Implementing DELETE processing in your IP is similar to that of SELECT processing, for finding matching rows. The difference is that when dam_isTargetRow returns DAM_TRUE, the IP should delete the corresponding row in the data source. The IP should separate the logic that finds the matching row from the logic that processes that row. If the current operation being done is DELETE, then after a row is matched the data in the source should be deleted. If transaction rollback is supported, then the backend should not commit these changes until the IP is called for END TRANSACTION.
