Yes, it was designed to do that:
CALCULATE (
MAX ( Table1[Package ID] ), <- returns package id (max can be any function)
ALLEXCEPT ( Table1, Table1[Student ID] ), <- filters table for just that packageid student id
FILTER ( Table1, Table1[Datetime] = MAX ( Table1[Datetime] ) ) <- filters to the max date in this new table
)
Essentially, it is looking at the table and figuring out what the package id is for the max date, given the current context of student ids. When you add it to a table with student ids, you are giving it a row context, saying do this for each student id. When trying to do a count, you are removing this row context. You are now saying, calculate the latest package ID, but not for each student.
Instead, you can add a column for last package ID using the following column:
Last Package ID =
VAR studID = Table1[Student ID]
VAR MaxDate =
CALCULATE(MAX(Table1[Datetime]),
FILTER(Table1,Table1[Student ID]=studID))
RETURN
CALCULATE(
MAX(Table1[Package ID])
,FILTER(Table1,Table1[Student ID]=studID)
,FILTER(Table1,Table1[Datetime]=MaxDate)
)
However, this will repeat the value a lot. This may be fine for you - just remember to use DO NOT SUMMARIZE when putting in a matrix. Depending on how else yoiu use your data, it may cause more problems.
A better way, create a table with a unique list of students:
New Table:
Students = VALUES(Table1[Student ID])
Join this table to your ogirinal one on Student IDs.
Then, you can add the last id column into this new table, making a very small change:
Last Package ID =
VAR studID = Students[Student ID]
VAR MaxDate =
CALCULATE (MAX ( Table1[Datetime] ),
FILTER ( Table1, Table1[Student ID] = studID ) )
RETURN
CALCULATE (
MAX ( Table1[Package ID] ),
FILTER ( Table1, Table1[Student ID] = studID ),
FILTER ( Table1, Table1[Datetime] = MaxDate )
)
