Question 100An employee leaves the company but a standalone Snowflake user remains active because identity deprovisioning does not sync. An engineer recommends: Integrate supported identity lifecycle provisioning or enforce an auditable user offboarding process. Which Snowflake architecture or data-governance principle best supports the recommendation? Identity revocation must propagate to each downstream service, including data platform accounts. Session role settings and connector support determine which granted privileges participate in access checks. Centralized identity policies must be applied to the relevant Snowflake user population rather than assumed from an IdP rollout. Object ownership carries administrative authority beyond ordinary read and write permissions. ExplanationCorrect Answer A:Identity revocation must propagate to each downstream service, including data platform accounts. This supports the stated action: Integrate supported identity lifecycle provisioning or enforce an auditable user offboarding process. Verification should show: The departed employee's old identity fails a login test and audit records show the disable operation. Incorrect Answers B:Session role settings and connector support determine which granted privileges participate in access checks. This does not establish the proposed intervention. The governing principle here is: Identity revocation must propagate to each downstream service, including data platform accounts. Apply it by: Integrate supported identity lifecycle provisioning or enforce an auditable user offboarding process. C:Centralized identity policies must be applied to the relevant Snowflake user population rather than assumed from an IdP rollout. This does not establish the proposed intervention. The governing principle here is: Identity revocation must propagate to each downstream service, including data platform accounts. Apply it by: Integrate supported identity lifecycle provisioning or enforce an auditable user offboarding process. D:Object ownership carries administrative authority beyond ordinary read and write permissions. This does not establish the proposed intervention. The governing principle here is: Identity revocation must propagate to each downstream service, including data platform accounts. Apply it by: Integrate supported identity lifecycle provisioning or enforce an auditable user offboarding process.
Question 101A producer publishes a secure dataset but the receiving team believes it can update the provider's source rows directly through the share. Which specific platform, security or business risk follows most directly if this issue remains unresolved? Wrong write-access assumptions can break collaboration processes and create data ownership disputes. Stale collaboration entitlements leave terminated partners with continued data access. Unplanned cross-region access can fail or violate location restrictions. An unsuitable consumer model can block onboarding or create unbudgeted provider charges. ExplanationCorrect Answer A:Wrong write-access assumptions can break collaboration processes and create data ownership disputes. Here the failing condition is: A producer publishes a secure dataset but the receiving team believes it can update the provider's source rows directly through the share. The specific correction is: Design a read-only shared consumption pattern and a separate authorized write-back workflow if needed. Verify through: A consumer SELECT succeeds while an attempted UPDATE to provider-owned source data is denied. Incorrect Answers B:Consumer access is controlled by explicit share and account grants that require lifecycle management. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Wrong write-access assumptions can break collaboration processes and create data ownership disputes. C:Data collaboration across locations has explicit service and governance requirements beyond same-region direct shares. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Wrong write-access assumptions can break collaboration processes and create data ownership disputes. D:Direct shares require supported consumer account arrangements and reader accounts have distinct provider responsibilities. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Wrong write-access assumptions can break collaboration processes and create data ownership disputes.
Question 102A team chains three transformation steps as separate scheduled tasks with overlapping schedules that sometimes execute out of order. After this response — Use supported task dependencies or a task graph with monitored predecessor-success conditions. — which result most directly proves successful correction? Task history proves the enrichment step runs after the source load and before the final publish. The corrected pipeline refreshes expected changed rows within target lag and cost bounds. A test approval step executes the required stored procedure while dependent output reconciles. A controlled update is consumed within retention and a stale-stream contingency is documented. ExplanationCorrect Answer A:Task history proves the enrichment step runs after the source load and before the final publish. This directly validates the action: Use supported task dependencies or a task graph with monitored predecessor-success conditions. The control reduces this exposure: Out-of-order task runs can produce incomplete or inconsistent data marts. Incorrect Answers B:Incremental materialization depends on supported query semantics and upstream change behavior. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: Task history proves the enrichment step runs after the source load and before the final publish. C:Dynamic tables primarily materialize declarative query transformations, not arbitrary procedural execution. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: Task history proves the enrichment step runs after the source load and before the final publish. D:A stream depends on source change tracking and cannot retain unconsumed changes indefinitely. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: Task history proves the enrichment step runs after the source load and before the final publish.
