The Adapter Agent checks that your data is ready for the next step in your workflow. It verifies that your data contains the expected columns and uses the correct data types before passing it to another agent. This helps prevent errors and ensures that every step in your workflow receives data in the expected format.
This article shows you how to add and configure an Adapter Agent from scratch.
Before you begin
Before you start, make sure you:
Have access to a Savant workspace.
Create or open an analysis.
Have a dataset available. You can upload a file or use an existing Source Agent.
Set up the Adapter Agent
Step 1. Add a source dataset
The Adapter Agent needs input data before you can configure it.
Open your analysis.
Add a Source Agent.
Choose one of the following options:
Select an existing Source Agent.
Select New Dataset to connect to a supported data source.
Select Upload to import a CSV or Excel file.
Complete the Source Agent configuration.
For this guide, use the Product Events as the Source Agent.
Confirm that your dataset contains the columns required for your workflow, such as:
user_id
user_email
event_id
event_type
event_type_name
event_time
You now have a Source Agent that provides data to the Adapter Agent.
Step 2. Add the Adapter Agent
Add an Adapter Agent to validate and organize your data before sending it to the next step.
Expand Rules Agents.
Select Adapter.
Connect the Source Agent to the Adapter Agent.
The Adapter Agent is connected and ready to be configured.
Step 3. Generate the field mappings
Generate the field mappings from your source data.
Select the Adapter Agent.
In the Field Mappings section, click Bootstrap.
Run Bootstrap whenever you connect a new dataset or make significant changes to your source data.
Wait while Savant generates the field mappings.
Bootstrap automatically reads your source data and creates a mapping for every column. This saves time because you don't need to configure each column manually.
Savant creates a field mapping for every column in the source dataset.
Step 4. Review the generated mappings
Review the generated mappings before continuing.
Confirm that every output column has the correct name.
Verify that each output column maps to the correct source column.
Check that every column uses the correct data type.
Mark important columns as Required if they must always appear in the output.
The Adapter Agent contains the correct field mappings for your workflow.
Step 5. Customize the output
Customize the generated mappings if your workflow requires different output columns.
Click the Lock icon to unlock the field mappings.
Rename output columns.
Change data types.
Update the source field mapping.
Add new columns.
Remove columns you no longer need.
Lock the field mappings when you're finished.
The Adapter Agent produces the output your workflow expects.
Step 6. Configure unmapped fields
Choose how the Adapter Agent handles columns that aren't included in the field mappings.
Locate Pass through Unmapped Fields.
Turn it on to include unmapped columns in the output.
Leave it turned off to output only the configured columns.
Use this option when downstream agents need additional columns that don't require validation.
The Adapter Agent handles extra columns according to your workflow.
Step 7. Apply the configuration
Apply the configuration and review the output.
Review your field mappings.
Click Apply.
Wait for the Adapter Agent to process the data.
Review the output.
If the output isn't correct, update the configuration and click Apply again.
The Adapter Agent validates the data and prepares it for the next step.
Step 8. Add a Destination Agent
Add a Destination Agent to receive the validated data.
Click Add Agent.
Under Destination Agents, choose a destination.
Connect the Adapter Agent to the Destination Agent.
Complete the destination configuration.
For example, to use the Email Destination Agent:
Enter the recipient email address.
Enter the sender name.
Enter the reply-to email address.
Enter the email subject.
Save the configuration.
The Destination Agent is ready to receive data.
Step 9. Verify the output
Run the workflow and verify the results.
Run the analysis.
Review the output generated by the Adapter Agent.
Confirm that the Destination Agent receives the expected data.
Verify that:
All expected columns appear.
Column names are correct.
Data types are correct.
Data appears as expected.
If necessary, return to the Adapter Agent, update the configuration, and run the workflow again.
The Adapter Agent validates and standardizes your data before passing it to the next step.
Next steps
Continue building your workflow by learning how to:
Use the Transform Agent
Use the Infer Agent
Build complete workflows using multiple agents