DP-900 Describe Core Data Concepts Practice Questions

Free DP-900 Describe Core Data Concepts practice test questions

8 questions from this domain with answers and explanations - different from the samples on the main DP-900 page. Sign up free to practice the full set.

1. Question 1Describe Core Data Concepts

   A hospital stores patient MRI scans, X-ray images, and handwritten doctor notes as scanned PDFs. None of these files follow a tabular schema. Which type of data classification best describes these files?

   • ARelational data
   • BUnstructured dataCorrect
   • CStructured data
   • DSemi-structured data
   ✓ Correct answer: B

   Unstructured data refers to information that does not conform to a predefined data model or schema. MRI scans and X-ray images are binary image files without a tabular format, and handwritten notes scanned as PDFs contain text and images without structured fields. Unstructured data requires different storage and processing approaches compared to data with defined columns and rows, often using object storage or specialized document management systems.

   Why the other options are wrong

   • ARelational data is incorrect because it must be organized into tables with rows and columns following a defined schema.
   • CStructured data is incorrect because it implies organization in fixed fields and records, which these medical files lack.
   • DSemi-structured data is incorrect because while JSON or XML documents have some structure with tags or key-value pairs, these hospital files do not have any inherent structure beyond being image or document files.
2. Question 2Describe Core Data Concepts

   A retail company processes millions of point-of-sale transactions every day in real time to detect fraudulent purchases immediately. Which type of data processing best describes this scenario?

   • ABatch processing
   • BStream processingCorrect
   • CAnalytical processing
   • DETL processing
   ✓ Correct answer: B

   Stream processing handles continuous data flowing from a source and processes it in real-time or near-real-time, enabling immediate analysis and action. The point-of-sale transactions require immediate fraud detection without waiting for batch collection, which demands real-time ingestion and processing of transaction streams. Stream processing technologies like Azure Stream Analytics or Kafka can consume transaction data as it arrives and trigger fraud alerts within seconds, making this the appropriate approach for immediate threat detection.

   Why the other options are wrong

   • ABatch processing is incorrect because it collects and processes data at scheduled intervals, introducing latency that would delay fraud detection unacceptably.
   • CAnalytical processing is incorrect because it is designed for historical trend analysis and complex queries over aggregated data, not real-time transaction monitoring.
   • DETL processing is incorrect because while ETL can handle both real-time and batch scenarios, it is more commonly used for periodic data integration rather than continuous fraud detection.
3. Question 3Describe Core Data Concepts

   A data engineer needs to build a pipeline that ingests data from on-premises SQL Server databases, transforms the data, and loads it into Azure Synapse Analytics. Which Azure service is specifically designed for orchestrating this type of data movement and transformation?

   • AAzure Stream Analytics
   • BAzure Monitor
   • CAzure Event Hubs
   • DAzure Data FactoryCorrect
   ✓ Correct answer: D

   Azure Data Factory is a cloud-based data integration service specifically designed for orchestrating and automating data movement and transformation workflows. It provides visual tools for designing pipelines that can extract data from on-premises and cloud sources, apply transformations, and load data into target systems like Azure Synapse Analytics. Data Factory handles scheduling, monitoring, and error handling, making it the purpose-built service for ETL/ELT orchestration at enterprise scale.

   Why the other options are wrong

   • AAzure Stream Analytics is incorrect because it is designed for real-time stream processing of continuous data flows, not for batch ETL pipelines that move data from on-premises databases to data warehouses.
   • BAzure Monitor is incorrect because it is a monitoring and diagnostics service used for observability and alerting, not for orchestrating data pipelines or data movement.
   • CAzure Event Hubs is incorrect because it is a message ingestion service designed for high-throughput event streaming scenarios, not for orchestrating batch data movement and transformation between different systems.
4. Question 4Describe Core Data Concepts

   Batch processing handles data in real time as each individual record arrives.

   • ATrue
   • BFalseCorrect
   ✓ Correct answer: B

   Batch processing handles data in scheduled intervals, collecting and processing large volumes of data together at predetermined times. It does not process individual records in real-time as they arrive. Real-time processing of individual records is the defining characteristic of stream processing, not batch processing. Batch processing is designed for scenarios where processing delay is acceptable and efficiency improves when handling data in groups.

   Why the other options are wrong

   • ATrue is incorrect. The statement is false. Batch processing handles data in scheduled intervals, collecting and processing large volumes of data together at predetermined times. It does not process individual records in real-time as they arrive.
5. Question 5Describe Core Data Concepts

   In a star schema data warehouse, what are dimension tables used for?

