To resolve duplicate records, you merge a target record with the duplicate records. The merge process creates a record that contains the most trustworthy data from all the participating records.
At the parent level, the merge process merges the data of the parent record.
At the child level, when the parent-to-child relationship is a one-to-one relationship, the merge process merges the child records. When the relationship is a one-to-many relationship, you can select the child records to merge from the
view. All records that are not selected for a merge are added to the merged record.
Developers define the parent-to-child relationships when they define the structure of the business entity.
After the merge process handles a child record, it tries to merge each level of descendant records using the same merge strategies. The merge process must run successfully on a level before the process can try to merge the next level.
The following image shows how the merge process works on three records, where each record has a parent record and two child-level relationships:
The merged record contains the merged parent record, the merged child records, and all the records that were not merged.