Snowflake SnowPro Core Exam Dumps and Practice Test Questions Set 5 Q81-100

Visit here for our full Snowflake SnowPro Core exam dumps and practice test questions.

Question 81 

What is the primary purpose of using a file format option during Snowflake COPY INTO operations?

A) To optimize the storage size of the table
B) To correctly interpret the structure and encoding of the incoming data
C) To automatically detect schema drift in source files
D) To increase the compute power during data loading

Answer: B

Explanation:

A file format parameter is used so that Snowflake receives explicit instructions on how incoming files should be parsed, interpreted, and processed. The first choice suggests that the intention is to reduce the storage footprint; however, file format definitions do not influence how data is ultimately compressed and stored internally. Snowflake automatically applies micro-partition compression, completely unrelated to file format settings. The next choice accurately represents the core function because the file format provides essential directives such as field delimiters, quote characters, data type interpretation behavior, null handling rules, and encoding details. 

These directives ensure that data is read and transformed correctly during ingestion, preventing misalignment or malformed records that might arise without such configuration. The following choice claims schema drift detection, yet this is not facilitated through file format configuration. Snowflake can detect some mismatches or failures based on load behavior, but file formats are not designed to monitor changing structures. Instead, they simply instruct Snowflake on how the file should be read. The last choice refers to compute scaling, but file format settings never increase or change compute resources. Virtual warehouse size determines compute scale, and modifying file formats cannot influence this factor. Snowflake uses the file format specification strictly for parsing logic and file interpretation behavior. 

Because the correct interpretation of field separators, escape characters, date formats, and encoding is crucial for accurate ingestion, the second choice is the correct answer. Without a proper file format definition, COPY INTO might misread structures, load corrupted values, or generate numerous load errors. The file format option gives Snowflake a deterministic rule set enabling consistent parsing across repeated ingestion operations. It also simplifies reusable pipelines since multiple COPY commands can reference the same reusable file format object. Consequently, the correct selection is the second answer because it aligns directly with Snowflake’s intended usage of file format parameters.

Question 82 

Which Snowflake feature allows organizations to calculate storage, compute, and cloud services usage at a granular level?

A) Snowflake Streams
B) Resource Monitors
C) Information Schema Views
D) Clustering Keys

Answer: C

Explanation: 

The first choice focuses on tracking data changes within a table, enabling incremental pipelines and providing CDC capabilities. While useful for ETL orchestration and capturing row-level deltas, it does not relate to cost breakdowns, billing details, or consumption reporting. The second choice allows administrators to enforce spending thresholds on virtual warehouse credit usage. It offers mechanisms to suspend or alert on excessive consumption but does not present granular cost details or historical breakdowns. It is mainly a governance tool rather than a reporting mechanism. 

The third choice refers to metadata views that expose deep insights across areas such as storage consumption, query history, warehouse credit usage, table sizes, user activity, and task execution patterns. These views enable precise reporting and detailed monitoring of how resources are used at different layers. They can be queried directly to create dashboards or cost governance tools, making them integral for auditing financial usage and developing internal chargeback models. The final choice involves a tuning mechanism used to influence micro-partition pruning. Clustering keys improve query efficiency for large tables with predictable filtering patterns, but they have no connection to financial reporting or monitoring costs. 

When analyzing each possibility, only the third choice acts as the primary mechanism administrators use to calculate detailed consumption metrics. It provides structured metadata designed for transparency into Snowflake’s resource utilization. Organizations rely on these views to understand which workloads consume the most credits, identify heavy users, compare storage patterns over time, and correlate compute activity with business demands. These insights support financial accountability by enabling cost allocation to business units or project teams. Because cost reporting is rooted in metadata querying and not in CDC, warehouse thresholds, or pruning optimization, the correct response is the third one.

Question 83 

Which Snowflake capability allows a warehouse to resume automatically when a query is submitted?

