Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!gem.mps.ohio-state.edu!ginosko!uunet!cyclops!csense!bote
From: bote@csense.UUCP (John Boteler)
Newsgroups: comp.databases
Subject: Re: Deleting duplicate records
Message-ID: <305@csense.UUCP>
Date: 17 Jul 89 15:25:20 GMT
References: <9426@phoenix.Princeton.EDU>
Organization: Common Sense Computing, McLean Va
Lines: 32

From article <9426@phoenix.Princeton.EDU>, by englandr@phoenix.Princeton.EDU (Scott Englander):
> I'd like to delete duplicate records in a FoxBase file.  I tried setting
> the index to unique and then copying the database to a new file, but the
> duplicates remained.  I ultimately did it with a simple and SLOW program
> that checked for duplicates and deleted them (as i'm typing, it's
> running in the background under Multifinder -- there are 4000 records!).
> But there must be a better way to do this using the unique index, since
> only the first record with a given value of the index field is indexed.

Assuming that a 'duplicate record' in your case is one which has
the same key as another, and the other fields do not enter into
the decision, here is one suggestion:

set unique on
use &tablename index &indexname
go top
do while .not. eof()
	delete	&& marks this record to be copied later
	skip	&& advances to next unique record in index
enddo
copy to &temp for deleted()	&& copy only those marked
!mv tablename.dbf tablename.old
!mv temp.dbf tablename.dbf
go home	&& relax!

Now, before you remove the old file, are you sure that those records
which did not make it into the new file are truly bogus? rm is forever.

-- 
Bote
uunet!cyclops!csense!bote
{mimsy,sundc}!{prometheus,hqda-ai}!media!cyclops!csense!bote