Transformations

Transformations turn the raw tables your data sources sync into models you can report on and send to other tools. You can transform in two ways:

1. In Weld, with the Weld Transform Engine: write SQL models in Weld's editor, or ask Ed to write them, and schedule them on their own or in an orchestration.
2. With another tool, such as dbt Cloud, dbt Core or SQL run directly in your warehouse. Weld keeps syncing your data either way. It can trigger dbt Cloud jobs when your syncs finish, and run other tools, such as Cube or SQLMesh, as webhook steps in an orchestration. Weld doesn't trigger dbt Core runs or SQL you run in your warehouse yourself.

You choose under Settings → Transform. The Transformations section has the full guides:

• Data models: drafts, publishing, and tables or views.
• Transformation scheduling: keep models up to date.
• Ed: Weld's AI agent writes and checks your SQL.
• Data lineage: see what each model depends on.
• Version history: compare and restore earlier versions of a model.
• GitHub sync: keep your models in a GitHub repository.
• Other transform tools: dbt Cloud, dbt Core and webhooks.