Oracle Analytics and AI

Jean-Pierre Hoedenaeken

When revisiting some of my datasets to put in production some workbooks and datasets, I have noticed that we can use in the "Logical SQL Request" (used to define an initial value) a dataset that is not present in the workbook !

I find that convenient since we don't need to add (blend) any dataset in that workbook that is not relevant for the end user (in my use case, I am using a calendar dimension to initialize some parameters used in an expression filter but that calendar dimension is not useful for the dataviz).

Does anyone knows if this is a "normal" behavior ?

Best regards.

Jean-Pierre

Gianni Ceresa

I would call it normal, and expected.

In DV they kind of merged into a single object (workbook) what was before split into various pieces (dashboard, dashboard pages, prompts, analyses etc.). And each piece before was independent and could link any source it wanted.

Each element still has it's own life and execute its own LSQL query. That's why you have no limit in what sources (datasets, semantic model subject areas) you uses inside your canvas. And they don't all have to be visible in the "data" part of your workbook, like the parameters that you are referencing now (mostly for this kind of objects that are "new", they have been implemented a bit differently, having their own LSQL expressions and queries detached from the criteria of the workbook).

It does just make your life a bit more complicated when you want to move objects between environments, because you will need to remember that there are more dependencies on your workbook than what you maybe think (except if you have adopted a proper lineage tool that can easily list all these dependencies fully).