Find Original and Dup then Delete Dup

[Melis]

[color:purple]I have a huge phone number database, I need to delete the duplicate records, tried (!) and it found the dups but also the original records. Any ideas on how to only mark the dup records for deletion? Read the other posts on deleting dups but didn't find one that addressed marking the dups.

Thanks!! Mel

[-Queue-]

Do a search in FM's online Help for finding duplicates using a self-join relationship.

[Melis]

Did it, but the results are much larger than what I get when I do a Find(!). Is it necessary to have a "creation date" field, so that it will automatically delete the oldest record?

Thanks for your help, Mel

Hmmm.. should I try this?

Post: A simple duplicate remover (Topic#162164)

http://www.fmforums.com/forum/showtopic.php?tid/162164/post/162164/hl/+John+Fowler/#162164

Edited October 4, 2005 by Guest

[-Queue-]

Did you set up the relationship and calculation to return 'unique'/'duplicate'? If so, do a search for 'duplicate' records only or find '!' and then constrain the found set with 'duplicate' or 'unique'.

No date field is necessary to designate the original record. A calculation of serial = relationship::serial will return 1 for each original record because a relationship naturally refers to the first created related record, if no sort order is specified for it.

[Melis]

Yay! I did it, it worked! Thank you so much!! Thats 8,000 plus records I was going to have to mark by hand if I couldn't figure this out. Thanks again.

~Mel