Getting started
Initial configuration
The following points are important to understand before starting to use the add-on:
Every table managed by the add-on is enriched with a data type that declares matching metadata. This metadata is required for add-on execution and is fully managed by the add-on. For instance, the matching metadata includes the matching score and the state of every record. The complete list of this metadata is defined in 'matching metadata' section (refer to 'Installation and first configuration' section to include this data type in a table).
Before using matching, it is necessary to configure matching policies. Many parameters are used to determine how matching is executed, to decide-depending on certain rules-how the automatic merging of records takes place (or not), which algorithm is used to find duplicate records, etc.
The first part of this user guide is dedicated to explaining how to master matching configuration. Depending on your business context, many different configurations could be created by:
changing the fields used for matching
switching which algorithms are employed
adjusting threshold levels at which automatic merging is triggered.
With the add-on, an unlimited number of matching policies can be configured and coexist in your environment. A matching configuration administrator should be identified in your organization to be responsible for these policies over time. In a given situation, once the most suitable policies have been identified, it should no longer be necessary to modify them. However, as business needs evolve over time, it is important to have the ability to smoothly adapt the policies with the changing needs.
Hence, before starting to configure matching policies, it is important to have a solid understanding of all parameters that the add-on uses.
To start right now you can refer to the Installation and first configuration section.
To use the data cleansing feature please refer to the Data Cleansing chapter that collects all information related to this functional domain.
Vocabulary
A basic understanding of the following terms is beneficial when working with the add-on:
Topic | Term | Definition |
|---|---|---|
Record state | Suspicious, Suspect, Pivot, Golden, Merged, Unmatched, To be matched, Deleted | Every record holds a state value. This value indicates a record's matching quality level. |
Score | Similarity score | A record holds a similarity score against another record. This score is a percentage of similarity. Scores are computed by matching policies configured with the add-on.
|
Group of records | Cluster | A cluster is a group of records that are suspect with each other. A record can be attached to only one cluster at a time. |
User Interface | EBX® regular view | This is the normal EBX® view, also known as the tabular view. The add-on does not impact this view. Two matching data are presented on every record to show the record's state and its cluster identifier. In order to hide this information and to display golden records only you can use a filtered view. EBX® services offer access to the light and full matching views. |
Light matching view | This view is provided by the add-on. It allows you to manage suspect records already identified and presented in a cluster. The light matching view can be integrated into a workflow layout. | |
Full matching view | This view is provided by the add-on. It allows you to interact with all available matching features. The user interface is divided into two frames. One to launch matching operations and a second to assess matching outcomes. The full matching view is self-sufficient and can be integrated into a workflow layout. | |
Simple matching view | This view is provided by the add-on. It allows you to decide what action to take on a suspicious record (directly set as golden, deleted, move it to a pivot record). | |
Stewardship | The stewardship process begins when a suspect record must be managed. | |
Survivorship | The survivorship process begins when a suspect record can be merged automatically into a pivot or golden record. | |
Relationship matching | Matching executes on records that have foreign keys referring to the current record. | |
Surrogate matching | When the returned matching result on the configured matching field is lower than 'Field stewardship min score (%)' on a record, the surrogate field values are taken to match instead. | |
Special notation: | |
|---|---|
| Important recommendation to use the add-on. |
| This feature is not yet available in the current release. |
