Adding a Data Source

Before Nexa can map your data, it needs to know what sources exist and which tables belong to each source. Use this page after connecting a source, when you want to group its tables and mark which ones to curate. There are two ways to set this up:

1. Pattern-based source creation — match tables by name or schema patterns.
2. Nomenclature upload — upload a JSON file defining your naming conventions so the AI can better understand your data structure.

Creating a source with patterns

Navigate to Data Flow → Add Source (or select the + button on the New Dataflows tab).

1. Name your source

   Enter a Source Name (letters, numbers, spaces, underscores, and hyphens allowed) and an optional Description.

2. Choose a pattern type

   Select how you want to match tables:

   Pattern Type	What It Matches	Example
   Table pattern	Individual table names using the * wildcard	crm_* matches crm_customers, crm_orders
   Schema pattern	Entire schemas by name	bronze_* matches all tables in matching schemas

3. Preview matches

   As you type your pattern, the preview panel shows matching entities highlighted so you can verify the pattern catches what you expect. The panel also shows a count of total matches (for example, “5 Undefined Entities” or “3 Schemas”).

4. Submit

   Select Submit to create the source and associate all matched tables. You’re redirected back to the Data Flow screen, where your new source appears in the New Dataflows tab.

Uploading nomenclature

Navigate to Data Flow → Upload Data Source to configure naming conventions that help the AI understand your schema structure. See also Nomenclature settings.

1. Upload the nomenclature file

   Drag and drop or browse to select a .json file containing your naming standards. The file is validated for correct JSON format before upload.

2. Review the configuration

   After a successful upload, a summary panel shows:

   • Naming Conventions — uploaded from your file
   • Source Buckets — managed via the Data Flow screen
   • Target — auto-infer the silver and gold layers

3. Download for reference

   You can download the saved nomenclature file at any time for reference or to share with team members.

Managing tables within a source

Once a source exists, select its name on the Data Flow screen to open the Tables & Entities panel. Each table shows:

• Table name with a data icon
• Curation status — tables excluded from curation show an “Excluded from Curation” badge

Actions per table

Action	What It Does
Send for Curation (toggle)	Include or exclude the table from AI mapping. Excluded tables won’t generate curated or consumption layer entities.
Change Source	Reassign the table to a different source. Useful if initial pattern matching grouped a table incorrectly. The current source shows a checkmark in the list.

Quick add/edit source

From the Data Flow screen, you can also open a lightweight Add/Update Source modal:

• Source Name — searchable dropdown of existing sources, or type to create a new one
• Description — auto-populated when you select an existing source

When editing an existing source, the name field is locked — only the description can be updated.

What happens next

After your sources and tables are configured:

1. Return to the Data Flow screen.
2. Select Generate Mapping to launch the AI mapping process.
3. The AI analyzes your raw schemas and generates curated and consumption layer mappings.
4. Review the results in the Curated Layer.