100% Latest & Updated Snowflake SnowPro Core COF-C03 Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 760
An ingestion-driven table refresh triggers costly queries that scan all historical data repeatedly Which technical action is most appropriate under this requirement?
- Disable refresh metrics to hide cost
- Compare an incremental design such as streams/tasks or dynamic tables against full recomputation
- Add more user roles without changing pipeline logic
- Recreate all source tables every refresh
Explanation
Correct Answer
B:Compare an incremental design such as streams/tasks or dynamic tables against full recomputation. This addresses the stated condition without bypassing the required boundary.
Incorrect Answers
A:The problem: An ingestion-driven table refresh triggers costly queries that scan all historical data repeatedly. The appropriate decision: Compare an incremental design such as streams/tasks or dynamic tables against full recomputation.
C:The problem: An ingestion-driven table refresh triggers costly queries that scan all historical data repeatedly. The appropriate decision: Compare an incremental design such as streams/tasks or dynamic tables against full recomputation.
D:The problem: An ingestion-driven table refresh triggers costly queries that scan all historical data repeatedly. The appropriate decision: Compare an incremental design such as streams/tasks or dynamic tables against full recomputation.
Question 761
A machine-learning feature pipeline repeatedly computes wide transformations and uses excessive remote spill Which proposal best pairs a technically correct change with the outcome that would validate it?
- Apply SQL DISTINCT blindly at every stage; no direct validation of The chosen design reduces spill while retaining validated feature outputs
- Break down the transformations and benchmark warehouse memory versus intermediate materialization; verify The chosen design reduces spill while retaining validated feature outputs
- Use only increased timeout to solve memory pressure; no direct validation of The chosen design reduces spill while retaining validated feature outputs
- Turn off query-profile access; no direct validation of The chosen design reduces spill while retaining validated feature outputs
Explanation
Correct Answer
B:Break down the transformations and benchmark warehouse memory versus intermediate materialization. Verify: The chosen design reduces spill while retaining validated feature outputs.
Incorrect Answers
A:Required correction: Break down the transformations and benchmark warehouse memory versus intermediate materialization. Evidence: The chosen design reduces spill while retaining validated feature outputs.
C:Required correction: Break down the transformations and benchmark warehouse memory versus intermediate materialization. Evidence: The chosen design reduces spill while retaining validated feature outputs.
D:Required correction: Break down the transformations and benchmark warehouse memory versus intermediate materialization. Evidence: The chosen design reduces spill while retaining validated feature outputs.
Question 762
A system requires migration from a legacy lakehouse with existing Iceberg metadata and cannot lose externally managed snapshots The team proposes: Plan a supported Iceberg table catalog/metadata integration before connecting workloads. Which test outcome is the strongest direct evidence that this change works?
- An unrelated compute counter improves with no correctness check
- The warehouse login succeeds but the required query fails
- Snowflake sees the correct registered snapshot with its table schema and column data
- A metadata row exists while no relevant data is tested
Explanation
Correct Answer
C:Acceptance evidence: Snowflake sees the correct registered snapshot with its table schema and column data. Completing configuration alone is insufficient.
Incorrect Answers
A:This observation is insufficient. The stronger validation is: Snowflake sees the correct registered snapshot with its table schema and column data.
B:This observation is insufficient. The stronger validation is: Snowflake sees the correct registered snapshot with its table schema and column data.
D:This observation is insufficient. The stronger validation is: Snowflake sees the correct registered snapshot with its table schema and column data.
Question 763
An audit requires identifying which roles actually queried regulated data over a period The team proposes: Use supported access history and relevant audit views with authorized privileges. Which test outcome is the strongest direct evidence that this change works?
- A role name looks correct while the effective grants remain unchecked
- An administrator can run the query but a standard role remains unauthorized
- Audit evidence identifies access events and objects at supported granularity
- The policy is listed in metadata but denied accesses are never tested
Explanation
Correct Answer
C:Acceptance evidence: Audit evidence identifies access events and objects at supported granularity. Completing configuration alone is insufficient.
Incorrect Answers
A:This observation is insufficient. The stronger validation is: Audit evidence identifies access events and objects at supported granularity.
B:This observation is insufficient. The stronger validation is: Audit evidence identifies access events and objects at supported granularity.
D:This observation is insufficient. The stronger validation is: Audit evidence identifies access events and objects at supported granularity.
Question 764
A workload only needs newly inserted rows and would like to avoid change-tracking overhead from updates The team proposes: Evaluate an append-only stream where its semantics fit the source. Which test outcome is the strongest direct evidence that this change works?
- A job status is green without comparing source and target record counts
- The stream surfaces intended inserts without pretending to capture deleted rows
- One connector starts but no expected event arrives
- The staged file appears in LIST but no destination row is validated
Explanation
Correct Answer
B:Acceptance evidence: The stream surfaces intended inserts without pretending to capture deleted rows. Completing configuration alone is insufficient.
Incorrect Answers
A:This observation is insufficient. The stronger validation is: The stream surfaces intended inserts without pretending to capture deleted rows.
C:This observation is insufficient. The stronger validation is: The stream surfaces intended inserts without pretending to capture deleted rows.
D:This observation is insufficient. The stronger validation is: The stream surfaces intended inserts without pretending to capture deleted rows.
Question 765
A query searches for rare substrings in very large text fields using supported predicates Which unresolved finding would be a valid reason to block acceptance after the change?
- Change and rollback evidence is retained with the release
- The relevant users retain exactly their approved access
- Unoptimized text searches scan far more data than needed
- The team confirms: Supported substring lookups improve in measured tests
Explanation
Correct Answer
C:Unresolved failure: Unoptimized text searches scan far more data than needed. Correct and test it before acceptance.
