DEA-C01 mock exam: 25 free practice questions

Q1 / 25

Why is applying `LIMIT` unsuitable as a replacement for representative sampling?

1. AIt does not add randomness or ensure distributed sampling.
2. BIt incrementally refreshes the selected aggregate values.
3. CIt computes an exact distinct count for the limited rows.
4. DIt converts a batch query into a stateful stream.

Q2 / 25

A scheduled Auto Loader batch sees a new column and must incorporate it after an automatic restart. Which configuration fits?

1. ATriggered execution with failOnNewColumns and no schema update
2. BContinuous execution with rescue mode and no schema evolution
3. CTriggered execution with addNewColumns and a schema location
4. DDirectory listing with a checkpoint used as the destination schema

Q3 / 25

An exploratory dashboard needs a fast estimate of customer cardinality on a very large dataset, and an exact answer is unnecessary. What is the best aggregate choice?

1. AUse `approx_count_distinct` for the customer identifier.
2. BUse a mean over the customer identifier values.
3. CUse `LIMIT` and count the returned customer rows.
4. DUse a stateful aggregate without a watermark.

Q4 / 25

A cleanup task must wait for three upstream tasks to finish and then run even if one of them failed. How should it be connected?

1. ALeave Depends on empty and set Run if to At least one failed
2. BSelect the failed task in Depends on and set Run if to None failed
3. CSelect all three in Depends on and set Run if to All done
4. DSelect the successful tasks in Depends on and set Run if to All succeeded

Q5 / 25

What is the recommended starting point when selecting an ingestion layer?

1. AThe most managed layer that supports the source and requirements
2. BA partner connector regardless of source compatibility or requirements
3. CThe most customizable layer even when managed support is sufficient
4. DA continuous pipeline for every ingestion cadence

Q6 / 25

A change must be peer-reviewed before execution and reproduced in two workspaces, while remaining inside the Unity Catalog privilege model. Which approach best satisfies the requirements?

1. AEdit each cloud storage policy outside the catalog
2. BUse Catalog Explorer separately with no recorded or reviewable command
3. CUse Spark UI to export and replay a stage configuration
4. DUse a reviewed SQL `GRANT` statement on the securable object

Q7 / 25

How can a materialized view maintain aggregate values as source data changes?

1. AIt can apply source changes incrementally during refresh.
2. BIt requires an unbounded stateful stream for the full dataset.
3. CIt converts the source into a temporary session view.
4. DIt samples the first source rows during every query.

Q8 / 25

How should a dependency on landed files affect trigger selection?

1. ASelect Scheduled so a clock event substitutes for file readiness
2. BSelect File arrival so new files provide the initiating event
3. CSelect Table update so source-table changes substitute for file arrivals
4. DSelect Table update because it monitors the storage location for files

Q9 / 25

Which pair correctly maps ingestion frequency to pipeline mode?

1. ARecurring batches use governance mode; arrival-driven streaming uses schema mode
2. BRecurring batches use partner mode; arrival-driven streaming uses catalog mode
3. CRecurring batches use triggered mode; arrival-driven streaming uses continuous mode
4. DRecurring batches use continuous mode; arrival-driven streaming uses manual mode

Q10 / 25

What does `GRANT` do in the Unity Catalog permissions model?

1. ARemoves every inherited privilege from a principal
2. BAssigns a privilege on a securable object to a principal
3. CTransfers ownership without changing any privileges
4. DCreates an explicit denial that overrides grants

Q11 / 25

Which Gold object stores a computed result and refreshes it from source data?

1. AA catalog stores and refreshes the result.
2. BA standard view stores and refreshes the result.
3. CA SQL warehouse stores and refreshes the result.
4. DA materialized view stores and refreshes the result.

Q12 / 25

Why should the task type be selected before artifact-specific fields are supplied?

1. AThe task type determines which configuration fields appear
2. BThe task type replaces Depends on and Run if settings
3. CThe task type removes the need to select compute
4. DThe task type converts every artifact into a notebook

Q13 / 25

What does Delta Lake schema enforcement validate?

1. AJob parameters against the schema evolution mode
2. BFile-event delivery against the batch trigger schedule
3. CWritten data against the destination table's schema requirements
4. DCheckpoint metadata against the cloud directory listing

Q14 / 25

What representation exposes task dependencies in a Lakeflow job?

1. AA directed acyclic task graph
2. BA chronological notification list
3. CA table of job owners
4. DA cost dashboard without tasks

Q15 / 25

Which object attribute drives the documented Unity Catalog ABAC policy model?

1. AA cloud file name
2. BA warehouse query duration
3. CA Spark task ID
4. DA governed tag

Q16 / 25

Which statement distinguishes a commit from the push performed by the combined UI action?

1. AThe commit creates a pull request; the push approves it
2. BThe commit updates the remote; the push creates a branch
3. CThe commit records work; the push updates the remote branch
4. DThe commit pulls changes; the push discards conflicts

Q17 / 25

Which combination is unsupported on serverless compute?

1. AR workloads and Spark RDD APIs
2. BPython and Spark Connect APIs
3. CSQL notebooks and automatic infrastructure management
4. DOn-demand scaling and query-profile inspection

Q18 / 25

A Silver table accepts numeric order amounts, but negative values are invalid for the business. Which control directly protects the rule?

1. AIncrease shuffle partitions for the pipeline query.
2. BUse a view name that mentions positive amounts.
3. CAdd a quality condition that rejects or flags negative amounts.
4. DReplace the amount column with an informational comment.

Q19 / 25

Which trigger fits a job whose required initiating signal is a configured time?

1. ATable update
2. BFile arrival
3. CData arrival
4. DScheduled

Q20 / 25

Which destination makes governance immediate for notebook-ingested data?

1. AA write to the trigger that schedules the job
2. BA direct write to a Unity Catalog-governed table
3. CA write to the notebook task's parameter collection
4. DA write to cloud storage with no governed table destination

Q21 / 25

A managed clustered table already has predictive optimization enabled and also runs an hourly scheduled OPTIMIZE job. What change follows the documented recommendation?

1. AReplace the scheduled OPTIMIZE job with VACUUM
2. BDisable predictive optimization and retain ZORDER
3. CConvert the table to external before each run
4. DDisable the scheduled OPTIMIZE job

Q22 / 25

When is `UNSET MANAGED` a supported route from a managed table to an external table?

1. AWhen rolling back a table previously converted with `SET MANAGED`
2. BWhenever a managed table owner wants to preserve files after a drop
3. CWhenever any native managed table has an unused external location
4. DWhen changing a managed table into a federated foreign table directly

Q23 / 25

How should a team prevent one member's branch switch from changing another member's view?

1. AKeep every member on the default branch
2. BGive each member a separate Git folder
3. CUse one shared Git folder with separate commit messages
4. DGive every member access to one shared folder

Q24 / 25

How are idle instances in a pool billed?

1. AThey incur no DBUs, but cloud-provider billing still applies
2. BThey incur both charges at the active-compute usage rate
3. CThey incur DBUs, but cloud-provider billing stops while idle
4. DThey incur neither DBUs nor cloud-provider infrastructure charges

Q25 / 25

A producer adds a string column where the target Delta table requires a numeric field, and the write fails. What is the best diagnosis?

1. AThe failure means the table should become a materialized view.
2. BThe failure proves that a Gold aggregate used the wrong mean.
3. CThe incoming data violates the table's schema requirements.
4. DThe failure proves that the driver requires more memory.