A) Multi-cluster Warehouse
B) Auto-resume
C) Result Cache
D) Fail-safe

Answer: B

Explanation: 

The first choice focuses on scaling compute resources horizontally across multiple clusters to handle concurrent workloads. While useful for large-scale environments requiring high throughput, this capability does not provide automated resumption behavior. Instead, it dynamically adds or removes clusters based on load but still requires the warehouse itself to be active. The second choice specifically enables a suspended warehouse to restart itself automatically when a query is issued. With this mechanism, Snowflake eliminates manual steps and reduces costs by suspending Warehouses when idle while preserving convenience for users. 

When a new request arrives, Snowflake seamlessly resumes operations with minimal delay. This automation is essential for cost optimization because warehouses incur credits only while running. The third choice refers to a caching mechanism that stores previous query results, allowing identical queries to return instantly without reusing compute resources. This improves performance but does not manage warehouse state transitions. The final choice deals with long-term data recovery beyond standard retention windows, providing an additional safety net for catastrophic failures. Although vital for recovery, it is unrelated to warehouse lifecycle behavior. 

Considering the differences, only the second choice directly controls automated warehouse activation. The ability to resume seamlessly ensures low overhead during periods of inactivity while preserving user experience. By combining auto-suspend and auto-resume settings, Snowflake offers an efficient compute consumption model where credits are consumed only during active workloads. Therefore, the correct answer is the second choice.

Question 84 

Which action ensures that data loaded through Snowpipe becomes available for querying as quickly as possible?

A) Using a large virtual warehouse to speed up Snowpipe ingestion
B) Configuring event notifications to trigger ingestion automatically
C) Increasing the retention period of the staging area
D) Using tasks to schedule Snowpipe execution

Answer: B

Explanation: 

The first choice implies that warehouse sizing can influence Snowpipe’s ingestion speed, yet Snowpipe does not use customer warehouses for loading. Instead, it relies on Snowflake-managed compute, which is independent of user warehouse configurations. Therefore, increasing warehouse size does not accelerate ingestion. The second choice describes event-based triggers such as cloud storage notifications that automatically begin the ingestion workflow the moment new files arrive. This mechanism ensures near-real-time pipeline execution by minimizing delay between file upload and ingestion. When notifications are configured, Snowpipe does not rely on manual initiation or periodic polling, making ingestion nearly immediate. The third choice focuses on extending retention periods, but retention only governs how long staged files remain accessible. It does not influence ingestion timeliness or latency within Snowpipe. 

The fourth choice mentions scheduled tasks, but Snowpipe does not operate through scheduled intervals. Snowpipe is inherently designed for continuous ingestion and does not require tasks to activate it. Scheduled tasks would introduce unnecessary delays rather than improve responsiveness. Upon analysis, only the second choice ensures minimal latency. Event notifications trigger ingestion automatically and reduce pipeline downtime, enabling Snowpipe to begin processing as soon as data arrives. This ensures the fastest availability of data for querying, making the second answer correct.

Question 85 

Which Snowflake mechanism helps reduce credit usage for queries that repeatedly request the same data?

A) Warehouse Cache
B) Query Acceleration Service
C) Result Cache
D) Materialized View

Answer: C

Explanation: 

The first choice describes local disk caching within a running warehouse. It accelerates processing by caching micro-partitions but does not eliminate compute costs because the warehouse must still run and consume credits. The second choice adds additional compute to accelerate certain workloads. While useful for complex queries, it increases rather than decreases compute spending. The third choice stores the results of previously executed queries that are identical in both text and underlying data conditions. 

If a repeated query qualifies, Snowflake returns the stored output instantly without using warehouse compute. This avoids credit consumption entirely during the response. The fourth choice uses precomputed results for improved performance but still requires maintenance compute, and its refresh costs often exceed the savings for repetitive simple queries. Because the third choice uniquely avoids compute consumption by serving results instantly from metadata storage when queries meet criteria, it represents the most accurate answer for reducing credits in repetitive query scenarios.

Question 86 

Which Snowflake feature ensures that external tables reflect the latest metadata from cloud storage?

