Merge behavior in Object Shopping

This topic describes merge behaviour for Object Shopping via an export file. For merge behaviour with copy-paste, go to "Conflict Resolution" in help topic: Object shopping via the clipboard (copy-paste).

When objects are imported in a target repository, the standard 'UPSERT' algorithm (a mix of UPDATE and INSERT) is applied:

If ...	Then:
The record exists in the target repository ...	All non-key columns are updated to the value found in the export file (UPDATE).
The record does NOT exist in the target repository ...	The record is added (INSERT).

An advantage of 'UPSERT' is that unique-key conflicts are avoided.

A notable consequence of 'UPSERT' is that records in the target repository are never DELETEd as part of the merge behavior during import.

For example, if you drop a column in the source repository, and your target repository contains the previous version of the table, then the column will NOT be dropped as part of the import routine.

Also, if you drop Job Tasks or change their Sequence Number, any Job Tasks in the target repository with Sequence Numbers that do not exist in the source will NOT be dropped from the target.

If the merge behavior does not raise errors for any of the objects in the export file, the import will be automatically committed. Otherwise, an error message will appear, the entire import attempt will fail, and the transaction will be rolled back.