfilling' data based on the business logic's event time rather than processing time. By using watermark-based processing in structured streaming or partitioning your lakehouse tables by the event date, you can ensure that late records are correctly appended to the historical partitions. This maintains reporting consistency because the data is attributed to the correct time period, regardless of when it arrived.

Answer

Delete and recreate the entire Lakehouse daily.

Answer

Increase the timeout threshold of the pipeline.

Answer

Reject all records that arrive after the window.

Question 11

Which tool in Microsoft Fabric allows you to visually explore and perform data transformations without writing code?

Accepted Answer

Data Wrangler. Data Wrangler is the built-in, low-code tool in Fabric that provides an interactive interface for exploring and cleaning datasets. It automatically generates the corresponding code (e.g., Pandas or PySpark) as you perform transformations. This is ideal for data engineers who want to quickly profile data, handle missing values, or apply complex transformations without manually writing every line of code, significantly accelerating the data preparation phase of a project.

Answer

Power Query

Answer

Azure Data Factory Mapping Data Flows

Answer

SQL Server Management Studio

Question 12

Which TWO factors are most important when choosing an partitioning strategy for a large Delta table in Microsoft Fabric?

Accepted Answer

Low to medium cardinality columns like Date.. Partitioning is essential for query performance in large tables. The two most important factors are cardinality and query patterns. Columns with low to medium cardinality, like 'Date' or 'Region', are ideal for partitioning to reduce the amount of data scanned. Furthermore, you must align the partitions with the most frequent query filters to ensure the query optimizer can prune unnecessary files, significantly improving performance and reducing the cost of query execution across the data lake.

Answer

High cardinality columns like Unique ID.

Answer

The number of total files in the storage account.

Answer

The color and naming convention of the folders.

Question 13

You are building a medallion architecture in a Microsoft Fabric Lakehouse. Raw CSV files land daily in the Files section of the lakehouse. You need to ingest these files into a Delta table in the Bronze layer while ensuring schema evolution is automatically handled if new columns appear in the source files. Which command should you use in a PySpark notebook?

Accepted Answer

df.write.format('delta').mode('append').option('mergeSchema', 'true').saveAsTable('bronze_sales'). Using format('delta').mode('append').option('mergeSchema', 'true').saveAsTable() ensures that any newly introduced columns in the incoming CSV files are automatically appended to the existing Delta table schema. This capability is essential for robust Bronze layer ingestion where upstream schema changes occur frequently, preventing ingestion pipeline failures without requiring manual intervention.

Answer

df.write.format('parquet').mode('overwrite').saveAsTable('bronze_sales')

Answer

df.write.format('delta').mode('errorIfExists').saveAsTable('bronze_sales')

Answer

df.write.format('csv').mode('append').saveAsTable('bronze_sales')

Question 14

You are designing a data ingestion pipeline in Microsoft Fabric. You need to land raw CSV files from an external source into the Lakehouse. Which storage location should you use to ensure the data is optimized for delta table conversion and subsequent Lakehouse processing?

Accepted Answer

The Files folder. The Files section of a Microsoft Fabric Lakehouse is the designated landing zone for raw, unstructured, or semi-structured data like CSVs. By landing files here, you can leverage Spark notebooks or Dataflows to perform ELT operations, enabling the transformation of raw files into high-performance Delta tables. This workflow is central to the Medallion architecture, ensuring data quality and lineage throughout the Fabric data lifecycle.

Answer

The Tables folder

Answer

An Azure SQL Database

Answer

The Semantic Model workspace

Question 15

You are cleaning data using a Dataflow Gen2 in Microsoft Fabric. You need to optimize the performance of the transformation process. Which TWO actions should you perform? Each correct answer represents a complete solution.

Accepted Answer

Enable staging on intermediate queries. Optimizing Dataflow Gen2 performance involves minimizing the amount of data moved and shifting computation to the source or the engine level. By enabling query folding, the engine pushes transformations back to the source system, reducing memory usage in Fabric. Additionally, staging reduces the number of times data is read, allowing the engine to cache intermediate results efficiently for subsequent transformations, which significantly improves throughput during large-scale data cleansing operations.

Answer

Disable query folding for all steps

Answer

Increase the timeout settings for the data source

Answer

Convert all data types to Text before processing