A) Snowflake Streams
B) External Table Refresh
C) Materialized Views
D) Snowpipe Auto-ingest

Answer: B

Explanation: 

The first choice captures data changes within Snowflake-managed tables and is intended for incremental ingestion workflows. It tracks row-level inserts, updates, and deletes but does not interact with external metadata stored in cloud object storage. Because external tables represent data managed outside Snowflake, they rely on a separate mechanism to update metadata rather than depending on change tracking inside the platform. The second choice refers specifically to a mechanism that synchronizes Snowflake’s metadata with the actual state of files stored in external locations. 

This process updates partitions, file listings, and other structural metadata to ensure that queries reflect the most current external data. When new files arrive or old files are removed, this refresh process enables Snowflake to incorporate those changes. The third choice precomputes data inside Snowflake to accelerate queries but does not manage external metadata. It focuses on internal optimization rather than connecting to external storage structures. The fourth choice automates ingestion of files into Snowflake-managed tables but does not update external table metadata. Instead, it moves or loads data inside Snowflake and is not concerned with synchronizing external metadata structures. 

Evaluating all the possibilities, only the second choice directly handles the reconciliation of metadata for files that reside outside Snowflake. Without the refresh process, newly uploaded files would not automatically appear in queries, and removed files could still appear in metadata. Thus, the refresh mechanism ensures accurate representation of external datasets. By analyzing its purpose relative to the other choices, the second answer is correct.

Question 87 

Which Snowflake privilege is required to allow a role to create a new warehouse?

A) MANAGE GRANTS
B) CREATE WAREHOUSE
C) USAGE
D) MONITOR

Answer: B

Explanation: 

The first choice provides the ability to modify privileges on objects but does not authorize the creation of compute resources. It is primarily used for privilege delegation and governance activities and cannot independently provision virtual warehouses. The second choice grants permission for warehouse creation. A role with this privilege can define new compute clusters, apply sizing configurations, set auto-suspend settings, and manage warehouse properties. Without this permission, even roles with high-level administrative capabilities cannot create warehouses. The third choice is used to allow operation of a warehouse, not creation. 

This includes running queries, using compute, and performing tasks requiring warehouse access. However, it does not allow provisioning of new warehouses. The fourth choice provides visibility into warehouse metrics and history such as credit usage and query load. While insightful for monitoring workloads, it does not grant the capability to create new compute resources.  The ability to create warehouses is explicitly controlled by the privilege in the second choice, making it the correct selection. Snowflake isolates warehouse creation as a specialized administrative action requiring explicit authorization, aligning with least privilege principles.

Question 88 

What is the purpose of using the COPY INTO <location> command?

A) To clone tables into another schema
B) To export query results to stage locations
C) To apply transformations to external tables
D) To refresh materialized views

Answer: B

Explanation: 

The first choice relates to the cloning mechanism that creates zero-copy clones of tables, schemas, or databases. Cloning is used to provide instantaneous duplication for testing or developmental workflows but is unrelated to exporting data. The second choice describes the ability to unload data from Snowflake tables or query outputs into an external stage or cloud storage destination. This mechanism enables external tools or downstream systems to consume Snowflake-generated datasets. It is commonly used in BI pipelines, data transfer processes, archival workflows, or cross-platform integrations. 

The third choice refers to reading external table data but does not represent export actions. External tables are designed for querying external storage; they do not undergo transformations through unload commands. The fourth choice refreshes precomputed views to ensure they remain synchronized with base tables, but it has no connection to extracting data to stages. When examining the objective of the command, the second choice stands out as the correct explanation. It handles the movement of processed data out of Snowflake for consumption elsewhere, aligning perfectly with the command’s intended function.

Question 89 

What happens when a Snowflake task is created without specifying a warehouse?

