Snowflake SnowPro Core COF-C03 Practice Test 35 - Prepaway
Practice Exams:

Pass Snowflake SnowPro Core COF-C03 Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers!

30 Days Free Updates, Instant Download!

Verified By Experts
Premium Bundle
$14.99
$74.99 $89.98

SnowPro Core COF-C03 Premium Bundle


  • Premium File 116 Questions & Answers. Last update: Nov 17, 2026
  • Study Guide 222 Pages

Snowflake SnowPro Core COF-C03 Practice Test 35

Question 340

During a change review for building event streams and supplier invoices at a facilities management group, the team finds: An integration user loses the ability to list cloud object files after a cloud IAM role policy change. Investigation shows: Snowflake storage integration remains defined but external cloud audit logs show denied object reads by the delegated principal. What is the most appropriate next Snowflake action?

  1. Define a scoped supported API integration with reviewed network rules and controlled invocation privileges.
  2. Set and audit approved client connection role, warehouse and namespace parameters with least privilege.
  3. Validate supported stage directory table refresh or notification configuration with file access privileges.
  4. Reconcile supported external cloud IAM trust and path permissions for the Snowflake storage integration without adding broad access.

Correct Answer: D

Explanation

A.

Define a scoped supported API integration with reviewed network rules and controlled invocation privileges. External network and API integrations should constrain egress destinations and execution privileges. For building event streams and supplier invoices at a facilities management group, Snowflake storage integration remains defined but external cloud audit logs show denied object reads by the delegated principal. Instead, reconcile supported external cloud IAM trust and path permissions for the Snowflake storage integration without adding broad access.

B.

Set and audit approved client connection role, warehouse and namespace parameters with least privilege. Snowflake drivers inherit or specify security and execution context separately from successful transport authentication. For building event streams and supplier invoices at a facilities management group, Snowflake storage integration remains defined but external cloud audit logs show denied object reads by the delegated principal. Instead, reconcile supported external cloud IAM trust and path permissions for the Snowflake storage integration without adding broad access.

C.

Validate supported stage directory table refresh or notification configuration with file access privileges. Directory metadata can require refresh or event configuration independent of files stored externally. For building event streams and supplier invoices at a facilities management group, Snowflake storage integration remains defined but external cloud audit logs show denied object reads by the delegated principal. Instead, reconcile supported external cloud IAM trust and path permissions for the Snowflake storage integration without adding broad access.

D.

Reconcile supported external cloud IAM trust and path permissions for the Snowflake storage integration without adding broad access. For building event streams and supplier invoices at a facilities management group, Snowflake storage integration remains defined but external cloud audit logs show denied object reads by the delegated principal. Snowflake connectivity spans internal roles and external provider identity policy. A confirming result should show that a controlled staged-file list succeeds for approved prefixes and a disallowed prefix remains inaccessible.

Question 341

An operations review of settlement reconciliation and fraud-monitoring datasets at a payments processor reports: A web analytics team doubles warehouse size repeatedly but interactive users still queue during simultaneous peak requests. The differentiating observation is: Query history reports queuing across many concurrent sessions rather than individual query resource starvation. Which proposed fix and test pair most directly addresses the failure?

  1. Benchmark a supported larger warehouse size and query optimization before adding clusters intended primarily for concurrency. — verification: The same query meets its resource target with less spill while total cost remains within budget.
  2. Review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely. — verification: A scheduled test query resumes the warehouse, succeeds and the warehouse later suspends again after inactivity.
  3. Inspect query planning, object statistics, metadata and query shape before changing compute sizing. — verification: A representative dashboard query compiles faster after a targeted approved change while result correctness is unchanged.
  4. Assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. — verification: Peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

Correct Answer: D

Explanation

A.

Benchmark a supported larger warehouse size and query optimization before adding clusters intended primarily for concurrency. Verify that the same query meets its resource target with less spill while total cost remains within budget. This action and its verification do not address the identified failure together. For settlement reconciliation and fraud-monitoring datasets at a payments processor, query history reports queuing across many concurrent sessions rather than individual query resource starvation. The appropriate correction is to assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. Confirm that peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

B.

Review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely. Verify that a scheduled test query resumes the warehouse, succeeds and the warehouse later suspends again after inactivity. This action and its verification do not address the identified failure together. For settlement reconciliation and fraud-monitoring datasets at a payments processor, query history reports queuing across many concurrent sessions rather than individual query resource starvation. The appropriate correction is to assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. Confirm that peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

C.

