Curated Layer

The curated layer (silver) is where raw data gets cleaned, typed, and standardized. Use it after a Generate Mapping run, when you want to review and adjust the AI’s proposed curated entities before generating pipeline code. Nexa’s AI produces the initial curated entity definitions automatically; the Curated Zone screen lets you finalize them.

The Curated Zone screen

Navigate to Data Flow → Raw to Curated (or continue from the AI process screen after a successful mapping run).

You’ll see a paginated card list of entity pairs:

Column	What It Shows
Raw Entity	The source table name, with an AI-generated comment describing what it contains
Curated Entity	The recommended curated entity name, also with an AI-generated description
Worked On	Badge showing the last user and date who modified this mapping

AI-generated comments are marked with an AI icon so you can distinguish them from manually written descriptions.

Pagination

Control how many entities you see per page: 10, 15, 25, or 50. The total entity count is displayed alongside the pagination controls.

Editing column mappings

Select the expand arrow on any entity row to open the inline column mapping editor. Each row represents one column mapping:

Element	Description
Raw Column	Dropdown listing all columns from the raw source table
Curated Column	The target column name — an editable text input for new columns, a dropdown for existing ones
Data Type	Dropdown selector: STRING, INTEGER, BIGINT, FLOAT, DOUBLE, BOOLEAN, DATE, TIMESTAMP, DECIMAL
Description button	Opens a side panel to view or edit the column description

System columns

Nexa automatically includes control columns (for example, _ingest_timestamp and _batch_id) in every curated entity. These appear in read-only alert boxes and cannot be deleted or renamed — they’re required for pipeline operation.

Adding a new column

Use the inline form at the bottom of the mapping editor:

1. Type a column name.
2. Select a data type from the dropdown.
3. Select Save.

New columns appear with a visual indicator. You can rename them by editing the name and selecting Save (which changes to Rename for saved columns).

Column descriptions

Select the description button on any column to open a side panel. Descriptions that were AI-generated show an A badge. Edit these to add business context — they’re included in your generated pipeline code as comments.

AI-powered features

The curated layer relies heavily on AI to accelerate mapping:

• Auto-mapping — suggests which raw columns map to which curated columns based on names, types, and semantic analysis
• Entity naming — recommends business-friendly curated entity names from raw table names
• Column descriptions — generates descriptions explaining what each column likely contains
• Data type inference — suggests appropriate curated data types based on raw column analysis

All AI suggestions are clearly marked and fully editable. They’re starting points, not locked decisions.

Entity actions

From the Curated Zone, each entity has these action buttons:

Action	What It Does
Edit Entity Name	Rename the curated entity via a modal dialog
Generate Artifacts	Creates schema files, SQL/PySpark code, and pipeline definitions for this entity, with step-by-step status updates
Check Diff	Shows schema changes between branches — useful for reviewing what changed before publishing
View 10 Records	Generates synthetic sample data so you can preview what the curated table will look like
Delete Entity	Soft-deletes the entity (with confirmation)

Generating artifacts

When you select Generate Artifacts, Nexa produces:

• Schema definition files
• SQL and/or PySpark transformation code
• Pipeline configurations

A progress tracker shows each generation step in real time. Once complete, the artifacts are ready for deployment through the Pipelines workflow.

What’s next

Once your curated mappings are finalized, select Continue at the bottom of the page to proceed to the Consumption Layer, where you build your gold-level business entities.