Pass Snowflake SnowPro Core Exam in First Attempt Easily
Real Snowflake SnowPro Core Exam Questions, Accurate & Verified Answers As Experienced in the Actual Test!

Verified by experts
3 products

You save $69.98

SnowPro Core Premium Bundle

  • Premium File 703 Questions & Answers
  • Last Update: Nov 22, 2026
  • Training Course 92 Lectures
  • Study Guide 413 Pages
$79.99 $149.97

Purchase Individually

  • Premium File

    703 Questions & Answers
    Last Update: Nov 22, 2026

    $76.99
    $69.99
  • Training Course

    92 Lectures

    $43.99
    $39.99
  • Study Guide

    413 Pages

    $43.99
    $39.99

SnowPro Core Practice Test 37

Question 360

The Snowflake implementation fails its pilot test. A customer table has historical duplicate customer IDs and joining it to orders unexpectedly multiplies invoice totals. Which evidence should be inspected first to isolate the likely cause?

  1. Inspect join type, row counts, missing keys and post-join WHERE filters.
  2. Inspect source values, conversion expressions and invalid-value count.
  3. Inspect join operator row counts, uniqueness constraints in data and sample matched keys.
  4. Review query result reuse status, query ID, data changes and warehouse cache conditions.

Explanation

Correct Answer

C:

Inspect join operator row counts, uniqueness constraints in data and sample matched keys. These checks diagnose Join explosion in query profile. Correct response: Fix join grain by constraining the customer version or deduplicating eligible keys before joining. Expected observation: The output cardinality and invoice sum match a controlled baseline rather than the exploded join.

Incorrect Answers

A:

Inspect join type, row counts, missing keys and post-join WHERE filters. Those checks would investigate LEFT JOIN preserves unmatched orders, not the current Join explosion in query profile failure. The targeted diagnostic here is: Inspect join operator row counts, uniqueness constraints in data and sample matched keys.

B:

Inspect source values, conversion expressions and invalid-value count. Those checks would investigate TRY_TO_NUMBER safe type conversion, not the current Join explosion in query profile failure. The targeted diagnostic here is: Inspect join operator row counts, uniqueness constraints in data and sample matched keys.

D:

Review query result reuse status, query ID, data changes and warehouse cache conditions. Those checks would investigate Result cache reuse versus warehouse local cache, not the current Join explosion in query profile failure. The targeted diagnostic here is: Inspect join operator row counts, uniqueness constraints in data and sample matched keys.

Question 361

A Snowflake implementation team encounters this specific requirement: An analyst wraps every event timestamp in a string function in a WHERE clause and a large table scans nearly every partition. Which action best addresses it?

  1. Use a LEFT OUTER JOIN from orders to the customer lookup while explicitly handling missing dimension rows.
  2. Evaluate a supported materialized view holding the frequently reused aggregate after checking maintenance trade-offs.
  3. Separate persisted query-result reuse from warehouse local disk-cache effects using controlled comparisons.
  4. Rewrite the filter to use a selective supported range predicate on the stored timestamp when semantically equivalent.

Explanation

Correct Answer

D:

Rewrite the filter to use a selective supported range predicate on the stored timestamp when semantically equivalent. This resolves Poor pruning from expression filters. Underlying reason: Predicate expression form and underlying micro-partition metadata influence pruning opportunities. Verify with: A matched query returns the same records while scanning substantially fewer eligible partitions. Boundary: Not all time-zone or type conversions are safe to remove without checking semantics.

Incorrect Answers

A:

Use a LEFT OUTER JOIN from orders to the customer lookup while explicitly handling missing dimension rows. That is suitable for a different requirement, LEFT JOIN preserves unmatched orders, with this separate condition: A sales report loses orders whose optional customer dimension record has not arrived yet.

B:

Evaluate a supported materialized view holding the frequently reused aggregate after checking maintenance trade-offs. That is suitable for a different requirement, Materialized view for repetitive aggregation, with this separate condition: Many dashboards repeatedly calculate the same supported grouped daily sales total and base-table scans dominate execution cost.

