This solves perfectly. Thank you very much for your responses. ... View more

I would rather have column names instead of mvarX. I tried as below but it didn't work. data _null_; set I_DATES; array t{*} _numeric_; do i=2 to dim(t); call symputx(vname(i), put(t{i},date9.)); end; run; ... View more

Thank you for your reply. This helps a lot.   But I do not want to hardcode column names. In the next run I may have additional columns like I_PAYMENT_DATE, I_RESPONSE_DATE, etc. I can already hardcode each column using SQL into statement. ... View more

Thank you very much for your response. Unfortunately I just created a simple sample but in fact I will use those macro variables in various steps of a very long process. I do not want to join or write another select for each. I would really prefer to convert them to macro variables.      ... View more

Let me give some further information. I wish to store my filters in a table with column names. Then I'm using into statement for using the values in later filtering. Here is how I use it:   /*Here I use into statement for taking filters*/ PROC SQL NOPRINT; select I_REPORT_DATE into: I_REPORT_DATE from WORK.I_DATES QUIT; PROC SQL NOPRINT; select I_END_DATE into: I_END_DATE from WORK.I_DATES QUIT;   /*Here I use filters within a SQL statement*/ PROC SQL; CREATE TABLE WORK.SAMPLE_OUTPUT AS SELECT t1.reporting_date, t1.approach   FROM prt.PORTFOLIO_TABLE t1 WHERE t1.reporting_date = "&I_REPORT_DATE"d AND t1.ending_date = "&I_END_DATE"d; QUIT;   With this I'm using two PROC SQL statements with INTO function for defining filters. I wish to have a dynamic code. If I input more filters with column names, it would define each of them.   Thank you.   ... View more