Date: Thu, 31 May 2007 23:21:22 -0700
Reply-To: cparatte <christophe@PARATTE.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: cparatte <christophe@PARATTE.NET>
Subject: Re: Dataset iteration
Content-Type: text/plain; charset="iso-8859-1"
On 1 juin, 04:46, davidlcass...@MSN.COM (David L Cassell) wrote:
> christo...@PARATTE.NET wrote:
> >Hi all
> >I have read a lot about dataset iteration here but i still have a
> >question :
> >is it possible to navigate in a dataset ? For example, i would like to
> >have a For Loop in which i can navigate to my dataset line[i], [i-1]
> >and [i+1] like in VB, C# or another language.
> >My problem is :
> >I have a dataset with multiple case_id which in turn may have multiple
> >offence_id with multiple item_id. I want to group them by case_id, a
> >certain type of offence_id and with a precise type of item_id in one
> >line... In fact changing the multiple lines in multiple fields for one
> >Thanks in advance
> >Christophe Paratte
> First, people often use PROC TRANSPOSE to take your originally
> data and turn it into a hideous nightmare that becomes harder and harder to
> work with as you progress through your process. I recommend that you not
> do that.
> Second, SAS actually provides a *ton* of ways to navigate through a
> data set and look for other records. The POINT= option is one thing
> you may already have seen. It lets you 'point' at records by record number.
> But there are a lot more ways of working with your data set, so POINT= might
> not be the best choice. It really depends on what you are trying to do.
> It may be that you are really looking for table look-up methods. If so,
> there is an entire paper or two by Art Carpenter in the SUGI archives at
> sas.com or lexjansen.com . Or perhaps you want a clever way to work
> thorugh your by-groups. by-processing might be what you want, or perhaps
> you would benefit from the intensive code of "The Magnificent DO" by Paul
> Dorfman, which can be found at Richard Devenezia's website. Then there's
> hashing, the Whitlock do-loop, the Schreier Self-Interleave, and any number
> of creative ways of using PROC SQL to access your data.
> So not only are there tons of choices here, but there's no way I can tell
> you what would be a really good choice for your data, unless you write back
> to SAS-L and explain your problems a lot more thoroughly.
> David L. Cassell
> mathematical statistician
> Design Pathways
> 3115 NW Norwood Pl.
> Corvallis OR 97330
> Don't miss your chance to WIN $10,000 and other great prizes from Microsoft
> Office Livehttp://clk.atdmt.com/MRT/go/aub0540003042mrt/direct/01/- Masquer le texte des messages précédents -
> - Afficher le texte des messages précédents -
Thank you very much for your answer.
I will read these documentations you gave me.
I am programming SAS for only 3 months so I do not have much
Have a good day