Tamr recommends the following workflow to incorporate enrichment into mastering projects.
- Use the unified dataset as the input for each enrichment project. In the example above, the unified dataset is used as input for the address and phone enrichment projects.
- Use a script transformation to join the enrichment project output datasets with the unified dataset.
LOOKUP get(address,0) AS address_standardized, get(verification_status,0) AS address_verification_status FROM enriched_address ON tamr_id = enriched_address.tamr_id; LOOKUP get(international_format,0) AS phone_standardized, get(valid,0) AS is_phone_valid FROM enriched_phone ON tamr_id = enriched_phone.tamr_id;
Important: Tamr recommends you either use
JOINto unify your data. Note
JOINchanges primary keys. See manual primary key management.
- Create the unified attributes added by enrichment. For example, after the above transformation you would create the attributes
phone_country_code, and so on. See Mapping Unified Attributes.
- Perform record mastering on the updated unified dataset and create golden records. See Mastering Project Workflow and Golden Records Workflow.
Note: If you make any changes to transformations, schema mapping, or enricher configuration after running the enriched mastering workflow, all records in that batch are enriched. In future batches, the model returns to incremental enrichment.
Updated about 1 year ago