Forum Discussion

Peter_'s avatar
Peter_
Icon for Resolver I rankResolver I
7 years ago
Solved

Filtering a measure doesn't work in Aggregated Context

Hi,   I've created a Measure which calculates another measure with an applied filter, but the filtering is incorrect in an aggregated context. Any suggestions? Note that the nested measure definiti...
Need Help
  • Peter_'s avatar
    Peter_
    7 years ago

    I think I've found a solution using an inner join which gives 1 in the test sampleset. Thanks Kristjan76 for giving me a hint why it wasn't working.

    [Enabled Devices: [Devices]] + Enabled 2] =
    VAR tbl = ADDCOLUMNS(SUMMARIZE( Table1,
    Table1[Device],
    "Number of Enabled Sessions", CALCULATE(SUM(Table1[Sessions]), Table1[Is Enabled] = true),
    "Number of All Sessions", SUM(Table1[Sessions])),

    "Percentage of Enabled Sessions", DIVIDE([Number of Enabled Sessions], [Number of All Sessions]))
    RETURN
    CALCULATE([Devices], FILTER(NATURALINNERJOIN(tbl, Table1), [Percentage of Enabled Sessions] >= 0.5))
Anonymous's avatar
Anonymous
Not applicable
7 years ago

[Enabled Devices: [Devices]] + Enabled] = VAR tbl = ADDCOLUMNS( SUMMARIZE( Table1; Table1[Device]; "Number of Enabled Sessions"; CALCULATE(SUM(Table1[Sessions]); Table1[Is Enabled] = true); "Number of All Sessions"; SUM(Table1[Sessions]) ); "Enable";DIVIDE([Number of Enabled Sessions]; [Number of All Sessions]) ) RETURN COUNTROWS( FILTER( tbl; [Enable]>0,5 ) )

  • Peter_'s avatar
    Peter_
    Icon for Resolver I rankResolver I
    7 years ago

    That gives the correct count, but is it possible to do this by calling a [Devices] measure within the parent measure and not calculate it in place? In other words, can I filter the table just to devices that pass the [Percentage of Enabled Sessions ] check and hand that off to the [Devices] measure to apply any further filters? So instead of doing the commented version from Kristjan76's solution, which gives 1, to the uncommented one, which gives 2,?

     

    [Enabled Devices: [Devices]] + Enabled 2] =
    VAR tbl = ADDCOLUMNS(SUMMARIZE( Table1,
    Table1[Device],
    "Number of Enabled Sessions", CALCULATE(SUM(Table1[Sessions]), Table1[Is Enabled] = true),
    "Number of All Sessions", SUM(Table1[Sessions])),
    "Percentage of Enabled Sessions", DIVIDE([Number of Enabled Sessions], [Number of All Sessions]))

    RETURN

    // COUNTROWS(FILTER(tbl, [Percentage of Enabled Sessions] >= 0.5))

    CALCULATE( [Devices],     FILTER(tbl, [Percentage of Enabled Sessions] >= 0.5))
    • Peter_'s avatar
      Peter_
      Icon for Resolver I rankResolver I
      7 years ago

      I think I've found a solution using an inner join which gives 1 in the test sampleset. Thanks Kristjan76 for giving me a hint why it wasn't working.

      [Enabled Devices: [Devices]] + Enabled 2] =
      VAR tbl = ADDCOLUMNS(SUMMARIZE( Table1,
      Table1[Device],
      "Number of Enabled Sessions", CALCULATE(SUM(Table1[Sessions]), Table1[Is Enabled] = true),
      "Number of All Sessions", SUM(Table1[Sessions])),

      "Percentage of Enabled Sessions", DIVIDE([Number of Enabled Sessions], [Number of All Sessions]))
      RETURN
      CALCULATE([Devices], FILTER(NATURALINNERJOIN(tbl, Table1), [Percentage of Enabled Sessions] >= 0.5))