topic Re: Proc Tabulate and summing measures in SAS Programming

Proc Tabulate and summing measures

robm — Mon, 17 Apr 2017 23:08:34 GMT

I have 2 Proc tabulates where I am trying to show 2 different measure summed over different acacdemic year terms basically i am trying to sum in one

-the Tuition from each ALLOCATED_COLLEGE/SUBJECT_DESC for each column TERM_CODE

-the Student count for each ALLOCATED_COLLEGE/SUBJECT_DESC for each column TERM_CODE

IE I have change the VAR for each but they still sum exactly the same ...any suggestions?

the output from below can be seen in the attached pdf

proc tabulate data=ROBM.TBIR1025 missing contents = ' ' format=COMMA6.0 S=[foreground=highlight.];
class ALLOCATED_COLLEGE SUBJECT_DESC TERM_CODE;
var INCLUDED_TUITION;
keylabel Sum = " ";
table (ALLOCATED_COLLEGE =' ') *(SUBJECT_DESC=' ' all='Total') all='Total', ((TERM_CODE=' ' all='Total')*((N=Amount))) /
contents = ' ' misstext=' ' box={label="ALLOCATED_COLLEGE/SUBJECT_DESC"};
run;
 
proc tabulate data=ROBM.TBIR1025 missing contents = ' ' format=COMMA6.0 S=[foreground=highlight.];
class ALLOCATED_COLLEGE 'SUBJECT_DESC'n 'TERM_CODE'n;
var STUDENT_COUNT;
keylabel Sum = " ";
table (ALLOCATED_COLLEGE =' ') *(SUBJECT_DESC=' ' all='Total') all='Total', ((TERM_CODE=' ' all='Total')*((N=Count))) /
contents = ' ' misstext=' ' box={label="ALLOCATED_COLLEGE/SUBJECT_DESC"};

Re: Proc Tabulate and summing measures

LaurieF — Tue, 18 Apr 2017 00:10:44 GMT

Can you provide some sample data? It's much easier to play around with the problem if there's something concrete to work on.

Re: Proc Tabulate and summing measures

Astounding — Tue, 18 Apr 2017 01:21:46 GMT

You will have to learn a little more PROC TABULATE, I'm afraid. You have never asked for the SUM of anything. You only asked for the N statistic, which gives you a count of the nonmissing values. So the count of the nonmissing values doesn't change. You're allowed to ask for many different statistical measures, but you only get what you ask for.

Re: Proc Tabulate and summing measures

robm — Tue, 18 Apr 2017 02:19:56 GMT

you aew right I building these of a EBI server

  proc tabulate data=ROBM.TABBS99991 missing contents = ' ' format=COMMA6.0 S=[foreground=highlight.];
   class &showvarvalue &showvarvalue1 &acrossvarvalue;
   var &measurevar;
   keylabel Sum = " ";
   table &showby1exp &showby2exp &showvarall, (&acrossexp(&measureexp)) /
                  contents = ' ' misstext=' ' box={label="&boxlabel"};
  run;

and (&acrossexp(&measureexp)) should have just been (&measurevar)

  proc tabulate data=ROBM.TABBS99991 missing contents = ' ' format=COMMA6.0 S=[foreground=highlight.];
   class &showvarvalue &showvarvalue1 &acrossvarvalue;
   var &measurevar;
   keylabel Sum = " ";
   table &showby1exp &showby2exp &showvarall, (&measurevar) /
                  contents = ' ' misstext=' ' box={label="&boxlabel"};
  run;

  proc tabulate data=ROBM.TBIR1025 missing contents = ' ' format=COMMA6.0 S=[foreground=highlight.];
   class 'ALLOCATED_COLLEGE'n 'SUBJECT_DESC'n 'TERM_CODE'n;
   var STUDENT_COUNT;
   keylabel Sum = " ";
   table ('ALLOCATED_COLLEGE'n =' ') *('SUBJECT_DESC'n=' ' all='Total') all='Total', (('TERM_CODE'n=' ' all='Total')*((STUDENT_COUNT))) /
                  contents = ' ' misstext=' ' box={label="ALLOCATED_COLLEGE/SUBJECT_DESC"};
  run;


  proc tabulate data=ROBM.TBIR1025 missing contents = ' ' format=COMMA6.0 S=[foreground=highlight.];
   class 'ALLOCATED_COLLEGE'n 'SUBJECT_DESC'n 'TERM_CODE'n;
   var INCLUDED_TUITION;
   keylabel Sum = " ";
  table (ALLOCATED_COLLEGE = ' ') *(SUBJECT_DESC = ' ' all = 'Total') all='Total', ((TERM_CODE = ' ' all = 'Total')*((INCLUDED_TUITION))) /
                  contents = ' ' misstext=' ' box={label = "ALLOCATED_COLLEGE/SUBJECT_DESC"};
  run;

Re: Proc Tabulate and summing measures

ArtC — Tue, 18 Apr 2017 05:36:57 GMT

Can we now assume that you are getting what you want?

Re: Proc Tabulate and summing measures

robm — Tue, 18 Apr 2017 13:20:32 GMT

yup sorry I should have spent more time examining and understanding my PROC TABULATE