Data authoring patterns

The data life cycle and authoring patterns vary depending on the type of entity. This page explains the various patterns applicable to different entity types.

Authoring basic entities

With basic entities:

  • Creating a new record generates a golden record through the certification process.

  • Updating a golden record modifies it and submits the changes to the certification process.

  • Importing a record either creates a new record if the imported record ID is not found in the hub, or updates the existing record that is found.

  • Updating an erroneous record consists in picking the record, modifying it, and resubmitting it to the hub.

Authoring basic-entity records in an application can be done through direct authoring or using workflows, leveraging the capabilities of steppers. Complex clusters of entity records (e.g., customers and contacts) can be managed within steppers.

Authoring ID- or fuzzy-matched entities

Matched entities yield two types of golden records:

  • Golden records consolidated from publishers' master records, which are referred to as master-based records.

  • Golden records created directly in an application, which are not derived from existing publisher data and are referred to as data-entry-based records, as they are not associated with any master records.

Authoring matched entities works on golden or master records.

When authoring golden records:

  • Creating a new golden record results in a data-entry-based record that only exists in the hub.

  • Updating or fixing an error in a golden record depends on the record type:

    • For data-entry-based records, the update is applied during the certification process. Such records can be entirely modified.

    • For master-based records, an override is applied to the consolidated golden record data. This override overwrites certain attributes according to the survivorship rules defined for the entity.

  • The import and mass-update actions on golden records automatically adapt to record types, using the golden-record update or override patterns as needed.

Authoring golden records in an application can be done through direct authoring or using workflows, leveraging the capabilities of steppers. Complex clusters of entity records (e.g., customers and contacts) can be managed within steppers.

When authoring master records, creating, updating, mass-updating, or importing master records or source errors results in the creation or modification of publisher records "on behalf of publishers". These records are subject to the same matching and consolidation rules that apply to new submissions by a publisher. Such master records may be overwritten later on by records with the same ID from the same publisher.

Authoring master records is restricted to direct authoring, using steppers limited to their root entity only. This means that master-record authoring is performed on an entity basis, not on a cluster of records. Users can manage records for only one entity at a time (e.g., customers or contacts).
Master-record authoring is more suitable for data stewards to create or rectify records before publishers in an interim period—​that is, before a publisher actually pushes those records—​while golden-record authoring is recommended for business users.