Assuming that you're using a version of Siebel that supports this feature 184.108.40.206 and later [preferably 220.127.116.11] then the high level steps should be as follows:
Using Siebel Tools:
- Modify DQ Integration Object to add 'Communication Address'
- Modify the 'DeDuplication' Business Service to add 'Communication Address' to the list of child dedup objects for Conatcts
- Compile the 'Contact Dedup IO' Integration Object and 'DeDuplication' Business Service
- Undeploy and the redeploy
- Modify third party vendor field mappings and add a new entry for Business Component Name: ‘Communication Address’, Operation: ‘DeDuplication’
- In the field mappings section, map the 'Address' field to EDQ field 'email'
- In vendor parameters add a new entry for 'Communication Address DeDup Record Type’
- If you're using UCM then you'll also need to modify the following Integration Object data maps to add a map for 'Communication Address':
- UCM Contact CIF IO to DQ IO (used by UCM Real-time Process)
- UCM Contact SDH IO to DQ IO (used by UCM Batch Process)
- UCM Contact SOAP IO to DQ IO (used by UCM Soap Process)
EDQ Siebel Connector:
- Modify dnd.properties and add new entries for ‘group.email’, 'group.email.concat' and 'group.email.delimiter' in the ‘Multiple child entity groupings’ section
Note that adding new child email addresses won't automatically trigger parent (contact) key gen and deduplication. In order to do this you'll need to automatically update a field on the parent, e.g. last updated.
We have done a detailed analysis, which is as below.
We have created a Contact and associated multiple addresses in Siebel.
Again we have created the same Contact and associated multiple addresses, however there is a common address in between them.
We found that,
1. In Siebel connector log, the candidate contact record is incomplete i.e. only a limited inforamtion such as Email Address,last Name, First name is present, however there is no individualid [i.e. ROW_ID of that contact] in the data.
2. We have captured the data in EDQ where externalid for the candidate record is null but it is present for driver record.
3. For your reference we are attaching the field mappings for Contact and personal address (for deduplciation and daat cleansing).
4. We have set the Support Multiple Child Entities Deduplication= Yes,
PS:- If we set Support Multiple Child Entities Deduplication= No, then externaid is populated as expected and realtime is working fine. This we have tested on dummy rule.