Performance Considerations
The majority of client processing during a Teradata MultiLoad job occurs when it is processing its input rows. The most efficient means of sending the row to Teradata Database would be a bulk move of the input row to the output row.
However, there are many cases where fields need to be evaluated and field data may need to be individually moved from the input to the output row. Note, however, that performance is affected whenever a field needs to be evaluated or individually moved.
The need for moving individual field data from the input to the output row occurs for any of the following scenarios:
Variable length fields, NULLIF in the layout, and APPLY WHERE clauses might require additional CPU consumption.