Since source entity ID is used to identify individual records within a particular universe it is not possible to have entities of the same type(model instance) that have the same source entity ID. What is the best practice for handling this?
The source has complex records that include fields that are parsed into the three models: Party, Phone, and Address.
A simple example source record may look like this:
Name: Joe Smith
Mailing Street Address: 123 Smith Drive
Mailing City: Columbus
Mailing State: Oh
Mailing Zip: 43210
After loading into MDM there will be one Party entity that references one Address entity and two separate Phone entities: one containing the phone and one containing the mobile phone. To prevent the "Mobile Phone" record from updating the already created "Phone" record when loading the data into MDM, an "M" is appended to the source entity ID(SEID) when mapping from the source profile to the upsert profile. This strategy allows the separate "Mobile Phone" entity to be created yet still be semi-linked to the source entity. When channeling updates from MDM to the source a custom JS function recognizes this "M", maps the MDM fields to the appropriate source fields, removes the "M" from the SEID, and then pass the update.