Proper syntax would be:   proc sql; create table want as select amtspent, count(distinct custid) as uniqueCustomers from grv group by amtspent; quit; (untested) ... View more

REMOVE the VALUE keyword from the ELSE clause of your CASE. Use valid boolean logic in the WHEN clauses case when amtspent > 100 then "below 100" when 100<= amtspent <=200 then "between100 to 200" when amtspent > 200 then "above 200" else "nothing" end REMOVE the extra semi-colons in the middle of your CREATE statement. Give your new variable that the CASE is generating a name. ADD a FROM clause so SQL knows what you are selecting from. SQL does not need RUN statement. Each statement runs immediately. Use QUIT statement to end the PROC SQL step. proc sql; create table new as select * ,case when amtspent > 100 then "below 100" when 100<= amtspent <=200 then "between100 to 200" when amtspent > 200 then "above 200" else "nothing" end as new_var from have ; quit;     ... View more

I'll ask for a third and final time: please post example data in a data step with datalines. See my footnotes for how to do it. ... View more

To do thing conditionally use an IF statement. If you want one data step to generate two datasets lists them both in the DATA statement. You can use dataset options to control which variables are kept in each dataset. Here is an sketch of a program. Fill in the details based on what it actually in you text file. data h(keep=type a b c ) c(keep=type x y z); infile 'myfile.txt' truncover; input type $ @; if type ='H' then do; input a b c ; output h; end; else do; input x y z ; output c; end; run; ... View more

You asked for a code, but you did not gave any example of input (few lines) nor of the rules of how to decide to which category should be.   Your note is  "from airline name we have to categorize data" - can you supply airline names and define their categories?     ... View more