TSQL Collapse Multiple Rows into a Single Row

Question

Here's an example of the data now:

Student ID  Type     Attendance Date    Attendance Name Given1  Attendance Name Given2 Test Type
445633      Student  2/04               Matt                    Smith                  BIO
445633      Student  2/04               Matt                    Smith                  HIST
445633      Student  2/04               Joe                     Smith                  BIO
445633      Student  8/09               Joe                     English                BIO
535485      Student  8/09               Joe                     English                HIST 
535485      Student  11/19              Joe                     English                BIO     
114687      Student  3/14               Frank                   William                GEN
114687      Student  11/10              Greg                    William                MATH
114687      Student  11/10              Greg                    William                CHEM
114687      Student  8/09               Dan                     Harris                 HIST
114687      Student  8/09               Dan                     Harris                 HIST

And these are the rows that I want (Ignoring the contents of Attendance Name Given1 and 2. Right now my group by is including unique rows that I do not consider unique bc the data entered into them was done by the user and is free text. I want my results to show these free text fields but do not want them to be considered a unique row strickly because of these 2 Name given fields. (Think of these two name given fields as possible lies).

So my results need to yield...

For 445633 I don't care if I keep 445633 Matt or Joe since Student ID is the same, Type is the same, Attendance Date is the same and Test is the same. The only fields that should make up a unique row should be Student ID, Attendance Date, and Test Type. Student Type will always be the same (Student).

445633 we loose a BIO entry... 535485 we keep both since the attendance dates are different 114687 we loose a HIST entry

Student ID  Type     Attendance Date    Attendance Name Given1  Attendance Name Given2 Test Type
445633      Student  2/04               Matt                    Smith                  BIO
445633      Student  2/04               Matt                    Smith                  HIST
445633      Student  8/09               Joe                     English                BIO
535485      Student  8/09               Joe                     English                BIO
535485      Student  11/19              Joe                     English                BIO
114687      Student  3/14               Frank                   William                GEN
114687      Student  11/10              Greg                    William                MATH
114687      Student  11/10              Greg                    William                CHEM
114687      Student  8/09               Dan                     Harris                 HIST

Thank you!

Answer 1

You can do this using group by , if you really don't care whether the two given names are from the same record:

select StudentID, Type, AttendanceDate,
       min(AttendanceNameGiven1), min(AttendanceNameGiven2),
       TestType
from table t
group by StudentID, Type, AttendanceDate, TestType;

Or, you can do this using row_number() if you want the values from the same record:

select StudentID, Type, AttendanceDate,
           AttendanceNameGiven1, AttendanceNameGiven2,
           TestType
from (select t.*,
             row_number() over (partition by StudentID, Type, AttendanceDate, TestType
                                order by newid()
                               ) as seqnum
      from table t
     ) t
where seqnum = 1;