Rank Based on measures with row context.

mim, ImkeF, Anonymous - interesting problem.

Anonymous

I agree with your logic in the last post, and for Excel 2013 I would write the measure like this:

=
IF (
    HASONEVALUE ( MyTable[ID] ),
    RANKX (
        FILTER (
            ALL ( MyTable[ID] ),
            [TOSTR FORECAST DATE]
                = CALCULATE ( [TOSTR FORECAST DATE], VALUES ( MyTable[ID] ) )
        ),
        MyTable[ID],
        VALUES ( MyTable[ID] )
    )
)

1. Added a HASONEVALUE check to allow evaluation only for single IDs
2. Within FILTER, the measure for the currently iterated ID is [TOSTR FORECAST DATE], and this is compared with the measure from the original filter context CALCULATE ( [TOSTR FORECAST DATE], VALUES ( MyTable[ID] ) ).
   Putting VALUES ( MyTable[ID] ) as a filter argument effectively undoes the context transition.
3. Ranking is just based on ordering of IDs, descending by default. Can be tweaked.

Sample workbook here in case useful.

Cheers,

Owen :)

Wait, are you trying to tell me you actually understand the 3rd argument to RANKX?  NOBODY DOES!!  :D

Totally based on Owen's really good thinking, *I* would probably do this slightly differently to avoid the 3rd param:

Rank of ID among those with same TOSTR FORECAST DATE 2:=IF (
HASONEVALUE ( MyTable[ID] ),
  RANKX (
    FILTER (
      ALL ( MyTable[ID] ),
      [TOSTR FORECAST DATE] = CALCULATE ( [TOSTR FORECAST DATE], VALUES ( MyTable[ID] ) )
    ),
    CALCULATE( VALUES( MyTable[ID] ) )
  )
)