Thank you for a very comprehensive answer. Maybe I just have to accept this loss of precision, but then I would expect it to be the same for all roundings on 276578736.46500, right? But if I use input(276578736.46500, 15.2) it rounds to .47, aswell as if I use "format=15.2" in PROC SQL? data rounding; d1=round(201578736.46500, .01); /*28 bit integer*/ d2=round(276578736.46500, .01); /*29 bit integer*/ d3=round(971578736.46500, .01); /*30 bit integer*/ d4=input(276578736.46500, 15.2); run; proc print data=rounding noobs; run;
... View more
