Ability to promote selected custom calculations to data set

TimVlamis

Organization Name

Vlamis Software Solutions

Description

It would be great to be able to selectively promote My Calculations from a Data Visualization project to a Data Set. Many times it's easiest to work directly in the project canvas and create new calculations. Some features don't work with custom calculations, so you need them to be in the proper data set. Some custom calculations you may want to promote to the data set, some you may not. For example, you may want different color assignments for different visualizations/canvases. This is project specific and there's no real need to have duplicate columns in the data set. Other calculations evolve as you write them and then you do want to promote them to the data set definition so that it's much easier to share the calculations between projects without having to replicate the project. (I know, I know, delete the canvases and save the project with a new name using "save as" and you can start with your custom calculations in a new project, but they're still not part of the data set.) 

Use Case and Business Need

This feature would also help reinforce the idea of pushing calculations down the stack. That is, from project to data set to subject area to data warehouse to ETL. Each step there's more process associated with a particular calculation. There's a natural trade off between the time it typically takes to add a data element and the availability, appropriate standards being applied, validation, security/access rules, conformed dimensions, and perhaps most importantly, system compute performance. Theoretically, every data element has its own "sweet spot". You don't want to burden developers and add to system overhead with seldom used calculations, but you also want calculations to be as consistent and performant as possible. 

More details

Takes time to recreate calculations in the data set editor or in a data flow or in BI Admin tool. 

Original Idea Number: 5fa771315d

Gabby Rubin-Oracle

Hi Tim, 

As you know, custom calculations done within the project are already part of the data set, but My Calcs are part of the project. Those are not the same and promoting to the data set might cause confusion and wrong results. 

• Data Set calculations - executed at the data set grain, this is what you want to use if your goal is to provide "governed" calculations with the data set.
• Project calculation - executed at the query result set grain

We are looking at ways to consolidate the concepts, but they are not equal and treating them as equals will create misunderstandings. 

The requirement itself is valid and we already have it on the backlog, the basic concept is that you want an easy way to bring My Calcs that you created in the past into other projects - it just not going to be delivered in the way that you described it. 

BTW, in 5.6 we made it easier to copy a My Calc expression so you can paste it quickly in the data set as a data set column.

TimVlamis

Thanks, Gabby, for your comments. Making My Calcs easier to copy is a help. The part about bringing My Calcs into other projects is true. We had custom calculations as catalog objects in the Classic interface (shared within a Subject Area). I understand there would be challenges with the data sets versus subject areas. It does bring up the issues with Subject Areas and local subject areas and data sets and how we keep all the similarities and differences in mind. Part is also that many of the legacy implementations of OBIEE that have migrated to OAC have very large complex subject areas. People are trying to develop in the Data Visualization interface in the same way they did in Answers, but without the same tool set. Thanks again for your insight and participation in this forum.  

Benjamin Arnulf-Oracle

I like the idea. I voted.