A) The task fails to run
B) The task uses Snowflake-managed compute
C) The task automatically uses the default warehouse
D) The task pauses until a warehouse is assigned manually

Answer: B

Explanation: 

The first choice incorrectly assumes tasks cannot operate without a warehouse. In reality, Snowflake allows the creation of tasks that use Snowflake-managed compute instead of a warehouse. These tasks can execute SQL statements without consuming warehouse credits. The second choice accurately captures the behavior where Snowflake provides internal compute resources dedicated to task execution when no warehouse is assigned. This capability is often preferred for lightweight automation routines, metadata operations, or small pipeline components. 

The third choice suggests that a default warehouse is automatically selected, but Snowflake does not assign compute resources implicitly. A warehouse must be explicitly defined for warehouse-based compute, and no default is applied in its absence. The fourth choice gives the impression that tasks remain inactive until assigned a warehouse, but tasks can still execute using Snowflake-managed compute as long as they qualify for that mode.
Given these distinctions, only the second choice correctly represents how Snowflake executes tasks that omit a warehouse specification.

Question 90 

Which Snowflake mechanism enables secure access to external cloud storage without sharing permanent credentials?

A) Secondary Roles
B) OAuth
C) Key Pair Authentication
D) Storage Integration

Answer: D

Explanation: 

The first choice pertains to role switching within a session, allowing users to assume additional roles temporarily. It has no relation to external cloud credential management. The second choice enables delegated authorization for applications but does not provide purpose-built integration with cloud storage credentials. The third choice authenticates users connecting via clients such as SnowSQL but does not manage access to cloud data storage. The fourth choice establishes a secure identity integration between Snowflake and cloud storage platforms, enabling Snowflake to access external data without embedding or sharing permanent keys. This mechanism leverages cloud IAM roles or temporary credentials to ensure secure, controlled access. Evaluating these differences demonstrates that only the fourth choice provides the correct mechanism for secure, credential-less access to external stages.

Question 91

 What is the main benefit of using reader accounts in Snowflake?

A) They allow unlimited compute at no cost
B) They let external parties query shared data without owning a Snowflake account
C) They provide automated ETL pipelines
D) They enable cloning of databases across regions

Answer: B

Explanation: 

The first choice suggests that compute is free within these environments, but reader accounts still consume credits from the provider account whenever queries run. They do not offer cost-free computation; instead, computation is billed back to the sharing provider rather than the consumer. The second choice accurately captures the purpose of these environments. Reader accounts allow organizations to share curated datasets with partners or clients who may not have their own Snowflake subscription. These environments operate under the provider’s billing structure while maintaining strict isolation and controlled data access. 

This creates a secure, governed mechanism for distributing analytics without requiring external consumers to provision or manage their own Snowflake deployments. The third choice implies automation pipelines, but reader accounts are not used for transformation or ingestion activities. They exist solely for consumption of shared data and do not replace Snowflake pipeline features such as tasks or streams. The fourth choice describes cross-region duplication capabilities, which are handled through replication features rather than reader accounts. By analyzing each possibility, it becomes clear that only the second choice aligns with the function and intent of these accounts. They provide a lightweight, secure option for external entities to interact with shared datasets, making that answer correct.

Question 92 

What is the main advantage of using the SEARCH OPTIMIZATION service?

A) It allows automatic warehouse scaling
B) It improves lookup performance on selective queries
C) It increases storage retention periods
D) It manages clustering keys automatically

Answer: B

Explanation: 

The first choice relates to warehouse behavior, but search optimization is unrelated to computing scale or warehouse adjustments. It does not manage cluster resizing or workload balancing. The second choice explains how the service accelerates queries that perform highly selective searches, such as point lookups or selective filtering on columns with low cardinality. This service builds additional search structures that allow Snowflake to minimize the amount of data scanned, enhancing performance for workloads where accessing small subsets of large tables is critical. The third choice concerns data retention policies, but search optimization does not influence these settings. Retention is configured separately at the table or database level. The fourth choice refers to key management for micro-partitioning, but search optimization does not replace clustering behavior. Clustering focuses on ordering data across micro-partitions, whereas search optimization enables efficient identification of qualifying data inside those partitions.
Among the various interpretations, the second choice precisely describes what this service provides: targeted performance acceleration for selective lookups.

