I am using proc tabulate to show demographic characteristics of a client population over several years. There are some missing values, and my understanding is that proc tabulate excludes the entire observation if a value for any variable in the class statement is missing. I want to keep all observations and show subtotals by demographic characteristic, without having a "Missing" or "Unknown" category for each group.    Here is what I have:   Here is what I want: Simplified version of code (I have several more class variables in the actual dataset): proc tabulate data=all; class raceeth gender year/missing; table (raceeth all='Subtotal')(gender all='Subtotal'), year*(N='No.' colpctn='%'*f=5.1); run; Is this possible in proc tabulate, or do I need to use a different method?  ... View more