C:

Separate persisted query-result reuse from warehouse local disk-cache effects using controlled comparisons. That is suitable for a different requirement, Result cache reuse versus warehouse local cache, with this separate condition: A report returns nearly instantly on a second identical run and engineers claim a warehouse resize caused the improvement.

Question 362

After implementing a Snowflake control, the reviewer must prove the business requirement was met. An analyst wraps every event timestamp in a string function in a WHERE clause and a large table scans nearly every partition. Which result is the most direct acceptance evidence?

  1. A sample query extracts the precise city field while preserving other JSON properties.
  2. Valid strings convert to numbers and unsupported strings produce controlled NULL or exception records.
  3. A matched query returns the same records while scanning substantially fewer eligible partitions.
  4. Each chosen service improves its targeted workload without unjustified duplicate maintenance cost.

Explanation

Correct Answer

C:

A matched query returns the same records while scanning substantially fewer eligible partitions. This directly demonstrates success for Poor pruning from expression filters. Implementation: Rewrite the filter to use a selective supported range predicate on the stored timestamp when semantically equivalent. Mechanism: Predicate expression form and underlying micro-partition metadata influence pruning opportunities.

Incorrect Answers

A:

A sample query extracts the precise city field while preserving other JSON properties. This outcome is evidence for the separate VARIANT nested path query case, not the described Poor pruning from expression filters requirement. For the current context, verify: A matched query returns the same records while scanning substantially fewer eligible partitions.

B:

Valid strings convert to numbers and unsupported strings produce controlled NULL or exception records. This outcome is evidence for the separate TRY_TO_NUMBER safe type conversion case, not the described Poor pruning from expression filters requirement. For the current context, verify: A matched query returns the same records while scanning substantially fewer eligible partitions.

D:

Each chosen service improves its targeted workload without unjustified duplicate maintenance cost. This outcome is evidence for the separate Clustering versus selective search optimization case, not the described Poor pruning from expression filters requirement. For the current context, verify: A matched query returns the same records while scanning substantially fewer eligible partitions.

Question 363

A Snowflake specialist explains why a proposed design works. An analyst wraps every event timestamp in a string function in a WHERE clause and a large table scans nearly every partition. Which technical principle is most relevant?

  1. Unstructured file metadata and content extraction are separate steps from ordinary relational row loading.
  2. Persisted result reuse can return eligible unchanged-query results without rerunning the underlying warehouse operators.
  3. Predicate expression form and underlying micro-partition metadata influence pruning opportunities.
  4. Bind parameters separate values from the SQL statement structure.

Explanation

Correct Answer

C:

Predicate expression form and underlying micro-partition metadata influence pruning opportunities. This explains Poor pruning from expression filters. Correct implementation: Rewrite the filter to use a selective supported range predicate on the stored timestamp when semantically equivalent. Acceptance evidence: A matched query returns the same records while scanning substantially fewer eligible partitions.

Incorrect Answers

A:

Unstructured file metadata and content extraction are separate steps from ordinary relational row loading. This mechanism concerns Unstructured file metadata and extraction. The mechanism specific to Poor pruning from expression filters is: Predicate expression form and underlying micro-partition metadata influence pruning opportunities.

B:

Persisted result reuse can return eligible unchanged-query results without rerunning the underlying warehouse operators. This mechanism concerns Result cache reuse versus warehouse local cache. The mechanism specific to Poor pruning from expression filters is: Predicate expression form and underlying micro-partition metadata influence pruning opportunities.

D:

Bind parameters separate values from the SQL statement structure. This mechanism concerns Parameterized SQL instead of concatenated filters. The mechanism specific to Poor pruning from expression filters is: Predicate expression form and underlying micro-partition metadata influence pruning opportunities.

Question 364

The project leaves the described Snowflake requirement unresolved. An analyst wraps every event timestamp in a string function in a WHERE clause and a large table scans nearly every partition. Which consequence is most directly related?

  1. Plain CAST over unreliable source text can stop a scheduled report.
  2. Uncontrolled inserts and source duplicates can create multiple current customer records.
  3. Parsing the array as one long string prevents correct item-level metrics.
  4. An unnecessarily non-sargable filter can turn a narrow inquiry into a large scan.