   • AHolding server-level configuration, security, and connection settings that the warehouse engine needs to run
   • BHolding the numeric measures and aggregatable metrics recorded for each individual business transaction event
   • CStoring descriptive attributes that provide context to facts (like product names, dates, locations)Correct
   • DJoining several separate fact tables together to act as a central bridge across the whole schema
   ✓ Correct answer: C

   Dimension tables in a star schema are the descriptive lookup tables that contain the attributes used to analyze facts. These tables store non-numerical information such as product details, customer information, dates, geographic locations, and other contextual attributes that provide meaning to the quantitative measures stored in the fact table. Dimension tables are joined with fact tables using foreign keys to enable multi-dimensional analysis and reporting.

   Why the other options are wrong

   • AConfiguration settings are not warehouse content at all; dimension tables hold descriptive business attributes that give facts meaning.
   • BNumeric measures and metrics live in the fact table; dimensions supply the descriptive context that those measures are grouped and filtered by.
   • DDimensions describe facts rather than link fact tables; fact tables reference dimensions through foreign keys, not the reverse.
6. Question 6Describe Core Data Concepts

   In a relational database, an Orders table has a CustomerID column, and the database rejects any order whose CustomerID does not match an existing row in the Customers table. Which data integrity concept enforces this rule?

   • AReferential integrity between related tablesCorrect
   • BEntity integrity within a single table's own key
   • CDomain integrity limiting a column's values
   • DDenormalization that adds redundancy for speed
   ✓ Correct answer: A

   Referential integrity ensures that a relationship between tables stays valid: a foreign key, here CustomerID in Orders, must point to an existing primary-key row in the referenced Customers table. The database rejects orphaned references, so you cannot record an order for a customer who does not exist. It is enforced by foreign-key constraints. This is distinct from rules that govern a single table's own key or the values allowed within one column.

   Why the other options are wrong

   • BEntity integrity guarantees each row is uniquely identifiable via a non-null primary key; it concerns one table's own key, not references to another table.
   • CDomain integrity restricts a column to valid data types, ranges, or formats; it does not verify that a value exists in a related table.
   • DDenormalization is a warehouse design choice that adds redundancy for query speed; it is not an integrity rule at all.
7. Question 7Describe Core Data Concepts

   An application constantly inserts and updates complete individual records and usually reads all fields of a record at once. Which storage layout best suits this transactional pattern?

   • ARow-based storageCorrect
   • BColumnar storage
   • CGraph-based storage
   • DColumn-family storage
   ✓ Correct answer: A

   In row-based (row-oriented) storage, all columns of a single record are stored together, so writing, updating, or reading an entire record touches one contiguous location. That is ideal for transactional (OLTP) workloads that constantly modify and fetch complete records. Columnar storage groups values by column and favors analytical aggregations, not frequent whole-record writes. Graph and column-family layouts serve relationship-heavy and wide sparse NoSQL scenarios, not this record-at-a-time transactional pattern.

   Why the other options are wrong

   • BColumnar storage groups values by column and is optimized for analytical scans, not frequent whole-record inserts and updates.
   • CGraph storage is designed for highly connected data and relationship traversal, not routine record-by-record transactions.
   • DColumn-family stores target wide, sparse NoSQL data; they are not the standard fit for simple transactional record updates.
8. Question 8Describe Core Data ConceptsSelect all that apply

   A data architect is choosing between OLTP and OLAP systems for two different workloads. Which TWO statements correctly describe an OLAP (analytical) system? (Choose two.)

   • AIt is optimized for complex queries that aggregate large volumes of historical data.Correct
   • BIt commonly uses a denormalized, dimensional schema such as a star schema.Correct
   • CIt is optimized for a high rate of small, concurrent insert and update transactions.
   • DIt enforces ACID guarantees primarily to support individual real-time business transactions.
   ✓ Correct answer: A, B

   OLAP (analytical) systems are designed for complex, read-heavy queries that summarize large amounts of historical data to support decision making, and they typically organize data in denormalized dimensional models such as star schemas to speed those queries. High-rate small transactions and individual real-time transaction guarantees are hallmarks of OLTP systems instead.

   Why the other options are wrong

   • COptimization for many small concurrent inserts and updates describes OLTP, not OLAP.
   • DEnforcing ACID for individual real-time transactions is an OLTP characteristic, not the analytical focus of OLAP.

How Describe Core Data Concepts is tested

This domain holds 199 of the 717 questions in the DP-900 bank, about 28%. The mix is 125 single-answer multiple choice, 33 multiple-response, 20 true/false, 9 yes/no scenario, 6 hotspot and 6 ordering, so it is worth practising the formats as well as the content.

Once you have a few attempts recorded, CertGrid scores every domain separately and points you at the weakest one, so you can drill Describe Core Data Concepts on its own rather than re-running full-length mocks.

Other DP-900 exam domains

• Identify Considerations for Relational Data177 questions
• Describe Considerations for Non-Relational Data137 questions
• Describe an Analytics Workload204 questions
• All DP-900 practice questions717 total
• Describe Core Data Concepts study notesKey concepts
• Microsoft practice examsAll Microsoft

DP-900 Describe Core Data Concepts FAQ