First off, you can do just about anything with SAS. Basically, I took your code and wrapped the parts that run the proc dbf and proc variogram in a macro, since that code runs repetitively for each dbf in the folder. I added a data step after the ods output to reformat your ods output then appended it to a results data set that captures the stats for each run. Also, instead of using the list with the file names, I opted to read the file names directly from the folder by piping the dir command. Here's the code. %let mydir = c:\data\test; libname mylib "&mydir"; DATA mylib.u107all; length CODE $ 56.; infile "&mydir\u107all.csv" dsd dlm=',' firstobs=2; input CODE $ thevariable; RUN; %macro analyze_dbf( fnm ); %let last4 = %substr( %scan( &fnm, 1, . ), 5, 4 ); filename m "&mydir\&fnm"; PROC DBF db5=m out=mylib.m&last4; run; PROC SQL; create table mylib.ma&last4 as select A.CODE, A.thevariable, B.CODE, B.POINT_X, B.POINT_Y from mylib.u107all A, mylib.m&last4 B where A.CODE = B.CODE; QUIT; ods output "Autocorrelation Statistics" = mylib.moran&last4; PROC VARIOGRAM data=mylib.ma&last4;  compute novar autoc (weights=distance);  coordinates xc=POINT_X yc=POINT_Y;  var thevariable; run; quit; ods output close; data stats; ma = &last4; do until ( done );    set mylib.moran&last4 end=done;    if coefficient =: 'Moran' then do;       morano = observedvalue;       morane = expectedvalue;       mz = z;       mp = pvalue;    end;    else if coefficient =: 'Geary' then do;       gearyo = observedvalue;       gz = z;       gp = pvalue;    end; end; output; keep ma morano--gp; run; proc append base=mylib.results data=stats; run; %mend; filename x pipe "dir /b &mydir\*.dbf"; data _null_; infile x; input fnm :$15.; call execute ( cats( '%analyze_dbf(', strip(fnm), ');' )); run; ... View more

Set up an array where the dimension is based on the columns for the shares outstanding for 1997 - 2011. Then use the years variable as the index to pick the correct value. data shr_out; set dat.shr_out; array shr_out(1997:2011) _1997--_2011; float = offer_amount / coalesce( shr_out( years ), shr_out( years+1 )); run; The float calculation returns some missing values because your sample does not always have the shares outstanding for the IPO year or the year after. ... View more

To build on data_null_'s reply, here's something to get you started. data _null_; rc = filename( 'mydir', 'c:\temp' ); did = dopen( 'mydir' ); count = dnum( did ); do i = 1 to count;    filenm = dread( did, i );    putlog filenm; end; run; Note that the dread function returns not only file names but directories too. You'll have to account for that. ... View more