Creating a Workflow from a Template

A workflow template is a standalone configuration file produced by saving a working Workflow. It captures the supporting and output Pipelines, the dataset requirements, and the required columns.

Applying a template creates a new Workflow with its Pipelines already configured. There is no prompt and no AI configuration step, so the same template produces the same result every time, and applying one consumes no tokens.

Your only job after applying a template is matching a Dataset to each requirement it declares.

---

Before You Start

Templates configure Pipelines. They do not create Datasets or Connections.

Note

The Datasets a template requires must already exist, with the correct columns, before you apply it. A template declares how many Datasets it needs and which columns each one must contain. Set those Datasets up first. If they are not in place, you will have nothing to select when the template asks you to match them.

You will need:

• The template file. Templates are distributed as configuration files, so you get one by having it shared with you. There is no in-product template library.

• One Dataset per requirement the template declares, each already created and carrying the columns the template expects. Nothing is added for you, so a column the pipelines need and the Dataset lacks has to be handled by a Pad Columns step in the template.

• The values for any workspace variables the template uses, such as your own tenant, model, or environment.

---

Step-by-Step Instructions

Step 1: Start a New Workflow and Choose From Template

Create a new Workflow as you normally would, then select From Template instead of building from scratch.

Step 2: Upload the Template File

Click the upload area, select the template configuration file from your machine, and click Open.

Step 3: Review the Template

Click Review Template. Nexadata reads the file and shows you what it declares: the Datasets it requires, the columns each one must contain, and any workspace variables you will be asked to set.

Read this screen before continuing. It tells you exactly what you need to have ready.

Step 4: Match a Dataset to Each Requirement

The template lists each Dataset it requires as its own slot.

For each one:

1. Click Choose Dataset on the slot.

2. Use Search by name to find the Dataset you want to supply.

3. Select the Dataset from the results.

4. Click Add to Workflow.

Note

Repeat this for every Dataset the template requires. A template with three Dataset requirements needs three separate matches. Work down the list until no slot is left unfilled. The Workflow will not be created while a required slot is empty.

Match each Dataset to the requirement it actually satisfies. The slot labels describe what the template expects that input to be, so read them rather than filling the slots in order.

Step 5: Set the Workspace Variables

If the template’s pipelines reference workspace variables, you can set their values here, before the Workflow is created. This is where you point the template at your own tenant, model, or environment rather than the one it was built against.

Step 6: Create the Workflow

Click Create Workflow. The supporting and output Pipelines arrive already configured, and the Workflow is ready to run.

---

Handling Datasets with Missing Columns

An input Dataset does not always carry every column a template expects, and that is often correct rather than broken. A smaller entity can legitimately return fewer columns than a larger one, simply because it does not use every segment or the corresponding activity has not happened yet.

Note

Templates do not add missing columns for you. If a Dataset is missing a column the pipelines expect, the template does not silently backfill it.

The way to handle this is a Pad Columns transformation in the pipeline, adding the columns that can legitimately be absent. Because the template captures every transformation in its pipelines, a Pad Columns step built into the source workflow travels with the template and applies everywhere it is used.

If you are building a workflow you intend to save as a template, and you know its sources vary in which columns they return, add Pad Columns before you export.

Dataset-count validation still runs. If a template requires three Datasets and you supply two, that fails with a specific reason.

---

What Travels With a Template

A template captures:

• The supporting and output Pipelines, including every transformation in them

• Mapping Group definitions and their rules, recreated on import and wired back to the transformation that uses them

• Dataset requirements, including how many Datasets the Workflow needs

• Required columns for each Dataset

• References to the workspace variables its pipelines use, so you can set them for your own environment on import

A template does not capture Datasets, Connections, or credentials. Those are yours and stay in your workspace.

---

Distributing Templates

Templates are configuration files, not a managed object. There is no template library and no in-product lifecycle for them in this release. You distribute a template by sharing the file, and you apply it by uploading it.

---

Summary

Applying a template turns a Workflow you have already proven into one you can stand up again anywhere. Set up the required Datasets first, upload the file, match a Dataset to every requirement the template declares, set any workspace variables, and create the Workflow. The Pipelines come configured, the result is identical every time, and no prompt is involved.