Explanation

Correct Answer

D:

An unnecessarily non-sargable filter can turn a narrow inquiry into a large scan. This is the direct exposure if Poor pruning from expression filters is not handled. Recommended protection: Rewrite the filter to use a selective supported range predicate on the stored timestamp when semantically equivalent. Validate by: A matched query returns the same records while scanning substantially fewer eligible partitions.

Incorrect Answers

A:

Plain CAST over unreliable source text can stop a scheduled report. This is an exposure in a separate TRY_TO_NUMBER safe type conversion scenario. The direct failure for Poor pruning from expression filters is: An unnecessarily non-sargable filter can turn a narrow inquiry into a large scan.

B:

Uncontrolled inserts and source duplicates can create multiple current customer records. This is an exposure in a separate MERGE for change-aware upsert scenario. The direct failure for Poor pruning from expression filters is: An unnecessarily non-sargable filter can turn a narrow inquiry into a large scan.

C:

Parsing the array as one long string prevents correct item-level metrics. This is an exposure in a separate LATERAL FLATTEN for order arrays scenario. The direct failure for Poor pruning from expression filters is: An unnecessarily non-sargable filter can turn a narrow inquiry into a large scan.

Question 365

The Snowflake implementation fails its pilot test. An analyst wraps every event timestamp in a string function in a WHERE clause and a large table scans nearly every partition. Which evidence should be inspected first to isolate the likely cause?

  1. Inspect query pattern, view eligibility, refresh billing and execution profile.
  2. Compare Query Profile partition scans and predicate text before and after rewriting.
  3. Inspect driver parameter markers, query history and adversarial test strings.
  4. Inspect FLATTEN input path, VALUE fields, parent join grain and empty arrays.

Explanation

Correct Answer

B:

Compare Query Profile partition scans and predicate text before and after rewriting. These checks diagnose Poor pruning from expression filters. Correct response: Rewrite the filter to use a selective supported range predicate on the stored timestamp when semantically equivalent. Expected observation: A matched query returns the same records while scanning substantially fewer eligible partitions.

Incorrect Answers

A:

Inspect query pattern, view eligibility, refresh billing and execution profile. Those checks would investigate Materialized view for repetitive aggregation, not the current Poor pruning from expression filters failure. The targeted diagnostic here is: Compare Query Profile partition scans and predicate text before and after rewriting.

C:

Inspect driver parameter markers, query history and adversarial test strings. Those checks would investigate Parameterized SQL instead of concatenated filters, not the current Poor pruning from expression filters failure. The targeted diagnostic here is: Compare Query Profile partition scans and predicate text before and after rewriting.

D:

Inspect FLATTEN input path, VALUE fields, parent join grain and empty arrays. Those checks would investigate LATERAL FLATTEN for order arrays, not the current Poor pruning from expression filters failure. The targeted diagnostic here is: Compare Query Profile partition scans and predicate text before and after rewriting.

Question 366

A Snowflake implementation team encounters this specific requirement: A report suddenly slows and the team cannot tell whether data volume, warehouse state or changed SQL caused it. Which action best addresses it?

  1. Use QUERY_HISTORY and Query Profile to separate compile, queue, scan and spill evidence before choosing a fix.
  2. Test Query Acceleration Service against the query's supported operators and workload evidence.
  3. Use TRY_TO_NUMBER or another appropriate TRY conversion and separately flag invalid source values.
  4. Add the correct join predicate or explicitly justified CROSS JOIN with controlled row grain.

Explanation

Correct Answer

A:

Use QUERY_HISTORY and Query Profile to separate compile, queue, scan and spill evidence before choosing a fix. This resolves Query history root-cause triage. Underlying reason: Query telemetry exposes resource and operator phases with different optimization levers. Verify with: The incident record points to a measured dominant latency component and a reproducible correction. Boundary: Performance comparisons should control for cache warmth and concurrent workloads.

Incorrect Answers

B:

Test Query Acceleration Service against the query's supported operators and workload evidence. That is suitable for a different requirement, Query acceleration on eligible scan-heavy query, with this separate condition: A long selective aggregation cannot meet latency targets after SQL tuning, and the team wants to offload eligible work without globally resizing.

