Combining rows of data

[FreelanceReinh]

Hi @Ksharp,

Thanks for rejoining the discussion. As far as I've understood, it's just about assigning observations (8 in the example, but presumably many more in reality) to 4 time slots (Week1, Week2, Week3, Week4). One variable (labelled "Total Samples" in the example) contains a number (the sum of ScoreA, ScoreB, etc.). The optimality criterion for the assignment is to achieve as similar totals of those numbers in the four time slots as possible. A precise definition of the criterion is pending. In the example the totals for Week1 - Week4 are 180, 170, 210 and 110, respectively.

[Ksharp]

So this question is the same as :

https://communities.sas.com/t5/General-SAS-Programming/Creating-Groups-of-Equal-Size/td-p/268862

[FreelanceReinh]

Yes, very similar to that one. Here, we have only a single variable, not three, with values to be summed within the groups. So, we might be very close to a solution already (requiring either SAS/OR or SAS/IML, though).

[Ksharp]

I can do that by IML . But OP should post another question at OR forum.

RobPratt is there . He could give some OR code .

[FreelanceReinh]

Unfortunately, they don't have a SAS/OR license. Not sure about a SAS/IML license (answer to question 4 is pending). I know that IML is included in SAS UE, but only for learning purposes, I guess.

[Ksharp]

hashexp:20,multidata:'y' =======>> hashexp:20,multidata : 'y'

[Ksharp]

hashexp:20,multidata:'y' =======>> hashexp:20,.... : 'y'