jingtail,

Don't forget

proc sql ; select ptid , max(slcond1 = 9) as scar , max(slcond1 = 10) as str , max(slcond1 = 11) as conj , max(slcond1 = 12) as oasa , max(slcond1 = 13) as extad from have group by ptid ; quit ;

and method 3 can be improved to one step

data Objective ( drop = slcond1 ) ; array w(9:13) SCAR STR CONJ OASA EXTAD; do slcond1 = 9 to 13 ; w(slcond1) = 0 ; end ; do until (last.ptid); set have; by ptid; if SLCOND1 in (9:13) then w(SLCOND1) = 1; end; run;

Ian Whitlock ================== Date: Mon, 3 Apr 2006 11:24:45 -0700 Reply-To: "jingtailan@gmail.com" <jingtailan@GMAIL.COM> Sender: "SAS(r) Discussion" From: "jingtailan@gmail.com" <jingtailan@GMAIL.COM> Organization: http://groups.google.com Subject: Array:Multiple obs into One obs. by PTID Comments: To: sas-l Content-Type: text/plain; charset="iso-8859-1" Dear all: Yesterday I posted a question on how to consolidate multiple observations into one observation by PTID. thank you for all your input and hints. Howard , "Data_Null_", Toby and David. I post all methods which i am aware of and hope you will learn it should there is a need. =============================================================== ORIGINAL DATA: data have ; infile cards ; input PTID SLCOND1 @@; cards ; 101 9 101 10 101 13 102 12 102 11 103 9 104 0 ; run; ============================================================== METHOD1:

data OBJECTIVE; set have; by ptid ; if slcond1 = '9' then scar = 1; else scar=0; if slcond1 = '10' then str = 1; else str= 0; if slcond1 = '11' then conj = 1; else conj= 0; if slcond1 = '12' then oasa = 1; else oasa=0; if slcond1 = '13' then extad = 1; else extad= 0; output objective; run;

proc means data=objective noprint; var scar str conj oasa extad; by ptid ; output out=objective1drop= _freq_ _type_) max= scar str conj oasa extad; run;

============================================================== METHOD2: data zeros; array _s(9:13) SCAR STR CONJ OASA EXTAD (5*0); output; run; data zhave(drop=slcond1); array _s(9:13) SCAR STR CONJ OASA EXTAD; point = 1; set zeros point=point; do until(last.ptid); set have; by ptid; if slcond1 then _s(SLCOND1) = 1; end; run; ======================================================== METHOD3: data Objective; array w(9:13) SCAR STR CONJ OASA EXTAD; do until (last.ptid); set have; by ptid; if SLCOND1 in (9-13) then w(SLCOND1) = 1; end; run; Data prep ( drop = I ) ; set objective ; array Nums (*) _numeric_ ; Do I = 1 to Dim(Nums) ; if ( Nums(I)=.) then Nums(I) = 0 ; end ; run ;