Question 103A vendor provides a JSON document per line but the loader treats every file as ordinary pipe-delimited text. What is the most appropriate Snowflake configuration or data-platform response? Correct the stage path or supported PATTERN and inspect matching file names before loading. Configure a compatible JSON file format and use VARIANT or explicit field extraction as appropriate. Create and test a CSV file format with the appropriate field delimiter and quoting settings. Create a supported external stage and authorize the bucket location before using COPY INTO. ExplanationCorrect Answer B:Configure a compatible JSON file format and use VARIANT or explicit field extraction as appropriate. This directly addresses the scenario: A vendor provides a JSON document per line but the loader treats every file as ordinary pipe-delimited text. Technical basis: Semi-structured JSON ingestion uses document-aware parsing distinct from CSV field splitting. Check success by confirming: A sample nested object loads with the correct path values and no truncated records. Incorrect Answers A:Stage path scoping and file selection determine which source objects enter bulk ingestion. That is relevant to a different intervention. For the stated situation, the required action is: Configure a compatible JSON file format and use VARIANT or explicit field extraction as appropriate. The governing reason is: Semi-structured JSON ingestion uses document-aware parsing distinct from CSV field splitting. C:Input parsing must reflect the real file's structure before mapping fields to target columns. That is relevant to a different intervention. For the stated situation, the required action is: Configure a compatible JSON file format and use VARIANT or explicit field extraction as appropriate. The governing reason is: Semi-structured JSON ingestion uses document-aware parsing distinct from CSV field splitting. D:Snowflake bulk loading reads data through defined stages with configured cloud storage access. That is relevant to a different intervention. For the stated situation, the required action is: Configure a compatible JSON file format and use VARIANT or explicit field extraction as appropriate. The governing reason is: Semi-structured JSON ingestion uses document-aware parsing distinct from CSV field splitting.
Question 104A university team publishes class performance statistics but a small subgroup query exposes one student's individual grade. What is the most appropriate Snowflake configuration or data-platform response? Test supported semi-structured data policy behavior and target the required sensitive nested fields. Inspect supported access history and query audit records with appropriately privileged monitoring roles. Apply an appropriate aggregation or privacy-preserving policy where supported for the released dataset. Define a supported row access policy that enforces authorized regions at the data boundary. ExplanationCorrect Answer C:Apply an appropriate aggregation or privacy-preserving policy where supported for the released dataset. This directly addresses the scenario: A university team publishes class performance statistics but a small subgroup query exposes one student's individual grade. Technical basis: Disclosure limitation may require restricting result granularity rather than only masking identifiers. Check success by confirming: A sufficiently aggregated report succeeds while a tiny cohort's identifying query is blocked or redacted. Incorrect Answers A:Nested data extraction and type-specific policy behavior must be validated, not inferred from the root object. That is relevant to a different intervention. For the stated situation, the required action is: Apply an appropriate aggregation or privacy-preserving policy where supported for the released dataset. The governing reason is: Disclosure limitation may require restricting result granularity rather than only masking identifiers. B:Effective governance needs observable use of data in addition to grants that authorize potential access. That is relevant to a different intervention. For the stated situation, the required action is: Apply an appropriate aggregation or privacy-preserving policy where supported for the released dataset. The governing reason is: Disclosure limitation may require restricting result granularity rather than only masking identifiers. D:Row access policies constrain visible rows independently of client application filters. That is relevant to a different intervention. For the stated situation, the required action is: Apply an appropriate aggregation or privacy-preserving policy where supported for the released dataset. The governing reason is: Disclosure limitation may require restricting result granularity rather than only masking identifiers.
Question 105A developer deploys a new reporting warehouse and assumes it already holds every file in the cache of an existing warehouse. After this response — Treat each warehouse-local data cache as distinct and benchmark workloads after warmup. — which result most directly proves successful correction? The new warehouse's cold and warm scan timings are measured separately. An instrumented run after resume identifies cold-cache scan time separately from result-cache behavior. A measured operating schedule balances cache-related latency and monitored credit spend. A controlled source update is reflected in a subsequent report with correct totals. ExplanationCorrect Answer A:The new warehouse's cold and warm scan timings are measured separately. This directly validates the action: Treat each warehouse-local data cache as distinct and benchmark workloads after warmup. The control reduces this exposure: Expecting cross-warehouse local cache sharing can lead to surprising latency and overbilling. Incorrect Answers B:Warehouse-local caching depends on active compute lifecycle and differs from reusable persisted query results. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The new warehouse's cold and warm scan timings are measured separately. C:Cache warmth is a performance trade-off with idle compute cost and depends on warehouse lifecycle. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The new warehouse's cold and warm scan timings are measured separately. D:Changes to underlying data can invalidate use of an otherwise reusable result set. This relates to a separate check and does not verify the specified repair. The relevant acceptance evidence is: The new warehouse's cold and warm scan timings are measured separately.
