So, I have the hardware in place, the software in place, I’ve had my training, now it’s time to play God and create a Universe.
The first thing to do is to open IDT and create a connection to the Oracle database. Once done the next stage is to create the Data Foundation. This is where you pull in the tables and fields from your data source, in my case the Oracle database.
Warning: If you don’t understand your business, its data, and your database at a technical level, do not attempt to do this bit yourself. You are setting yourself up to fail. It is critical that you understand these things before you attempt to create the Data Foundation. Your Universe performance and performance hinges on what you do here.
I have worked with the Oracle database through SAP Crystal Reports for many years so I understand the structures and nuances of the database. Even with that it took me six months to create a single-source universe. Ok, I have a full time job running a company as well, so I had limited time, but don’t underestimate the time needed to create the Data Foundation.
A word of caution about the Data Foundation, you must choose whether you want a single-source universe or a multi-source universe. It’s crucial you choose the right one for the job as there is no going back once chosen. Single-source by its name means a single source of data only. A multi-source universe allows you to introduce further data sources as and when needed. There is currently no way of converting one to the other. I made the mistake of creating a single-source universe and had then to create a new multi-source one so that I could add other data sources.
Another point to bear in mind is that single-source universes use the native database language, in my case Oracle SQL. However, when using a multi-source universe IDT uses SQL-92 standard SQL. There is a radio button that allows you to switch to “native” SQL but I have found it’s better to leave everything as “SQL-92”.
Next, you create the Business Layer. This is where the fun begins. You have to decide on how you will structure the data. Do you use a familiar “Customer”, “Contract” folder approach, or do you hold all the “Dimensions” together and all the “Measures” together?
Please watch this space for future blogs for more about what this means to you and your organisation!