LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (October 2005, week 5)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Mon, 31 Oct 2005 16:44:20 -0500
Reply-To:     Chang Chung <chang_y_chung@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Chang Chung <chang_y_chung@HOTMAIL.COM>
Subject:      Re: Trying to avoid division by zero...
Comments: To: oseithedude@GMAIL.COM

On Mon, 31 Oct 2005 12:59:33 -0800, oseithedude@GMAIL.COM wrote:

>Hello, I have around 40 or so numeric variables in a small program I am >writing. >Basically, the variables are being created as follows: > >var1 = (a/(b*c)); >var2 = (f/(h*d)); >var3 = (dog/(cat*mouse)); >var4 = (k/(l*var1)); >etc.... > >and so on for about 40 variables. My problem is that, even though the >final data seems to be coming out correctly, I am getting thousands of >"Division by zero detected..." errors for the cases where the >denominator in the above equations equal zero. It turns out that if the >denominator for any of the equations is equal to 0, then the variable >actually does not need to be created anyway. Is it posible to write >simple code that prevents the above mathematical operations from being >performed if the denominator equals zero or do I have to encase each of >the 40 variables I am creating with an if... then statement for the >case that the variables in the denominator are equal to zero? >Appreciate any help in this matter, thanks!!!

Hi, Julie, I think you answered your own question in the part: "It turns out that if the denomicator for any of the equations is equal to 0, then the variable ..." This would actually partition the data into two: observations that are used in the calculations, and observations which are not. I would distinguish these two groups of observations first and do the calculations on the former group only like:

data one; set two; /* do something similar like this one to create a flag var */ inSample = not ((b=0) or (c=0) or (h=0) or (cat=0) or (mouse=0) or (L=0)); if inSample then do; /* do calculations */ end; else do; /* results set to missing */ end; run;

If you save the inSample flag, then it will be useful later.

Cheers, Chang


Back to: Top of message | Previous page | Main SAS-L page