Snowflake CLI: Where It Wins Over Snowsight

2026-06-22 · Source: Data Engineering on Medium · Field: Technology & Digital — Software Development & Engineering, Data Science & Analytics, Artificial Intelligence & Machine Learning · Depth: Intermediate, medium

Summary

The Snowflake CLI (the "snow" command) offers a robust development environment for Snowpark Python UDFs and Streamlit apps, addressing limitations of the Snowsight web interface for complex projects. While Snowsight excels at ad hoc SQL and data exploration, the CLI provides essential features like version control, local testing, and repeatable builds. The article demonstrates this by detailing the process of building, testing, and deploying a Python UDF for phone number cleaning, which requires the `phonenumbers` package. Key steps include installing the CLI, scaffolding a project with `snow init --template example_snowpark`, defining the UDF and its dependencies in `snowflake.yml` and `requirements.txt`, performing local unit tests with `pytest` in 16 milliseconds, and deploying via `snow snowpark deploy`. This workflow enables environment swapping via named connection profiles and ensures dependencies are resolved against Snowflake's Anaconda channel, contrasting with Snowsight's manual, UI-driven approach.

Key takeaway

For MLOps Engineers or Software Engineers shipping Snowpark pipelines or Streamlit apps, you should adopt the Snowflake CLI for your development workflow. It enables critical practices like local unit testing, declarative dependency management via `requirements.txt`, and version control, which Snowsight lacks. This approach significantly reduces debugging cycles and ensures repeatable, scalable deployments across environments, making it indispensable for collaborative projects and CI/CD integration.

Key insights

The Snowflake CLI provides a robust, scalable development environment for Snowpark and Streamlit, surpassing Snowsight for complex, collaborative projects.

Principles

• Development environments require version control.
• Local testing reduces debug cycles and costs.
• Declarative manifests enable repeatable deployments.

Method

The method involves installing the CLI, scaffolding a project with `snow init`, defining functions and dependencies in `snowflake.yml` and `requirements.txt`, local unit testing with `pytest`, and deploying via `snow snowpark deploy`.

In practice

• Use `snow init` for project scaffolding.
• Define dependencies in `requirements.txt`.

Topics

• Snowflake CLI
• Snowpark
• Python UDFs
• Streamlit
• Local Testing
• Dependency Management
• CI/CD

Best for: Machine Learning Engineer, MLOps Engineer, Software Engineer

Related on AIssential

• Snowflake CLI: Where It Wins Over Snowsight-Part2(With Git&GitHub) — Automating Snowpark UDF deployments via Git and GitHub Actions enhances structure, safety, and collaboration.
• Integrating Streamlit with Snowflake for Live Cloud Data Apps (Part 1) — Integrate Streamlit with Snowflake for dynamic, cloud-ready data applications using secure credentials and modular query logic.
• Google Launches Colab CLI for Developers, Automation, and AI Agents — The Google Colab CLI enables direct terminal interaction with remote Colab runtimes for developers and AI agents.
• Databricks vs Snowflake: Choosing the Right Platform for Modern Data and AI Workloads — Databricks excels in AI/ML and data engineering, while Snowflake dominates SQL-based analytics and business intelligence.
• Connections, Roles, and Warehouses: Getting CoCo Desktop Production-Ready from Day One — Proper Snowflake CoCo Desktop connection setup, including authentication and configuration files, is crucial for reliable agent operation.
• Day 09 of MLOps: From Localhost to Production-Ready ML Deployment on AWS — Production ML deployment on AWS requires specific tools and practices for scalability, reliability, and automation beyond localhost.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Data Engineering on Medium.