Inspect query planning, object statistics, metadata and query shape before changing compute sizing. Verify that a representative dashboard query compiles faster after a targeted approved change while result correctness is unchanged. This action and its verification do not address the identified failure together. For settlement reconciliation and fraud-monitoring datasets at a payments processor, query history reports queuing across many concurrent sessions rather than individual query resource starvation. The appropriate correction is to assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. Confirm that peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

D.

Assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. Verify that Peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters. For settlement reconciliation and fraud-monitoring datasets at a payments processor, query history reports queuing across many concurrent sessions rather than individual query resource starvation. Multi-cluster warehouses address concurrency pressure differently from single-query compute capacity.

Question 342

The Snowflake data engineering team at an engineering design firm investigating project time-series and contract expenditure models notes: An audit team adds generic sensitivity tags to tables but fails to identify columns actually containing new personal identifiers. Evidence: Data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. They plan: Apply supported sensitive-data classification and governed object tags with review and ownership of findings. Which technical principle justifies the correction?

  1. Dynamic masking can enforce column-value protection according to access context.
  2. Data lineage supports governed interpretation and change impact beyond final result values.
  3. Reading data and delegating object privileges are separate capabilities in Snowflake DAC and RBAC.
  4. Metadata tags provide governance value only when classification and lifecycle controls are reliable.

Correct Answer: D

Explanation

A.

Dynamic masking can enforce column-value protection according to access context. This choice does not explain the observed failure mechanism. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. The controlling principle is that metadata tags provide governance value only when classification and lifecycle controls are reliable.

B.

Data lineage supports governed interpretation and change impact beyond final result values. This choice does not explain the observed failure mechanism. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. The controlling principle is that metadata tags provide governance value only when classification and lifecycle controls are reliable.

C.

Reading data and delegating object privileges are separate capabilities in Snowflake DAC and RBAC. This choice does not explain the observed failure mechanism. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. The controlling principle is that metadata tags provide governance value only when classification and lifecycle controls are reliable.

D.

Metadata tags provide governance value only when classification and lifecycle controls are reliable. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. The documented change is to apply supported sensitive-data classification and governed object tags with review and ownership of findings. A successful test shows a seeded new sensitive column is detected, labeled appropriately and included in the audit inventory.

Question 343

For legal matter metrics and billing data marts at a regional law practice, engineers propose: Attach an approved dynamic data masking policy and verify role-aware evaluation for intended consumers. Their incident states: A reporting view exposes full national identifiers to a broad analyst role despite a requirement to show only redacted values. What evidence should be verified before implementing the change?

  1. A low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column.
  2. A controlled session can query another unit's case rows because row access policy predicates do not use the correct tenant mapping.
  3. Data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory.
  4. GRANTS inventory shows table SELECT is present but warehouse USAGE is missing from active role hierarchy.

Correct Answer: A

Explanation

A.

A low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column. Dynamic masking can enforce column-value protection according to access context. To correct it, attach an approved dynamic data masking policy and verify role-aware evaluation for intended consumers. Verification should establish that restricted roles receive redacted test values and a specifically authorized role receives only approved detail; the reported setting is: for legal matter metrics and billing data marts at a regional law practice, engineers propose: attach an approved dynamic data masking policy and verify role-aware evaluation for intended consumers.

B.

A controlled session can query another unit's case rows because row access policy predicates do not use the correct tenant mapping. Row-level security requires correct identity-to-row entitlement logic. For legal matter metrics and billing data marts at a regional law practice, a low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column. For this incident, attach an approved dynamic data masking policy and verify role-aware evaluation for intended consumers.

C.

Data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. Metadata tags provide governance value only when classification and lifecycle controls are reliable. For legal matter metrics and billing data marts at a regional law practice, a low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column. For this incident, attach an approved dynamic data masking policy and verify role-aware evaluation for intended consumers.

D.

GRANTS inventory shows table SELECT is present but warehouse USAGE is missing from active role hierarchy. Snowflake role authorization must cover both referenced data and chosen execution warehouse. For legal matter metrics and billing data marts at a regional law practice, a low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column. For this incident, attach an approved dynamic data masking policy and verify role-aware evaluation for intended consumers.

Question 344

The sales aggregates and supplier order tracking environment at a wholesale retailer has an incident: A finance warehouse auto-suspends overnight but scheduled morning jobs fail because the service cannot resume it. The difference-making observation is: Warehouse state is suspended, the scheduling role has usage permissions, and auto-resume is disabled for the chosen warehouse. Choose the most defensible corrective response.

  1. Assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency.
  2. Benchmark a supported larger warehouse size and query optimization before adding clusters intended primarily for concurrency.
  3. Review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely.
  4. Configure an appropriate supported notebook execution environment and approved warehouse access rather than granting broad table rights.

Correct Answer: C

Explanation

A.

Assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. This choice does not explain the observed failure mechanism. For sales aggregates and supplier order tracking at a wholesale retailer, warehouse state is suspended, the scheduling role has usage permissions, and auto-resume is disabled for the chosen warehouse. Instead, review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely.

B.

Benchmark a supported larger warehouse size and query optimization before adding clusters intended primarily for concurrency. This choice does not explain the observed failure mechanism. For sales aggregates and supplier order tracking at a wholesale retailer, warehouse state is suspended, the scheduling role has usage permissions, and auto-resume is disabled for the chosen warehouse. Instead, review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely.

C.

Review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely. For sales aggregates and supplier order tracking at a wholesale retailer, warehouse state is suspended, the scheduling role has usage permissions, and auto-resume is disabled for the chosen warehouse. Warehouse auto-suspend and auto-resume must be configured together for intermittent workloads. A confirming result should show that a scheduled test query resumes the warehouse, succeeds and the warehouse later suspends again after inactivity.

D.

Configure an appropriate supported notebook execution environment and approved warehouse access rather than granting broad table rights. This choice does not explain the observed failure mechanism. For sales aggregates and supplier order tracking at a wholesale retailer, warehouse state is suspended, the scheduling role has usage permissions, and auto-resume is disabled for the chosen warehouse. Instead, review warehouse auto-resume and approved execution role settings without disabling cost-saving auto-suspend entirely.

Question 345

To correct this issue at a hospitality chain involving booking data marts and occupancy forecasts: A team downloads millions of rows to a workstation to apply a complex Python transformation even though a supported Snowpark function could run near data. The team applies: Evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data. What outcome should close the incident?

  1. A permitted viewer sees only approved aggregates and cannot retrieve underlying restricted rows.
  2. A representative transformation completes without bulk workstation export and produces matching approved results.
  3. The seeded net-revenue query matches a reviewed SQL baseline across sample periods.
  4. An idle warehouse suspends, compute charges stop for idle time, and an authorized new query resumes access to unchanged table rows.

Correct Answer: B

Explanation

A.

A permitted viewer sees only approved aggregates and cannot retrieve underlying restricted rows. Application interfaces do not replace authorization checks at the data and execution boundary. For booking data marts and occupancy forecasts at a hospitality chain, network egress and local memory limits dominate processing while Snowflake compute is underutilized. A meaningful check confirms that a representative transformation completes without bulk workstation export and produces matching approved results.

B.

A representative transformation completes without bulk workstation export and produces matching approved results. For booking data marts and occupancy forecasts at a hospitality chain, network egress and local memory limits dominate processing while Snowflake compute is underutilized. Snowpark enables supported programming-language computations within the Snowflake processing environment. The corresponding fix is to evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data.

C.

The seeded net-revenue query matches a reviewed SQL baseline across sample periods. Natural language analytics accuracy depends on governed metric definitions and source semantics. For booking data marts and occupancy forecasts at a hospitality chain, network egress and local memory limits dominate processing while Snowflake compute is underutilized. A meaningful check confirms that a representative transformation completes without bulk workstation export and produces matching approved results.

D.

An idle warehouse suspends, compute charges stop for idle time, and an authorized new query resumes access to unchanged table rows. Snowflake separates independently scalable query compute from persistent managed table storage. For booking data marts and occupancy forecasts at a hospitality chain, network egress and local memory limits dominate processing while Snowflake compute is underutilized. A meaningful check confirms that a representative transformation completes without bulk workstation export and produces matching approved results.

Question 346

Administrators at a facilities management group plan: Assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency. They are addressing the following problem in building event streams and supplier invoices: A web analytics team doubles warehouse size repeatedly but interactive users still queue during simultaneous peak requests. Which test demonstrates resolution?

  1. The same query meets its resource target with less spill while total cost remains within budget.
  2. A scheduled test query resumes the warehouse, succeeds and the warehouse later suspends again after inactivity.
  3. An idle warehouse suspends, compute charges stop for idle time, and an authorized new query resumes access to unchanged table rows.
  4. Peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

Correct Answer: D

Explanation

A.

The same query meets its resource target with less spill while total cost remains within budget. Warehouse scale-up and multi-cluster scale-out address different performance bottlenecks. For building event streams and supplier invoices at a facilities management group, query history reports queuing across many concurrent sessions rather than individual query resource starvation. The outcome should establish that peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

B.

