Question 790A long-running transformation joins source tables before restricting to the day's changed records. An engineer proposes to push selective filters before expensive joins where semantically safe. Which explanation best describes why this is appropriate? Source evolution can silently change downstream mappings and expectations. Early reduction of rows limits intermediate join work when optimizer and predicates allow. Table lifecycle choices govern persistence, recovery overhead and maintenance behavior. Task graphs provide ordered execution and per-step operational visibility. ExplanationCorrect Answer B:Early reduction of rows limits intermediate join work when optimizer and predicates allow. This is testable: query Profile reports fewer scanned or joined rows for the same correct output. Incorrect Answers A:Source evolution can silently change downstream mappings and expectations. For the current requirement, the decisive mechanism is: Early reduction of rows limits intermediate join work when optimizer and predicates allow. The related action is: Push selective filters before expensive joins where semantically safe. C:Table lifecycle choices govern persistence, recovery overhead and maintenance behavior. For the current requirement, the decisive mechanism is: Early reduction of rows limits intermediate join work when optimizer and predicates allow. The related action is: Push selective filters before expensive joins where semantically safe. D:Task graphs provide ordered execution and per-step operational visibility. For the current requirement, the decisive mechanism is: Early reduction of rows limits intermediate join work when optimizer and predicates allow. The related action is: Push selective filters before expensive joins where semantically safe.
Question 791An upsert pipeline inserts duplicate customer IDs because it runs INSERT for every replayed change. After the team decides to use a correctly keyed MERGE with matched-update and not-matched-insert logic, which observation most directly validates the result? The top-ranked partition rows match a reviewed expected output. Exactly one latest record per customer is returned under a deterministic tie-breaker. A replay updates existing customers without adding duplicate identifiers. A three-record test yields the expected monotonic cumulative sum. ExplanationCorrect Answer C:A replay updates existing customers without adding duplicate identifiers. This matters because mERGE applies conditional DML using source-target match rules in one statement. Incorrect Answers A:QUALIFY filters results of window expressions without wrapping every query in a subquery. This is not the decisive validation for the stated requirement. The relevant evidence is: A replay updates existing customers without adding duplicate identifiers. B:A window ordering assigns ranks while preserving original row attributes. This is not the decisive validation for the stated requirement. The relevant evidence is: A replay updates existing customers without adding duplicate identifiers. D:Window order and frame rules determine which earlier values contribute to a running total. This is not the decisive validation for the stated requirement. The relevant evidence is: A replay updates existing customers without adding duplicate identifiers.
Question 792A test warehouse consumes unexpected credits overnight and no credit threshold has been configured. If the responsible team does not correct this condition, which outcome is the most relevant remaining risk? Latency is mistaken for an audit-record loss. The team cannot connect high-cost SQL to its executing identity. Unmonitored storage and other services exceed budget while warehouse quotas appear healthy. Idle workloads exceed their agreed compute budget undetected. ExplanationCorrect Answer D:Idle workloads exceed their agreed compute budget undetected. The direct prevention is to add an appropriate resource monitor with approved warehouse quota actions. Resource monitors help track supported warehouse credit usage and trigger notification or suspension actions. Incorrect Answers A:Some historical account usage views are delayed relative to live query monitoring. The immediate omission risk in this case is: Idle workloads exceed their agreed compute budget undetected. B:Account usage views provide historical activity records useful for account-level attribution. The immediate omission risk in this case is: Idle workloads exceed their agreed compute budget undetected. C:A monitor's coverage does not replace comprehensive billing analysis across products. The immediate omission risk in this case is: Idle workloads exceed their agreed compute budget undetected.
Question 793An operations group wants to enable search optimization on every column without determining which predicates users run. Which Snowflake configuration or operational action most directly addresses this situation? Analyze actual query predicates and apply optimization only to justified targets. Investigate multi-cluster scale-out for queue-bound queries before adding acceleration. Evaluate Query Acceleration Service for supported queries and measured workload behavior. Confirm eligible selective predicates and expected benefit before enabling search optimization. ExplanationCorrect Answer A:Targeted access paths have maintenance cost and predicate-support constraints. The matching action is to analyze actual query predicates and apply optimization only to justified targets. Selected columns show improved lookup latency at acceptable maintenance cost. Incorrect Answers B:Query acceleration targets eligible execution work, not warehouse admission contention. The operative distinction here is: Targeted access paths have maintenance cost and predicate-support constraints. The targeted change is: Analyze actual query predicates and apply optimization only to justified targets. C:Query Acceleration Service can offload eligible portions of a query to additional compute. The operative distinction here is: Targeted access paths have maintenance cost and predicate-support constraints. The targeted change is: Analyze actual query predicates and apply optimization only to justified targets. D:Search optimization is not a general replacement for full-table scan processing. The operative distinction here is: Targeted access paths have maintenance cost and predicate-support constraints. The targeted change is: Analyze actual query predicates and apply optimization only to justified targets.
