Microsoft DP-203 Practice Test

DP-203 practice test: what's covered

• Azure Data Engineer Associate

Free DP-203 sample questions

1. You are provisioning an Azure Data Lake Storage account that will serve directory-scoped POSIX permissions and atomic folder renames for a Synapse Spark analytics workload. Which configuration must be enabled on the storage account?

   • A. The hierarchical namespace feature on a StorageV2 (general-purpose v2) account
   • B. Blob soft delete with a 7-day retention policy on the container
   • C. The SFTP endpoint with a local user mapped to the container
   • D. Read-access geo-redundant storage (RA-GRS) replication only
   Show answer

   Correct: A. ADLS Gen2 is enabled by turning on the hierarchical namespace on a StorageV2 account, which provides directory/file POSIX ACLs and atomic directory operations needed by analytics engines. Soft delete only protects against accidental deletion and is not required. The SFTP endpoint is a transfer protocol, not what gives ACLs, and RA-GRS is a replication choice that does not enable the namespace, so those options are incorrect.

2. Your team is loading a 4 TB fact table into a Synapse dedicated SQL pool. Queries join it heavily on CustomerKey. To minimize data movement during those joins, which table distribution should you choose?

   • A. Hash distribution on the CustomerKey column
   • B. Round-robin distribution across all 60 nodes
   • C. Replicated distribution copied to every node
   • D. Heap storage with no clustered index defined
   Show answer

   Correct: A. Hash distributing a large fact table on the frequent join key co-locates matching rows on the same distribution, eliminating shuffle (data movement) during the join. Round-robin spreads rows randomly and forces redistribution on joins, replicated is only suitable for small dimension tables under ~2 GB because it copies the whole table to every node, and a heap is an indexing choice that does not control distribution, so those are incorrect.

3. A data architect is designing partitioning for time-series Parquet data in a data lake to maximize partition pruning while avoiding the small-file problem. Which practices should be applied? (Choose TWO)

   • A. Partition by a date hierarchy such as year/month/day that matches query filters
   • B. Target file sizes roughly between 100 MB and 1 GB after compaction
   • C. Create one partition folder per individual transaction record
   • D. Store every table as a single uncompressed CSV file
   • E. Partition by a high-cardinality GUID column for even spread
   Show answer

   Correct: A and B. Date-hierarchy partitioning lets engines prune folders that do not match query predicates, and right-sizing files to roughly 100 MB-1 GB balances parallelism against per-file overhead. One folder per record or partitioning on a high-cardinality GUID creates millions of tiny partitions (the small-file problem), and a single uncompressed CSV defeats columnar pruning and compression, so those choices are incorrect.

4. An analytics lakehouse needs ACID transactions, time travel, and schema enforcement layered over files in ADLS Gen2 so concurrent Spark writers do not corrupt tables. Which storage format meets these requirements?

   • A. Delta Lake, which adds a transaction log over Parquet files
   • B. Plain CSV files written with a header row per partition
   • C. Line-delimited JSON exported from the source application
   • D. Fixed-width text files compressed with gzip on write
   Show answer

   Correct: A. Delta Lake stores Parquet data plus a JSON transaction log (_delta_log) that delivers ACID guarantees, optimistic concurrency, time travel, and schema enforcement. CSV, JSON, and fixed-width text are row-oriented or semi-structured formats with no transaction log, so they cannot provide atomic multi-writer commits, versioning, or enforced schemas, which is why those options are incorrect.

5. You must secure folder-level access in an ADLS Gen2 lake so a service principal can read only the /curated/finance directory while inheriting nothing from sibling folders. Which mechanism provides this fine-grained control?

   • A. POSIX-style access control lists (ACLs) applied at the directory level
   • B. A storage account access key shared with the principal
   • C. A subscription-level Azure Policy assignment for tagging
   • D. Transparent data encryption configured on the file system
   Show answer

   Correct: A. ADLS Gen2 ACLs grant read/write/execute at the directory and file level, so you can authorize one folder without granting sibling folders. An account access key gives full account access with no folder scoping, Azure Policy governs resource configuration and tagging rather than data-path permissions, and transparent data encryption protects data at rest but does not control who can read a folder, so those options are incorrect.

Last updated: 2026-10-19

DP-203 practice test FAQ

Related practice tests

• Microsoft DP-700 practice test
• Microsoft DP-600 practice test
• Microsoft DP-900 practice test
• Microsoft DP-300 practice test
• Databricks Data Engineer Associate practice test
• AWS DEA-C01 practice test