Forum Discussion
Calculated column with condition returns wrong results (DAX)
Hi Anonymous,
To get the result you want, you can either
- Change the calculated column to:
Column = IF ( Sheet11[Test1] >= CALCULATE ( [Measure2], ALL ( Sheet11 ) ), "Success" )
or - Change the definition of Measure2 to:
Measure2 = CALCULATE ( AVERAGE ( Sheet11[Test2] ), ALL ( Sheet11 ) )
The reason for the original behaviour is:
- All measures are wrapped in an implicit CALCULATE()
- In a row context (such as in a calculatd column), CALCULATE results in context transition of the row context to filter context.
So in your original calculated column, the current row became filter context and [Measure2] ended up returning the average of Test2 in across all rows matching the current row, i.e. in this case just the value of Test2 in the current row. So you got "Success" in rows where Test1 >= Test2.
The fixes above use ALL(Sheet11) to ignore the filter context introduced by context transition, so that [Measure2] is evaluated in the context of an unfiltered table Sheet11.
Hope that helps,
Owen :)
- Change the calculated column to:
Hi Anonymous,
To get the result you want, you can either
- Change the calculated column to:
Column = IF ( Sheet11[Test1] >= CALCULATE ( [Measure2], ALL ( Sheet11 ) ), "Success" )
or - Change the definition of Measure2 to:
Measure2 = CALCULATE ( AVERAGE ( Sheet11[Test2] ), ALL ( Sheet11 ) )
The reason for the original behaviour is:
- All measures are wrapped in an implicit CALCULATE()
- In a row context (such as in a calculatd column), CALCULATE results in context transition of the row context to filter context.
So in your original calculated column, the current row became filter context and [Measure2] ended up returning the average of Test2 in across all rows matching the current row, i.e. in this case just the value of Test2 in the current row. So you got "Success" in rows where Test1 >= Test2.
The fixes above use ALL(Sheet11) to ignore the filter context introduced by context transition, so that [Measure2] is evaluated in the context of an unfiltered table Sheet11.
Hope that helps,
Owen :)
@OwenAuger Beautiful explanation of context transition in DAX!
If the only purpose of Measure2 is to help calculating the desired column (or you are interested in a language comparison) the M-solution has advantages:
let Source = #table({"Test 1", "Test 2"}, {{10, 10}, {9, 14}, {8, 4}, {7, 3}, {6, 4}, {5, 6}, {4, 3}, {3, 2}, {2, 1}, {1, 3}}), Measure2 = List.Average(Source[Test 2]), Column = Table.AddColumn(Source, "Column", each if [Test 1] >= Measure2 then "Success" else null) in ColumnThere is no need for a separate measure as the average will be calculated as a temporary step that can be referenced but must not be shown anywhere (and maybe later hidden from client tools).
Or even more compact:
let Source = #table({"Test 1", "Test 2"}, {{10, 10}, {9, 14}, {8, 4}, {7, 3}, {6, 4}, {5, 6}, {4, 3}, {3, 2}, {2, 1}, {1, 3}}), Column = Table.AddColumn(Source, "Column", each if [Test 1] >= List.Average(Source[Test 2]) then "Success" else null) in ColumnThis can be done easier using the UI only. Simply add a custom column and paste the following code into the custom column formular window:
if [Test 1] >= List.Average(Source[Test 2]) then "Success" else null
Within a custom column in M/the query editor a reference to a column like [Test 1] is equivalent to an evaluation in row context in DAX, so only the value of the current row will be considered. The trick to adress the column "Test 2" as a whole lies in referencing it with a preceding table name like this: Source[Test 2].
Thanks a million Owen for the solution and the great explanation! This did the trick!
George.
Anonymous An alternative would be to use AVERAGEX like this...
Column = IF ( Sheet11[Test 1] >= AVERAGEX ( ALL ( Sheet11 ), Sheet11[Test 2] ), "Success" )