C:

Use TRY_TO_NUMBER or another appropriate TRY conversion and separately flag invalid source values. That is suitable for a different requirement, TRY_TO_NUMBER safe type conversion, with this separate condition: A feed supplies numeric amounts as VARCHAR but occasionally contains the word UNKNOWN, causing CAST errors in a report.

D:

Add the correct join predicate or explicitly justified CROSS JOIN with controlled row grain. That is suitable for a different requirement, Avoid Cartesian product from missing join condition, with this separate condition: A report joining 100 stores to 12 periods returns far too many rows when a third table is added without its key condition.

Question 367

After implementing a Snowflake control, the reviewer must prove the business requirement was met. A report suddenly slows and the team cannot tell whether data volume, warehouse state or changed SQL caused it. Which result is the most direct acceptance evidence?

  1. The incident record points to a measured dominant latency component and a reproducible correction.
  2. A sample query extracts the precise city field while preserving other JSON properties.
  3. Special characters in user input remain data and do not alter the intended SQL logic.
  4. Every original transaction row remains present alongside its intended cumulative amount.

Explanation

Correct Answer

A:

The incident record points to a measured dominant latency component and a reproducible correction. This directly demonstrates success for Query history root-cause triage. Implementation: Use QUERY_HISTORY and Query Profile to separate compile, queue, scan and spill evidence before choosing a fix. Mechanism: Query telemetry exposes resource and operator phases with different optimization levers.

Incorrect Answers

B:

A sample query extracts the precise city field while preserving other JSON properties. This outcome is evidence for the separate VARIANT nested path query case, not the described Query history root-cause triage requirement. For the current context, verify: The incident record points to a measured dominant latency component and a reproducible correction.

C:

Special characters in user input remain data and do not alter the intended SQL logic. This outcome is evidence for the separate Parameterized SQL instead of concatenated filters case, not the described Query history root-cause triage requirement. For the current context, verify: The incident record points to a measured dominant latency component and a reproducible correction.

D:

Every original transaction row remains present alongside its intended cumulative amount. This outcome is evidence for the separate Window running total without row collapse case, not the described Query history root-cause triage requirement. For the current context, verify: The incident record points to a measured dominant latency component and a reproducible correction.

Question 368

A Snowflake specialist explains why a proposed design works. A report suddenly slows and the team cannot tell whether data volume, warehouse state or changed SQL caused it. Which technical principle is most relevant?

  1. Different Snowflake optimization features target distinct access patterns and incur separate charges.
  2. MERGE supports conditionally matched updates and not-matched inserts in one DML operation.
  3. Query telemetry exposes resource and operator phases with different optimization levers.
  4. FLATTEN expands array elements into rows within a correlated lateral relationship.

Explanation

Correct Answer

C:

Query telemetry exposes resource and operator phases with different optimization levers. This explains Query history root-cause triage. Correct implementation: Use QUERY_HISTORY and Query Profile to separate compile, queue, scan and spill evidence before choosing a fix. Acceptance evidence: The incident record points to a measured dominant latency component and a reproducible correction.

Incorrect Answers

A:

Different Snowflake optimization features target distinct access patterns and incur separate charges. This mechanism concerns Clustering versus selective search optimization. The mechanism specific to Query history root-cause triage is: Query telemetry exposes resource and operator phases with different optimization levers.

B:

MERGE supports conditionally matched updates and not-matched inserts in one DML operation. This mechanism concerns MERGE for change-aware upsert. The mechanism specific to Query history root-cause triage is: Query telemetry exposes resource and operator phases with different optimization levers.

D:

FLATTEN expands array elements into rows within a correlated lateral relationship. This mechanism concerns LATERAL FLATTEN for order arrays. The mechanism specific to Query history root-cause triage is: Query telemetry exposes resource and operator phases with different optimization levers.

Question 369

The project leaves the described Snowflake requirement unresolved. A report suddenly slows and the team cannot tell whether data volume, warehouse state or changed SQL caused it. Which consequence is most directly related?

