Snowflake SnowPro Core Practice Test 28
Question 270
For an original SnowPro Core educational scenario, consider: A developer copies a large Snowflake table into local pandas solely to perform a supported filter and aggregation. Select TWO complementary steps: the targeted correction and its strongest prechange validation.
- Check: Reproduce creation, querying and session close using two independent connections.
- Assume every DataFrame method automatically downloads the whole table to the workstation.
- Check: Inspect execution history before and after actions such as collect or write and compare data transfer size.
- Change: Express supported transformations with Snowpark DataFrames and materialize results using an execution action on Snowflake compute.
- Change: Design a supported Snowflake Iceberg table with its external cloud data and metadata location and appropriate catalog configuration.
Explanation
Correct Answer
C:Snowpark DataFrame transformations are generally lazily composed into executable plans and can avoid pulling full datasets into a local client. Required intervention: Express supported transformations with Snowpark DataFrames and materialize results using an execution action on Snowflake compute.
D:Snowpark DataFrame transformations are generally lazily composed into executable plans and can avoid pulling full datasets into a local client. Evidence: Inspect execution history before and after actions such as collect or write and compare data transfer size.
Incorrect Answers
A:Check: Reproduce creation, querying and session close using two independent connections. A developer copies a large Snowflake table into local pandas solely to perform a supported filter and aggregation.
B:For this select two assessment, the shortcut changes or bypasses the necessary correctness, governance or operational boundary rather than correcting it. Snowpark DataFrame transformations are generally lazily composed into executable plans and can avoid pulling full datasets into a local client.
E:Change: Design a supported Snowflake Iceberg table with its external cloud data and metadata location and appropriate catalog configuration. Snowpark DataFrame transformations are generally lazily composed into executable plans and can avoid pulling full datasets into a local client.
Question 271
A Snowflake engineer is examining this platform situation: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API. A proposed shortcut is: Mark all Cortex-generated text as verified business records without checking its content. What should the Snowflake engineer recommend as the narrowly scoped correction? Select ONE best answer.
- Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation.
- Mark all Cortex-generated text as verified business records without checking its content.
- Configure and evaluate an appropriate multi-cluster warehouse and scaling policy to absorb concurrent sessions.
- Use permanent tables for durable recovery requirements and transient tables for suitable short-lived data with reduced recovery guarantees.
Explanation
Correct Answer
A:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. Applied case: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API.
Incorrect Answers
B:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. For this best change assessment, the shortcut changes or bypasses the necessary correctness, governance or operational boundary rather than correcting it.
C:Configure and evaluate an appropriate multi-cluster warehouse and scaling policy to absorb concurrent sessions. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
D:This option addresses a different Snowflake technical topic (Permanent versus transient table retention) rather than the documented Snowflake Cortex AI SQL functions decision.
Question 272
A Snowflake implementation review finds: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API. The team proposes: Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation. Which platform principle most strongly supports the recommended response? Select ONE best answer.
- Snowflake's legacy notebook experience is being superseded by Notebooks in Workspaces; availability and workflows depend on account rollout and current documentation.
- Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
- Snowflake-managed cloud services coordinate metadata, access decisions and query optimization independently of warehouse compute.
- Distinct Snowflake warehouses have independent compute resources, so separating workloads reduces resource contention without copying database storage.
Explanation
Correct Answer
B:Demonstrable success requires: Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth. Applied case: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API.
Incorrect Answers
A:This option addresses a different Snowflake technical topic (Notebooks in Workspaces versus legacy notebooks) rather than the documented Snowflake Cortex AI SQL functions decision.
C:Snowflake-managed cloud services coordinate metadata, access decisions and query optimization independently of warehouse compute. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
D:Distinct Snowflake warehouses have independent compute resources, so separating workloads reduces resource contention without copying database storage. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
Question 273
The data platform team has not yet approved changes. Current evidence indicates: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API. Which evidence best tests the technical diagnosis before approving a disruptive change? Select ONE best answer.
- Test permissions, sample outputs, cost consumption, sensitive-data handling and false assertions against known cases.
- Inspect execution history before and after actions such as collect or write and compare data transfer size.
