The UNION CORR ALL method is probably the best.  (without the ALL keyword UNION will remove duplicate rows). If you try to do it using a data step then you can get metadata attributes like FORMAT settings overriding the attributes set in the template. data template; input A B C; cards; ; data sample; input B A ; format a 4. ; cards; 4 1 5 2 6 3 ; proc transpose data=sample (obs=0) out=names ; var _all_; run; proc sql noprint ; select _name_ into :names separated by ' ' from names ; quit; data almost; set template(keep=&names) sample; run; proc sql noprint; create table want as select * from template union corr all select * from sample ; describe table almost ; describe table want ; quit; 144 describe table almost; NOTE: SQL table WORK.ALMOST was created like: create table WORK.ALMOST( bufsize=65536 ) ( A num format=4., B num ); 145 describe table want ; NOTE: SQL table WORK.WANT was created like: create table WORK.WANT( bufsize=65536 ) ( A num, B num );   ... View more