Mark Rittman's blog
In this article Mark explais how one can drill from a "Discoverer for OLAP worksheet" to a "Discoverer Plus Relational worksheet" using Discoverer 10.1.2.
"I have a question about drilling from an OracleBI Discoverer for OLAP 10.1.2 worksheet to a Discoverer Plus Relational worksheet. When you pass values from an OLAP worksheet you pass either the dimension name or the dimension value to the associated parameter in the relational worksheet. Obviously, in OLAP this dimension is treated as an object, and we have no idea which level the user may have picked before he drills out. On the other hand, in the relational world, each level of the dimension would be split out as a separate parameter. Could you run through a simple example where you drill from an OLAP worksheet to a relational worksheet and show how this is done?"
In this article, Mark explains how the SQL MODEL clause can be used to generate rather complex financial statements.
One of the most powerful features of the multidimensional engine behind analytic workspaces is the ability to create formulas. Formulas, or "calculated measures" as they're referred to in AWM10g, are measures that are derived from other measures. Using AWM, you can create simple formulas that reference other measures in a cube, allowing you for example to create a "margin" measure derived from sales and costs measures. If you're an old Express hand though, you'll know that this simple type of formulas is just the tip of the iceberg, and what you often used to end up doing was creating for example a three dimensional formula based on measures from four and five dimensional variables, rolling up unneeded dimensions and pulling in variables held in what would now be referred to as "cubes".
If you're looking to tune an SQL statement or a batch job, a common way to find out what happened during the execution of the SQL is to run an extended SQL trace and examine the wait events. But what happens if you are using parallel execution, and all your trace file contains is the parallel execution wait events that are generally considered idle events? Your trace file shows how long your query took to run, and the work involved in controlling the PQ slaves, but the real details of what took up all the execution time are actually to be found in the corresponding PQ slave trace files in the BDUMP directory.
I've recently been reading Bert Scalzo's "Oracle DBAs Guide To Data Warehousing And Star Schemas", and he's got an interesting take on the ETL process that I thought it worth looking at in more detail.
If you've been developing applications on Oracle for a while, you've no doubt come across the concept of 'Bind Variables'. Bind variables are one of those Oracle concepts that experts frequently cite as being key to application performance, but it's often not all that easy to pin down exactly what they are and how you need to alter your programming style to use them. With this in mind, I've tried to pull together the key information about bind variables and why they are a 'good thing' when building Oracle applications.
With the recent release of Oracle Database 10g, OTN have recently been running a series of articles by Arup Nanda entitled "Oracle Database 10g: The Top 20 Features for DBAs".
With this in mind, I thought it worth putting something similar but this time focusing on those features that apply to Data Warehousing and Business Intelligence. So here's my Top 10...