- Validate path selection, output row counts, INDEX and VALUE on arrays including empty and missing cases.
- Verify table type, historical retention expiration and documented support/recovery process before planning restoration.
Explanation
Correct Answer
A:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. Confirm afterward that: Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth.
Incorrect Answers
B:This option addresses a different Snowflake technical topic (Snowpark DataFrame execution) rather than the documented Snowflake Cortex AI SQL functions decision.
C:Validate path selection, output row counts, INDEX and VALUE on arrays including empty and missing cases. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
D:This option addresses a different Snowflake technical topic (Fail-safe purpose and operational limits) rather than the documented Snowflake Cortex AI SQL functions decision.
Question 274
A candidate correction has been tested and the platform team requires a post-change acceptance check. Original problem: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API. The service owner needs positive and negative results and evidence that legitimate data workflows still function. Which acceptance result demonstrates correct platform behavior without disrupting authorized workflows? Select ONE best answer.
- Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth.
- Operational lookups and constraints behave as required while bulk analytics use fit-for-purpose storage.
- Only authorized cross-boundary data is accessible and recovery assumptions match configured replication behavior.
- Cost analysis and architecture diagrams identify cloud services and warehouse activity as distinct components.
Explanation
Correct Answer
A:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. The result must be validated using the documented dataset or access context.
Incorrect Answers
B:This option addresses a different Snowflake technical topic (Hybrid tables and transactional constraints) rather than the documented Snowflake Cortex AI SQL functions decision.
C:Only authorized cross-boundary data is accessible and recovery assumptions match configured replication behavior. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
D:Cost analysis and architecture diagrams identify cloud services and warehouse activity as distinct components. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
Question 275
An implementation sign-off request identifies the following problem: A support team wants to summarize ticket text stored in governed Snowflake tables without exporting all records to a third-party API. Which paired technical change and validation should close the implementation review? Select ONE best answer.
- Change: Evaluate Snowpark-optimized warehouse configurations only for supported memory-intensive Snowpark operations with measured benefit. Verify: Compare the job's memory failures, execution time, supported workload requirements and credits on both warehouse types.
- Change: Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation. Verify: Test permissions, sample outputs, cost consumption, sensitive-data handling and false assertions against known cases.
- Change: Set workload-suitable AUTO_SUSPEND and AUTO_RESUME values and validate startup behavior and billing. Verify: Review warehouse configuration, idle duration and metered credit history around overnight gaps.
- Change: Use LATERAL FLATTEN against the appropriate VARIANT array and select the expanded VALUE fields. Verify: Validate path selection, output row counts, INDEX and VALUE on arrays including empty and missing cases.
Explanation
Correct Answer
B:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. The change is successful only when: Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth.
Incorrect Answers
A:Change: Evaluate Snowpark-optimized warehouse configurations only for supported memory-intensive Snowpark operations with measured benefit. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
C:Change: Set workload-suitable AUTO_SUSPEND and AUTO_RESUME values and validate startup behavior and billing.
D:Change: Use LATERAL FLATTEN against the appropriate VARIANT array and select the expanded VALUE fields. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
Question 276
A Snowflake engineer is examining this platform situation: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation. A proposed shortcut is: Mark all Cortex-generated text as verified business records without checking its content. What should the Snowflake engineer recommend as the narrowly scoped correction? Select ONE best answer.
- Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation.
- Check the current account edition and feature-specific entitlement before promising governance, recovery or security behavior.
- Mark all Cortex-generated text as verified business records without checking its content.
- Design a supported Snowflake Iceberg table with its external cloud data and metadata location and appropriate catalog configuration.
Explanation
Correct Answer
A:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. Applied case: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation.
Incorrect Answers
B:This option addresses a different Snowflake technical topic (Snowflake editions and feature entitlements) rather than the documented Snowflake Cortex AI SQL functions decision.
C:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. Use this instead: Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation.
D:This option addresses a different Snowflake technical topic (Apache Iceberg table architecture) rather than the documented Snowflake Cortex AI SQL functions decision.
Question 277
A Snowflake implementation review finds: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation. The team proposes: Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation. Which platform principle most strongly supports the recommended response? Select ONE best answer.
