| Date: | Thu, 20 Sep 2001 23:35:06 GMT |
| Reply-To: | pAyp <pauladx@HOME.COM> |
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
| From: | pAyp <pauladx@HOME.COM> |
| Organization: | Excite@Home - The Leader in Broadband http://home.com/faster |
| Subject: | Re: from proc fastclus to proc cluster |
|---|
This problem has been resolved, thanks to Zack Matthew.
Paula D
"pAyp" <pauladx@home.com> wrote in message
news:zadq7.23613$L%5.18888501@news1.rdc1.sfba.home.com...
> I follow SAS ondoc to run Proc fastclus first, so I can use its output as
> input into Proc cluster. The Var statement under Proc fastclus is like
this
>
> ..
>
> var v1, v2, v3....;
>
> run;
>
> At the proc fastclus line, I use out=lib.file1 to hold all the variables
> including a new field named Cluster and another new variable named
Distance,
> among others.
>
> Now I start to use the lib.file1 as input data set for proc cluster
>
> proc cluster data=lib.file1 method=ward ......................;
>
> The problem, and my question, is how to specify the VAR statement. SAS
> support consultant said I should copy and paste
>
> var v1, v2, v3..........;
>
> Which I tested. But if I do that, the whole thing runs as slowly as if I
did
> not run proc fastclus in the first place. Remeber: the reason that proc
> fastclus is recommended to run before running proc cluster is to speed up.
I
> don't know if I should specify
>
> VAR cluster distance;
>
> instead.
>
> Paula D
>
>
>
|