Question 794A support analyst needs to create application users but should not administer warehouses. An engineer proposes to delegate approved user management through a suitably scoped USERADMIN or custom role. Which explanation best describes why this is appropriate? ACCOUNTADMIN is a powerful administrative role and should not be an application default. Creating principals is different from modifying compute or data objects. Snowflake separates security administration from data and warehouse object administration. Organization scope and one account's administrative scope are distinct security boundaries. ExplanationCorrect Answer B:Creating principals is different from modifying compute or data objects. This is testable: the designated operator creates a test user but cannot resize production warehouses. Incorrect Answers A:ACCOUNTADMIN is a powerful administrative role and should not be an application default. For the current requirement, the decisive mechanism is: Creating principals is different from modifying compute or data objects. The related action is: Delegate approved user management through a suitably scoped USERADMIN or custom role. C:Snowflake separates security administration from data and warehouse object administration. For the current requirement, the decisive mechanism is: Creating principals is different from modifying compute or data objects. The related action is: Delegate approved user management through a suitably scoped USERADMIN or custom role. D:Organization scope and one account's administrative scope are distinct security boundaries. For the current requirement, the decisive mechanism is: Creating principals is different from modifying compute or data objects. The related action is: Delegate approved user management through a suitably scoped USERADMIN or custom role.
Question 795A warehouse has MAX_CLUSTER_COUNT set to one and engineers expect it to scale out to five. After the team decides to raise max cluster count within licensed, cost-approved limits and observe scale behavior, which observation most directly validates the result? Warehouse Load History shows fewer extra clusters while waits stay within policy. Peak queued load falls while median execution time per query stays similar. The warehouse adds a second cluster during a controlled concurrency test. Peak dashboard requests show the intended queue-latency reduction. ExplanationCorrect Answer C:The warehouse adds a second cluster during a controlled concurrency test. This matters because the maximum cluster count is a hard upper bound on simultaneous warehouse clusters. Incorrect Answers A:Economy waits longer before provisioning another cluster to favor utilization. This is not the decisive validation for the stated requirement. The relevant evidence is: The warehouse adds a second cluster during a controlled concurrency test. B:Additional clusters provide concurrency capacity without requiring faster individual queries. This is not the decisive validation for the stated requirement. The relevant evidence is: The warehouse adds a second cluster during a controlled concurrency test. D:Standard responds more aggressively to workload queue pressure. This is not the decisive validation for the stated requirement. The relevant evidence is: The warehouse adds a second cluster during a controlled concurrency test.
Question 796A team wants declarative derived results that refresh to an assigned staleness target without manually recreating a table. If the responsible team does not correct this condition, which outcome is the most relevant remaining risk? A recoverable accidental drop is treated as permanent data loss. The design either pays for unneeded protection or promises unsupported recovery. A proprietary-only table design blocks planned cross-engine access. A manually rebuilt table becomes stale between runs. ExplanationCorrect Answer D:A manually rebuilt table becomes stale between runs. The direct prevention is to design a dynamic table with a supported target lag and refresh configuration. Dynamic tables maintain query-defined results through managed refresh scheduling. Incorrect Answers A:Permanent tables can retain historical versions for supported recovery operations. The immediate omission risk in this case is: A manually rebuilt table becomes stale between runs. B:Transient tables omit Fail-safe and have more limited recovery than permanent tables. The immediate omission risk in this case is: A manually rebuilt table becomes stale between runs. C:Iceberg provides an open table format with governed catalog integration across engines. The immediate omission risk in this case is: A manually rebuilt table becomes stale between runs.
