Data Source Editor
After creating a data source, you are automatically taken to the editor. The editor organizes configuration into tabs, each covering a different aspect of the data source.
Tabs Overview
| Tab | Purpose |
|---|---|
| General | Update name, connection settings, description, and status flags (enabled, locked) |
| Filters | Control which schemas, tables, and columns are included in semantic extraction |
| Operations | Trigger and schedule semantic extraction jobs |
| Parameters | Override the AI model used for semantic extraction on this data source |
| Suggestions | Manage AI-generated example prompts tied to this data source |
| Hierarchies | Define and manage dimensional hierarchies for grouped analysis |
| Descriptions | Review and edit table and column metadata generated during semantic extraction |
| Relationships | Review, add, and correct the table connections used to build joins |
| Datasource Map | Visualize the tables and relationships detected during semantic extraction |
| Audit Log | View a history of all configuration changes made to this data source |
Navigating to the Editor
- Open the workspace and click Data in the top navigation.
- Hover over the data source you want to configure.
- Click the pencil icon (Edit) to open the editor.
Workflow
For a new data source, the typical workflow is:
- General — confirm connection settings and add a description.
- Filters — narrow the schema, tables, and columns to only what is relevant for analysis.
- Parameters — optionally override the AI model used for semantic extraction.
- Operations — trigger the initial semantic extraction.
- Descriptions — review generated metadata and add missing descriptions or synonyms.
- Hierarchies — define dimensional groupings if needed.
- Relationships — review detected table connections and correct or add any the process missed.
- Datasource Map — review the detected tables and relationships visually.
- Suggestions — review and curate AI-generated example prompts.
A data source binds exactly one source. To combine a database with uploaded files, create a separate File Upload data source for the files and combine both in a Custom Model.
After changes to filters or connection settings, re-run semantic extraction from the Operations tab to update the semantic layer.