I'm not sure that will solve my problem. With it I can change each variable into an observation, but then I have 100k new variables from all of the observations. I would need some way of consolidating all of those variables into a smaller range of variables, one variable containing a count for each specific value or value range that appeared, correct? And even if I do that, how do I account for the invalid values? In general, I can't create a generic variable that contains invalid counts, as there are times when it is important to know what the invalid values actually are.
I guess my original question may have been this. Is it possible to use proc report or proc tabulate to create a table where each row is a different variable and the columns are predefined ranges and any invalid values in the variables (the predefined ranges I believe would come from a proc format)? Each cell on the table would then contain a count of how many observations contained that value or range for that variable. The output from running proc freq on 100-200 similar variables (3-4 times) is lengthy and cumbersome to analyze. The hope was that we could just use another procedure to consolidate the information, and to save paper.
I apologize, I am fairly new to SAS. I had great success teaching myself macros and a lot of other things, but this seems like it should be easy and is really giving me a difficult time. Thanks again for any help!