You can join or unjoin a table pair. Join tables if you want to run a VALUE or an OUTER_VALUE test. Data Validation Option ignores joins for all set and aggregate tests.
To create a joined table pair, define one or more conditions based on equality between the tables or files. For example, if both tables in a table pair contain employee ID numbers, you can select EMPLOYEE_ID as the join field for one table and EMP_ID as the join field for the other table. Data Validation Option performs an inner equijoin based on the matching ID numbers.
You can create a join with one set of fields or with multiple sets of fields from each table if the table requires this to produce unique records. Note that additional sets of fields increase the time necessary to join two sources. The order of the fields in the join condition can also impact the performance of Data Validation Option tests. If you use multiple sets of fields in the join condition, Data Validation Option compares the ports in the order you specify.
When you create a join, you can select a field from each table or enter an expression. Enter an expression to join tables with key fields that are not identical. For example, you have two customer tables that use different cases for the LAST_NAME field. Enter the following expression for one of the join fields:
lower(LAST_NAME)
.
When you enter an expression for one or both join fields, you must specify the datatype, precision, and scale of the result. The datatype, precision, and scale of both join fields must be compatible. The expression must use valid PowerCenter expression syntax. Data Validation Option does not check the expression syntax.