Date:         Mon, 18 Mar 2002 16:17:13 -0500
Reply-To:     clmakinson@GOODYEAR.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
Comments:     To: Bruce Butts <BBUTTS%internet@akr.goodyear.com>
From:         clmakinson@GOODYEAR.COM
Subject:      Re: MDDB vs. SAS datasets
Comments: To: SAS-L@LISTSERV.VT.EDU
Content-type: text/plain; charset=us-ascii

MDDB is ideal if you want to use any of the EIS tools.   To query the MDDB
with PROC SQL you must resort to MDX; but this will eliminate the need to
maintain both MDDB's and  Datasets.  If you don't need EIS then the Summary
Datasets may be easier to build and maintain.    Big problem here was that
some of our 'facts' were ratios of other columns (like Btu/lb) which are
easily handled by adding computed columns to the EIS Metabase.  However,
since these do not actually exist in the MDDB they can't be queried with
MDX.  Proc Summary doesn't understand ratios either so all the levels above
the lowest were wrong.  Our solution was post process the  Proc Summary
output dataset to fake the summary of ratios.

***********************************************************
Contains confidential and/or proprietary information.
May not be copied or disseminated without express
consent of The Goodyear Tire & Rubber Company.
***********************************************************



                    Bruce Butts
                    <BBUTTS@GW.ODH.ST        To:     SAS-L@LISTSERV.VT.EDU
                    ATE.OH.US>               cc:
                    Sent by: "SAS(r)         Subject:     [SAS-L] MDDB vs. SAS datasets
                    Discussion"
                    <SAS-L@listserv.v
                    t.edu>


                    03/18/02 03:35 PM
                    Please respond to
                    Bruce Butts






Greetings SAS-L,
I am  considering creating MDDBs to summarize and store about 2 million
records of Birth and Death data summarized on year, county, age, race, sex
and other demographic variables.  I am also considering storing the same
data in SAS data sets as the output from Proc Summary at the same levels I
would build the data cubes with Proc MDDB.  We will be using the datasets
to respond to external requests through a browser (SAS/Intrnet & App Dev
Studio applications) and for internal users to access for research
purposes.

Does anyone have evidence or opinions on the relative merits of creating
MDDBs vs. summary SAS data sets?

Thanks in advance for your responses.

Bruce Butts,
Ohio Department of Health