Proc tabulate. No print missing

[Elena]

Hi,

I have this dataset and I'd like to do a proc tabulate to produce a tabulate like the one below but without the column missign.

data com;
infile datalines;
   input pop   sot    
         ;
datalines;
1  1
2  .
3  1
1  .
2  1
3  .
1  .
2  1
3  1
1  .
2  .
3  1
;
run;


PROC TABULATE  data=com missing;
CLASS  pop sot ;
TABLE (pop all='Totale'),
     sot * (rowpctn) ;
RUN;

 thanks!

[Shmuel]

Just run:

PROC TABULATE  data=com   /* missing  */;
CLASS  pop sot ;
TABLE (pop all='Totale'),
     sot * (rowpctn) ;
RUN;

[Elena]

Run but not well.

The table whithout missing in this

[Astounding · Accepted solution]

When you want the % of a flag, you are better off using 0 instead of missing in your data:

data com;

infile datalines;

input pop sot;

if sot=. then sot=0;

datalines;

..........

;

Then you have choices, but the basic idea is to compute the MEAN of the variable.  You can label it as the Row % if you want, but calculate the mean.

proc tabulate data=com;

class pop;

var sot;

tables (pop all='Totale'), sot*mean='Row %';

run;

This will get you a decimal fraction, such as 0.25.  It you really want to convert this, you could apply a format:

sot*mean='Row %' * f=percent9.2;

[ballardw]

Try

PROC TABULATE  data=com ;
CLASS  pop;
class sot /missing ;
TABLE (pop all='Totale'),
     sot * (rowpctn) ;
RUN; 

Though I preferr to use 0/1 coded variables as Var variables as N, Sum and Mean become very useful as a group: N= total valid responses (either 0 or 1), Sum= number of 1 valued responses and Mean with a percent format is the percentage of 1 responses.