Trillium includes a Quality process called Transformer Unified Output in a name and address project. The Transformer Unified Output generates cleansed and standardized attributes in a common format in the same encoding (UTF8) for all country-specific flows including a no-match flow, which can then be merged into a single output entity.
Guidelines
Use the following guidelines when you are working with the Transformer Unified Output process:
- The Transformer Unified Output is the default process for a multi-country name and address project.
- Projects that were created in the previous version will use the Transformer Label Lines and new projects will use the Transformer Unified Output. You can remove and replace these processes manually if necessary.
- The Transformer Unified Output process does not automatically merge outputs. If you want to merge outputs from multiple countries, you still need to select, in addition to the Transformer Unified Output, the Merge Country Flows option in the Create Project Wizard or in the Global Data Router's configuration window.
- The Merge process for multi-country merge is designed for batch only and cannot be deployed in a real-time environment. If a real-time project includes the merge process, the process will be excluded when the project is deployed for use in real time.
- You cannot edit the schema directly in the Multi Country Merge process. If you need to change the schema, for example, to add user-defined attributes, you must do it in the prior Transformer Unified Output process.
- The Multi Country Merge process expects all inputs and the merged output to have the same schema or format. That means when adding attributes, you must add the same attributes to each country’s Transformer Unified Output including the no-match flow.
Transformer Unified Output vs. Transformer Label Lines
When you create a name and address project, you must select either Transformer Unified Output or Transformer Label Lines for the final process in each flow. This section compares the features in each process.
The Transformer Unified Output generates the same properly formatted names and addresses as the Transformer Label Lines. In addition, the Transformer Unified Output produces a variety of output attributes including derived attributes from the Transformer Data Reconstruction and attributes from the Relationship Linker.
The Transformer Label Lines has the advantage of producing a smaller output entity as there are fewer attributes in the output schema. The Transformer Unified Output is useful when you are building a global project and generating a consolidated output.
The following table shows the differences between the Transformer Unified Output and Transformer Label Lines process:
|
Feature |
Transformer Unified Output Process |
Transformer Label Lines Process |
|---|---|---|
|
Output |
|
|
|
Encoding for output schema |
UTF8. See Working with UTF8 data for more information on how the program handles UTF8 encoding. |
Encoding specified for each country |
|
Merge country flows |
You can merge outputs to create a single output entity | You cannot merge outputs |
|
Available projects |
|
|
| Available countries | All countries | All countries |
- Example
-
Your name and address project includes process flows for the US (United States), GB (United Kingdom), and no-match. You wanted to merge the final output so you selected the Transformer Unified Output process and the Merge Country Flows option during a project creation. When the project is run, the output of the Multi-Country Merge process, gb_mrgsplit_p31, is propagated and contains rows from the outputs of the Transformer Unified Output processes of US, GB and no-match flows.
Default Rules
The default rules can be viewed and edited under the Output Conditionals tab. The conditionals are displayed, and buttons allow you to view and edit the associated attribute scans and table recodes.
To modify the default rules