Date:         Sun, 14 Mar 1999 01:51:54 +0000
Reply-To:     John Whittington <medisci@POWERNET.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         John Whittington <medisci@POWERNET.COM>
Subject:      Re: data manipulation
Comments: To: "Christian F. G. Schendera" <cschende@IX.URZ.UNI-HEIDELBERG.DE>
Content-Type: text/plain; charset="us-ascii"

At 19:10 13/03/99 +0100, Christian F. G. Schendera wrote:
 
>Aaaaaaaaah, sorry for unnecessary confusion!!!!!!!!!!!!
>Of course the dataset looks like:
>
>ID VAR1 VAR2 var3 var4 var5.........
>01     1          0      1     0     0
>02     0         1      0     0     1
>03     1          0      0     1     0
>.........................................
>I will try to be more precise. All colums contain single, and separated
>levels of categorial variables (I have changed the example above). The VARs
>stand for GENDER (var1, and 2), and US states. That guy infiled e.g. data
>for men (var1) and women(var2), but forgot to define the variable GENDER
>(having two levels, var1 and var2), or infiled data columes for the singles
>US states (say: Arkansas, var3,
>Washinton, var4, etc.) but missed to define the var USSTATES (having here 7
>levels).
 
Right, Chris, that makes a lot more sense :-).  I think what you therefore
need is something like:
 
data new (keep = id gender state) ;
    array gend(*) var1 var2 ;
    array stat(*) var2 - var8 ;
    set yourdata ;
    do i = 1 to dim(gend) ;
       if gend(i) = 1 then gender = i ;
    end ;
    do i = 1 to dim(stat) ;
       if stat(i) = 1 then state = i - 2 ; * or whatever - this gives state
                                             numbers starting at 1 ;
       if state = 1 then leave ; * saves a little time if there are many
states ;
    end ;
    * ... etc etc. for other group of variables ;
end ;
 
I coded the gender items in that way to show the general principle (as with
teh States).  With only two levels, a simpler solution for gender would be:
 
   if var1 = 1 then gender = 1 ; else gender = 2 ;
 
Regards,
 
John
 
----------------------------------------------------------------
Dr John Whittington,        Voice:    +44 (0) 1296 730225
Mediscience Services        Fax:      +44 (0) 1296 738893
Twyford Manor, Twyford,     E-mail:   medisci@powernet.com
Buckingham  MK18 4EL, UK              mediscience@compuserve.com
----------------------------------------------------------------