Proc Transpose wide to long

[BTAinRVA]

I'm trying to transpose a table using the code below. The problem is that a person can have multiple encounters on the same day. The output from the code below puts same-day info into separate columns. Is there some trick to getting SAS to stack all the Diagnosis_1-Diagnosis_10 info into a single column?

proc transpose data=TBI_temp out=TBI_transposed;

  by Person_ID Encounter_Date;

  var Diagnosis_1-Diagnosis_10 ;

run;

Thanks,

Brian

[PGStats · Accepted solution]

Proc Transpose doesn't have a mechanism to transpose by OBS when there is no obs unique ID. It is simpler to do the transpose explicitly in a data step.

data TBI_transposed;
set TBI_temp;
array d  Diagnosis_1-Diagnosis_10;
do i = 1 to dim(d);
	Diagnosis = d{i};
	output;
	end;
keep Person_ID Encounter_Date Diagnosis;
run;

(untested)

[BTAinRVA]

Thanks!