SOL-C01 Snowflake SnowPro Associate: Platform Certification Exam Free Practice Exam Questions (2025 Updated)

Snowflake’s architecture separates storage and compute, enabling:

• Compute scaling (up/down or multi-cluster) without changing storage

• Storage expansion without affecting compute

• Cost optimization by paying for compute only when needed

This separation does not impact governance, latency, or data loading requirements.

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

Snowflake supports several numeric data types, includingINTEGER(and its aliases INT, BIGINT, SMALLINT, TINYINT) andFLOAT(a floating-point type). INTEGER represents exact whole numbers, and FLOAT supports approximate numerical values with fractional components. BOOLEAN is a logical data type used for TRUE/FALSE values, not numeric. VARCHAR is a string data type, also not numeric. Thus, INTEGER and FLOAT are the correct numeric data types supported by Snowflake.

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

Scaling in/outrefers tohorizontal scaling, which is available only inmulti-cluster warehouses. When Snowflake scalesout, additional compute clusters are added automatically to handle high concurrency, preventing query queuing. When load decreases, Snowflake scalesin, reducing cluster count to save compute credits.

This differs fromscaling up/down, which refers to changing warehouse size (X-Small → Small → Medium). Vertical scaling increases compute resources for a single cluster but does not improve concurrency for simultaneous workloads.

Scaling in/out is valuable for BI dashboards, workloads with many simultaneous users, peak-hour query loads, and applications requiring consistent performance under high concurrency.

Changing regions is an account-level configuration and unrelated to warehouse scaling.

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

Snowflake supports bothTOPandLIMITas valid SQL mechanisms for restricting the number of rows returned by a query. They function similarly but are used in different syntactic positions.

LIMITappears at the end of the query:

SELECT * FROM my_table LIMIT 10;

This tells Snowflake to return only the specified number of rows.

TOPis used in the SELECT clause:

SELECT TOP 10 * FROM my_table;

Both approaches are equivalent in functionality and are supported to maintain compatibility with ANSI SQL and other SQL dialects.

The other options are not correct row-limiting mechanisms:

•SELECTis a required keyword but does not limit rows on its own.

•GETis not a SQL keyword for row restriction; GET is used for downloading files from internal stages.

•ROW_NUMBERis a window function used for ranking rows, not constraining row count.

Therefore, TOP and LIMIT are the only correct SQL keywords in Snowflake for controlling how many rows are returned by a query.

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

COPY INTO <title> loads data from staged files (internal or external stages) into a Snowflake table.

It supports pattern matching, file formats, error-handling modes, and parallel loading.

It does not move files, does not load to a warehouse (warehouses only supply compute), and table-to-table copying uses INSERT INTO…SELECT.

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

The Public schema is created by default within every Snowflake database and serves primarily as a shared workspace that is accessible to all users who have the necessary privileges. One of its main roles is to act as adefault schemafor objects created without explicitly specifying a schema name. This ensures usability and prevents object creation errors when users forget to define a schema. Additionally, the Public schema is commonly used to storeshared or common resourcesthat multiple users or teams need to access. However, it should not be used for sensitive or user-specific data because access can be broad. It does not store account-level metadata; that information resides in INFORMATION_SCHEMA views or account-level metadata functions. Overall, the Public schema acts as both a convenient default location and a shared-access area within a Snowflake database.

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

The correct SQL command for creating a new database in Snowflake isCREATE DATABASE database_name;. This initializes a new logical container for schemas and objects. USE DATABASE switches the active database context but does not create one. Snowflake does not support the shorthand CREATE DB or USE DB syntax. Therefore, CREATE DATABASE is the correct and only valid command for creating a database.

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

As a fully managed cloud data platform, Snowflake is responsible for infrastructure provisioning, hardware, software installation, platform upgrades, scaling, and internal metadata management such as micro-partitions and statistics. Customers do not manage physical hardware or install Snowflake software.

Customers are responsible for their data and its lifecycle within Snowflake. This includes loading data into tables from internal and external sources, unloading data when required, organizing data structures (databases, schemas, tables), defining access controls, and managing how data is used, transformed, and governed. They design schemas and workloads but do not manage the underlying engine. Therefore, “Loading, unloading, and managing data” correctly describes the customer’s responsibility.

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

Standard permanent tables in Snowflake have a default1-day Time Travel retention period. During this period, users may query historical data, restore dropped tables, or clone past states. Higher editions allow extending this period up to 90 days. Temporary and transient tables have 0-day retention by default.

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

Snowflake’s architecture is built on three primary layers that separate and optimize storage, compute, and services. TheData Storagelayer stores all structured and semi-structured data in Snowflake’s optimized, compressed columnar format. Snowflake manages file organization, metadata, and optimizations automatically, ensuring efficient storage management. TheQuery Processinglayer is implemented through virtual warehouses that independently execute queries using MPP (massively parallel processing) compute clusters. These warehouses scale elastically based on workload requirements. TheCloud Serviceslayer manages metadata, authentication, access control, query optimization, and infrastructure orchestration. This separation allows independent scaling of compute and storage, optimizing performance while simplifying administration.