Incorrect Answers
A:“Change and rollback evidence is retained with the release” describes a target or safeguard, not the release-blocking failure. The defect: Unoptimized text searches scan far more data than needed.
B:“The relevant users retain exactly their approved access” describes a target or safeguard, not the release-blocking failure. The defect: Unoptimized text searches scan far more data than needed.
D:This is a legitimate target or safeguard, not the unresolved failure described in the question: Unoptimized text searches scan far more data than needed.
Question 766
Storage cost estimates assume every warehouse duplicates all table data Select the THREE statements identifying the correct action, direct evidence, and unresolved release-blocking risk. (Choose THREE.)
- The capacity model counts shared table storage once and compute consumption separately
- Keep one warehouse for all users regardless of concurrency
- Block release while: The design double-counts storage merely because warehouses are isolated
- Distinguish centrally maintained compressed table storage from independent warehouse compute
- Copy every base table into a separate schema per warehouse
- Set MAX_CLUSTER_COUNT equal to stored terabytes
Explanation
Correct Answers
A:Acceptance evidence: The capacity model counts shared table storage once and compute consumption separately. Completing configuration alone is insufficient.
C:Unresolved failure: The design double-counts storage merely because warehouses are isolated. Correct and test it before acceptance.
D:Distinguish centrally maintained compressed table storage from independent warehouse compute. This addresses the stated condition without bypassing the required boundary.
Incorrect Answers
B:The problem: Storage cost estimates assume every warehouse duplicates all table data. The appropriate decision: Distinguish centrally maintained compressed table storage from independent warehouse compute.
E:The problem: Storage cost estimates assume every warehouse duplicates all table data. The appropriate decision: Distinguish centrally maintained compressed table storage from independent warehouse compute.
F:The problem: Storage cost estimates assume every warehouse duplicates all table data. The appropriate decision: Distinguish centrally maintained compressed table storage from independent warehouse compute.
Question 767
A user drops a table and promptly needs it restored while its data is still eligible Which unresolved finding would be a valid reason to block acceptance after the change?
- The team confirms: The original table is accessible again with its expected data
- Change and rollback evidence is retained with the release
- The object becomes unrecoverable because the team fails to use the supported window
- The relevant users retain exactly their approved access
Explanation
Correct Answer
C:Unresolved failure: The object becomes unrecoverable because the team fails to use the supported window. Correct and test it before acceptance.
Incorrect Answers
A:This is a legitimate target or safeguard, not the unresolved failure described in the question: The object becomes unrecoverable because the team fails to use the supported window.
B:“Change and rollback evidence is retained with the release” describes a target or safeguard, not the release-blocking failure. The defect: The object becomes unrecoverable because the team fails to use the supported window.
D:“The relevant users retain exactly their approved access” describes a target or safeguard, not the release-blocking failure. The defect: The object becomes unrecoverable because the team fails to use the supported window.
Question 768
A batch import must reject invalid records rather than silently write partial bad data Which technical action is most appropriate under this requirement?
- Set AUTO_RESUME to FALSE as error handling
- Choose an ON_ERROR behavior appropriate to the business contract and inspect load errors
- Disable COPY_HISTORY after each run
- Always force CONTINUE without checking invalid rows
Explanation
Correct Answer
B:Choose an ON_ERROR behavior appropriate to the business contract and inspect load errors. This addresses the stated condition without bypassing the required boundary.
Incorrect Answers
A:The problem: A batch import must reject invalid records rather than silently write partial bad data. The appropriate decision: Choose an ON_ERROR behavior appropriate to the business contract and inspect load errors.
C:The problem: A batch import must reject invalid records rather than silently write partial bad data. The appropriate decision: Choose an ON_ERROR behavior appropriate to the business contract and inspect load errors.
D:The problem: A batch import must reject invalid records rather than silently write partial bad data. The appropriate decision: Choose an ON_ERROR behavior appropriate to the business contract and inspect load errors.
Question 769
Security wants password strength and expiration rules applied without editing each application table The team proposes: Configure supported account-level or user-level password policies where relevant. Which test outcome is the strongest direct evidence that this change works?
- The policy is listed in metadata but denied accesses are never tested
- An administrator can run the query but a standard role remains unauthorized
- A role name looks correct while the effective grants remain unchecked
- Sign-in behavior follows documented password requirements
Explanation
Correct Answer
D:Acceptance evidence: Sign-in behavior follows documented password requirements. Completing configuration alone is insufficient.
Incorrect Answers
A:This observation is insufficient. The stronger validation is: Sign-in behavior follows documented password requirements.
B:This observation is insufficient. The stronger validation is: Sign-in behavior follows documented password requirements.
C:This observation is insufficient. The stronger validation is: Sign-in behavior follows documented password requirements.
Question 770
A JSON stage includes multiple variant event types and downstream needs only one event category Which technical action is most appropriate under this requirement?
- Use supported COPY selection or staged transformations to filter and map relevant records
- Force all staged data into a VARCHAR table without filtering
- Change query result format to remove unrelated events
- Use a network policy as a row-level ingestion filter
Explanation
Correct Answer
A:Use supported COPY selection or staged transformations to filter and map relevant records. This addresses the stated condition without bypassing the required boundary.
Incorrect Answers
B:The problem: A JSON stage includes multiple variant event types and downstream needs only one event category. The appropriate decision: Use supported COPY selection or staged transformations to filter and map relevant records.
C:The problem: A JSON stage includes multiple variant event types and downstream needs only one event category. The appropriate decision: Use supported COPY selection or staged transformations to filter and map relevant records.
D:The problem: A JSON stage includes multiple variant event types and downstream needs only one event category. The appropriate decision: Use supported COPY selection or staged transformations to filter and map relevant records.
SnowPro Core COF-C03 Practice Tests