No, you were clear – I misread it! ... View more

Can also be done in a DATA step, but PROC SQL might be better since you don't have to sort.     data mydata; input patient_id $ medication :$11. claim; datalines; ID1 medication . ID1 medication . ID1 medicationX 1 ID1 medication . ID2 medication . ID2 medication . ID2 medication . ID3 medicationX 1 ID3 medication . ID3 medicationX 1 ID3 medication . ; run; proc sort data = mydata; by patient_id descending claim; run; data want; set mydata; by patient_id; retain new_claim; if first.patient_id then call missing(new_claim); if not missing(claim) then new_claim = claim; run; Obs patient_id medication new_claim 1 ID1 medicationX 1 2 ID1 medication 1 3 ID1 medication 1 4 ID1 medication 1 5 ID2 medication . 6 ID2 medication . 7 ID2 medication . 8 ID3 medicationX 1 9 ID3 medicationX 1 10 ID3 medication 1 11 ID3 medication 1 This is commonly called last observation carried forward or LOCF.   ... View more

I would recommend marking @PaigeMiller's response as the correct one. ... View more

@PaigeMiller's solution was untested, so this was expected.   Use LIBNAME = "WORK" instead. Always use caps when you use dictionary.tables/columns. ... View more

Are you able to show your actual code?    You say you use two numeric variables to create the third, but you're only showing one? I'm just a little confused. Maybe I'm missing something. ... View more

So the value would contain dollar amounts?   Any reason it isn't numeric? ... View more

Ah, my apologies. ... View more

I haven't used it in this way, but this gives you the green boxes for SASHELP.CLASS.   proc report data=sashelp.class; column age name sex height weight; define name / order; define age / order; define sex / display; define height / display; define weight / display; compute weight; if weight >95 and height>63 then do; call define('height','style','style={background=lightgreen}'); end; endcomp; run; Changed height from analysis to display, and removed the .sum in the call define.   Again, not sure if that's what you need. ... View more

INTCK and INTNX (INTNX is what you would want) takes care of leap years.   https://blogs.sas.com/content/iml/2017/05/15/intck-intnx-intervals-sas.html ... View more

Are you getting an error?   Code is untested, but you can just move that COALESCE to the join.   PROC SQL; CREATE TABLE date_naiss AS SELECT pers.date_naiss , sousc.no_pol , COALESCE(ctant.ctant_pere, ctant.contrac) AS coal FROM dossier sousc , dossier2 ctant , personne pers WHERE sousc.countr = COALESCE(ctant.pere, ctant.contrac) AND ctant.pers = pers.pers ; QUIT; Again, I think that works but I can't tell without data.   I'd assume the calculated doesn't work in the join. ... View more

Okay, echoing @PaigeMiller, do you have any code you can post? Can you also post your log of what isn't working?   For code, click the insert sas code icon in your post and paste it in. For your log, you can click the "insert code" icon (</>) and paste it in there. ... View more

Are you putting quotes around year in the INTNX function? ... View more

There is a pinned tab on the main page of the SAS Programming forum that discusses this. You may have to modify some things, but it should get you started. ... View more