Question 93 

Which Snowflake feature allows organizations to unify structured and semi-structured data under a single engine?

A) Hybrid Tables
B) Snowpipe
C) Dynamic Tables
D) Snowflake’s native VARIANT data type

Answer: D

Explanation: 

The first option points to a recently introduced table type designed for workloads that emphasize rapid point lookups and high-performance transactional behavior. This table design is particularly valuable for applications that require fast single-row reads and writes or scenarios resembling operational workloads. While this table format enhances performance for such use cases, it does not attempt to unify multiple categories of data within a single system. Its purpose is focused on transactional efficiency rather than enabling a universal framework capable of integrating structured, semi-structured, and hierarchical data types. Because of that narrow focus, it cannot be the mechanism responsible for combining diverse formats into a single analytical environment.

The second option represents a service centered on continuous ingestion of files placed in cloud storage. It automates the detection and loading of incoming files, which is highly useful for pipelines that handle streaming-like data or frequently arriving batches. However, despite its ingestion automation, it does not play any role in integrating or harmonizing different data structures. Its responsibility stops once files are loaded, and it does not define how Snowflake interprets or unifies diverse formatting approaches. Consequently, it is not aligned with the concept of a universal layer that supports multiple data types with a consistent query model.

The third option provides functionality for maintaining derived tables or aggregated datasets in an efficient and incremental way. These tables store precomputed results and can automatically refresh when underlying data changes. Such a feature is extremely advantageous for workloads involving repeated analytical queries, but its purpose is to improve performance rather than unify data formats. It does not influence how Snowflake handles structured versus semi-structured data, nor does it offer a framework that consolidates different data types into a single engine for uniform querying.

The fourth option enables Snowflake to treat semi-structured data formats such as JSON, Parquet, XML, ORC, and Avro with the same SQL engine used for traditional tabular data. This capability allows nested, schema-flexible content to be stored efficiently while still being directly queried using familiar SQL syntax. Rather than requiring external transformation, preprocessing, or schema enforcement before analysis, this data type allows Snowflake to seamlessly integrate heterogeneous content. It effectively creates a unified environment where structured and semi-structured data coexist and can be accessed through a common set of operations. Because this mechanism directly supports the unification of diverse formats within one analytical platform, it is the only choice that satisfies the requirement described.

Question 94 

Why would an organization use multi-cluster warehouses?

A) To parallelize data loading into external tables
B) To increase concurrency handling during heavy workloads
C) To improve search optimization performance
D) To enhance security rule enforcement

Answer: B

Explanation: 

The first choice implies that loading external tables requires parallel compute scaling. External tables do not ingest data into Snowflake; instead, they reference data stored externally. Therefore, parallel loading is not applicable. The second choice describes how multi-cluster warehouses automatically add or remove compute clusters to handle periods of high concurrency. When many users submit simultaneous queries, additional clusters activate to maintain performance. During light periods, the warehouse contracts to reduce credit consumption. This dynamic scaling behavior is the essence of multi-cluster capabilities. 

The third choice states that search optimization performance is enhanced by multi-cluster warehouses, which is incorrect. Search optimization uses metadata structures unrelated to warehouse scale. The fourth choice focuses on security enforcement, but multi-cluster functionality does not influence access checks or privacy controls. Analyzing the intent behind this feature, the second choice clearly presents the purpose: scaling compute to accommodate concurrent workloads.

Question 95 

Which Snowflake resource is billed when running queries?

A) Virtual warehouse credits
B) Cloud services credits
C) Database storage credits
D) Fail-safe credits

Answer: A

Explanation:

The first option refers to the type of credit that Snowflake bills when compute operations are performed, and this accurately reflects how Snowflake handles query-related charges. When a user executes queries, loads data, performs transformations, or runs analytic workloads, Snowflake relies entirely on virtual warehouses to supply the compute power required to process those tasks. A virtual warehouse is a collection of compute resources that scales independently of storage, and each second of active usage consumes a portion of warehouse credits. These credits directly measure compute consumption and represent the primary cost driver for any operation that requires CPU, memory, or execution time. Because compute processing is the centerpiece of query execution, warehouse credits serve as the billing unit that accurately accounts for all workloads evaluated by the engine. As long as a warehouse is running, even if lightly, Snowflake applies billing based on warehouse credits, making this option the correct interpretation of how compute is charged.