A scheduled test query resumes the warehouse, succeeds and the warehouse later suspends again after inactivity. Warehouse auto-suspend and auto-resume must be configured together for intermittent workloads. For building event streams and supplier invoices at a facilities management group, query history reports queuing across many concurrent sessions rather than individual query resource starvation. The outcome should establish that peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

C.

An idle warehouse suspends, compute charges stop for idle time, and an authorized new query resumes access to unchanged table rows. Snowflake separates independently scalable query compute from persistent managed table storage. For building event streams and supplier invoices at a facilities management group, query history reports queuing across many concurrent sessions rather than individual query resource starvation. The outcome should establish that peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters.

D.

Peak multiuser queries complete within agreed wait times without unnecessary idle warehouse clusters. For building event streams and supplier invoices at a facilities management group, query history reports queuing across many concurrent sessions rather than individual query resource starvation. Multi-cluster warehouses address concurrency pressure differently from single-query compute capacity. The documented change is to assess supported multi-cluster warehouse configuration, scaling policy and workload partitioning under realistic concurrency.

Question 347

A Snowflake analyst supporting application case metrics and protected citizen records at a benefits agency intends: Inspect query planning, object statistics, metadata and query shape before changing compute sizing. The reported condition is: A dashboard experiences long compilation times but engineers keep increasing its warehouse size without changing query planning. Identify the most decisive diagnostic record.

  1. Query History attributes most elapsed time to compilation rather than execution and warehouse resources remain underutilized.
  2. Query history shows overlapping workloads and queuing despite healthy underlying cloud storage availability.
  3. Consumption reports show storage bytes rising while warehouse execution credits remain relatively stable.
  4. Query history reports queuing across many concurrent sessions rather than individual query resource starvation.

Correct Answer: A

Explanation

A.

Query History attributes most elapsed time to compilation rather than execution and warehouse resources remain underutilized. Cloud services coordinate SQL compilation and metadata separately from virtual warehouse execution. To correct it, inspect query planning, object statistics, metadata and query shape before changing compute sizing. Verification should establish that a representative dashboard query compiles faster after a targeted approved change while result correctness is unchanged; the reported setting is: a snowflake analyst supporting application case metrics and protected citizen records at a benefits agency intends: inspect query planning, object statistics, metadata and query shape before changing compu.

B.

Query history shows overlapping workloads and queuing despite healthy underlying cloud storage availability. This choice does not explain the observed failure mechanism. For application case metrics and protected citizen records at a benefits agency, query History attributes most elapsed time to compilation rather than execution and warehouse resources remain underutilized. For this incident, inspect query planning, object statistics, metadata and query shape before changing compute sizing.

C.

Consumption reports show storage bytes rising while warehouse execution credits remain relatively stable. This choice does not explain the observed failure mechanism. For application case metrics and protected citizen records at a benefits agency, query History attributes most elapsed time to compilation rather than execution and warehouse resources remain underutilized. For this incident, inspect query planning, object statistics, metadata and query shape before changing compute sizing.

D.

Query history reports queuing across many concurrent sessions rather than individual query resource starvation. This choice does not explain the observed failure mechanism. For application case metrics and protected citizen records at a benefits agency, query History attributes most elapsed time to compilation rather than execution and warehouse resources remain underutilized. For this incident, inspect query planning, object statistics, metadata and query shape before changing compute sizing.

Question 348

The Snowflake data engineering team at an engineering design firm plans to: Apply supported sensitive-data classification and governed object tags with review and ownership of findings. The issue affecting project time-series and contract expenditure models is: An audit team adds generic sensitivity tags to tables but fails to identify columns actually containing new personal identifiers. Which finding best distinguishes this cause from other Snowflake failures?

  1. A low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column.
  2. The dashboard's numeric output exists but approved traceability tooling and source-to-target lineage are incomplete.
  3. A dashboard retains fewer days than the agreed investigation window even though account-level usage/access views can support longer analysis.
  4. Data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory.

Correct Answer: D

Explanation

A.

A low-privilege test session selects cleartext identifiers because no applicable masking policy protects the column. Dynamic masking can enforce column-value protection according to access context. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. Instead, apply supported sensitive-data classification and governed object tags with review and ownership of findings.

B.

The dashboard's numeric output exists but approved traceability tooling and source-to-target lineage are incomplete. Data lineage supports governed interpretation and change impact beyond final result values. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. Instead, apply supported sensitive-data classification and governed object tags with review and ownership of findings.

C.

A dashboard retains fewer days than the agreed investigation window even though account-level usage/access views can support longer analysis. Operational dashboards and governed audit history have different retention and access requirements. For project time-series and contract expenditure models at an engineering design firm, data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. Instead, apply supported sensitive-data classification and governed object tags with review and ownership of findings.

D.