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

The COMPLETE function enables natural language generation within Snowflake Cortex. It accepts a prompt and produces a text-based completion using an underlying large language model (LLM). COMPLETE can generate single-turn or multi-turn responses, depending on how the prompt is structured. It is used for tasks such as drafting summaries, generating recommendations, writing structured output such as JSON, or extending existing text. It is not a classification tool—that role is served by CLASSIFY_TEXT—and it does not perform linguistic translation (TRANSLATE) or document extraction (PARSE_DOCUMENT). COMPLETE allows developers and analysts to embed AI-driven generative capabilities directly within SQL pipelines or Snowpark processes, eliminating the need for external LLM services.

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

TheCLASSIFY_TEXTfunction performscategory assignmentfor free-form text. Users provide a list of predefined labels, and the function predicts the most appropriate category using a Snowflake-optimized LLM. It is ideal for automatically tagging customer feedback, routing support tickets, labeling product reviews, classifying email content, and segmenting operational logs.

The function outputs both the predicted label and an associated confidence score, enabling downstream filtering, dashboarding, and automated processing.

Other options belong to different Cortex functions:

TRANSLATEhandles multilingual translation.

COMPLETEgenerates natural language completions.

PARSE_DOCUMENTextracts structured information from documents.

Thus, CLASSIFY_TEXT is task-specific and optimized specifically for classification workloads.

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

Standard Edition provides all essential Snowflake capabilities:

• SQL data warehousing

• Secure data sharing

• Time Travel (1 day)

• Internal/external stages

• Basic governance features

It does not include HIPAA/PCI compliance (Business Critical), nor unlimited Time Travel, nor advanced private connectivity.

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

CREATE STAGE is the correct Snowflake DDL command for creating a named internal stage. Stages hold files for loading/unloading. The command defaults to internal storage unless external parameters are provided.

Other options are not valid Snowflake syntax.

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

In Snowflake Notebooks, Python and SQL operate in an integrated environment. Python variable substitution is performed usingstring placeholders or f-string interpolationwithin SQL statements. This allows dynamic query construction where Python values are embedded directly into SQL code executed through the Snowpark session.

Example:

table_name = "CUSTOMERS"

session.sql(f"SELECT * FROM {table_name}").show()

This method enables parameterized, context-aware SQL generation—ideal for iterative development, automation, and pipeline construction. It avoids manual text editing and provides consistency across Python and SQL execution layers.

Incorrect responses:

Network settings and HTML are unrelated to variable handling.

Editing data files does not apply to dynamic query parameterization.

Variable substitution is essential for combining logic, iteration, and conditional flows within notebooks.

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

To monitor virtual warehouse performance in Snowsight, the correct approach is using theQUERY_HISTORYview or its Snowsight equivalent under the Activity section. QUERY_HISTORY provides detailed metrics on execution time, queueing, credit consumption, compilation time, and warehouse utilization for every executed query. By analyzing this information, users can diagnose issues such as concurrency bottlenecks, under-sized warehouses, long-running queries, or excessive I/O.

SHOW WAREHOUSES returns metadata about warehouse configurations (size, state, scaling policy) but does not provide performance data. LOAD_HISTORY tracks COPY INTO operations only and is not representative of general warehouse workload. DESCRIBE WAREHOUSE shows configuration details but no runtime or historical performance metrics.

Thus, the most accurate and complete way to assess warehouse performance—particularly execution delays, queuing, or excessive credit usage—is by using QUERY_HISTORY.

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

Asharein Snowflake is a securable object that contains all the metadata required to provide access to selected data—typically tables, views, and secure views—to other Snowflake accounts. The share object defines which objects are included and what privileges (e.g., SELECT, USAGE) consumers receive.

A share does not store data itself; it provides a reference to the provider’s underlying micro-partitioned storage, enabling real-time, zero-copy access.

Reader accounts (option A) can consume shares but do not define them.

Replication objects (option C) relate to database replication, not data sharing.

Option D is incorrect because Snowflake does not generate share-specific metrics objects.

Thus, the purpose of a share is to encapsulate all metadata required for secure, controlled data sharing between Snowflake accounts.

Snowflake supports zero-copy cloning at thedatabase,schema, andtablelevels. Cloning creates a new metadata reference to existing data without physically duplicating micro-partitions, significantly reducing storage cost and clone creation time. Cloning adatabasereplicates all schemas and tables within it at that moment in time. Cloning aschemareplicates all objects inside that schema, including tables and views. Cloning atablecreates an isolated version of that table for development, testing, or analysis. Importantly,rolescannot be cloned because they are account-level security objects, not data objects. Privileges must be manually granted or recreated for cloned objects. Cloning enables rapid, efficient creation of consistent environments for workload isolation and testing.

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

To create a new stage in Snowsight, users navigate through:Data → Database → Schema → Create Stage. Within the schema, selecting the "+" icon opens the option to create a stage. Users may configure internal or external stages, encryption settings, and directory table options. Other paths listed reference creating schemas rather than creating stages and are therefore incorrect.

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