Skip to main content
Column-Level Lineage Diff shows the blast radius of a pull request directly in the PR. It identifies structural column changes in your dbt™ models and posts an automated comment listing every impacted downstream node, so reviewers see exactly what’s affected before they merge.
Prerequisites
  • A connected git provider so Paradime can post PR comments.
  • A production connection with access to your sources and models, so Paradime can run information schema queries and build field-level lineage.
  • At least one Turbo CI schedule configured, which is required to generate field-level lineage.
Estimated time: 15 minutes.

Steps

1

Confirm your production connection and Turbo CI

Make sure your production connection is in place and at least one Bolt Turbo CI schedule runs successfully. Together they let Paradime build the field-level lineage the diff relies on.
2

Connect your git provider

Connect the git provider that hosts your dbt™ repository so Paradime can read pull requests and post comments back:
  • GitHub: install the Paradime GitHub App on your repository. Paradime posts the report as a PR comment.
  • GitLab: connect GitLab on gitlab.com or GitLab Self-Managed, with your dbt™ project as one of the connection’s projects. Paradime posts the report as one merge request note and updates it on each commit.
  • Azure DevOps or BitBucket: add a pipeline that asks Paradime for the report and posts it.
Follow the reference page for your provider in the section below.
3

Connect your BI tools

Connect BI tools like Looker, Tableau, or ThoughtSpot so the report also surfaces affected dashboards and other downstream assets, not only dbt™ models.
Connecting BI tools gives you the widest impact coverage, so reviewers see the full downstream picture.
4

Open a pull request with a column change

Open a test PR that renames, removes, or adds a column in a SELECT statement, for example renaming customer_name to full_name. These structural SELECT changes are what the diff detects.
The diff tracks structural changes to SELECT statements (column renames, removals, and additions). It does not track logic changes in WHERE, JOIN, GROUP BY, or ORDER BY clauses, or changes to how a column value is calculated.
5

Review the impact report

Paradime posts an automated comment on the PR listing all impacted downstream dbt™ models and BI nodes. Use it to assess the potential impact of the change before approving.
You’ll know it worked when your test PR receives an automated comment listing the affected downstream nodes. If the comment reports it can’t find your public git email address, complete the git provider setup for the user opening the PR on the provider reference page.

Set up your git provider

Finish the connection on the reference page for your provider. BI integrations for Looker, Tableau, and ThoughtSpot are covered from the same section.

GitHub

Native, through the Paradime GitHub App.

GitLab

Native, through the GitLab connection.

Azure DevOps

BitBucket

Next steps

Set up Turbo CI

Run and test dbt™ changes on every pull request.

Lineage Diff reference

Supported changes, BI integrations, and setup details.