COF-C03 cheat sheet

Exam at a glance

Domain weightings

How much of the exam each domain covers. Spend your study time in proportion — the heavier the domain, the more questions you'll see.

Key terms

Virtual warehouse

A virtual warehouse is a cluster of compute resources in Snowflake that executes queries, loads data, and performs DML. Warehouses can scale up (larger size) for complex queries or scale out (multi-cluster) for concurrency, and they auto-suspend and auto-resume to control cost.

Micro-partition

A micro-partition is the small, contiguous unit of storage (roughly 50-500 MB uncompressed) into which Snowflake automatically divides table data. Snowflake stores columnar metadata per micro-partition to enable pruning, which skips partitions that cannot match a query.

Time Travel

Time Travel is a Snowflake feature that lets you query, clone, or restore data as it existed at a point in the past within a defined retention period. It supports recovering dropped objects and comparing historical data, and precedes the Fail-safe period.

Fail-safe

Fail-safe is a non-configurable 7-day period after Time Travel during which Snowflake can recover historical data only through Snowflake support. It is a disaster-recovery measure and is not intended for routine data access by users.

Zero-copy cloning

Zero-copy cloning is a Snowflake operation that creates a copy of a database, schema, or table without duplicating the underlying storage. The clone shares existing micro-partitions and only consumes additional storage as the clone or source diverges.

Stage

A stage is a Snowflake object that references a location where data files are stored for loading or unloading. Stages can be internal (user, table, named) or external (cloud storage buckets), and are used with the COPY INTO command.

Snowpipe

Snowpipe is Snowflake's continuous, serverless data-ingestion service that loads files as soon as they arrive in a stage. It supports auto-ingest via cloud notifications; classic Snowpipe bills for the compute used plus a per-file overhead (the Snowpipe Streaming variant bills differently).

RBAC

RBAC (role-based access control) is Snowflake's security model in which privileges are granted to roles, and roles are granted to users or other roles. Access is determined by the active role, and system-defined roles form a hierarchy such as SYSADMIN, SECURITYADMIN, and ACCOUNTADMIN.

Result cache

The result cache is a Snowflake caching layer that returns the stored result of an identical query for 24 hours without using a virtual warehouse. It applies when the query and underlying data are unchanged, saving compute cost.

Clustering key

A clustering key is a subset of columns that Snowflake uses to co-locate related rows in the same micro-partitions. Defining one improves pruning and query performance on very large tables, and automatic clustering maintains it over time.

Secure data sharing

Secure data sharing is a Snowflake capability that provides live, read-only access to data across accounts without copying it. Providers create shares that consumers access directly, so there is no data movement and no additional storage on the consumer side.

Snowflake Marketplace

The Snowflake Marketplace is a platform where data providers publish listings that consumers can access through secure data sharing. It supports public and private listings and lets consumers query shared datasets without ETL.

Snowsight

Snowsight is Snowflake's web interface for running queries, exploring data, and administering the account. It provides worksheets, charts, and usage/monitoring views, and is the primary UI alongside the Snowflake CLI and drivers.

Cortex

Snowflake Cortex is the suite of AI and machine-learning features in the Snowflake AI Data Cloud. It offers LLM functions and ML capabilities that run directly on governed data inside Snowflake, without moving data out.

Query Profile

The Query Profile is a Snowflake diagnostic view that shows the execution plan and per-operator statistics for a query. It reveals pruning effectiveness, spillage to disk, and expensive operators, guiding performance optimization.