Here's one way: data have; input @1 ID $3. Symptom $30.; datalines; 101 1-Burning 102 1-Burning2-Itching 103 1-Burning2-Itching3- swelling ; run; data want; drop i words Symptom; set have; words = countw(Symptom, '-123 '); do i = 1 to words; Symptom_Single = scan(Symptom,i, '-123 '); output; end; run; Please post your data like I've done in a DATALINES statement in the future to save us time in answering. ... View more

Found it helpful! ... View more

I'm not going to format each individual value, but I'm assuming this is what you wanted.   data want; set have; length acnoth $200.; array v [*] ther -- other; do i = 1 to dim(v); if not missing(v[i]) then acnoth = catx(', ', acnoth, vname(v[i])); end; run; In future posts, please provide more information - we can only make assumptions. If you have actual character values associated with variables, please provide those too. ... View more

Define what "first IP administration" is or how we can tell in the data that occurs. I hate to guess and since there is no Variable named either IP or Administration or anything resembling either of those words that is what I would be doing.   It appears like a "find the value per Subjid (NOT mentioned in your request) and merge that value back onto the whole data set by subjid.   One potential problem, what if there is break in the data, such as long interval related to whatever this activity is, and the same subjid reappears later and should not be treated as the same sequence of visits? ... View more

@tarheel13 wrote: is it the last non missing screening visit?  Do you have any missing "Epoch" values? (If that is how to tell if a visit is screening or not)   ... View more

There is seldom a good reason to change the data. Unless it is erroneous, always use a format as explained by @Patrick . ... View more

Please show us some examples of the desired output. ... View more

v_have is raw data, by using this data create AVISIT AVISITN VISIT VISITNUM   for exp:could you please  look at subject 13-0036 LBORRES is Quantity not sufficient after that LBORRES  is still having values so we skip the visitnum =1 and  take visitnum =1.1     look at subject : 16-0066 visitnum  have 1 and 1.1 so we skip the 1.1 but  we want only visitnum 1  ... View more

Native and American as separate races???? Native what? Is this the Native Hawaiian/ Other Pacific Islander or Alaska Native/American Indian (two typical race categories in the US OMB playbook)     ... View more

Hi @teja5959    Could you please explain the rules that should apply? Thank you very much.   Best, ... View more

Please post the data in usable form and add a description of the rules that you want applied. ... View more

Try this SAS Code: data have; input number texts $80.; datalines; 001 NEW OTC OR RX DRUG ADDED, HOSPITALIZATION (INCLUDES ER VISITS) 002 NEW OTC OR RX DRUG ADDED, Non-drug therapy, HOSPITALIZATION (INCLUDES ER VISITS) 003 STUDY TREATMENT DISCONTINUED PERMANENTLY, Non-drug therapy, HOSPITALIZATION (INCLUDES ER VISITS) ;;;; run; data want; set have; texts=scan(texts,1,','); run; ... View more

@teja5959 : If I get your pattern correctly, then try: data have ; input subjid visitnum visitdate :date. visit & :$32. ; cards ; 101 1 27NOV2018 VISIT 1 / BASELINE / DAY 1 101 2 28NOV2018 VISIT 1 / BASELINE / DAY 1 101 3 29NOV2018 VISIT 1 / BASELINE / DAY 1 101 9.00 06DEC2018 UNSCHEDULED VISIT 1 101 9.00 06DEC2018 UNSCHEDULED VISIT 1 101 9.00 06DEC2018 UNSCHEDULED VISIT 1 102 1 20NOV2018 VISIT 1 / BASELINE / DAY 1 102 2 04DEC2018 VISIT 2 / INTERIM / DAY 15 102 3 18DEC2018 VISIT 3 / END OF STUDY 102 9.00 01DEC2018 UNSCHEDULED VISIT 1 102 9.00 01DEC2018 UNSCHEDULED VISIT 1 102 9.00 01DEC2018 UNSCHEDULED VISIT 1 102 9.00 01DEC2018 UNSCHEDULED VISIT 1 run ; proc sql ; create view hs as select distinct * from have order subjid, visitdate ; quit ; data want (drop = _:) ; do until (last.subjid) ; set hs ; by subjid ; if find (visit, "UNSCHEDULED") then visitnum = sum (1, divide (_vn, 10)) ; output ; _vn = visitnum ; end ; run ; Though I concur with @andreas_lds that when you ask a question, you should present sample data in a usable form, from which the transition of the data\patterns from HAVE to WANT must be clear and unambiguous and easily detected without unnecessary guessing.   Kind regards Paul D.  ... View more

@teja5959: See also this thread from August 2018: assign unscheduled visits, in particular my post regarding numeric representation issues. In your example, the new visit number "1.7" (most likely) requires rounding, otherwise it will not be recognized in an IF or WHERE condition like visitnum=1.7. ... View more