- Semi-structured paths and types must match document structure; missing keys and invalid casts need intentional treatment.
- Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
- Temporary tables exist only for their creating session and are removed when the session ends.
- Information Schema and Snowflake shared-account usage views provide different metadata scopes and refresh characteristics.
Explanation
Correct Answer
B:Demonstrable success requires: Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth. Applied case: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation.
Incorrect Answers
A:This option addresses a different Snowflake technical topic (Semi-structured path navigation) rather than the documented Snowflake Cortex AI SQL functions decision.
C:This option addresses a different Snowflake technical topic (Session-scoped temporary tables) rather than the documented Snowflake Cortex AI SQL functions decision.
D:Information Schema and Snowflake shared-account usage views provide different metadata scopes and refresh characteristics. A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation.
Question 278
The data platform team has not yet approved changes. Current evidence indicates: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation. Which evidence best tests the technical diagnosis before approving a disruptive change? Select ONE best answer.
- Compare QUERY_HISTORY warehouse identifiers, queued time and load before and after changing session warehouse assignments.
- Test permissions, sample outputs, cost consumption, sensitive-data handling and false assertions against known cases.
- Inspect refresh history, dependencies, actual lag and warehouse or serverless resource configuration.
- Create a controlled clone, compare initial storage and query results, then modify only the clone.
Explanation
Correct Answer
B:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. This check tests the actual data, compute or access boundary before an operationally risky change.
Incorrect Answers
A:Compare QUERY_HISTORY warehouse identifiers, queued time and load before and after changing session warehouse assignments. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
C:Inspect refresh history, dependencies, actual lag and warehouse or serverless resource configuration. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
D:Create a controlled clone, compare initial storage and query results, then modify only the clone. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
Question 279
A candidate correction has been tested and the platform team requires a post-change acceptance check. Original problem: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation. The service owner needs positive and negative results and evidence that legitimate data workflows still function. Which acceptance result demonstrates correct platform behavior without disrupting authorized workflows? Select ONE best answer.
- Downstream work starts only after its required predecessor succeeds and failures are visible to operators.
- File inventory and authorized downstream processing work without misrepresenting staged objects as parsed records.
- Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth.
- The targeted memory-intensive job succeeds economically without changing unrelated dashboard infrastructure.
Explanation
Correct Answer
C:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. The approved implementation step was: Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation.
Incorrect Answers
A:Downstream work starts only after its required predecessor succeeds and failures are visible to operators. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
B:File inventory and authorized downstream processing work without misrepresenting staged objects as parsed records. A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation.
D:The targeted memory-intensive job succeeds economically without changing unrelated dashboard infrastructure. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
Question 280
For an original SnowPro Core educational scenario, consider: A data steward assumes generative AI responses are guaranteed accurate and safe to store as audited facts without evaluation. Select TWO complementary steps: the targeted correction and its strongest prechange validation.
- Check: Test permissions, sample outputs, cost consumption, sensitive-data handling and false assertions against known cases.
- Check: Inspect refresh history, dependencies, actual lag and warehouse or serverless resource configuration.
- Mark all Cortex-generated text as verified business records without checking its content.
- Change: Set explicit current database and schema or use fully qualified identifiers for deployment-sensitive operations.
- Change: Evaluate supported Snowflake Cortex AI functions for text tasks with role access, privacy controls, cost monitoring and output validation.
Explanation
Correct Answer
A:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. Acceptance: Approved use cases generate useful drafts without bypassing classification or uncritically treating model output as factual truth.
E:Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism. Evidence: Test permissions, sample outputs, cost consumption, sensitive-data handling and false assertions against known cases.
Incorrect Answers
B:Check: Inspect refresh history, dependencies, actual lag and warehouse or serverless resource configuration. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
C:For this select two assessment, the shortcut changes or bypasses the necessary correctness, governance or operational boundary rather than correcting it. Managed AI functions can operate within supported Snowflake services, but model output requires governance and validation rather than assumption of determinism.
D:This option addresses a different Snowflake technical topic (Database, schema and object hierarchy) rather than the documented Snowflake Cortex AI SQL functions decision.
