Snowflake Core

Before you create anything you need to know where it lives. You compare editions and cloud regions, see how organizations hold accounts, and understand the object hierarchy and the parameters that set behavior at each level.

• Editions and what each includes
• Cloud providers and regions
• Organizations and multiple accounts
• The hierarchy of objects
• Parameters at account, object and session level
• Snowsight, SnowSQL and the CLI at a glance

You build A summary sheet matching the edition and region choices to a company’s stated needs.

Access in Snowflake is built from roles. You learn the system roles and how they inherit, how privileges are granted and who owns objects, and set up simple access for a small team.

• System roles and the role hierarchy
• Privileges and grants
• Ownership of objects
• Users and authentication basics
• Network policies at a glance
• Switching roles and secondary roles

You build A role and grant setup for a small team, tested by switching between roles.

Warehouses supply compute and drive most of the bill. You choose sizes, set auto-suspend and auto-resume, use multi-cluster scaling for concurrency, and set resource monitors to keep spend in view.

• Warehouse sizes and credit use
• Auto-suspend and auto-resume
• Multi-cluster warehouses and scaling policies
• Resource monitors
• Compute, storage and serverless costs
• Choosing a warehouse for a workload

You build A warehouse setup for two workloads with suspend rules and a resource monitor.

Data has to arrive before it can be queried. You use internal and external stages and file formats, load in bulk with COPY INTO, load continuously with Snowpipe, and read load metadata to avoid duplicates.

• Internal and external stages
• File formats
• COPY INTO for bulk loading
• Snowpipe for continuous loading
• Validation and error options
• Load metadata and avoiding reloads

You build A bulk load and a continuous load into the same table, with a check on what loaded.

Data also has to leave and change shape. You unload tables to files, transform data while loading and afterward, choose among table types and views, and meet streams and tasks as the basis of automated transformation.

• Unloading tables to files
• Unload options and file formats
• Transforming during load
• Permanent, temporary, transient and external tables
• Views, materialized views and dynamic tables
• Streams and tasks in outline

You build An unload to files and a view that reshapes the loaded data for reporting.

Most real data is not a tidy table. You work with Snowflake’s data types, store JSON and similar formats as VARIANT, query nested values with paths and FLATTEN, and let schema detection do the first pass.

• Data types and their behavior
• VARIANT, OBJECT and ARRAY
• Paths and FLATTEN for nested data
• JSON, Parquet and Avro files
• Schema detection
• Casting and handling missing values

You build A query set that flattens a nested JSON feed into typed relational columns.

Documents, images and audio sit beside tables in the same platform. You catalog files with directory tables, give controlled access through URLs, and process files with functions instead of moving them out.

• What counts as unstructured data
• Directory tables and file metadata
• Scoped, file and pre-signed URLs
• Processing files with functions
• Access control for files

You build A stage of files catalogued by a directory table and reachable through controlled URLs.

A slow query has a cause you can see. You read query history and the query profile, understand pruning and clustering, know what search optimization and acceleration offer, and watch usage through the standard views.

• Query history and the query profile
• Pruning and clustering
• Search optimization and query acceleration
• Materialized views and result reuse
• Spilling and queuing
• Usage views in ACCOUNT_USAGE and INFORMATION_SCHEMA

You build A diagnosis of one slow query with the profile evidence and a chosen fix.

Data protection in Snowflake is mostly built in, but you have to know how it works. You use Time Travel and cloning to recover and test, understand what Fail-safe is for, and see how encryption, masking and replication add protection.

• Time Travel and its retention
• Fail-safe and what it is not for
• Cloning and shared storage
• Encryption and key management at a glance
• Masking and row access policies at a glance
• Replication and failover at a glance

You build A recovery exercise that restores a dropped table and tests a change on a clone.

Data becomes more valuable when it is shared without being copied. You create secure shares, serve consumers with and without Snowflake accounts, publish listings, and understand what a consumer can and cannot do.

• Secure shares and consumers
• Reader accounts
• Listings and the Marketplace
• Sharing across regions and clouds
• What consumers can and cannot do with shared data

You build A secure share of one table, consumed from a second account.

People and applications reach Snowflake in many ways. You use Snowsight and SnowSQL, choose drivers and connectors for code and BI tools, understand authentication for clients, and see how private network paths work.

• Snowsight and SnowSQL
• JDBC, ODBC and language drivers
• The Python connector and Snowpark
• Partner and BI tool connections
• Authentication options for clients
• Network paths and private connectivity at a glance

You build The same query run through three clients with a note on how each authenticated.

You build a small working warehouse from scratch: accounts and roles, warehouses, loading, semi-structured and file data, sharing and protection. Then you use what you learned to explain each design choice.

• Planning the objects, roles and warehouses
• Loading structured, semi-structured and file data
• Tuning and monitoring the workload
• Protecting and sharing the result
• Explaining the design choices

You build A small, documented warehouse with loaded data, access rules, a share and a tuned query.