Using Proc format to code multiple values to missing

[HyunJee]

Thanks again for all your responses.

It appears there is not easy way to "fix" my problem. I was really hoping I could use proc format to recode all the variables, but it appears that I cannot do that if there are not actual missing data to begin with; which, unfortunately, several of my variables do not have missing values.

The only solution for me at this point is to recode every variable with the values I want as missing to missing.

Is there a way that SAS can fix this issue? Seems like it defeats the purpose of using a proc format to recode variables if in the end you have to end of having to recode each and every single one if they do not have any missing. Then you have to figure out which variables have no missing in the original data or just assume that all may possibly not have missing and recode each one, either way each option takes up time, which proc format is supposed to save you, but apparenlty not in the case with my current data.

[Tom]

There are probably cases where users would like values formatted to blank to NOT be treated as a missing category by PROC FREQ, so I am not sure there is change that SAS could make that would satisfy all potential users.

Using values like 77 or 99 to represent a missing value is more the way that SPSS works.  In SAS you would use special missing values like .A or .M. 

So in your particular case I think you will benefit from recoding the actual stored values to special missing values instead.  Either by making a copy of the dataset with the values recoded. Or a copy with new variables that have the special missing values.