Find only one record amonst duplicates

[cabsandy]

Hi all,

I want to find/keep only one record of recurring duplicates-please see attatched.

In column RF node, the first 3 records are all the same-I want to get to a point where I only keep the first one of the 3.The next 3 again are all the same-I only want to keep one.The next 8 are the same-I only want to keep one-and so on.

Could somebody advise me on the best way to do this-the aim is to have an exportable summary file that contains only one record of each, rather than multiples.

Thanks

Sandy

testfile.zip

[Ugo DI LUCA]

A must see here...

calcing a field to purge duplicate

and then jump to Andy's site to see it at work.

[cabsandy]

Ok-I've found out what my mistake was.Thanks for the help :-)

Sandy

[Søren Dyhr]

What if you choose the last of the 3 identical records instead??? Then could this be used:

http://www.fmforums.com/threads/download.php?Number=133778

The problem with the other method is that the search is unindexed due to the involved relation, and if you tag every record in the file, is it definately going to take some time. The method I suggest works best when the number of dupes are large, since my script only handles the record on both sides of the break values - But often are you not allowed to make scripted replaces in a networked solution!

--sd

[Søren Dyhr]

the aim is to have an exportable summary file that contains only one record of each

Ah I've just figured out, that Export summaries STILL exists with 7 (called group by) but you're on 6 so it's probably called Summarize by..., if you have a summary field in you file like I have - will the export dialog have an extra feature if the data to get exported has been sorted - So the recipee is make a summaryfield Count( on the RF node field. I made a test and it boils you data of 200 records down to 52 records.

So Ugo excellent solution ....unfortunaly is it to the wrong task!

--sd

[cabsandy]

Soren,

Many thanks for the response-I've been playing with the solutions offered above and it all seems to work.But I'd like to throw down a challenge-is it possible to use or adapt any of these to get, where possible, 2 duplicates, rather than the one? The reason is , as you've mentioned, the " boil" number can be quite low, especially on those sheets where the original record number is low to start with.And I only have FP6 so I dont know if the above file will work for me :-(

Thanks

Sandy

[-Queue-]

Relationship::field refers to the first related field according to a relationship's sort order. So if you define a relationship to sort in a specific order, say by boil number descending, then the record marked as unique will be the one with the largest boil number value. Perhaps this will help you.

The solution works the same in 6 as it does in 7. serial = selfrelationship::serial is the calculation for both.

[Fenton]

This saves 1 & 2.

testomit1_2.zip

[cabsandy]

Fenton,

Works like a charm-thank you very much.Thanks to all who responded.

Sandy

[Søren Dyhr]

I could imagine the thrill it is to make a bunch of crafty developer jumping hoops, but what the h... is the purpose?? ....and what if no dupes exists - is it the idea to dupe the singled out record?? a sync'ing of the two is an apparent problem. Is this a system like the one Jon Rosen had up in Advisor several years ago, to prevent recordlocking ...say when a collegue leaves his computer for lunchbreak with the cursor right in a records field???

--sd