Data profiling finds untagged sensitive fields and a generic table tag provides no reliable column inventory. Metadata tags provide governance value only when classification and lifecycle controls are reliable. To address this, apply supported sensitive-data classification and governed object tags with review and ownership of findings. Confirm that a seeded new sensitive column is detected, labeled appropriately and included in the audit inventory; the reported setting is: the snowflake data engineering team at an engineering design firm plans to: apply supported sensitive-data classification and governed object tags with review and ownership of findings.

Question 349

For customer risk analytics and branch transaction reports at a regional bank, engineers propose: Evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data. Their incident states: A team downloads millions of rows to a workstation to apply a complex Python transformation even though a supported Snowpark function could run near data. What evidence should be verified before implementing the change?

  1. Network egress and local memory limits dominate processing while Snowflake compute is underutilized.
  2. A restricted test viewer can retrieve sensitive rows through the application that should not be in their authorized scope.
  3. Generated SQL and semantic-model metrics select the wrong field even though source invoice rows are accurate.
  4. Benchmarks show ordinary SQL aggregation does not use the specialized memory-oriented workload features and cost rises.

Correct Answer: A

Explanation

A.

Network egress and local memory limits dominate processing while Snowflake compute is underutilized. Snowpark enables supported programming-language computations within the Snowflake processing environment. The corresponding fix is to evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data. A successful test should show that a representative transformation completes without bulk workstation export and produces matching approved results.

B.

A restricted test viewer can retrieve sensitive rows through the application that should not be in their authorized scope. Application interfaces do not replace authorization checks at the data and execution boundary. For customer risk analytics and branch transaction reports at a regional bank, network egress and local memory limits dominate processing while Snowflake compute is underutilized. To address this, evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data.

C.

Generated SQL and semantic-model metrics select the wrong field even though source invoice rows are accurate. Natural language analytics accuracy depends on governed metric definitions and source semantics. For customer risk analytics and branch transaction reports at a regional bank, network egress and local memory limits dominate processing while Snowflake compute is underutilized. To address this, evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data.

D.

Benchmarks show ordinary SQL aggregation does not use the specialized memory-oriented workload features and cost rises. Warehouse type should be chosen from actual supported workload requirements, not marketing terminology. For customer risk analytics and branch transaction reports at a regional bank, network egress and local memory limits dominate processing while Snowflake compute is underutilized. To address this, evaluate a supported Snowpark DataFrame or UDF design with governed package access and verify execution near Snowflake data.

Question 350

A Snowflake analyst supporting student analytics and research-grant reporting at a public university intends: Review join/sort memory needs, SQL shape and approved warehouse capacity before changing clustering keys. The reported condition is: A large sort query takes longer as datasets grow and analysts blame partition pruning without checking memory pressure. Identify the most decisive diagnostic record.

  1. Query Profile shows high local and remote spill even though the scan filters relevant partitions correctly.
  2. Query Profile shows much higher scanned micro-partition counts after the change with identical date filter intent.
  3. Query history contains query IDs but service account and tagging conventions do not map runs to owning workloads.
  4. Query Profile shows a data scan and row access predicates, not a metadata-only computation.

Correct Answer: A

Explanation

A.

Query Profile shows high local and remote spill even though the scan filters relevant partitions correctly. Query Profile distinguishes memory spill from excessive table scans and pruning inefficiency. The documented change is to review join/sort memory needs, SQL shape and approved warehouse capacity before changing clustering keys. The acceptance check is whether a representative query reduces spill while preserving results and stays within cost limits.

B.

Query Profile shows much higher scanned micro-partition counts after the change with identical date filter intent. This choice does not explain the observed failure mechanism. For student analytics and research-grant reporting at a public university, query Profile shows high local and remote spill even though the scan filters relevant partitions correctly. The better-matched action is to review join/sort memory needs, SQL shape and approved warehouse capacity before changing clustering keys.

C.

Query history contains query IDs but service account and tagging conventions do not map runs to owning workloads. This choice does not explain the observed failure mechanism. For student analytics and research-grant reporting at a public university, query Profile shows high local and remote spill even though the scan filters relevant partitions correctly. The better-matched action is to review join/sort memory needs, SQL shape and approved warehouse capacity before changing clustering keys.

D.

Query Profile shows a data scan and row access predicates, not a metadata-only computation. This choice does not explain the observed failure mechanism. For student analytics and research-grant reporting at a public university, query Profile shows high local and remote spill even though the scan filters relevant partitions correctly. The better-matched action is to review join/sort memory needs, SQL shape and approved warehouse capacity before changing clustering keys.

SnowPro Core COF-C03 Practice Tests