Question 106A sales analyst aggregates customer revenue but groups by a nonunique display name and merges unrelated customers. Which specific platform, security or business risk follows most directly if this issue remains unresolved? Grouping by unstable labels can materially misstate customer performance. Uncontrolled divide-by-zero behavior can break monthly executive reports. Cross-joining dimensions can multiply amounts and misstate accounting totals. Expensive self-joins can miss the reporting SLA and complicate correctness reviews. ExplanationCorrect Answer A:Grouping by unstable labels can materially misstate customer performance. Here the failing condition is: A sales analyst aggregates customer revenue but groups by a nonunique display name and merges unrelated customers. The specific correction is: Group by stable business keys and include display labels only with defined relationship semantics. Verify through: Two customers sharing a name retain separate totals in the corrected report. Incorrect Answers B:Analytic percentage formulas need explicit handling for zero and missing baselines. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Grouping by unstable labels can materially misstate customer performance. C:SQL relational joins must preserve intended cardinality and business grain. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Grouping by unstable labels can materially misstate customer performance. D:Window functions can express running aggregates without an unnecessary quadratic self-join. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Grouping by unstable labels can materially misstate customer performance.
Question 107A customer service team wants sentiment classification on stored ticket text but designs a costly external model API for simple supported analysis. An engineer recommends: Assess an appropriate Snowflake Cortex AI SQL function with data access and usage controls. Which Snowflake architecture or data-governance principle best supports the recommendation? Retrieval quality for unstructured knowledge requires indexed content and proper access controls, not file names alone. Model lifecycle governance separates experimental artifacts from controlled production inference versions. AI inference can incur additional metered usage and quality risk beyond ordinary SQL processing. Cortex AI capabilities can run supported AI inference close to governed data with distinct cost and permission requirements. ExplanationCorrect Answer D:Cortex AI capabilities can run supported AI inference close to governed data with distinct cost and permission requirements. This supports the stated action: Assess an appropriate Snowflake Cortex AI SQL function with data access and usage controls. Verification should show: Representative ticket labels pass a human-checked sample without unnecessary raw-data export. Incorrect Answers A:Retrieval quality for unstructured knowledge requires indexed content and proper access controls, not file names alone. This does not establish the proposed intervention. The governing principle here is: Cortex AI capabilities can run supported AI inference close to governed data with distinct cost and permission requirements. Apply it by: Assess an appropriate Snowflake Cortex AI SQL function with data access and usage controls. B:Model lifecycle governance separates experimental artifacts from controlled production inference versions. This does not establish the proposed intervention. The governing principle here is: Cortex AI capabilities can run supported AI inference close to governed data with distinct cost and permission requirements. Apply it by: Assess an appropriate Snowflake Cortex AI SQL function with data access and usage controls. C:AI inference can incur additional metered usage and quality risk beyond ordinary SQL processing. This does not establish the proposed intervention. The governing principle here is: Cortex AI capabilities can run supported AI inference close to governed data with distinct cost and permission requirements. Apply it by: Assess an appropriate Snowflake Cortex AI SQL function with data access and usage controls.
