You are reusing the counter variable (i) in your inner loop. So, it will be more than two the second time it tries to do the outer loop. 🙂   Also, I would use  p1=tranwrd(original, 'r/t',''); instead of the compress. Compress removes all chars, not just that exact string. So, if you have "t-u", then the code will remove "t" as well.   ... View more

This is a start, don't know your exact rules.    data want; length colout1 $1 _coltemp $ 11; set have; do _i = 1 to countw(col1,":,"); *loop all the words; _coltemp = strip(scan(col1,_i,":,")); *Three different cases, a list, a span or a single value.; if index(_coltemp,"-")>0 then do; *Span, Might be problematic if the casing is different from start to end. ; _start = scan(_coltemp,1,"-"); _end = scan(_coltemp,2,"-"); *Rank gets the ASCII value for us to loop from start to end.; do _j = rank(_start) to rank(_end); *Use byte function to go from ASCII to string. ; colout1 = byte(_j); output; end; end; else if index(_coltemp,":")>0 then do; *A list of values separated (in this case with : ). ; do _j =1 to countw(_coltemp,":"); *Loop all of the values. ; colout1 = scan(_coltemp,_j); output; end; end; else do; *Single value version; colout1 = _coltemp; output; end; end; drop _:; run; ... View more

At one time or another I had some problems with missing values and quotes. So, I started to use the quotes to help this. I don't know if this is the best way to do it. It might have been when having x ne &y or similar and not the = . ... View more

Not exactly what you want, but close maybe a start?  data in; input Patient $ VisitDate $ Value Unit $ Type $ Visit $ bflag $; cards; A Jan12019 1 m Height baseline Y A Jan12019 50 kg Weight baseline Y A Jan52019 2 m Height visit1 . A Jan52019 55 kg Weight visit1 . B Jan12019 1 m Height baseline Y B Jan12019 50 kg Weight . . B Jan52019 55 kg Weight baseline Y ; proc sort data=in; by patient descending VisitDate; run; data want; set in; by patient descending VisitDate; retain height_ height_; if first.VisitDate then do; height_=.; weight_=.; end; if type='Height' then height_=Value; else if type='Weight' then weight_=Value; if last.VisitDate then do; if weight_ = . then weight_ = lag(value); if Visit = . then Visit = lag(Visit); if bflag = . then bflag = lag(bflag); bmi=weight_/height_**2; end; output; if bmi ne . then do; type='BMI'; value=bmi; unit='kg/m2'; output; end; drop height_ weight_ bmi; run; proc sort data=want; by patient VisitDate; run; ... View more

As other has said, remove the two extra output statements. But, my main issue with this is the instruction, 82×3,172×793 = 206,262,472 not as the assignment says 206,262,372. 🙂   DATA WOMAN; DO AGE=18 TO 99; DO PETS=4 TO 3175; DO ADDRESS=1 TO 793; OUTPUT; END; *OUTPUT; END; *OUTPUT; END; RUN;     ... View more

Totally missed your problem. One more go. So, you want to find duplicate rows that also have TBI =1 and Abuse = 1?    data have; input INC_KEY TBI Abuse; datalines; 12345 0 1 12346 0 1 12346 1 1 12347 0 1 12347 1 0 ;;; run; proc sql; create table want as select *, count(INC_KEY) as duplicates from have group by INC_KEY having duplicates > 1 and TBI = 1 and abuse = 1; quit; NB! This will only give you the rows where all of these things are true. So, this will only give you one row. INC_KEY TBI Abuse duplicates 12346   1   1     2   You can also do a subquery: proc sql; create table want as select *, count(INC_KEY) as duplicates from have where INC_KEY in (select distinct INC_KEY from have where TBI = 1 and abuse = 1) group by INC_KEY having duplicates > 1; quit; INC_KEY TBI Abuse duplicates 12346   0   1     2 12346   1   1     2   ... View more

Your first data step creates some empty variables, but that is easy fixed. So, for the next step there is a bit of if cases in the end. I assume that you only want to output the last row for each ID or if gap is larger than 10.  data have; length ID 8 item $ 1 gap 8; input ID item gap ; datalines; 1 O 3 1 O 4 1 O 5 1 O 15 1 O 3 1 T 4 2 O 3 2 O 5 2 O 7 2 O 11 2 O 3 2 T O ;;; run; data want; set have; by ID; retain O T string; length O T $1. string $50.; if first.ID then do; O='0'; T='0'; string=' '; end; if item='O' then O ='1'; else if item='T' then T='1'; if last.ID then do; *If this is the last row with this ID, then add item to the string then output it.; string=catx('-',string,item); output; end; else if gap > 10 then do; *If the gap is larger than 10, then output it first, then "reinit" string with item. ; output; string=item; end; else string=catx('-',string,item); *If neither of those cases, add item to string and then, without output, read the next line. ; run; ... View more

I read your code a bit to fast. Yes, I1-I5 does not exist anymore, they have been transposed to the variable Instrument.  ... View more