Thank You again. I guess I should back track. I am creating a chapter in a book. Each chapter takes one factor and splits it by a number of different factors. The data for each chapter was created using a BY statement. For one chapter it took 50 different BY statements for 13 different variables. The BY statement creates a firstdim and a seconddim variable. This created a problem because the data table it created was huge and unreadable. So a new variable was created with lables for each category of the variable but created a 'MISSING" label for the rest. So we started with something like this from the BY statements (X = firstdim Y = 2ndDim) X Y Z a 6 g b 7 h c 8 i d 9 j e 7 k f 1 l And ended up with something like this were X was broken to create two new variables with lables and the same for Z X Y Z W A B C a 6 g a ' ' g ' ' b 7 h b ' ' h ' ' c 8 i c ' ' i ' ' d 9 j ' ' d ' ' j e 7 k ' ' e ' ' k f 1 l ' ' f ' ' f Now using the PROC TABULATE statement with MISSING to create tables Var Y class W A B C; table W*B, Y; table W*C, Y; table A*B, Y; table A*C, Y; The problem is that this creates a value for the MISSING variable (which is the sum of the Y for the missing) when I am not interested in it. If I use a where startment to select only A's with values than I end up with W's with no values. Because I am creating so many tables and I would like each chapter to be in one spreadsheet. I could modify the data but it would be easier to create a MACRO in excel....which I guess is where I am at. Thank you for your help.
... View more