Objects in OWB

Question

The OWB documentation tells you how to 'drive' objects - but is not clarify when to use them - or give examples of how to use
So, can anybody tell me
Why and when would I use:
- a Mapping Input Parameter
- a Mapping Output Parameter
- an External Process object
- a Mapping Sequence
- A Pre Mapping Process and Post Mapping Process
and finally
- a Key Lookup
Sorry if it is long question....
Much obliged
Chris McGill

Answer 1

Hi,
- a Mapping Input Parameter A mapping becomes a PL/SQL package, and mapping input parameters become part of the signature of the MAIN procedure in this package. So if you want to feed information into the package you add a mapping Input parameter to the mapping. This then allows you to funnel information from the outside world into the pl/sql.
- a Mapping Output Parameter This is a similar thing as the input paramter, however this becomes an OUT parameter on the main signature, and can be used to give information to the outside world from the PL/SQL
- an External Process object You can use this to call an external process. For example if you need to launch an OS call from within PL/SQL you will use this external process.
- a Mapping Sequence is a database sequence. In other words it is a number generator which you can use to create surrogate keys while loading data into for example a dimension.
- A Pre Mapping Process and Post Mapping Process The OWB package that holds the mapping consists of the following:
Pre mapping process
exec <your procedure>
Main mapping process
exec mapping diagram in SQL
Post mapping process
exec <your procedure>
As you can see the mapping consists of 3 parts. The pre mapping process runs before you run the actual mapping. The post mapping one runs after the ETL mapping is complete. You can use these entry points to perform any activity just before or just after the mapping runs. An example of a post mapping process could be to truncate a table of data that you just loaded.
and finally
- a Key Lookup If you use surrogate keys in a dimension, you will have to do a key-lookup to determine which original record belongs to the surrogate key. The most common example is when loading a fact table (let's say orders with records that have product and time keys). You do have the product code, but you need to find the product_id (surrogate key). So you do a lookup to the dimension to find the product_id that belongs to the product code. Then insert that number into the fact to form the FK to the dimension.
For more information on surrogate keys I would suggest to read Ralph Kimball's book on data warehousing.
Hope this helps,
Jean-Pierre