Date:         Thu, 13 Jan 2000 11:04:21 -0500
Reply-To:     Howard Schreier <Howard_Schreier@ITA.DOC.GOV>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Howard Schreier <Howard_Schreier@ITA.DOC.GOV>
Subject:      Re: Tabulate question
Content-Type: text/plain; charset=US-ASCII

You can do it by normalizing the data set. PROC TRANSPOSE is my
preferred method. In this case it's necessary to generate a BY variable
so that each observation is transposed individually. I would do this in
a view, just for efficiency.

Tested code:

data test;
input var1 var2 var3;
cards;
1 2 3
4 . 5
2 3 4
;
data totransp / view = totransp;
set test;
* Add a suitable BY variable;
__by_obs = _n_;
run;

proc transpose data=totransp out=to_tabul;
by __by_obs;
var v:;
* Should pick up all variables beginning with letter "v";
run;

proc tabulate data=to_tabul;
class _name_ col1;
table _name_, col1*n*f=4.;
run;



Date:    Thu, 13 Jan 2000 08:43:08 -0500
From:    dennis.catley@CAPITALONE.COM
Subject: Tabulate question

     Greetings O wise SAS wizards near and far:

     a quick aside to my question below.....Pudge, to me, will always be
     remembered as a Boston Red Sock...the big home run is just too
     engrained, him waving the ball fair at Fenway.

     On to the question....

     I had a user ask the following, and for the life of me I just can't
     seem to get tabulate to produce a good likeness of his report desires.
     Thoughts anyone? And thanks in advance for all efforts!!!

     Dennis Catley
     Capital One Financial Services, Inc.

     -----------------------
     Do you know is it possible, or how to generate a table like the
     following in SAS?

     In the following table, 1-5 are the values of the variables, and n are
     the count/frequency of corresponding values.


     Var_name   1       2       3       4       5
       var1     n       n       n       n       n
       var2     n       n       n       n       n
       var3     n       n       n       n       n