I'm always doing this.  This particular way I can have data set options applied to each dataset. . proc sql noprint;   select compress(libname || "." ||memname||' &DSoptions.')     into: datasets separated by ' '     from Dictionary.tables     where libname="LIBNAME"  ; *<-- needs to be all caps; quit; ... View more

Glad to help, you don't need to remove the format, but when you run PROC FREQ you can issue a FORMAT statement so that you may see the unformatted values.  proc freq data = probably;   FORMAT ACEDIVRC  ;   tables ACEDIVRC  test1 test2 test3 /MISSING; run; ... View more

I have never used Youden's index personally.  It should be easy to calculate inside the data step I provided above. ... View more

Well we technically did not answer the question pertaining to the combinations of factors... so... what can we do? ... View more

I'll add if you want to do this with more than one variable you can do this: data work.given;   do i=1 to 300; drop i;     A=ranuni(0)>0.5;        W=ranuni(0)>0.5;        x=ranuni(0)>0.5;        Y=ranuni(0)>0.5;     Z=ranuni(0)>0.5;     output;   end;   stop; run; PROC TABULATE DATA=&syslast               OUT=work.Want  ;   CLASS A W--Z ;   TABLE W--Z,A * N     ;   ; RUN; ... View more

Actually PROC MEANS is what I've used and it involved another step, but PROC TABULATE can be used using one step. My question about predictive models is still relevant:  I believe you need to choose dichotomous values to represent your independent variables to do this correctly.  If both MALNUT and another variable, _x_ are dichotomous then this code will produce a dataset with up to four rows, each corresponding to a cell of the confusion matrix.  Sensitivity, specificity and Youden's index can be calculated from these values. PROC TABULATE DATA=Given_DS               OUT=Want_X_vs_Malnut  ;   CLASS _x_     ;   CLASS Malnut  ;   TABLE _x_,Malnut * N     ;   ; RUN; ... View more

So you want sensitivity and specificity with respect to what predictive model?  Is 'MALNUT' the result of your predictive model, or is it the known state of the patient?  If you have established a predictive model then its outcome would be compared with the clinical data, this is how I know this type of statistics. I can refer you to https://en.wikipedia.org/wiki/Confusion_matrix.  I don't know anything in SAS_STATS that will do this, but perhaps this question should be posed in that forum: https://communities.sas.com/community/support-communities/sas_statistical_procedures . ... View more

We can communicate a solution more effectively if you give us a sample of your data.  Then we can speak in terms of your data's variables names. ... View more

PROC MEANS/SUMMARY can be used to calculate a confusion matrix. ... View more

This turned out to be something that was more crazy than I first thought...  I don't see STEP1 being solved with out a doubly-nested loops and multiple passes through the algorithm until all time periods are concatenated. data STEP1 (Keep=id   drug  refill begin_date   end_date over_Lap);   array b_date [&MAX_ARRAY_SIZE] b_date1-b_date&MAX_ARRAY_SIZE;   array e_date [&MAX_ARRAY_SIZE] e_date1-e_date&MAX_ARRAY_SIZE;   array ovrLap [&MAX_ARRAY_SIZE] ovrLap1-ovrLap&MAX_ARRAY_SIZE;   do until (eof);     do until (last.id);       /*Read Input*/       do N=1 by 1 until (last.drug);         set have end=eof ;by id drug;         b_date[N]=begin_date;         e_date[N]=end_date;         ovrLap[N]=0;       end;       NUM_ITEMS =N;       /*STEP-1 Concatenate Allowed Periods*/       NOT_FINISHED=1;       do while (NOT_FINISHED);NOT_FINISHED=0;         do M=1 to NUM_ITEMS ;           if not missing(b_date ) then           do P=M+1 to NUM_ITEMS ;             if not missing(b_date ) then             do;               if    -1 <= e_date -b_date <=5                  OR -1 <= e_date -b_date <=5 then               do;                 b_date =min(b_date , b_date );                 e_date =max(e_date , e_date );                 b_date =.;e_date =.;                 NOT_FINISHED=1;                 M=NUM_ITEMS;P=NUM_ITEMS;               end;             end;           end;         end;       end;       /*STEP-2 (prep) Find and flag Overlaps*/       do M=1 to NUM_ITEMS ;         if not missing(b_date ) then         do P=M+1 to NUM_ITEMS ;           if not missing(b_date ) then           do;             if    e_date -b_date >5                AND e_date -b_date >5 then             do;               ovrLap =1;               ovrLap =1;             end;           end;         end;       end;       /*Output*/       do N=1 to NUM_ITEMS ;         begin_date=b_date[N];         end_date  =e_date[N];         over_Lap  =ovrLap[N];         refill=N;         if not missing(b_date[N])           then output;       end;     end;/*do until (last.id);*/   end;/*do until (eof);*/   stop; run; data STEP2 (Keep=id overlap);   id=0;   overlap=0;   do until (last.id);     set STEP1; by ID;     if over_Lap then overlap=1;   end;   output; run; ... View more