How to make hierarchical path use visual filters

Hi DSenkovic 

Here's an initial idea - PBIX attached.

I'm working on the assumption that you have a data model with:

1. Nodes table, with columns Node, Parent and Path (created using PATHITEM)
2. NodeValue table containing Node and Value columns.
3. Relationship between NodeValue[Node] and Nodes[Node]

Path = 
PATH ( Nodes[Node], Nodes[Parent] )

Then you can create a measure like the one below (note that this can most likely be optimized).

Assumptions I made in writing the measure:

• You only want the measure evaluated for a single node at a time.
• The measure calculates the value sum for the topmost node in the path for the currently filtered node, within the overall selection of nodes (using ALLSELECTED).

Topmost node value = 
VAR CurrentNode =
    SELECTEDVALUE ( Nodes[Node] )
RETURN
    IF (
        NOT ISBLANK ( CurrentNode ),
        -- Only return result if single node selected
        VAR CurrentNodePath =
            -- Get path of current node
            SELECTEDVALUE ( Nodes[Path] )
        VAR CurrentNodePathLength =
            PATHLENGTH ( CurrentNodePath )
        VAR CurrentNodePathTable =
            -- Table containing Node/Position pairs for CurrentNodePath
            SELECTCOLUMNS (
                GENERATESERIES ( 1, CurrentNodePathLength ),
                -- Table of values {1,2,...,CurrentNodePathLength}
                "@Node", PATHITEM ( CurrentNodePath, [Value] ),
                "@Position", [Value]
            )
        VAR VisibleNodesAllSelected =
            -- List of all nodes selected overall
            CALCULATETABLE (
                VALUES ( Nodes[Node] ),
                ALLSELECTED ()
            )
        VAR CurrentNodePathTableFiltered =
            -- Filter path to just those nodes selected overall
            FILTER (
                CurrentNodePathTable,
                [@Node] IN VisibleNodesAllSelected
            )
        VAR TopNode =
            -- Get the top Node (i.e. with minimum position in the path)
            -- and treat as Nodes[Node] value
            TREATAS (
                SELECTCOLUMNS (
                    TOPN ( 1, CurrentNodePathTableFiltered, [@Position], ASC ),
                    "@TopNode", [@Node]
                ),
                Nodes[Node]
            )
        RETURN
            CALCULATE ( [Value Sum], TopNode, ALL ( Nodes ) )
    )

For example, if you select N1, N11, N12, N2, N14, then:

• N1, N11 and N12 have Topmost node value = 7
• N2 and N14 have Topmost node value = 8

Hopefully this is of some use and can be adapted to your model.

Regards,

Owen