KB45424: Element Browsing and View Filter behavior for reports based of Intelligent Cubes in MicroStrategy Analytics Enterprise 9.4.1 and newer releases

The Strategy Analytics Enterprise 9.4.1 and newer releases include a new feature - the ability to add multiple objects from different datasets in a document to a single grid, graph or widget in that document. This feature is documented in KB44944 "Multiple datasets in a single grid/graph/widget object in Strategy Web 9.4."
 
To support these changes there have been changes to the Analytical Engine in the Intelligence Server. This document attempts to describe the impact of these changes to the View Filter and Element browsing functionality below:
 
1. Changes to View Filter functionality:
 
When a user adds a View Filter to a grid in a document in Strategy Web Analytics Enterprise 9.4.x, all the attributes and metrics in all datasets for the document can be set as "Filter On" criteria. In earlier versions of Strategy, only the attributes and metrics that in the grid source dataset in the document can be set as "Filter On" criteria.
In the image below, the view filter definition interface is shown - only the template units from grid "R1" are available when defining the View Filter on the grid for R1.

 
The corresponding elements for browsing the grids in the view filter in Strategy Analytics Enterprise 9.4.1 are shown below. As seen, users may define the View Filter on any attribute or metric no matter which grid is used.
 
 

 
2. Changes to Element Browsing:
 
For the element browsing discussion, a new Analytical Engine VLDB setting is relevant. This setting is called "Document Grids from Multiple Datasets" and is set in the Project Configuration -> Advanced -> Analytical Engine VLDB settings section. The possible values of this setting are "Allow objects in document grids to come from multiple datasets .." and "Objects in document grids must come from the grid's source dataset only" (i.e. restrict behavior to be similar to 9.3.x and earlier releases). 
 

• If the document has already been run by the user:
   
  
• If the Analytical Engine VLDB setting for Document Grids from Multiple Datasets is set to "Allow.." (i.e. the default 9.4.1 and newer behavior), then the elements displayed for element browsing will be the union of the elements available from all datasets in the document for that attribute. For the example above, it can be seen that the element browser window shows elements for the attribute "Month" from both reports:
   
  

• With the Analytical Engine VLDB setting for Document Grids from Multiple Datasets set to "Objects..must come from grids source dataset only", only the elements for the attribute from that grids data source will be displayed. This is the same behavior as can be observed in Strategy 9.3.x and earlier releases.
   
  If the grids data source is set to "None", then the elements will be displayed from the first dataset added to the document.
  NOTE: The first dataset to be added to the document is shown in bold in the Document Objects browser in Strategy Web.
   
• If the document has not been run by the user:
  
• If the Analytical Engine VLDB setting for Document Grids from Multiple Datasets is set to "Allow.." (i.e. the default 9.4.1 and newer behavior), then the Analytical Engine will do the following:
   
  
• Step 1: Identify the dataset to use for the elements: 
   
  
• If the grid source for the grid has the attribute then the elements will be obtained from the dataset used as the grid source
   
• If the grid source does not contain the attribute, the Analytical Engine will examine the metrics in the grid. If there is only one metric, the Analytical Engine will check the dataset that the metric belongs to and obtain the attribute elements from that dataset.
  NOTE: For this check, Smart metrics defined on an attribute, and 'Aggregate from Base' metrics defined on an attribute on the document with the property 'UseLookupForAttribute' set to 'True' will be ignored.
   
• If the grid source does not contain the attribute and there are multiple metrics in the grid, the Analytical Engine will obtain the elements for the attribute from the first dataset to be added to the document that contains the attribute (the sequence in which the datasets is added to the document is saved with the document and  known to the engine).
   
• If the grid source does not contain the attribute and none of the datasets that the metrics on the grid belong to contain the attribute either, then the Analytical Engine will fetch the elements from the first primary dataset to be added to the document that contains this attribute (Similar to the previous condition but the primary dataset(s) are selected by the user). 
   
• If the grid source, all metric datasets and all primary datasets do not contain this attribute, then the first secondary dataset added to the document that is found to have the attribute is used for the attribute elements.
   
• Step 2: Obtain the elements from the dataset:
   
  
• If the dataset is an Intelligent  Cube (also from Data Import)/Report based of an Intelligent Cube that has been published we obtain the elements from the cube instance. If the cube is not published the document cannot be opened.
   
• If the report is a normal warehouse report/MDX report, and there is a report cache, the elements are obtained from the cache if present, otherwise the elements are obtained from the corresponding warehouse/MDX data source.
   
• If the Analytical Engine VLDB setting for Document Grids from Multiple Datasets is set to "Objects..must come from grids source dataset only" then a view filter can only use the grid source dataset. If the grid source is set to "None" then the attribute elements must come from the first dataset added to the document. The logic to obtain the elements is the same as 2.b.i.B above.
   
  NOTE: This behavior differs from 9.3.x and earlier releases. In those versions if the document has not been run, the elements are retrieved from the warehouse if the grid data source is a normal report, or from the cube instance if the grid data source is an Intelligent  Cube or report based of an Intelligent Cube.