I want to create 2 flags: one is if there is change of more than 5 points between recs and another flag if there is no change in score from recs to another 3 consecutive recs.   I am not figuring it out on how to do this and need help.   Required flag1 and flag2 and need something like below PID recs score flag1 flag2 121 1 11     121 2 23 Y   121 3 23   Y 121 4 23   Y 121 5 23   Y 122 1 11     122 2 14     122 3 24 Y   122 4 34 Y   122 5 22     122 6 11     ... View more

How to resize sas datasets. I found a program in sas but it gives errors. one of the error is  http://support.sas.com/resources/papers/proceedings13/206P-2013.pdf   ERROR: File WORK.MAXX_T.DATA does not exist.   May be i am missing something.  Any one knows program which does the resizing. please let me know   %macro chglen (inp=sl); proc contents data = &inp. noprint out = charvar (where =(type=2) keep = memname name type length label); run; data charvar2; length name1 name2 name3 name4 name5 $100; set charvar; name1 = trim(name)||'1=length('||trim(name)||')'; name2 = 'max('||trim(name)||'1) as '||trim(name); name3 = trim(name)||'x = '||trim(name); name4 = trim(name)||' = '||trim(name)||'x'; name5 = trim(name)||' $'||strip(put(length, best.)); run; proc sql noprint; select trim(name1) into: newvar separated by '; ' from charvar2; select trim(name2) into: maxvar separated by ', ' from charvar2; select trim(name3) into: tname separated by '; ' from charvar2; select trim(name4) into: tnamex separated by '; ' from charvar2; select trim(name) into: cvar separated by ' ' from charvar2; select trim(name)||'x' into: cvarx separated by ' ' from charvar2; select trim(name5) into: cvarlen separated by ' ' from charvar2; quit; data varlen; set &inp.; &newvar; run; proc sql noprint; create table maxx as select &maxvar from varlen; quit; data maxx_t2; length name6 $100; set maxx_t; if col1<1 then col1=1; name6 = trim(name)||' $'||strip(put(col1, best.)); run; proc sql noprint; select strip(name6) into: newlen separated by ' ' from maxx_t2; quit; data temp (drop = &cvar. ); length &cvarlen; set &inp.; &tname.; run; data newone (drop= &cvarx.); length &newlen.; set temp; &tnamex.; run; %mend chglen; %chglen; ... View more

Hi, I want to create a dummy randomisation list to randomly assign 2 treatments for pat id's. i found many examples but if i get a new pat id then the program should randomly assign one treatment and not change the existing treatment groups. is there any existing program which does or an example would be appreciated. assigning trt A and trt B like below pat_id       trt 101           trt A 102          trt B 103          trt B 104          trt A   ... View more

I have few datasets and i am splitting them to different folders based on conditions. The problem is if there is one folder where there is no condition satisfied then it just creates a dataset with 0 observations. Can i remove them or have some condition for not creating them if condition is not satisfied.?   I am using the belwo code: proc datasets nolist; copy in=analysis out=work; quit; run; %*split the data into different folders based on fold; data _null_; set sashelp.vtable (where=(libname="WORK")); call execute('data sa.'||strip(memname)||'; set WORK.'||strip(memname)||'; where fold="19A"; run;'); call execute('data s1a.'||strip(memname)||'; set WORK.'||strip(memname)||'; where fold="10A"; run;'); call execute('data s2a.'||strip(memname)||'; set WORK.'||strip(memname)||'; where fold="12A"; run;'); run; ... View more