Thanks.  So I can't just use style= in the proc line, I'd have to manually change the elements?  That's too bad, but at least this is something. ... View more

So if I wanted to create a PDF of this report is there anyway to do such a thing?  Automatically that is, rather than manually going to the export button. ... View more

Ah sorry, I misread what you were doing there. Here's the\A solution for the original problem, should anyone ever come here looking for how to get it. data real; do numbers = 1,2,3,4;   do letters = 'A','B','C','D';   output;   end; end; run; proc format; value $L 'A'='A' 'B'='B' other = 'X'; run; data dummy; do letters = 'A','B','C','D','E'; numbers = 0;   output;   end; run; data out; set real dummy; run; option missing = 0; proc tabulate data=dummy missing; where letters = 'E'; class letters/preloadfmt; var numbers; tables (letters='')*(   numbers * sum='' *f=comma24.0 )   /printmiss;   format letters $l.; run; ... View more

PGStats:  you're so close to what works though!  Just need the x=0 line and tabulate will play. data myClassData; do page = 1, 2;     do line = 1 to 4;         do cat = "A","B","C","D","E"; x= 0;             output;             end;         end;     end; run; ... View more

If letters = 'D', there the table is three rows tall as written in the original tabulate A                    B                 X numbers     numbers     numbers .                     .                  10 To your latest reply: I'm using tagsets.excelxp.  There's a dozen tabs with 3 tables each, each created as the result of a different where.  These tabs are then linked to an excel book to create a dashboard because SAS can't create the appropriate layout that was requested.  As a result, I'm relying on the tables to be in the same location on each tab.  So if table 1 populates and takes up rows 1-6, and table 2 is supposed to be rows 8-12, but doesn't populate or output then table 3 is going to be at row 8 instead of row 14. For me, this point is moot now because I've created the solution using the dummy data, I was just hoping there was an option or something that would tell sas or tabulate to create the table, even if there is no data for it.  The solution ended up being to make sure that I had a value for each Where possibility and it had a value for the summary variable of zero, rather than missing.  Tabulate has no problem outputting a table of zeros. ... View more

X-5 was my "real data" example. Here, it'd be X-2ish.  That is one row for the values, one row for the A\B\Other header and then some number of rows for the page spacing in the excel book (rows between tables). ... View more

The reason I needed the empty table was for the output so that Row X is always Row X and not X-5 or whatever.      ... View more

As much as I enjoy the snark, let's discuss what the issue is and what solution I've had to use. If we use [where letters = 'A'] (or B or C or D) then we get a table with three columns, no matter which letter we pick in the data: A   B   Other X    .      . However, with [where letters = 'E'], we get no output.  What I was looking for was to still get the 3 column table with everything missing. A     B     Other   .      .         . What I have done as a workaround is create a dummy table for each of my where\class values and set numbers = 0 for each of these dummy values.  Then when I get to the E table, I get a table with three columns output, zeros as each value. A     B     Other 0      0       0 This works enough for my purposes, though if someday another letter (F?  J?  Who knows?) enters the fold, I'd have to add it to the list in my dummy data. data dummy; do letters = 'A','B','C','D','E'; output; end; run; So silly. ... View more

I can make a dummy set to merge into the real data for my output, but I still have to deal with this: WARNING: A logical page containing only missing values has been deleted from the output. ... View more

Neither of you are listening to the problem.  The problem is that I want the table to exist, even if E doesn't exist in the data. ... View more