/
Loading...
Loading...
Chonks are the building blocks of your dbt project. Each chonk type serves a specific purpose in the data transformation pipeline.
DataChonk follows the principle that dbt projects should have a clear, predictable structure. Each layer of transformation has a specific purpose:
Define your raw data sources with freshness checks and column documentation.
Clean, rename, and cast columns from your sources into a consistent format.
Build dimension tables that represent business entities like customers or products.
Create fact tables that capture business events and transactions.
Define semantic layer metrics with dimensions and calculations.
| If you need to... | Use this chonk |
|---|---|
| Document a raw table | Source |
| Clean and rename columns | Staging |
| Build a dimension table | Entity |
| Create a transactional fact | Fact |
| Define a business metric | Metric |
Chonks can reference each other to build a complete data pipeline:
Chonk automatically understands these relationships and will suggest appropriate references when generating code.