The second option mentions credit usage associated with metadata services, authentication support, optimization processes, and similar system-level functions. These operations fall under Snowflake’s cloud services layer. While this layer plays an important role in query planning, metadata management, access control, and various support functions, it is not directly billed for normal query execution as long as usage stays within Snowflake’s 10% monthly allowance relative to compute consumption. Only when this cloud services activity exceeds that threshold would additional cloud services credits be billed. Because normal querying rarely pushes these functions beyond allowed limits, this credit type does not represent the main billing mechanism for standard query execution and therefore cannot be considered the correct answer.

The third option highlights storage-related credits, which are charged based on the compressed volume of data persisted within Snowflake. Storage is billed independently of compute and does not fluctuate with query activity. Whether a user queries data once or a thousand times, the storage cost remains determined purely by how much data exists in the system. Since storage credits are not connected to query execution cycles, they are not applicable as the resource billed specifically for running queries.

The fourth option concerns the long-term recovery feature intended for exceptional restoration scenarios. Fail-safe ensures historical recoverability but does not associate any billing with query execution. Because it does not charge per query and does not contribute to compute, it cannot be the correct selection.

Evaluating all four choices shows that only virtual warehouse credits directly measure compute workloads, making the first option correct.

Question 96 

What is one primary benefit of using Snowflake Managed Accounts for data sharing?

A) They eliminate the need for replication
B) They allow consumers to access shared data without managing infrastructure
C) They automatically refresh all shared tables without configuration
D) They store data in a separate cloud provider by default

Answer: B

Explanation:

The first option suggests that this mechanism allows data providers to avoid the need for replication altogether, but this interpretation does not align with how such accounts operate. Replication decisions still depend on the sharing design established by the provider. If a provider wants to share data across regions or with numerous external consumers, they may still choose to replicate data, but the presence of a managed environment does not inherently remove that requirement. Replication remains a separate architectural consideration rather than a function directly tied to these accounts.

The second option reflects the genuine purpose of these managed environments. They are designed to let an external consumer use shared data without having to deploy, configure, or administer their own Snowflake system. The provider sets up the environment, manages operational aspects, controls compute access, and defines the data available to the consumer. The recipient gains the ability to query the shared datasets immediately without incurring the responsibilities associated with infrastructure ownership. This creates a seamless experience for external partners, customers, or organizations that need access to analytical data but would prefer not to maintain a Snowflake account. This ease of adoption is one of the strongest benefits of these managed setups and is exactly why they are considered a powerful feature for frictionless data sharing.

The third option implies that shared objects automatically refresh themselves without any action from the provider. In reality, updates are made available based on how the provider maintains and refreshes the underlying shared tables. There is no built-in automation within the managed environment itself that takes responsibility for refreshing data. The mechanism relies entirely on the provider’s maintenance of the source data, not on autonomous behavior from the account that receives it.

The fourth option proposes that the data would automatically reside in a separate cloud provider by default. This is not accurate because the environment lives in the same cloud and region chosen by the provider. Cross-cloud placement does not occur automatically and requires deliberate configuration.

Given all considerations, the second option accurately describes the intended benefit, making it the correct answer.

Question 97 

Which Snowflake feature can reduce the need for frequent manual file processing when working with semi-structured data?

A) File Format Cloning
B) Automatic Schema Detection
C) Snowpipe
D) Reader Accounts

Answer: C

Explanation: 

The first option refers to cloning of file formats, but this feature does not influence how often files need to be manually processed or loaded. A file format in Snowflake only defines instructions for interpreting data files such as CSV, JSON, Avro, or Parquet. Cloning such a format simply duplicates the configuration, not the data itself, and certainly does not automate ingestion or reduce operational tasks. Because it has no capability to detect new files, launch ingestion jobs, or coordinate processing events, this option does not meaningfully reduce the manual workload associated with loading semi-structured files into Snowflake.

