Measure result returns all rows in filter context

[sbuster]

I have a measure [Time in State (ISO)] where I am simply trying to convert minutes to ISO8601 display format (PTxDxHxM) so i can see the duration a particulate items was in a given state.  The measure itself is working fine but I am confused as to the behavior of the visual display.  I simply have two visual tables for each data table, and when I click on Reco column it should filter State column.  The weird behvior is when debugging the measure, if I return the actual string result (r1), then resulting visual table for State display all ows, but only some rows have he proper value returned.  When the measure returns a number (like v) the filtering happens correctly.

Below is the result when I return a number from the meausre (again, just for debugging pursposes the meausre simply returns sumx(State, State[Time In State]).  The row A is selected in the left table and the right table is filtered correctly.

This is the result when the measure returns the string value r1 and row A is selected in the left table.  Notice the right table contains all rows in State table, and only the A rows have correct value from measure.  Here I would rows for B/C not to be displayed at all.  Am I missing something in a filter context?

Simplified table definitions are as follows:

Reco[ID]

State[RecoID, State, TimeInState]

There is a 1..Many relationship between Reco->State.

Measure[Time in State (ISO)]  = 

       VAR v = sumx(state,State[TimeInState])

       VAR minutesInDay = 1440

        VAR minutesInHour = 60

        VAR d1 = ROUNDDOWN(DIVIDE(v,minutesInDay),0)

        VAR d1Mod = v - (d1*minutesInDay)

        VAR h1 = ROUNDDOWN(DIVIDE(d1Mod,minutesInHour),0)

        VAR h1Mod = d1Mod - (h1*minutesInHour)

        VAR r1 = "PT"&d1&"D"&h1&"H"&h1Mod&"M"

       return r1

[sbuster · Accepted solution]

After watching videos on data linage I believe it defintely has something to do with that, although I'm still not sure.  Closing this quesiton and opening a new one to be more specific about questions related to data linage.