OBIEE ENABLE_DIMENSIONALITY variable — Oracle Analytics

Oracle Analytics Cloud and Server

Welcome to the Oracle Analytics Community: Please complete your User Profile and upload your Profile Picture

OBIEE ENABLE_DIMENSIONALITY variable

Received Response
300
Views
6
Comments
SPowell42
SPowell42 Rank 5 - Community Champion

I see release notes for 11.1.1.7.16 and 11.1.1.9 that mention an "ENABLE_DIMENSIONALITY" variable, which when set to 1 seems to make queries across facts with non-conformed dimensions to map to "total" levels of the non-conformed dimensions without having to set lots of level metrics in the RPD. In general I like this option, but two quick questions:

1. Other than the checkbox on the advanced tab in analyses editor, is there another way to turn this on (possibly as a default)? I like to keep end users away from the advanced tab

2. I can't find any documentation on exactly what setting ENABLE_DIMENSIONALITY = 1 does...other than the 1 sentence on the screen. Are there more technical details about this somewhere?

Thanks,

Scott

Answers

  • [Deleted User]
    [Deleted User] Rank 2 - Community Beginner

    Hi Scott,

    If you look at DV/VA/DVD queries in detial you'll see that it's ON by default for all queries. In those tools it makes sense but in on-prem OBI I'm not a fan of just setting it to "smooth over RPD deficiencies".

    Still it's just a session variable like any other so you can set it with request variables just like any other.

  • SPowell42
    SPowell42 Rank 5 - Community Champion

    Hi Christian, not sure I consider this as "smoothing over RPD deficiencies", but more of just adapting the principle of saying that any time you have multiple facts with non-conformed dimensions, we'd want total levels. Using out of the box campus solutions and there are a ton of dimensions....as we add additional ones and more facts, it would sure be nice not to have to go in to each fact LTS for every dimension it's not hitting and set to the total levels.

    Scott

  • Thomas Dodds
    Thomas Dodds Rank 8 - Analytics Strategist

    A seasoned modeler wouldn't assume and would go in to each one any way ... providing the BI server with explicit instructions removes a lot of the potential for unexpected behavior.

  • [Deleted User]
    [Deleted User] Rank 2 - Community Beginner

    I hear you :-) It was a more sarcastic statement as I've seen it literally used as a "oh let's brush this under a carpet" solution. Obviously dangerous when dimensionalities aren't just black/white but have nuances.

  • SPowell42
    SPowell42 Rank 5 - Community Champion

    Gotcha. This is actually a very unusual situation...first time I've run in to a case where 5+5 doesn't equal 10...

    Appreciate the help - thanks!

    Scott

  • [Deleted User]
    [Deleted User] Rank 2 - Community Beginner

    I suggest you try some cubes from financial institutions or finance depts of multinationals. 5+5 can be anything from 7, 13, 10.5, fish and a towel ;-)