For example, you configure a Lookup transformation to perform a dynamic lookup on the employee table, EMP, matching rows by EMP_ID. You define the following lookup SQL override:

SELECT EMP_ID, EMP_STATUS FROM EMP ORDER BY EMP_ID, EMP_STATUS WHERE EMP_STATUS = 4

When you first run the mapping, the

mapping

task builds the lookup cache from the target table based on the lookup SQL override. All rows in the cache match the condition in the WHERE clause,

EMP_STATUS = 4

.

The

mapping

task reads a source row that meets the lookup condition you specify, but the value of EMP_STATUS is 2. Although the target might have the row where EMP_STATUS is 2, the

mapping

task does not find the row in the cache because of the SQL override. The

mapping

task inserts the row into the cache and passes the row to the target table. When the

mapping

task inserts this row in the target table, you might get inconsistent results when the row already exists. In addition, not all rows in the cache match the condition in the WHERE clause in the SQL override.

To verify that you only insert rows into the cache that match the WHERE clause, you add a Filter transformation before the Lookup transformation and define the filter condition as the condition in the WHERE clause in the lookup SQL override.

You enter the following filter condition in the Filter transformation and the WHERE clause in the SQL override:

EMP_STATUS = 4