Question 797A troubleshooting script assumes a session variable persists across independent new connections. Which Snowflake configuration or operational action most directly addresses this situation? Initialize required session variables in each new connection before executing dependent SQL. Bind the workload to the approved warehouse explicitly in its session configuration. Set and verify the session's authorized primary role before execution. Inspect session, user and account parameter precedence before changing account defaults. ExplanationCorrect Answer A:Session variables are scoped to the connection rather than durable account metadata. The matching action is to initialize required session variables in each new connection before executing dependent SQL. A newly opened session runs correctly after the variable initialization step. Incorrect Answers B:Current warehouse context determines where eligible SQL statements execute. The operative distinction here is: Session variables are scoped to the connection rather than durable account metadata. The targeted change is: Initialize required session variables in each new connection before executing dependent SQL. C:The current role determines object ownership on CREATE and contributes effective privileges. The operative distinction here is: Session variables are scoped to the connection rather than durable account metadata. The targeted change is: Initialize required session variables in each new connection before executing dependent SQL. D:More specific supported parameter scopes can override broader account values. The operative distinction here is: Session variables are scoped to the connection rather than durable account metadata. The targeted change is: Initialize required session variables in each new connection before executing dependent SQL.
Question 798An operator wants to retry a corrupted batch but cannot tell what COPY loaded yesterday. An engineer proposes to review load history, staged file metadata and target row counts before choosing a reload. Which explanation best describes why this is appropriate? A SELECT transform during COPY can align and cast source expressions into target columns. Snowflake records file-loading outcomes that can support safe reconciliation. Pre-load validation surfaces parsing and column conversion problems. COPY error-handling options determine whether loading aborts, skips or continues. ExplanationCorrect Answer B:Snowflake records file-loading outcomes that can support safe reconciliation. This is testable: the restart plan names exactly which files need correction and reprocessing. Incorrect Answers A:A SELECT transform during COPY can align and cast source expressions into target columns. For the current requirement, the decisive mechanism is: Snowflake records file-loading outcomes that can support safe reconciliation. The related action is: Review load history, staged file metadata and target row counts before choosing a reload. C:Pre-load validation surfaces parsing and column conversion problems. For the current requirement, the decisive mechanism is: Snowflake records file-loading outcomes that can support safe reconciliation. The related action is: Review load history, staged file metadata and target row counts before choosing a reload. D:COPY error-handling options determine whether loading aborts, skips or continues. For the current requirement, the decisive mechanism is: Snowflake records file-loading outcomes that can support safe reconciliation. The related action is: Review load history, staged file metadata and target row counts before choosing a reload.
Question 799An idle ad-hoc warehouse restarts every few seconds and users report high minimum billing overhead. After the team decides to tune auto-suspend against session burst patterns instead of suspending between every query, which observation most directly validates the result? Before/after query history differentiates in-flight jobs from queries started after resize. A controlled run succeeds from the suspended state without manual intervention. Startup counts and billable time fall after the revised threshold. Reported credits reconcile with warehouse usage history rather than an all-month estimate. ExplanationCorrect Answer C:Startup counts and billable time fall after the revised threshold. This matters because too-short suspend windows cause repeated startups with minimum usage charges and cache loss. Incorrect Answers A:New capacity is available for subsequently started or queued work rather than rewriting already-running computation. This is not the decisive validation for the stated requirement. The relevant evidence is: Startup counts and billable time fall after the revised threshold. B:Suspended warehouses require a supported start path before executing compute work. This is not the decisive validation for the stated requirement. The relevant evidence is: Startup counts and billable time fall after the revised threshold. D:Per-second usage after the startup minimum depends on running duration and resource size. This is not the decisive validation for the stated requirement. The relevant evidence is: Startup counts and billable time fall after the revised threshold.
Question 800A nested numeric value is used in arithmetic but arrives as a VARIANT string and comparison behaves unexpectedly. If the responsible team does not correct this condition, which outcome is the most relevant remaining risk? Ad hoc string parsing yields brittle or missing nested values. Nested items remain inaccessible to relational aggregation. Queries return NULL for fields that exist under different key spelling. Lexical comparison produces incorrect monetary totals. ExplanationCorrect Answer D:Lexical comparison produces incorrect monetary totals. The direct prevention is to apply explicit TRY_CAST or appropriate conversion to a supported numeric type. Typed conversion clarifies semantics for numeric predicates and calculations. Incorrect Answers A:VARIANT preserves semi-structured types and supports documented element-path access. The immediate omission risk in this case is: Lexical comparison produces incorrect monetary totals. B:FLATTEN expands array or object elements to relational rows for downstream SQL. The immediate omission risk in this case is: Lexical comparison produces incorrect monetary totals. C:Semi-structured element paths follow JSON key semantics independently of SQL identifier rules. The immediate omission risk in this case is: Lexical comparison produces incorrect monetary totals.