For the sake of simplifying sample data generation (in the spoiler) diagnoses are just integers. Likewise simplifying the assignment of the macro variable whose value is a list of the 'values of interest' needed for some study.
%let study_diagnoses = 2,5,11,17,23,31,43,53,61,71,79,89;
For the case of character diagnoses it might look like
%let study_diagnoses = `F341', 'F41', 'F340', 'F998', 'F1234', 'F123', 'F12';
If the list (i.e the relevant ones) has a more patterned or complex origin you might be using DATA / SYMPUT('<macrovar>', or SQL / INTO :<macrovar>
dx is implicitly added to the PDV as a number, and is the host variable for the hash key and will be used to 'extract' values from the dxs array when interacting with the hash object.
call missing(dx);
For the case of character diagnoses ( DIAG: ) the host variable type needs to correspond. Replace the call MISSING with a statement such as
length dx $8;
The elements of array dxs are the variables whose names start with diag
array dxs diag:;
The diagnoses is my generated sample data are numeric, so the original sample code won't be appropriate to your actual data until the dx type is changed to character.
When looping over the dxs array pull out a diagnosis, check for it's relevance, and track its presence as a hash entry.
dx = dxs(index);
if dx in (&study_diagnoses)
then rc = study_dx.add();
else rc = other_dx.add();
Hope this explanation clarifies the technique and makes it applicable to your use case.
... View more
