The most skilled task in the world of Oracle Business Intelligence is to design and code the RPD. While we are all aware of the general concepts of Business models, logical tables, joins and hierarchies, the RPD is essentially a query generating tool. The query generated by the tool depends on a sound understanding of the business model and data warehousing concepts. The most trickiest part of RPD development is to ensure that OBIEE generates a tuned physical query. I wish to write a series of articles that show the impact of coding something in the RPD on the Physical query and would hence enable the developers to find out the best things to do from a performance stand point. Please note that this article does not recommend one query over the other as the efficiency of the query will depend on the data structures themselves. The intention of this blog is to study the impact of changing the RPD on the Physical SQL. This is the first article in this series and we will be talking about the scenario where we pull columns from 2 different physical sources and 1 common dimension in the same analysis.
We know that we can use the evaluate function to call any function stored on the database and pass arguments to it. What if we have a requirement to use different database functions for the same column based on different selections made by the user. This article talks about a method to handle this requirement.