DocsChonk TypesSource Chonks

DocsChonk TypesSource Chonks

Chonk Types4 min readBeginnerUpdated Jan 2026

Source Chonks

Source chonks define your raw data tables and create the foundation for your dbt project's data lineage. They're the entry point for all your data.

What They Generate

A Source chonk analyzes your raw tables and generates comprehensive documentation and data quality tests.

sources.yml

Source definition with table and column documentation

Freshness Tests

Automated checks for data staleness

Column Docs

AI-generated descriptions for each column

When to Use

Create a Source chonk when you want to establish the starting point for your data pipeline.

Document a raw table from your data warehouse
Set up freshness monitoring for data pipelines
Establish the starting point for your DAG
Create a reference for downstream staging models

Example Output

Here's what DataChonk generates when you create a Source chonk for e-commerce order data.

models/staging/sources.yml

version: 2

sources:
  - name: raw_ecommerce
    description: Raw e-commerce data from the production database
    database: analytics
    schema: raw
    
    freshness:
      warn_after: {count: 12, period: hour}
      error_after: {count: 24, period: hour}
    
    tables:
      - name: orders
        description: Raw orders table containing all customer orders
        loaded_at_field: _loaded_at
        columns:
          - name: order_id
            description: Unique identifier for the order
            tests:
              - unique
              - not_null
          - name: customer_id
            description: Foreign key to the customers table
            tests:
              - not_null
          - name: order_date
            description: Date when the order was placed
          - name: total_amount
            description: Total order amount in USD
          - name: status
            description: Current order status (pending, shipped, delivered, cancelled)

Configuration Options

Customize how your Source chonk generates code with these options.

Property	Type	Default	Description
include_freshness	boolean	true	Add freshness tests to detect stale data
include_tests	boolean	true	Add column-level tests (unique, not_null)
generate_descriptions	boolean	true	Use AI to generate column descriptions
freshness_warn	number	12	Hours before freshness warning
freshness_error	number	24	Hours before freshness error

Best Practices

One source per system

Group tables by their origin system (e.g., Stripe, Shopify, Salesforce). This makes it clear where data comes from.

Always set freshness

Freshness tests catch data pipeline issues early. Set warn/error thresholds based on your SLAs.

Document everything

Future you will thank present you. Add descriptions to sources, tables, and columns.

Test primary keys

Always add unique and not_null tests to ID columns. This catches data quality issues early.

Prompting Tips

Get better results from Chonk by being specific about your source requirements.

Example Prompt

"Create a source for our Stripe data. Include the charges, customers, and subscriptions tables. Set freshness to warn after 6 hours since we sync hourly. Add tests for primary keys and foreign key relationships."

Things to mention:

Which tables to include
The source system name
Any specific freshness requirements
Known data quality issues to test for

Source Chonks

Source chonks define your raw data tables and create the foundation for your dbt project's data lineage. They're the entry point for all your data.

What They Generate

A Source chonk analyzes your raw tables and generates comprehensive documentation and data quality tests.

sources.yml

Source definition with table and column documentation

Freshness Tests

Automated checks for data staleness

Column Docs

AI-generated descriptions for each column

When to Use

Create a Source chonk when you want to establish the starting point for your data pipeline.

Document a raw table from your data warehouse
Set up freshness monitoring for data pipelines
Establish the starting point for your DAG
Create a reference for downstream staging models

Example Output

Here's what DataChonk generates when you create a Source chonk for e-commerce order data.

models/staging/sources.yml

version: 2

sources:
  - name: raw_ecommerce
    description: Raw e-commerce data from the production database
    database: analytics
    schema: raw
    
    freshness:
      warn_after: {count: 12, period: hour}
      error_after: {count: 24, period: hour}
    
    tables:
      - name: orders
        description: Raw orders table containing all customer orders
        loaded_at_field: _loaded_at
        columns:
          - name: order_id
            description: Unique identifier for the order
            tests:
              - unique
              - not_null
          - name: customer_id
            description: Foreign key to the customers table
            tests:
              - not_null
          - name: order_date
            description: Date when the order was placed
          - name: total_amount
            description: Total order amount in USD
          - name: status
            description: Current order status (pending, shipped, delivered, cancelled)

Configuration Options

Customize how your Source chonk generates code with these options.

Property	Type	Default	Description
include_freshness	boolean	true	Add freshness tests to detect stale data
include_tests	boolean	true	Add column-level tests (unique, not_null)
generate_descriptions	boolean	true	Use AI to generate column descriptions
freshness_warn	number	12	Hours before freshness warning
freshness_error	number	24	Hours before freshness error

Best Practices

One source per system

Group tables by their origin system (e.g., Stripe, Shopify, Salesforce). This makes it clear where data comes from.

Always set freshness

Freshness tests catch data pipeline issues early. Set warn/error thresholds based on your SLAs.

Document everything

Future you will thank present you. Add descriptions to sources, tables, and columns.

Test primary keys

Always add unique and not_null tests to ID columns. This catches data quality issues early.

Prompting Tips

Get better results from Chonk by being specific about your source requirements.

Example Prompt

Things to mention:

Which tables to include
The source system name
Any specific freshness requirements
Known data quality issues to test for

Source Chonks

What They Generate

sources.yml

Freshness Tests

Column Docs

When to Use

Example Output

Configuration Options

Best Practices

One source per system

Always set freshness

Document everything

Test primary keys

Prompting Tips

Things to mention:

Related Documentation

Staging Chonks

Knowledge Sync

Source Chonks

What They Generate

sources.yml

Freshness Tests

Column Docs

When to Use

Example Output

Configuration Options

Best Practices

One source per system

Always set freshness

Document everything

Test primary keys

Prompting Tips

Things to mention:

Related Documentation

Staging Chonks

Knowledge Sync