The Missing Option
If an observation has any variable missing, the default is to exclude it from the table and all calculations. You can override this by using the missing option.
Suppose you have a data set with 28 observations on hair color (brown or blond) and eye color (blue or brown), but one person is missing hair color and another is missing eye color.
input haircolor $ eyecolor $;
proc freq;tables haircolor*eyecolor;run;
Using the missing option will display "missing" as a level of the variable. Those observations will count in totals and percentages.
proc freq;tables haircolor*eyecolor/missing;run;