Skip to content

Snowflake Connector

Key Concepts

The Snowflake Connector block enables secure, scalable, and intelligent integration between your workflow and a Snowflake data warehouse. It is designed for high-performance SQL-based data access, leveraging structured table indexing and intelligent prompt generation to allow smooth interaction between large datasets and large language models (LLMs).

This block includes workflow nodes for everything from connection initialization to query execution and response generation. It's suitable for analytical workloads, operational dashboards, and intelligent automation tasks that require direct access to Snowflake data.

Key Definitions

Term Definition
Snowflake Connector A block that connects to a Snowflake data warehouse and enables LLM-powered querying through workflows.
Warehouse A virtual compute cluster used by Snowflake to process queries.
Schema The structure that defines tables, views, and other objects within a Snowflake database.
Workflow Node A defined processing step within the Snowflake workflow (e.g., initialize, index, execute).

Setup Guide: Configuring a Snowflake Connection

To set up a connection to your Snowflake environment, complete the following fields in the “Set a new Snowflake connection” screen:

1. Account

  • The full Snowflake account name or URL.
  • Example: xyz12345.eu-central-1.snowflakecomputing.com

2. User

  • The username authorized to access the Snowflake account.

3. Password

  • The corresponding password for the Snowflake user account.
  • Use the visibility toggle to verify during entry.

4. Database

  • Name of the Snowflake database to connect to.
  • Make sure the database exists and is accessible by the user.

5. Schema

  • Name of the schema within the selected database.

6. Warehouse

  • Specify the warehouse to use for executing SQL operations.
  • Ensure the warehouse is configured and active.

7. Role

  • Enter the role (e.g., SYSADMIN, ANALYST) that defines access permissions.

8. DB Tables

  • Enter the name(s) of the table(s) you want to query.
  • Click Add to DB Tables to include them in the configuration.
  • This scoping helps optimize performance and maintain data security.

9. Description

  • Provide a description to document the purpose of this connection.

  • Example: Data warehouse for financial analytics used by BI team.

  • This is optional in custom workflows but recommended for clarity and routing support.

10. Save

  • Click Save to finalize the connection and return to the workflow editor.

Workflow Nodes

The Snowflake Connector block provides a robust set of workflow nodes for retrieving and processing data efficiently:

initialize_snowflake_workflow (First Node)

  • Sets up the workflow environment by initializing credentials, configuration, and resources for Snowflake access.

index_all_snowflake_tables

  • Scans and indexes all specified tables in the Snowflake database.
  • Enables efficient retrieval and structure awareness for downstream operations.

prepare_snowflake_table_retrieval

  • Analyzes the selected tables and schemas to ensure compatibility with workflow logic.

get_snowflake_vector_index_dict

  • Generates a dictionary of vector-based indexes for the specified tables.
  • Supports advanced querying with semantic or similarity-based retrieval techniques.

retrieve_snowflake_table_context

  • Retrieves relevant context from selected Snowflake tables based on a given query or condition.

generate_snowflake_prompt

  • Creates a structured and context-aware prompt to guide the LLM in generating valid SQL queries.

parse_snowflake_from_response

  • Extracts and validates the SQL query from the LLM’s response.
  • Ensures syntax is valid and conforms to Snowflake’s SQL dialect before execution.

setup_snowflake_retriever

  • Configures the connection to execute SQL operations within the Snowflake environment securely and efficiently.

execute_snowflake_and_generate_response (Last Node)

  • Executes the validated SQL query on the Snowflake warehouse.
  • Processes and formats the result for further use or direct output to the user or system.

Best Practices for Successful Snowflake Integration

  • Use Valid Credentials: Verify your account, role, and schema access in Snowflake.
  • Use Scoping: Restrict access to specific tables using the DB Tables field for performance and security.
  • Assign Roles Carefully: Ensure the user role has the correct level of permissions for read/query operations.
  • Enable Warehouses: Ensure that the selected warehouse is running or auto-resumes on demand.

Example Use Case

You want to enable business analysts to query customer churn metrics directly from Snowflake via a natural language interface:

  1. Configure the Snowflake Connector to the customer_analytics database with tables like churn_metrics and user_activity.
  2. When a user submits a query like “Show churn rate for Q1 by subscription type”, the workflow:
  3. Uses generate_snowflake_prompt to build a SQL query,
  4. Parses the query with parse_snowflake_from_response,
  5. Executes it with execute_snowflake_and_generate_response,
  6. Returns the results in a readable format to the user or connected application.