  1. Warm-cache claims can hide expensive cold-query performance and overstate tuning benefits.
  2. Guessing based on wall-clock time can worsen costs and hide query regressions.
  3. Independent queries with different filters can give inconsistent denominators.
  4. A directory listing alone does not parse PDF content or grant safe document access.

Explanation

Correct Answer

B:

Guessing based on wall-clock time can worsen costs and hide query regressions. This is the direct exposure if Query history root-cause triage is not handled. Recommended protection: Use QUERY_HISTORY and Query Profile to separate compile, queue, scan and spill evidence before choosing a fix. Validate by: The incident record points to a measured dominant latency component and a reproducible correction.

Incorrect Answers

A:

Warm-cache claims can hide expensive cold-query performance and overstate tuning benefits. This is an exposure in a separate Result cache reuse versus warehouse local cache scenario. The direct failure for Query history root-cause triage is: Guessing based on wall-clock time can worsen costs and hide query regressions.

C:

Independent queries with different filters can give inconsistent denominators. This is an exposure in a separate Conditional aggregation in one scan scenario. The direct failure for Query history root-cause triage is: Guessing based on wall-clock time can worsen costs and hide query regressions.

D:

A directory listing alone does not parse PDF content or grant safe document access. This is an exposure in a separate Unstructured file metadata and extraction scenario. The direct failure for Query history root-cause triage is: Guessing based on wall-clock time can worsen costs and hide query regressions.

Question 370

The Snowflake implementation fails its pilot test. A report suddenly slows and the team cannot tell whether data volume, warehouse state or changed SQL caused it. Which evidence should be inspected first to isolate the likely cause?

  1. Inspect query pattern, view eligibility, refresh billing and execution profile.
  2. Inspect source values, conversion expressions and invalid-value count.
  3. Inspect query ID, warehouse, credits, compilation, execution and profile operators.
  4. Review GROUP BY keys, conditional predicates and reconciliation to source totals.

Explanation

Correct Answer

C:

Inspect query ID, warehouse, credits, compilation, execution and profile operators. These checks diagnose Query history root-cause triage. Correct response: Use QUERY_HISTORY and Query Profile to separate compile, queue, scan and spill evidence before choosing a fix. Expected observation: The incident record points to a measured dominant latency component and a reproducible correction.

Incorrect Answers

A:

Inspect query pattern, view eligibility, refresh billing and execution profile. Those checks would investigate Materialized view for repetitive aggregation, not the current Query history root-cause triage failure. The targeted diagnostic here is: Inspect query ID, warehouse, credits, compilation, execution and profile operators.

B:

Inspect source values, conversion expressions and invalid-value count. Those checks would investigate TRY_TO_NUMBER safe type conversion, not the current Query history root-cause triage failure. The targeted diagnostic here is: Inspect query ID, warehouse, credits, compilation, execution and profile operators.

D:

Review GROUP BY keys, conditional predicates and reconciliation to source totals. Those checks would investigate Conditional aggregation in one scan, not the current Query history root-cause triage failure. The targeted diagnostic here is: Inspect query ID, warehouse, credits, compilation, execution and profile operators.

SnowPro Core Practice Tests

Purchase Individually

  • Premium File

    703 Questions & Answers
    Last Update: Nov 22, 2026

    $76.99
    $69.99
  • Training Course

    92 Lectures

    $43.99
    $39.99
  • Study Guide

    413 Pages

    $43.99
    $39.99

SPECIAL OFFER: GET 10% OFF
This is ONE TIME OFFER

You save
10%

Enter Your Email Address to Receive Your 10% Off Discount Code

SPECIAL OFFER: GET 10% OFF

You save
10%

Use Discount Code:

A confirmation link was sent to your e-mail.

Please check your mailbox for a message from support@examlabs.com and follow the directions.

Download Free Demo of VCE Exam Simulator

Experience Avanset VCE Exam Simulator for yourself.

Simply submit your email address below to get started with our interactive software demo of your free trial.

  • Realistic exam simulation and exam editor with preview functions
  • Whole exam in a single file with several different question types
  • Customizable exam-taking mode & detailed score reports