Date: Fri, 9 May 2003 11:06:42 -0400
Reply-To: Sigurd Hermansen <HERMANS1@WESTAT.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Sigurd Hermansen <HERMANS1@WESTAT.COM>
Subject: Re: Combining variables
Content-Type: text/plain; charset="iso-8859-1"
SAS PROC SUMMARY uses binary numbers to accomplish the same result more
efficiently for each class of values. It will also handle n-ary variables,
although the number of distinct states of ten n-ary variables can add up to
very large numbers.
The number of distinct states of ten variables often amounts to a small
fraction of the maximum number possible. Structural dependencies eliminate
many possibilities.
Sig
-----Original Message-----
From: Gianluca [mailto:gianluca.marchi@ITALY.ACNIELSEN.COM]
Sent: Wednesday, May 07, 2003 3:55 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Re: Combining variables
"Brent Baxter" <bax@U.WASHINGTON.EDU> wrote in message
news:200305061955.h46JtOY00992@listserv.cc.uga.edu...
> Is there an easy way to combine several character variables into one
> variable for the purpose of viewing composite frequencies? I have cases
> (drivers) with varying numbers of traffic violations (violations codes
> stores in ViolCode01-ViolCode10). I'd like to produce a frequencies table
> that show all the violation codes together, rather than the frequencies in
> ViolCode01 through ViolCode10. Thanks for any help!
Hi Brent,
you can use this simple way to codify composite characters:
suppose you have 3 characters: A, B, C; they can be present or not and can
combine each other.
You would face 8 cases:
none, only A, only B, only C, A+B, B+C, A+C, A+B+C
now we define a new variable X which codifies the events.
X=A*2^0+B*2^1+C*2^2
So the previous cases will resolve in:
none => X=0
only A => X=1
only B => X=2
only C => X=4
A+B => X=3
B+C => X=6
A+C => X=5
A+B+C => X=7
Obviously you can use this system to manage much more character.
Hope this can help you.
Bye
Gianluca