The second option highlights the platform’s ability to infer structure from semi-structured data naturally when loaded into the VARIANT column type. While this flexibility avoids the need to define rigid schemas, it does not automate ongoing ingestion cycles or help process new files as they arrive. The system may interpret hierarchical structures dynamically, but that alone does not remove the need for repeated manual commands or scheduled tasks to load files. Schema flexibility enhances usability and querying but does not replace the need for automated ingestion orchestration.

The third option represents a mechanism specifically designed to automate loading as soon as files appear in the cloud storage stage. This feature continuously monitors storage locations such as Amazon S3, Azure Blob Storage, or Google Cloud Storage and, using event notifications or periodic checks, triggers ingestion automatically. This reduces operational overhead because users no longer need to repeatedly execute manual COPY commands. It is especially suited for semi-structured data, which often arrives in frequent small batches from streaming systems, APIs, or event generators. The system leverages serverless compute to handle ingestion, providing scalability and reliability without requiring warehouse scheduling. By eliminating routine manual processing, improving ingestion latency, and streamlining data flow, this option directly addresses the need stated in the question.

The fourth option references the mechanism for granting external consumers secure access to provider-shared data. These accounts are purely about controlled consumption and do not interact with file ingestion or transformation pipelines. They have no involvement in automating loading tasks or managing semi-structured file workflows.

After understanding the roles of each feature, it becomes clear that only the third option provides continuous, automated ingestion that prevents the need for repetitive manual processing, making it the correct answer.

Question 98 

What is the main benefit of using zero-copy cloning for analytics development?

A) Each clone requires its own physical storage blocks
B) Clones allow isolated experimentation without altering the original
C) Clones automatically refresh to stay in sync with source data
D) Clones reduce compute usage during transformation

Answer: B

Explanation: 

The first option states that each clone requires its own dedicated physical storage blocks, but this conflicts with the foundational principle behind zero-copy cloning in Snowflake. Cloning works by referencing existing micro-partitions rather than duplicating them. Only when changes occur within either the source or the clone does Snowflake create new micro-partitions using its copy-on-write strategy. This design significantly reduces storage consumption and speeds up the creation of cloned environments. Because the clone initially requires no additional physical storage, the assertion in the first option is inaccurate and does not describe any actual benefit related to cloning in Snowflake.

The second option correctly describes the essential value that clones provide, especially in development and analytical workflows. A clone allows a user to create a separate environment that mirrors the state of the original object at the moment the clone is created. This enables teams to test transformations, design new analytical models, run validation checks, or experiment with data without affecting production systems. Such isolation is crucial for analytics development, as developers can freely modify data, try different processing approaches, and evaluate the effects of transformations safely. If mistakes occur or experiments produce unexpected results, the original data remains fully intact. This creates a reliable, low-risk development workflow that ensures data integrity while providing operational flexibility.

The third option claims that clones automatically refresh in order to stay synchronized with the source, but this does not align with how cloning operates. A clone captures the source data as a snapshot and does not maintain a continuous connection or synchronization mechanism. If up-to-date data is needed, the user must either recreate the clone or use features such as replication or streams to track changes. Therefore, the idea of an automatically refreshed clone misrepresents its behavior.

The fourth option implies that cloning somehow reduces compute usage during transformations. Compute consumption in Snowflake depends entirely on the workload and the size of the operations performed. Cloning influences storage efficiency, not compute optimization. Executing queries or transformations on a clone still requires virtual warehouse resources, and no compute savings are inherently provided by the cloning mechanism.

Considering all explanations, the second option accurately captures the primary benefit of zero-copy cloning—safe, isolated experimentation—making it the correct answer.

Question 99 

Which type of view stores precomputed results to accelerate repetitive queries?

