The Normalizer transformation is used to "normalize" data, or to pivot columns into rows. How can you do the opposite (de-normalize, denormalize), or pivot multiple rows into multiple columns in a single row? Suppose you have a source table with this data that is a record of monthly expenses for each of your Sales Reps:
To accomplish this do the following:
1. Create an Aggregator transformation with the following ports and expressions:
2. Connect the SALES_ID, MONTH and the AMOUNT ports from the Source Qualifier to the Aggregator.
3. Connect the JAN_AMT, FEB_AMT, MAR_AMT, etc. ports to the target.
This approach can be used anytime the source has a column containing finite, repeating values that are known in advance such as the names of months used in this example.
Minnu's Blog on Informatica & Data warehouse concepts
Monday, June 9, 2008