Candidate and client groups are identified by their IDs, which are numeric and assigned in increasing sequence when the new group is established. During incremental updating of the record set and rearranging of groups (caused by adding or deleting records or changes of record keys or other attributes), there is an effort to retain already used group IDs as much as possible. For this reason, only one record of each group is called the Merge survivor and becomes the carrier of the group ID.

When the new group (candidate or client) is formed, the following cases can occur:

• The group does not contain a carrier.

  The group obtains a new ID from the sequence and a carrier is determined.

• The group contains just one carrier (inherited from a previous group).

  The group obtains the ID of this carrier.

• The group contains two or more carriers (inherited from previous groups).

  The best carrier, depending on a selection rule, is chosen, and the group obtains its ID. Other carriers lose their IDs and from this point on are not carriers.

There are two strategies for determining which record is assumed to be the Merge survivor (that is, the carrier of the ID):

• When a new ID is assigned to a group, one record is selected based on certain Merge survivor selection rules. This record is marked as the Merge survivor.
• The record marked as the center of the group is assumed to be the ID carrier. When the center of some previous group has moved to a newly formed group, it can carry the previous group ID, even if it is not selected as the center of the new group. Simultaneously, the previous group loses its group ID.

The switch useCenterAsSurvivor of unification defines the strategy to be used.

Even if the record carrying the group ID is currently deleted from the repository, it can still give its ID to the group to which it could have belonged if it was not deleted.