Databricks Data Engineer Associate Practice Test

Answer choices

1. A. Create a separate metastore per notebook so lineage and privileges remain isolated at the code level, as the proposed databricks intelligence platform approach.
2. B. Attach the workspaces to the same Unity Catalog metastore so governance is centralized while compute and workspace collaboration remain separate.
3. C. Place all classic compute in the Databricks control plane so the metastore can see the data, within the documented operational, security, ownership, and validation requirements.
4. D. Merge the three workspaces into one because a Unity Catalog metastore cannot govern more than one workspace, under the documented operational and governance requirements.

Wrong-answer review

• A. Create a separate metastore per notebook so lineage and privileges remain isolated at the code level, as the proposed databricks intelligence platform approach.: Governance is organized at metastore/catalog/schema/object scope, not per notebook.
• C. Place all classic compute in the Databricks control plane so the metastore can see the data, within the documented operational, security, ownership, and validation requirements.: Classic compute runs in the customer's AWS account, not in the Databricks control plane.
• D. Merge the three workspaces into one because a Unity Catalog metastore cannot govern more than one workspace, under the documented operational and governance requirements.: A single metastore can be linked to multiple workspaces in the same region, so merging the workspaces is not required.

Answer choices

1. A. Use a column mask to transform nested JSON into relational columns, for the described technical objective and its associated operational control requirements, as proposed.
2. B. Discard the raw nested payload as soon as the first parse succeeds, for the described technical objective and its associated operational control requirements, for evaluation.
3. C. Make Gold the raw landing layer and Bronze the business aggregate layer, for the required business outcome.
4. D. Keep the raw/nested representation in Bronze and perform deterministic parsing/normalization into Silver, for the stated data ingestion and loading requirement.

Answer choices

1. A. The table is automatically converted to a managed table before the metadata is removed, for the stated security, delivery, and accountability requirements.
2. B. Databricks deletes the external data files because each Unity Catalog DROP owns the underlying lifecycle, as the organization’s selected response.
3. C. The cloud storage path is moved into Databricks-controlled managed storage, as the recommended response to this scenario.
4. D. Unity Catalog removes the table metadata, while the externally managed data files remain at the storage location, in practice.

Answer choices

1. A. Asset Bundles now refer to Unity Catalog tables and cannot define jobs, for the described technical objective and its associated operational control requirements, for evaluation.
2. B. Declarative Automation Bundles are the current name for the capability formerly called Databricks Asset Bundles, within the stated policy framework.
3. C. Declarative Automation Bundles are a Git provider and do not deploy Databricks resources, under the organization’s defined implementation and exception-management process.
4. D. The reviewer is correct; Asset Bundles were replaced by Dashboard tasks, within cross-functional operational-accountability boundaries.

Answer choices

1. A. Use a `For each` task with the SQL text as the input array, for the described technical objective and its associated operational control requirements, under this approach.
2. B. Use a Dashboard task because SQL files and dashboards use the same task semantics, under the described working with lakeflow jobs criteria.
3. C. Use a SQL task configured to run the SQL file on a supported SQL warehouse and depend on the transformation task, within the stated policy framework.
4. D. Store the SQL in a notebook solely so a Notebook task can execute it, even though a SQL task directly supports files, within the defined security and accountability boundaries.

Answer choices

1. A. A Git folder, because source control prevents concurrent data modifications, for the described technical objective and its associated operational control requirements, within this context.
2. B. Unity Catalog, because lineage alone serializes all writes to the physical files, for the described technical objective and its associated operational control requirements.
3. C. A SQL warehouse, because query compute is what creates table transaction history, as the proposed design for the complete governed operational workflow.
4. D. Delta Lake, because its transaction log provides ACID semantics and supports both batch and Structured Streaming against the same table.

Answer choices

1. A. Choose clustering behavior based on observed filter/query patterns and measured benefit, within the described context.
2. B. Typically cluster on each table column so all possible queries are covered, for the specified implementation requirement.
3. C. Choose keys alphabetically because liquid clustering ignores query access patterns, for the stated scenario.
4. D. Typically cluster on the highest-cardinality column regardless of query predicates, as described.

Answer choices

1. A. Deduplicate or aggregate to one row per order before summing order totals by region, as the primary proposed approach.
2. B. Use `approx_count_distinct` on order total and multiply it by region count, within this context.
3. C. Increase the shuffle partition count because inflated totals indicate reducer skew, within the proposed design.
4. D. Convert `order_total` to string before summing so repeated values compare exactly, within the documented operational, security, ownership, and validation requirements.

Answer choices

1. A. Use one checkpoint if both sources are JSON, for the required data ingestion and loading outcome.
2. B. Do not share it; each source ingestion workload requires a separate streaming checkpoint, in the described situation.
3. C. Share the checkpoint because a common target table makes the streams one logical source, within the stated policy framework.
4. D. Checkpoints are unnecessary when the target is Delta, for the described technical objective and its associated operational control requirements, when applied.

Answer choices

1. A. Review and test a policy-driving tag assignment as a security change, within the defined security and accountability boundaries.
2. B. The primary risk is increased Spark shuffle because tags add a partition column, within the proposed design.
3. C. No concern; tag assignments does not affect whether an ABAC policy applies, within this design.
4. D. Tag changes require a cluster restart but cannot change authorization, for this requirement.