@madara155 wrote:
Thanks everyone for your thoughts.
As per everyone, it seems I can't avoid using a Data step. Will proceed to do that with the help of some SAS generate code.
If you use Proc Import to generate the data step code it will generate separate code bits for every single variable.
A table with 2000 variables normally got variable names that follow some naming pattern like myvar_1 to myvar_2. If you've got such patterns then your SAS datastep code can become much simpler to "type" or potentially generate via some script.
data demo;
infile datalines dsd dlm=',' truncover;
informat
key best32.
col1-col6 $1.
;
input key col1-col6;
datalines;
1,a,b,c,d,e,f
2,g,h,i,,k,l
;
... View more