Often you have an EIM prozess where you get source data, which includes new and updated data. In the data flow, you can use the table comparison transformation, to identify the changes.
If you have a large dimension table or a large fact table and only little load data, the process often have a long processing time. This depends on the big comparsion table in the table-comparison component.
You can tune this, if you reduce the comparsison data to the minimum. This can be done for a cutomer dimension like this:
- source table for new/changed customer data: imp_customer
- target table for customer: dim_customer
- for the comparison create a view like:
- create view comp_customer as
select d.* from dim_customer d inner join imp_customer i on (d.cutomer_num = i.customer_num);
- Chance the table-comparison transformation and set the comparison table to the view comp_customer
Now, the processing time of the transformation is improved.