Skip to content

Snowflake-Labs/sf-cheatsheets

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

70 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Snowflake Developer Cheatsheets

Single-file, copy-paste-ready quick references for Snowflake features developers reach for daily. Not docs. Not tutorials. The 5–10 things you actually use, verified once, formatted to fit in your peripheral vision.

Community resource — not official Snowflake documentation. For authoritative reference, see docs.snowflake.com.

Cheatsheets

Cheatsheet What it covers
Executor Role EXECUTE AS OWNER/CALLER, task/dynamic table execution context, masking policy context functions, troubleshooting permission errors
Snowflake CLI snow CLI commands, connection management, SQL execution, object management
SPCS Snowpark Container Services setup, compute pools, service specs, image registry
Warehouses Warehouse sizing, auto-suspend, multi-cluster, resource monitors, cost controls
Iceberg Iceberg table creation, catalog integration, storage options, time travel
Snowflake Postgres Snowflake Postgres instances, pg_lake, CLD setup, managed storage
GitHub Authentication OIDC-first auth, GitHub Actions integration, PAT fallback
RAG Evaluation RAGAS metrics, Cortex evaluation functions, evaluation dataset setup
Cortex Code Cortex Code CLI commands, sessions, skills, context management

Why a cheatsheet, not just asking an agent?

A cheatsheet is the answer you already verified, ready the moment you need it. An AI agent is the answer you hope is right, arriving after a round-trip.

Agents re-generate every time. That means hallucination risk (confidently wrong flags, deprecated syntax, commands that don't exist), 2–10 second round-trip latency, and no offline use. A cheatsheet has been manually verified once, survives a plane flight or an air-gapped VPC, and can be pinned in Slack, bookmarked, or pasted into onboarding docs. It also builds mental models in a way that prompting doesn't.

The mistake is using an agent as a faster cheatsheet. Agents win on reasoning over your specific context — debug this error, explain why this query is slow, generate this pipeline from a description. That is not what a cheatsheet is for.

Issues

File all issues at https://github.com/Snowflake-Labs/sf-cheatsheets/issues

License

Apache License

About

Developer quick-reference cheatsheets for Snowflake products, CLIs, and features

Topics

Resources

License

Stars

10 stars

Watchers

2 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors