On Apr 2, 2:56 pm, HERMA...@WESTAT.COM (Sigurd Hermansen) wrote: > Since this question arises about a often as any other on the 'L, I'll > chant some of the usual caveats. First, you are defining what amount to > key, as opposed to tuple (row) duplicates. If you extend one of the many > solutions that you'll see, it may not work as expected. For example, a > missing value of the Date attribute would force most of the solutions to > choose it over others with actual date values. Second, you are asking a > machine to make arbitrary decisions about content. What if the record > with the earlier date occurs by error? > > I'll stop there for the moment. Just remember that duplicates generally > display a symptom of a data integrity problem, not the data integrity > problem itself. Deleting one of a set of duplicates amounts to much the > same thing as throwing out data that don't fit a statistical model. Not > a good thing to do blindly .... S > > > > -----Original Message----- > From: owner-sa...@listserv.uga.edu [mailto:owner-sa...@listserv.uga.edu] > > On Behalf Of sdlenter > Sent: Monday, April 02, 2007 4:41 PM > To: s...@uga.edu > Subject: Remove duplicates keeping only the earliest date > > How do I Remove duplicates keeping only the earliest date > > Example: > > Field1 field2 Field 3 Date > 1212 df gt12 20070405 > 1212 df gt12 20070409 > 1312 df gt12 20050407 > > So after removing duplicates only two will be left > 1212 df gt12 20070405 (Earliest date) > 1312 df gt12 20050407- Hide quoted text - > > - Show quoted text -

Very apt reply Sigurd Hermansen.