Date:         Fri, 2 Jun 2006 14:51:19 -0700
Reply-To:     "Pardee, Roy" <pardee.r@GHC.ORG>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Pardee, Roy" <pardee.r@GHC.ORG>
Subject:      Re: Question on pattern matching between two dissimilar sas
              datasets
Content-Type: text/plain; charset="us-ascii"

The transpose rubs me wrong too--I don't think you need it.

If the "straight merge" results you list below consist of one copy of
dset2 for each row in dset1 (aka a 'cartesian product') then that's
good--that's what you want.  You just throw a subsetting if in there to
strain out any rows that don't meet your criteria.  Something like:

Data gnu ;
        merge dset1 dset2 ;
        by <<something>>? ;

        if prxmatch(<<something clever>>) > 0 ;

Run ;

(Told you I couldn't remember how to do datastep merges...)

-----Original Message-----
From: sunspark@gmail.com [mailto:sunspark@gmail.com] On Behalf Of Peter
Constantinidis
Sent: Friday, June 02, 2006 2:38 PM
To: Pardee, Roy
Subject: Re: Question on pattern matching between two dissimilar sas
datasets

Hi thanks for writing, though I don't know how to use proc sql so I
guess I should read up and see if it can be of assistance because I
don't entirely understand your example code.

But you bring up a good point, I should describe a simplified form of
the datasets better.

Dataset 1: 10,000 observations, 2 variables- name and 30 digit serial
number.

Dataset 2: 5,000 observations, 1 variable- 5 digit number.

I can't do a straight merge because then I would have like:

Bob 123456890123456890123456890 99991
Jay 123456890123456890123456890 99992
etc..

That's why I thought transpose might work because then it'd be like:

Name Serial                                       var1 --- var5000
Bob 123456890123456890123456890 94991 -- 99999 Jay
123456890123456890123456890 94991 -- 99999 ...etc.

But it's not ideal to make it wide. Hmm.