A) Secure View
B) Materialized View
C) Reader View
D) Masking View

Answer: B

Explanation: 

The first option refers to a view type that is designed to protect underlying logic, metadata, and SQL definition details by preventing consumers from seeing the full text of the view’s query. Secure views emphasize governance and privacy, ensuring that sensitive structures or logic remain hidden from unauthorized users. However, these views do not store any precomputed results; instead, each time they are queried, they execute the underlying SQL logic on the base tables. Because they always rely on runtime computation, they do not provide any performance advantages in terms of storing or reusing previous results. Their purpose is entirely centered on security rather than query acceleration.

The second option correctly identifies a structure that stores precomputed results derived from a query. These views keep the results physically materialized so that when a user queries them, Snowflake can retrieve data much faster than if it had to recalculate the results every time. This is especially useful for workloads involving complex aggregations, frequent filtering of large datasets, or repeated computations that do not change often. Materialization reduces the time needed to process these repetitive queries and decreases the computational load on virtual warehouses. Although Snowflake automatically maintains materialized views by refreshing them when underlying tables change, these refreshes are incremental and optimized for efficiency. Because they store precomputed information and significantly speed up analytical operations, materialized views directly fulfill the purpose described in the question.

The third option refers to accounts used by external parties to access shared data. These accounts are part of Snowflake’s secure data-sharing framework and allow consumers to query shared datasets without requiring their own Snowflake subscription. However, they do not introduce any alternative type of view and have no role related to stored computations. Their function revolves around providing secure and isolated access rather than accelerating performance through precomputation.

The fourth option applies dynamic data masking rules to restrict the visibility of sensitive fields based on user roles or policies. Masking views are useful in regulatory compliance scenarios by automatically altering data exposure at query time. Yet they do not store results, accelerate queries, or rely on precomputation. They simply enforce conditional transformations during runtime.

When comparing all options, only the second one accurately describes a construct that physically stores precomputed data, enabling faster performance for repeated queries. This makes it the correct answer.

Question 100 

Which Snowflake feature best supports building incremental data transformation pipelines?

A) Fail-safe
B) Streams
C) Snowpipe
D) Replication

Answer: B

Explanation:

The first option describes a mechanism designed for long-term data recovery rather than operational data processing. Fail-safe exists as a final protective layer that allows Snowflake to recover data that has been lost due to extreme or unexpected events. It is not meant for day-to-day data engineering tasks, nor does it participate in tracking row-level changes or supporting incremental transformations. It cannot detect inserts, updates, or deletes, and does not integrate with transformation logic. Because it focuses exclusively on historical data recovery and disaster resilience, it does not contribute to the construction of incremental pipelines.

The second option provides a mechanism that records changes occurring in a table, capturing inserts, updates, and deletes in near real time. This change data capture capability allows transformation pipelines to process only the new or modified rows instead of scanning full tables repeatedly. This greatly improves performance and efficiency, especially for large datasets where full refresh processing is costly and unnecessary. Streams maintain a continuously updated record of table modifications by exposing these changes as a relational construct. When a downstream process consumes the changes, the stream automatically advances its offset, ensuring that each modification is processed exactly once. This workflow enables low-latency, incremental data transformations and supports modern ELT architectures where data freshness and efficiency are critical. Because it is specifically built for tracking changes and powering incremental workloads, it directly aligns with the goal described in the question.

The third option represents an automated loading service that continuously ingests files from cloud storage. While highly useful for ingesting semi-structured or frequently arriving datasets, it does not track modifications to existing tables or identify which rows have changed. Its role is limited to initial ingestion, not incremental transformation of data already stored in Snowflake. Therefore, although valuable for data loading, it does not satisfy the requirement for an incremental transformation mechanism.

The fourth option enables the movement and synchronization of data across regions or accounts for business continuity or multi-region availability. While replication ensures copies of objects remain consistent across locations, it does not identify row-level changes for transformation processes. Its purpose is disaster recovery and high availability rather than powering incremental data pipelines.

Considering all functionality, only the second option provides the change-tracking foundation necessary for incremental transformations, making it the correct answer.