To understand the implications of unmerging a row, consider the following example:
Start with four base object rows: A, B, C, and D. Merge A and B, and merge C and D. You later merge AB and CD. This gives you one row: ABCD.
If you unmerge C, you have two rows: C and ABD. Even though D was initially merged with C, it remains part of the merged base object ABD.
The new base object contains all of the cells in the current cross-reference row for C, and has the same value for ROWID_OBJECT that C originally had. If C does not contain some columns, the base object contains either the default value or NULL.
Informatica MDM Hub
will reinstate the original ROWID_OBJECT value of the cross-reference record. When you merge two or more rows, the merged row gets the same value for ROWID_OBJECT as one of the original rows. If you unmerge this row,
Informatica MDM Hub
merges all of the remaining cross-reference records into a ROWID_OBJECT that originally belonged to one of the remaining cross references.
Every user-defined foreign key that pointed to base object ABCD continues to point to ABD, even if that foreign key originally pointed to C.
All influence of the values from C is removed from the existing base object ABD. For columns that have trust rules defined, the values with the highest trust from A, B, and D will be reflected in the base object. For columns that do not have trust rules defined, the values from the last record merged will be reflected in the base object.