1. Administering and Working with Oracle Enterprise Data Management Cloud
  2. Administration
  3. Working with Matching and Merging

26 Working with Matching and Merging

Matching and merging refers to the process of comparing incoming nodes to existing nodes and identifying nodes that are the same, and then merging in the results.

When you confirm that a record from a particular data source is a match to an existing node in specific node type, the system stores information about that match so that the next time you bring data in from that data source, the incoming node is automatically matched to the existing node

Use Cases

There are two main uses cases for matching and merging nodes:

Matching and merging multiple records with different identifiers from the same data source.

For example, suppose you have an existing entity node with the name Oracle Incorporated, and you are merging in a file with a node name Oracle Inc. If you were to submit the request without matching, the Oracle Inc. node would be added as a new node because there isn't an exact match on the existing node name. But by setting up matching rules and reviewing the match results, you can specify that the incoming node is a match for the existing node. That way, when the request is processed instead of adding a new node the incoming node is merged with the existing node.

Matching and merging multiple records with different identifiers across multiple data sources.

For example, let's say you have a customer record, and you have multiple data sources that each contain different information about that customer. You might get credit card information for that customer from a point of sale device, and biographical information such as a name and email address from an online ordering system. If you were to submit a request to merge the records from these two data sources together, you would get two different customer records, each with different information about that customer.

With matching, you can set up a match rule that looks at, for example, the credit card number as a way to match the two records. Then, when you merge the data sources together, you can use the matching workbench to confirm that the two records are a match, and thus create a master record for that customer with all of the information from both data sources.

Terminology

The following terms can help you understand the matching and merging process:

  • Data source: An object that represents the source for the incoming data to be matched and linked in Oracle Enterprise Data Management Cloud. This can be either another Oracle Enterprise Data Management Cloud application (called a registered data source) or an external system whose data is not being managed in Oracle Enterprise Data Management Cloud (called an unregistered data source). See Understanding Data Sources.
  • Matching rule: Controls how nodes are matched from an incoming data source to nodes that already exist in a node type. See Creating, Editing, and Deleting Matching Rules.
  • Survivorship rule: Specifies which properties and relationships from the incoming source node get merged into the target nodes in a viewpoint after a match has been confirmed. See Creating, Editing, and Deleting Survivorship Rules.
  • Matching workbench: Enables you to review match candidates based on the criteria from the matching rules and accept the ones that you want to merge into the existing nodes. See Matching and Merging Request Items.

Process Overview

Configuring and using matching and merging follows this general process:

  1. A metadata manager on a dimension owner performs the initial setup of the matching process:
    • Creates the data sources if you are bringing in data from a source other than an Oracle Enterprise Data Management Cloud application.
    • Defines the matching rules that determine how nodes from other data sources are matched with existing nodes based on their property values.
    • Defines the survivorship rules that specify how the properties and relationships from the incoming nodes are merged with the existing nodes.
  2. A business user creates and submits a request with nodes to merge into existing nodes. The request items can be manually created or loaded from a file.

    Note:

    The matching process compares incoming nodes to existing nodes in a node type. Therefore, matching takes place in the context of a request.
  3. A metadata manager runs matching against the incoming nodes, reviews the matching candidates, and accepts or rejects the matches. See Matching and Merging Request Items.
  4. The metadata manager applies the match results to the incoming request items. For those nodes that are accepted as a match, the action in the request item is changed from an Add to an Update, Insert, or Move on the existing target nodes. See Reviewing Match Results and Applying Changes.

    Accepted match results are saved for each data source so that the next time request items are brought in from that data source they are matched automatically.