Background: Data editor rebuilds the underlying result set after a row has been inserted in order to be able to edit the data in the newly inserted row. This approach may need review, but so far it is the only way to accomplish that, since result set can not scroll to the newly inserted rows. Using CachedRowSet was not an option, since this does not provide interactive input validation against database constraints. When rebuilding the data editor model, the system tries to match the old model rows, to be able to highlight the changes. If the table or view has primary or unique key columns, the matching is done only on these columns, which ensures the right row is matched. If the dataset has no primary or unique key, the matching is done on all values of the row, matching which typically fails in case the row has been updated by a trigger after insert.