Issue 18416

Irish occurrence dataset published as metadata-only

Reporter: kylecopas
Type: Feedback
Summary: Irish occurrence dataset published as metadata-only
Description: 3,371 occurrences at the source—the publishing bug continues.
Resolution: Fixed
Status: Closed
Created: 2016-04-19 17:51:03.547
Updated: 2017-10-06 10:32:48.961
Resolved: 2017-10-06 10:32:48.937

Author: rdmpage
Created: 2016-04-19 19:07:10.714
Updated: 2016-04-19 19:07:10.714
It's a little hard to investigate while indexing on the portal is suspended, but a simple fix might be to ensure that the

field in the meta.xml file is accompanied by


Author: kylecopas
Created: 2016-04-20 08:32:11.528
Updated: 2016-04-20 08:32:11.528
Thanks, Rod—it's a known bug in the NBDC platform. Until they fix it (as they've said they plan to), I'm simply trying to catch anything new as it arrives and flag it up for informatics.


Created: 2016-04-20 08:45:58.884
Updated: 2016-04-20 08:45:58.884
To elaborate a little:

When registering a dataset one declares it's type - in this case through the API from their system.  Currently we don't alter that programmatically (i.e. we don't look at possible attached data and infer it).  As we progress we should do so, but also consider that it is becoming increasingly meaningless - sample based data include occurrences, checklists include observations etc. and perhaps we should be considering multiple categories by which a dataset falls under - e.g. long term monitoring, checklist, observation, with preserved evidence (e.g. specimens), undigitised but with descriptive metadata, descriptive metadata but data available only on request etc.

It makes counting things less easy, but I don't think that should be our key goal - besides, it is easy to get high counts currently if that is one's wish (consider Plazi for example with 1000s of tiny recordsets).


Comment: The dataset has manually been changed to type OCCURRENCE.
Created: 2016-04-20 09:48:22.638
Updated: 2016-04-20 09:48:22.638