> ## Documentation Index
> Fetch the complete documentation index at: https://docs.k16solutions.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Create a blueprint

> Build a documented DataX entity from validated SQL and publish it to Lakehouse.

A blueprint turns a query into a documented, versioned entity. Create one when a stable dataset should be discoverable and reusable outside a personal worksheet.

<Note>
  **New Blueprint** appears only when you have blueprint management access. Ask an administrator to
  review **Manage blueprints** under **Lakehouse** in your role.
</Note>

## Before you begin

Prepare:

* a unique blueprint name;
* a concise description of the entity's purpose and grain;
* validated SQL that uses catalog entities available to your role;
* descriptions and sensitivity classifications for important output columns;
* tags that help other people find the entity.

## Create the blueprint

<Steps>
  <Step title="Open the form">
    Go to **Documentation** > **Catalog**, then select **New Blueprint**.
  </Step>

  <Step title="Add documentation">
    Enter **Blueprint name**, **Description**, and **Tags**. The name must be valid and unique.
  </Step>

  <Step title="Write the SQL">
    Build the query in the SQL editor. Search the **Source explorer** and select an entity or column to insert it into the editor.
  </Step>

  <Step title="Validate the output">
    Run the SQL. DataX validates the query and fills the **Columns** and **Results** tabs.
  </Step>

  <Step title="Document columns">
    Review inferred data types, add column descriptions, and apply available masking classifications where appropriate.
  </Step>

  <Step title="Review role access">
    If you can manage data access, use **Role Access** to choose which roles can query the new entity.
  </Step>

  <Step title="Save or publish">
    Select **Save Draft** to save the definition without deploying it, or **Publish & Deploy** to make the validated version live.
  </Step>
</Steps>

## Write maintainable SQL

* Select only the columns the entity needs.
* Give calculated columns clear, stable names.
* Make joins explicit and verify their effect on row counts.
* Avoid depending on an entity that is disabled or scheduled for replacement.
* Use the editor's format action before review.

Running SQL validates the current text. If you change the SQL afterward, run it again before publishing.

## Draft and publish requirements

**Save Draft** requires a valid name and SQL. Publishing also requires a successful SQL validation and all required fields.

During **Publish & Deploy**, DataX creates the version, stages it, and applies it. Keep the page open while the progress steps run. If version creation succeeds but deployment does not finish, the entity may still contain a saved version. Open its detail or version-control page before retrying.

<Warning>
  A published blueprint becomes a dependency for other DataX assets. Confirm descriptions, column
  names, data types, constraints, and access before you deploy it.
</Warning>

After publishing, review [entity details](/datax/catalog/entity-details) and [blueprint versions](/datax/catalog/blueprint-versions).