Date:         Mon, 8 Jan 2001 08:31:28 -0800
Reply-To:     "Huang, Ya" <ya.huang@AGOURON.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Huang, Ya" <ya.huang@AGOURON.COM>
Subject:      Re: Help: PROC FREQ with 1000 variables..
Comments: To: Victor Aina <aina@SFU.CA>
Content-Type: multipart/alternative;

Victor,

A little bit of modification of my old post might
get what you want:

data xx (keep=vname y vvalue);
  set xx;
length vname $ 8;
array v inquiry lien judgemnt trade sex;
do over v;
vvalue=v;
call vname(v,vname);
output;
end;

proc sort;
by vname;

proc freq data=xx noprint;
output out=yy chisq;
table y*vvalue / chisq;
by vname;
run;

proc print data=yy;
run;

HTH

Ya Huang

-----Original Message-----
From: Victor Aina [mailto:aina@SFU.CA]
Sent: Sunday, January 07, 2001 11:10 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Help: PROC FREQ with 1000 variables..


Hello everyone:

I need to run the FREQ procedure on more
than 1000 variables with names like INQUIRY,
LIEN, JUDGEMENT, TRADE, SEX etc.
The objective is to compute the relationship
of these variables with one dependent variable,
say Y.
I would like to extract  some of the statistics
generated by the PROC FREQ e.g. Chi Square value
and the associated p-value for each of the variables.

Then combine all these 1000 together and sort them
according to say the p-value or any criteria.

I've written a little macro that I can only
call 1000 or more times e.g.

%cfreq(dsn=data_res, dependvar =Y , indep_var =LIEN );
%cfreq(dsn=data_res, dependvar =Y , indep_var =SEX );
.
.
.
%cfreq(dsn=data_res, dependvar =Y , indep_var =XMED );


Grateful if anyone could point me to a more
efficient way of handling the programming.

--


+----------------------------------------------------------+
| victor aina  |  e-mail: aina@sfu.ca | fax:(972) 255-6955 |

[text/html]