Date: Fri, 7 Jan 2011 14:13:59 -0500
Reply-To: Arthur Tabachneck <art297@ROGERS.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Arthur Tabachneck <art297@ROGERS.COM>
Subject: Re: Calculating Readmissions
Since your macro variables aren't data dependent, as I recall, move them to
the start of your code. They don't have to be in a datastep.
Then, after running your code, look at the files created and see if the data
at least looks like you would expect.
Do some frequency distributions on some of them, especially the site and
other variables that for which you've already indicated observing
The extra dates in the array shouldn't be too much of a bother but, without
sample data, don't really know.
On Fri, 7 Jan 2011 14:06:25 -0500, HIMS <hims.sperritt@SYMPATICO.CA> wrote:
>Thanks very much for replying so quickly.
>You are correct, I only want when chartno isn't missing.
>The portion about dates was for what was provided to me in the original
>response. The data actually only has data from 4/1/2005 to 3/31/2010 so I'm
>assuming I can just change that?
>The macro variables were also provided in the original post so what should
>they be? Note that the original database has 8938 abstracts and one would
>assume that I should have the same volumes resulting but I actually only
>6019 which is why some cases aren't showing up. Any assistance in fixing
>this would be greatly appreciated.
>From: Arthur Tabachneck [mailto:art297@ROGERS.COM]
>Sent: Friday, January 07, 2011 1:34 PM
>To: SAS-L@LISTSERV.UGA.EDU; Shelley Perritt
>Subject: Re: Calculating Readmissions
>Without some sample data it's really difficult to say. However, I did
>notice some things that are/may be contributing factors.
>First, you use: where chartno NE " ";
>Is that really what you want or do you just want:
> where not missing(chartno)
>Second, you define your array to include all dates from Jan 1 1990 thru Dec
>Is that what you intended to do?
>Finally, and probably the cause, you create macro variables bh and eh, and
>then attempt to use them in the same datastep.