Santosh,

Okay I would suggest an array approach. However, before we go there lets stop and think about what you are doing. Why do you want the data in such a denormalized form? Perhaps if you write back to the list explaining that we could better help you. I have just spent the last week working with denormilized data sets. This denormalized data form will make your code extremely hard to work with, cause you to hard code variables in, and finally create holy heck and havoc when you need to do something very simple.

Toby Dunn

From: "santosh.k" <santosh.kothamsu@CHASE.COM> Reply-To: "santosh.k" <santosh.kothamsu@CHASE.COM> To: SAS-L@LISTSERV.UGA.EDU Subject: Re: Data transformation question Date: Mon, 5 Dec 2005 08:47:46 -0800 Let me transform my question:

I have a SAS dataset in the following form. cat cycle field pass miss dm 5008 bal 10 15 dm 5009 bal 12 20 dm 5010 bal 15 9 ib 5009 chg 5 12 ib 5010 chg 6 11 ob 5008 bal 10 15 ob 5009 bal 12 20 ob 5010 chg 12 13

I need to create a SAS dataset in the following form: 1) grouped by 'cat' and 'field' 2) ' pass' and 'miss' values spread across as observations for each field Attached is the template

cat field p5008 p5009 p5010 m5008 m5009 m5010 dm bal 10 12 15 15 20 9 ib chg . 5 6 . 12 11 ob bal 10 12 . 15 20 . ob chg . . 12 . . 13

Thanks, Santosh