This is something I could just avoid, but I would rather try to understand it.  EXAMPLE: The early years of the NHANES data do not calculate a ratio of two variables, but later years do, and it is URDACT.  If I try to replace the missing in the DATA when I combine them with the SET statement, I get a single value.  If I do the same replacement in a DATA step after the combining, I get the correct results.  What logic am I missing?  THANKS! data alb_cr; set dat_.LAB16 dat_.L16_b dat_.L16_c dat_.alb_cr_d /*I removed the names of the rest of the data sets*/... ; if missing(URDACT) then URDACT=(URXUMA*100/URXUCR);   proc print data=alb_cr(obs=10); var seqn URDACT; run;   data alb_cr; set dat_.LAB16 dat_.L16_b dat_.L16_c dat_.alb_cr_d  /*I removed the names of the rest of the data sets*/... ; data alb_cr; set alb_cr; if missing(URDACT) then URDACT=(URXUMA*100/URXUCR);   proc print data=alb_cr(obs=10); var seqn URDACT; run;   Obs SEQN URDACT 1 2 6.27586 2 3 6.27586 3 5 6.27586 4 6 6.27586 5 7 6.27586 6 8 6.27586 7 9 6.27586 8 10 6.27586 9 11 6.27586 10 12 6.27586 Obs SEQN URDACT 1 2 6.2759 2 3 8.2540 3 5 3.5465 4 6 4.0323 5 7 5.2344 6 8 5.0413 7 9 4.1379 8 10 3.9785 9 11 2.6506   ... View more