Snowflake’s database storage layer is responsible for encrypting data at rest. All persisted data—whether structured, semi-structured, or unstructured—is encrypted using strong encryption algorithms such as AES-256. This process is automatic and transparent to users, ensuring that files and micro-partitions stored in Snowflake-managed cloud storage are encrypted by default. Data in transit is protected separately through TLS.

Virtual warehouses provide compute resources to execute queries and do not perform storage-level encryption. Data loading utilities (such as COPY INTO and client tools) orchestrate data movement but do not handle at-rest encryption. The Cloud Services layer manages metadata, session management, security policies, and query optimization, but the actual encryption of stored data is part of the storage subsystem.

==================

AFILE FORMATobject defines how Snowflake interprets files during loading and unloading operations. It includes settings such as:

Type (CSV, JSON, Parquet, Avro)

FIELD_DELIMITER, SKIP_HEADER, NULL_IF

Compression (GZIP, BZ2)

Time and date format handling

Encoding and escape character rules

By encapsulating these parsing rules, organizations standardize ingestion logic across stages and COPY INTO commands.

FILE FORMATs improve data consistency, reduce errors, and simplify maintenance by centralizing file parsing configurations. They do not manage compute resources, create backups, or define data security policies—those belong to warehouses, fail-safe/Time Travel, and RBAC respectively.

====================================================

The Snowsight Monitoring tab provides a centralized view of virtual warehouse usage metrics, enabling administrators and developers to evaluate how compute resources are being consumed. This includes critical insights such as credit usage, query load, concurrency levels, average queue times, execution durations, and auto-scaling activity (for multi-cluster warehouses). These metrics help determine whether a warehouse is correctly sized, whether concurrency issues are occurring, or whether workloads require scaling up or adding clusters.

Query history is available in a different section—“Activity → Query History”—not under Monitoring. Time Travel snapshots are not visualized within Monitoring; Time Travel is controlled via retention parameters and accessed with SQL (AT/BEFORE clauses). Schema change history is also not part of Monitoring and instead is discoverable through ACCOUNT_USAGE or specific metadata views.

The Monitoring tab exists specifically to help evaluate warehouse performance and resource consumption, enabling optimization of compute spending and better workload management.

==================

SnowflakeCortex functionsenable AI-powered operations directly within SQL and Python, supporting multiple languages via translation, classification, summarization, and document extraction functions. They operate seamlessly onstructuredandsemi-structureddata stored in VARIANT columns, allowing LLMs to process text, JSON, documents, and rich content.

Cortex functions can be combined within SQL pipelines—for example, using PARSE_DOCUMENT to extract data, followed by CLASSIFY_TEXT to categorize extracted fields, and then SUMMARIZE to produce narrative summaries. This composability enables sophisticated multi-stage workflows for document intelligence, support ticket analysis, compliance automation, and semantic enrichment.

They arenot limited to a single domain; Cortex functions apply to finance, retail, support operations, healthcare, legal analysis, and general AI-assisted data processing.

Thus, Cortex functions are versatile, multi-domain, multi-language tools integrated with Snowflake’s compute and governance layers.

====================================================

Snowflake Notebooks provide an integrated, interactive development environment that enables users to write SQL and Python code, visualize results, and document analysis—all in the same workspace. They support data visualizations, charts, tables, and Streamlit-based interactions. Notebooks run entirely within Snowsight and require no manual hardware configuration. They are not CLI-based and support much richer displays than plain text. This integration streamlines the entire data exploration, transformation, and modeling workflow.

=======================================

In Snowflake’s Secure Data Sharing model, the accountsharingdata is known as thedata provider. Providers create shares that include selected objects (tables, views, secure views, etc.) and grant access to consumer accounts.

Adata consumerreceives the shared data.Data readeranddata administratorare not defined terms in Snowflake’s sharing architecture.

==================

ThePARSE_JSON()function converts a valid JSON string into aVARIANTvalue. This allows Snowflake to store and query nested, hierarchical data using dot and bracket notation.

Example:

SELECT PARSE_JSON( ' { " name " : " John " , " age " :30} ' ) AS data;

After parsing, fields can be accessed like:

data:name or data[ ' age ' ]

Incorrect options:

TO_JSON()converts VARIANT to a JSON string (opposite direction).

CONVERT_JSONandEXTRACT_JSONare not Snowflake functions.

PARSE_JSON is essential for dynamically loading, transforming, or analyzing JSON content from files, streams, or external applications.

====================================================

Schemas in Snowflake serve asorganizational containersfor database objects such as tables, views, sequences, stages, file formats, and functions. They existwithin a database, forming Snowflake’s logical hierarchy: account → database → schema → object.

Schemas are not limited to a single role; access depends on granted privileges. Schemas are not immutable—administrators can rename, alter, or drop them. Schemas do not contain virtual warehouses; warehouses exist independently at the account level.

Thus, the correct characteristics are that schemas organize objects and are contained within a database.

==================

Snowflake Marketplace serves as a centralized catalog where providers publishready-to-use datasets, models, and applicationsacross industries. It includes:

Public datasets, such as demographics, weather, healthcare statistics, and government data.

Financial market datasets, including equities, commodities, macroeconomic indicators, and pricing feeds from major financial data vendors.

Third-party datasetssupplied by external organizations specializing in geospatial intelligence, retail analytics, marketing insights, healthcare, census data, and more.

“Proprietary datasets” are internal to organizations and typically not published on the public Marketplace unless the owner chooses to list them. Marketplace listings are curated and governed.

====================================================

The FLATTEN table function expands semi-structured data such as arrays or objects stored in VARIANT columns. Its output includes several metadata columns, and the most important among them isVALUE, which contains the extracted element from the array or object. VALUE is the column that developers typically join on or reference to access the flattened data.

In addition to VALUE, FLATTEN returns KEY (for object keys), INDEX (position in arrays), SEQ (internal ordering), and PATH (hierarchical reference path). These other columns support analytical or transformation logic, but VALUE remains the primary output used for downstream SQL.

The other listed options—OUTPUT, ROWNUM, and LEVEL—are not part of the FLATTEN function’s output schema. ROWNUM and LEVEL may appear in other SQL contexts but not in FLATTEN results. Therefore, the correct answer is VALUE.

==================