Question 108An operations group uses an economy scaling policy while expecting subsecond response for bursty business-critical dashboards. Which specific platform, security or business risk follows most directly if this issue remains unresolved? Excessive tolerance for queues can make real-time dashboards unusable during peak hours. Unbounded ad hoc processing can consume credits and delay essential reporting. Assuming one control covers every charge can leave the account budget unprotected. Scaling only one cluster can leave queues long and overprovision per-query execution. ExplanationCorrect Answer A:Excessive tolerance for queues can make real-time dashboards unusable during peak hours. Here the failing condition is: An operations group uses an economy scaling policy while expecting subsecond response for bursty business-critical dashboards. The specific correction is: Choose the cluster scaling policy that reflects required response time and validated query demand. Verify through: Burst tests keep dashboard queue latency within the agreed service objective under the chosen policy. Incorrect Answers B:Query execution limits and resource isolation bound the impact of runaway statements on shared compute. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Excessive tolerance for queues can make real-time dashboards unusable during peak hours. C:Resource monitors have defined scope and cannot be assumed to meter every type of Snowflake consumption. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Excessive tolerance for queues can make real-time dashboards unusable during peak hours. D:Adding clusters can address concurrency while increasing the size of one cluster chiefly changes per-query compute resources. That does not describe the principal exposure from the stated failure. The direct avoidable risk is: Excessive tolerance for queues can make real-time dashboards unusable during peak hours.
Question 109A warehouse receives Parquet with embedded column types but the project manually stringifies every field before downstream analysis. What is the most appropriate Snowflake configuration or data-platform response? Use an authorized external stage with scoped write permissions and permitted encryption settings. Evaluate supported column-name matching or a controlled explicit projection before inserting into the target. Benchmark supported file size and compression choices with practical stage and warehouse parallelism. Use supported Parquet loading and schema mapping appropriate to the intended target table. ExplanationCorrect Answer D:Use supported Parquet loading and schema mapping appropriate to the intended target table. This directly addresses the scenario: A warehouse receives Parquet with embedded column types but the project manually stringifies every field before downstream analysis. Technical basis: Columnar files carry structured type metadata that can simplify compatible ingestion. Check success by confirming: A representative sample preserves numeric and timestamp meanings in the target records. Incorrect Answers A:Unloading data to cloud storage must obey destination access and data handling policies. That is relevant to a different intervention. For the stated situation, the required action is: Use supported Parquet loading and schema mapping appropriate to the intended target table. The governing reason is: Columnar files carry structured type metadata that can simplify compatible ingestion. B:Schema drift requires deliberate field mapping rather than assuming stable column positions. That is relevant to a different intervention. For the stated situation, the required action is: Use supported Parquet loading and schema mapping appropriate to the intended target table. The governing reason is: Columnar files carry structured type metadata that can simplify compatible ingestion. C:Bulk ingest efficiency depends on file shape and parallel work distribution as well as warehouse size. That is relevant to a different intervention. For the stated situation, the required action is: Use supported Parquet loading and schema mapping appropriate to the intended target table. The governing reason is: Columnar files carry structured type metadata that can simplify compatible ingestion.
Question 110A support engineer retries the same previously loaded stage file with identical settings and assumes it always inserts duplicate rows. An engineer recommends: Review COPY load history and use deliberate force or replay policy only when reprocessing is authorized. Which Snowflake architecture or data-governance principle best supports the recommendation? Bulk load validation can identify format and conversion errors before production ingestion. Bulk ingest efficiency depends on file shape and parallel work distribution as well as warehouse size. Snowflake tracks previously loaded files for supported COPY duplicate prevention behavior. Snowflake bulk loading reads data through defined stages with configured cloud storage access. ExplanationCorrect Answer C:Snowflake tracks previously loaded files for supported COPY duplicate prevention behavior. This supports the stated action: Review COPY load history and use deliberate force or replay policy only when reprocessing is authorized. Verification should show: A standard retry creates no duplicate rows while a controlled intentional replay follows the documented process. Incorrect Answers A:Bulk load validation can identify format and conversion errors before production ingestion. This does not establish the proposed intervention. The governing principle here is: Snowflake tracks previously loaded files for supported COPY duplicate prevention behavior. Apply it by: Review COPY load history and use deliberate force or replay policy only when reprocessing is authorized. B:Bulk ingest efficiency depends on file shape and parallel work distribution as well as warehouse size. This does not establish the proposed intervention. The governing principle here is: Snowflake tracks previously loaded files for supported COPY duplicate prevention behavior. Apply it by: Review COPY load history and use deliberate force or replay policy only when reprocessing is authorized. D:Snowflake bulk loading reads data through defined stages with configured cloud storage access. This does not establish the proposed intervention. The governing principle here is: Snowflake tracks previously loaded files for supported COPY duplicate prevention behavior. Apply it by: Review COPY load history and use deliberate force or replay policy only when reprocessing is authorized.