Learning Objectives
By the end of this chapter, you will be able to:
- Outline Snowflake AI Data Cloud key features and benefits
- Describe the three layers of the Snowflake architecture (cloud services, compute / virtual warehouses, centralized storage)
- Identify Snowflake user interfaces, including Snowsight, Snowflake Notebooks, and Worksheets
- Use Snowsight to navigate the platform, load data, view query history, browse objects, and create objects
- Work with Snowflake Notebooks: run SQL and Python cells, interpret cell execution status, visualize data using Streamlit, and use Python variable substitution
- Describe the Snowflake object hierarchy, including databases, schemas, tables, and views
- Identify Snowflake data types, including numeric, string, semi-structured (VARIANT, OBJECT, ARRAY), GEOGRAPHY, GEOMETRY, and VECTOR
Executive Summary
- The Snowflake AI Data Cloud is a managed multi-cloud platform with three independently scaling layers: cloud services, compute, and centralized storage. Decoupling storage from compute is the architectural decision that drives almost every behavior the exam tests.
- Three interactive surfaces ship inside the Snowsight web UI: SQL Worksheets for ad-hoc queries, Snowflake Notebooks for cell-based SQL and Python work with optional Streamlit visualization, and Snowsight navigation pages for object browsing, data loading, and query history.
- The object hierarchy is account, then database, then schema, then schema-scoped objects (tables, views, stages, functions, procedures). Every fully qualified name uses dot notation in the form
database.schema.object. - Snowflake's data type catalog covers numeric (NUMBER, FLOAT), string (VARCHAR, STRING), date and time, semi-structured (VARIANT, OBJECT, ARRAY), geospatial (GEOGRAPHY, GEOMETRY), and the newer VECTOR type used by AI features such as the Cortex LLM functions.
Assumptions
- The reader has general SQL literacy and at least three months of hands-on Snowflake exposure, in line with the recommended profile for the exam.
- All examples assume the Snowsight web UI and a standard Snowflake account on AWS, Azure, or GCP. Edition gating is called out only where it changes the answer to a likely exam question.
- Terminology follows the official Snowflake documentation. Where a feature has a legacy and a current name, the chapter uses the current name and notes the legacy form once.