Standard referential integrity is a form of referential integrity that incurs modest performance overhead because it tests, on a row-by-row basis, the referential integrity of each insert, delete, and update operation on a column set with a defined referential integrity constraint. Each such operation is checked individually by AMP software, and if a violation of the specified referential integrity relationship occurs, the operation is rejected and an error message is returned to the requestor.
You cannot use bulk data loading utilities like FastLoad, MultiLoad, and Teradata Parallel Transporter to load rows into tables defined with standard referential integrity.
Because referential integrity is checked at the granularity of individual rows, there is no opportunity for so-called dirty reads to occur during an insert, delete, or update operation on a table.