Date:         Wed, 13 Mar 2002 08:53:35 -0500
Reply-To:     Charles Patridge <Charles_S_Patridge@PRODIGY.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Charles Patridge <Charles_S_Patridge@PRODIGY.NET>
Subject:      Re: Many many to many merge/joins
Comments: To: Paul Dorfman <paul_dorfman@HOTMAIL.COM>

Dear Nigel,

My dear boy!  Do you really want to compare names and addresses from two
files by doing a many-to-many sql join????

I do not know the specifics of what you are asking of the Computer God but
I suspect he will immediately send you to the Under God of Below the Lowest
Priority Possible on any computer system.

Again, I am not sure of your QUEST but maybe a suggestion will help:

1.  How about merging the two files and when you have a perfect match, save
the record and then delete it from future matches.  Hopefully, you will get
something on the order of 30% (maybe even higher) matches.  Then you will
only deal with those records that did not match (hopefully reducing the
task you asked of the Computer God).

2.  You are welcome to borrow the Fuzzy Merge Routine from SCONSIG.com to
make your QUEST a bit more manageable - still large but at least there will
not be the need to do a MANY-to-MANY SQL JOIN which is VERY MASSIVE in your
case.

I am willing to provide a little of my time to guide you through the code
if needed, or make suggestions when requested.

HTH,
Charles Patridge
Email: Charles_S_Patridge@prodigy.net

PS, I have found doing such tasks much better on FAST PCs than the BIG IRON
machines provided you have the hard drive space and speed.