Date:         Wed, 22 Feb 2012 10:18:39 -0600
Reply-To:     Robin R High <rhigh@UNMC.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Robin R High <rhigh@UNMC.EDU>
Subject:      Re: Proc means to proc report
Comments: To: anon <rhian.pilling@GMAIL.COM>
In-Reply-To:  <201202221549.q1MFctfd001278@waikiki.cc.uga.edu>
Content-Type: text/plain; charset="US-ASCII"

Have you entered the missing option in TABULATE? also add the MISSING
statement in a DATA step


PROC FORMAT;
VALUE incm 1='less than 15000'  2='15000 to 30000'
           3='30000 to 45000'   4='greater than 45000'
          .n="no response"     .r="refused to answer"
          .u="unemployed"       .="unknown"  ;
RUN;

DATA thr;
MISSING n r u;  * the letters n r u for numeric data refer to
                  specific reasons for missing values of income listed in
PROC FORMAT;
INPUT id income @@;
age = 50+5*rannor(91092);
CARDS;
 1 1  2 .  3 r  4 4
 5 2  6 n  7 u  8 1
 9 2 10 r 11 1 12 .
13 2 14 3 15 n 16 .
;

PROC PRINT data=thr NOobs Label;
VAR id income age;
RUN;

ods output table=tbl4f(drop=_TYPE_ _PAGE_ _TABLE_);

PROC TABULATE data=thr NOseps missing;
CLASS income; var age;
TABLE income, age*(n='Count'*f=7.0 mean*f=5.2) / RTS=20;
FORMAT income incm. ;
RUN;
ods output close;

proc print data=tbl1; run;




Robin High
UNMC





From:
anon <rhian.pilling@GMAIL.COM>
To:
SAS-L@LISTSERV.UGA.EDU
Date:
02/22/2012 09:51 AM
Subject:
Proc means to proc report
Sent by:
"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>



Dear all,

I am struggling to distinguish between different codings for missing
values
in proc means and proc report.

I have been asked to produce a report for a variable showing median/iqr
etc
statsitics. I have therefore used proc means to achieve this (I have also
looked at proc univariate)as I know that this is not achievable by using
proc report alone.

The continuous variable I am using however has two different formats for
coding missing values. The missing data has been coded as . and also .n. I
need to separate the two out in proc report, however proc means (and proc
univariate when output to dataset) combine these two into one variable
missing.

Is there a way that I can separate the two formats and represent each in
the
proc report.

Thank you in advance

BW