Free SnowPro Core Exam Flashcards

Snowflake separates storage, compute, and cloud services. Storage holds managed data, virtual warehouses run query work, and cloud services coordinate authentication, optimization, metadata, transactions, and result caching.

The control layer that parses SQL, optimizes queries, manages metadata, checks authentication, and coordinates transactions. It is separate from the warehouse that performs query execution.

The virtual warehouse layer that supplies processing power for SQL execution, loading with warehouse compute, transformations, and other workload operations.

The managed layer where Snowflake stores table data in compressed columnar micro-partitions. Users do not manually size files, indexes, or partitions for internal table storage.

A database contains schemas, and schemas contain objects such as tables, views, stages, functions, streams, and tasks. Fully qualifying names helps avoid ambiguity across environments.

Snowflake editions determine access to some security, governance, continuity, and warehouse features. Exam questions often test whether the feature is a platform concept or edition-dependent capability.

A named compute cluster used to run queries and data operations. It is independent from databases, so any permitted warehouse can query any permitted data object.

Warehouse size controls the compute resources available to each cluster. Increasing size can help resource-heavy queries, while concurrency problems often call for queue analysis or multi-cluster behavior.

A cost-control setting that stops a warehouse after it has been idle for the configured interval. It reduces credit use when teams forget to shut down compute.

A warehouse setting that starts suspended compute automatically when a submitted statement needs it. It improves usability but can surprise teams if warehouse access is too broadly granted.

A warehouse configuration that can add clusters to handle concurrent query demand. It addresses queue pressure, not poor SQL design or missing pruning.

A credit governance tool that tracks consumption and can notify users or suspend warehouses when configured thresholds are reached.

An immutable internal storage unit that records metadata such as value ranges. Snowflake uses this metadata to skip irrelevant data during scans.

The process of skipping micro-partitions that cannot contain rows needed by a filter. Effective pruning usually depends on how data is naturally ordered or clustered relative to common predicates.

A table definition choice that guides Snowflake to organize data around selected columns. It is useful for large tables with repeated selective filters that do not prune well naturally.

A metadata-based copy that initially shares existing storage with the source object. Changed data later uses copy-on-write behavior, so clones are fast and storage-efficient for testing or recovery workflows.

A recovery and historical-query feature that lets users access data as it existed within the configured retention window. It supports query, clone, and undrop workflows.

A Snowflake-managed recovery period after Time Travel retention ends. Users do not query it directly; recovery requires Snowflake support and is meant for exceptional data recovery.

The standard persistent table type for durable managed data. It supports Time Travel and Fail-safe according to the account and object retention settings.

Transient tables persist until dropped but skip Fail-safe, while temporary tables are scoped to a session. Both are useful when recovery requirements are lower than for permanent data.

A Snowflake SQL clause used to filter rows after window functions are calculated. It avoids wrapping window-function logic in an extra subquery just to filter ranked rows.

A window function that assigns a unique sequence within each partition. It is commonly used to pick one row per group, remove duplicates, or implement page-like result slices.

Both assign the same value to tied rows. RANK leaves gaps after ties, while DENSE_RANK continues with the next consecutive rank.

A named result set defined with WITH for use inside a single SQL statement. CTEs improve readability and can help express multi-step transformations clearly.

A SQL command that applies conditional inserts, updates, or deletes from a source into a target. It is a common pattern for incremental loads and dimension maintenance.

A Snowflake data type for semi-structured values such as JSON-like objects and arrays. It lets teams load nested data first and extract paths during query or transformation steps.

A table function that expands arrays or objects from semi-structured data into rows. It is often paired with LATERAL so each source row can produce multiple output rows.

Snowflake grants privileges to roles, then grants roles to users or other roles. This separates job responsibilities from individual users and supports least-privilege access.

When one role is granted to another, the parent role inherits the child role's privileges. This allows teams to build layered access models instead of duplicating every grant.

A container access privilege. To query a table, a role needs access through the warehouse, database, and schema plus the object-level privilege such as SELECT.

The controlling privilege for an object. The owner can manage the object and grant privileges, so transferring ownership is a sensitive administrative operation.

A powerful privilege for granting and revoking privileges beyond normal object ownership paths. It should be limited to trusted administrative roles.

Multi-factor authentication adds another verification step beyond a password. It is especially important for administrative users and broad-privilege roles.

An authentication method that uses key material instead of an interactive password flow. It is commonly used for service accounts, connectors, and automated jobs.

A control that restricts logins by allowed or blocked network ranges. It adds a perimeter-style check before users or services can connect to the account.

A column-level governance feature that returns different values depending on policy logic and user context. It protects sensitive values without creating separate copies of the table.

A governance policy that filters which rows a role can see. It is useful for tenant, region, department, or other context-based data access rules.

A view designed to protect underlying query logic and support secure sharing patterns. Use it when exposing the SQL definition would reveal sensitive rules or data structure.

An account usage view for auditing which objects and columns were accessed. It helps with governance investigations, compliance reviews, and least-privilege cleanup.

A location used to hold files before loading into Snowflake or after unloading from Snowflake. Stages can be internal to Snowflake or point to external cloud storage.

A reusable object or inline setting that tells Snowflake how to interpret staged files, including delimiters, headers, compression, and semi-structured format handling.

The bulk-load command that reads files from a stage into a target table. It pairs with file formats, validation, and error-handling options.

The unload pattern that writes query or table results from Snowflake to staged files. It is used for exports, downstream exchange, and archival workflows.

A COPY option that controls what happens when a load encounters bad rows or files. The right setting depends on whether data quality or uninterrupted ingestion is the priority.

A continuous ingestion service that loads new staged files as they arrive. It is best for event-driven file ingestion rather than large manual batch runs.

A diagnostic view of query execution that shows operator costs, scan behavior, pruning, spilling, and bottlenecks. It helps decide whether to tune SQL, data layout, or warehouse resources.

A cloud-services cache that can return results for repeated identical queries when the underlying data and context still qualify. It avoids warehouse work for eligible repeat reads.

A view whose results are stored and maintained by Snowflake for faster reads. It can improve repeated query patterns but adds maintenance and storage considerations.

A collaboration feature that lets a provider share live data with consumers without copying it into the consumer account. Consumers query shared objects using their own compute.

Streams track table changes for incremental processing, while tasks run scheduled or dependent SQL work. Together they support automated ELT pipelines inside Snowflake.