I've just has an update of a Recorder dataset and have successfully imported it, but it appears that it has generated duplicates in the TAXON_OCCURRENCE table - nearly 500 of them. At least that is how it appears when I try to update my snapshot databases. Creating the custom 'denormalised' view of the data fails if I try to use TAXON_OCCURRENCE_KEY as a primary key. Taking this constraint off and performing a simple count query shows that there are now coming out twice.

At first I though that this was probably my fault. However, if I use the 'Go_To_Key' add-in, I find that four records come up. All the more surprising is that there are slight differences in the four records, all revolving around the 'Location' used. There appear to be two locations involved, and the four records shown are the four combinations of these locations used in the SURVEY_EVENT and SAMPLE tables.

So I've traced the problem now to the LOCATION_NAME table, and it appears that somehow some locations have ended up with two preferred LOCATION_NAME entries. It looks as though I can edit and correct these Locations (even though I am not custodian), so I can fix the issue on my system, although obviously I would be breaking the data ownership model by doing so!

I am still on 6.17 - I wanted to get these imports in before upgrading!