Why Document Your dbt™ Project?Thorough documentation is a critical component of any successful dbt™ project. Well-documented models and columns provide numerous benefits:
- Knowledge Sharing: Makes your data models accessible to stakeholders across your organization
- Onboarding: Helps new team members understand your data models quickly
- Self-Service Analytics: Empowers business users to explore data confidently
- Governance: Helps track sensitive data and ownership for compliance
Creating and Editing Documentation
Step-by-Step Guide
- Click Catalog to view and edit your dbt™ project’s documentation in real time. By default, you’ll see a catalog preview; expand it to utilize all catalog features.
- Make your desired edits:
-
Model Classification
- Edit existing classifications in the meta section
- Add new classifications
-
Model Description
- Type your model description manually
- Use the text formatting toolbar for better readability
- Alternatively, use DinoAI to autogenerate an in-depth, context-specific description
-
Model Classification

- Column Details
- Edit individual column descriptions
- Add or modify column classifications (e.g.,
[primary_key, true])
- Click Save Changes to apply your edits
- When finished, commit and push your git branch
- The edits you made in the UI will be reflected in the corresponding
.ymlfile
- The edits you made in the UI will be reflected in the corresponding
Available Documentation Features
Refresh your development catalog
Click Refresh in the header row of the Catalog tab, beside the Description and Columns tabs, to rebuild your development catalog. Paradime runsdbt docs generate on your current branch with your development credentials, then builds the catalog from the result. The button reads Refreshing while it runs. You need editor access to refresh, and you can run one refresh at a time.
You can keep working while it runs. The status line beside the button shows each stage:
- Refreshing · starting…
- Refreshing · running dbt™ docs generate, with the time so far. After your first refreshes, it adds how long a refresh usually takes, for example usually about 3 min.
- Refreshing · building the catalog…
- Last updated just now, with a notification that the catalog was refreshed and how long it took.
When a refresh fails
The status line reads Refresh failed, with the time of the last good build if there is one. The failure leads with its cause and a note on whether it needs a change on your side or should clear on the next refresh.- If the tab has docs to show, the failure is one line above them, with Show error to open the dbt™ output.
- If there is nothing else to show, the failure fills the tab, with the dbt™ output open.
When there are no docs to show
If the Catalog tab has nothing to display, it says why:- Nothing is open in the editor: No file is open right now. Open a dbt™ model to see its docs.
- The open file isn’t one the catalog documents, for example a YAML file: the tab reminds you that the catalog covers dbt™ models, sources, and seeds.
- Your development catalog hasn’t been built: Your catalog hasn’t been built yet. Click Refresh to build it. The first build takes a few minutes.
- The first build is running: Building your catalog. You get a notification when it’s ready.
- The model isn’t in your development catalog yet, for example because it’s new: No docs for the file. Click Refresh to pick it up.
- The lookup itself failed: Could not load docs, with the error underneath.
The Catalog is built using the current version of your dbt™ repository from the git branch currently selected in the Paradime IDE and Information schema from your Development Environment default connection.
For maximum efficiency, use Dino AI to autogenerate comprehensive data documentation, then refine it to meet your specific needs.