Date:         Mon, 3 Jun 2002 16:56:03 -0400
Reply-To:     "Fehd, Ronald J." <rjf2@CDC.GOV>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Fehd, Ronald J." <rjf2@CDC.GOV>
Subject:      Re: efficiency expertise needed
Comments: cc: Karen Olson <karen.olson@TCH.HARVARD.EDU>
Content-Type: text/plain

> From: Karen Olson [mailto:karen.olson@TCH.HARVARD.EDU]

from your earlier posts I remember that you're working with proc FREQ.

Consider the possibility of creating one cross-tab with the maximum number
of variables:

PROC FREQ data    = LIBRARY.ENORMOUS;
          tables    M * N * O * P * Q ...
          /         noprint
           out    = LIBRARY.XTAB
          (drop   = Percent
           rename = (Count = Weight);

Any lesser cross-tab can be produced from XTAB
using the weight statement:

PROC FREQ data    = LIBRARY.XTAB
          tables    ...;%*any vars, any order from above list;
          weight    Weight;

This is drill-down in action.

hth

Ron Fehd  the macro maven  CDC Atlanta GA USA RJF2@cdc.gov

By using your intelligence
you can sometimes make your problems twice as complicated.
-- Ashleigh Brilliant