No need for the format in this case for the missing values. The Tabulate table option MISSTEXT will suppress the . for the N statistic:   proc tabulate data = have out = want; class year category; keylabel n=' '; table year, category*N / misstext=' '; run; @Patrick wrote: Not sure if I understand but are you asking for something as simple as below? If so then the main thing that's missing in your code is a comma in the table statement after variable year - used to separate the x from the y axis. data have; input id category $ year; cards; 1 X 2009 2 U 2009 3 X 2009 4 Y 2010 5 Y 2010 6 Z 2011 7 Z 2011 8 X 2012 9 U 2013 10 U 2013 ; proc format; value missasblank .=' ' other=[best16.] ; run; proc tabulate data = have out = want; class year category; keylabel n=' '; table year, category*N*f=missasblank.; run;     ... View more

proc freq data=have; tables main_category*state; run; If you want to combine failed and cancelled, you can do that with custom formats, or in a DATA step.   In the future, please provide example data in a usable form, as a SAS data step (instructions). Data in a screen capture is not usable. ... View more

Alternatywą dla CARDS4 jest DATALINES4, schemat jest następujący: data test; infile DATALINES4 dlm=';'; input a b c; DATALINES4; 1;2;3 4;5;6 ;;;; run;   Bart. ... View more