|
I don't think you are going to be happy waiting for your program to
finish with 118 PROC SURVEYSELECTS plus the other bits putting the
data back together and whatever else needs to be done. I think you
should back up and lets figure out a way to create data so that you
can get all the samples in one call to surveyselect.
Can you supply some sample data. The have and need scenario works
well on SAS-L. Perhaps the sample data from your first post. Plus a
mock up what you need from surveyselect.
I have been trying to understand your proceess but I could not follow
the process in the first post.
Perhaps the answer is in the Cassell paper.
On Tue, Mar 4, 2008 at 2:31 PM, maryidahosas <maryidahosas@yahoo.com> wrote:
> I am trying to get a monte carlo process working to build some
> coincidence factors by number of customers. I have had success in
> getting Proc SurveySelect to process the replications, but I still
> need my surveyselect to process several times with different
> sampsize . My code is below. I tried the macro loop for the
> processing, but it resolves before the survey select even operates, so
> it only processes for the J=End valuse of 118. Is there an easy way
> to execute this looping.?? Thanks for your help.
> Mary
>
>
> /*************************************************/
> %LET BEGIN = 1;
> %LET END = 118;
> %macro surv;
> %Do J = &begin %to &end;
> proc surveyselect data=mg1.resjul07bins
> out=mg1.sample1
> method=pps
> sampsize=&j
> rep=3
> seed=1;
> size binwidth;
> run;
> %end ;
> %mend surv;
>
> %surv
> run;
>
| 
