allow hiding/disabling copies of datasets when users download from OpenData

07-12-2021 02:57 PM
Status: Under Consideration
Labels (2)
Occasional Contributor II

Apparently every time someone downloads an item from OpenData, it creates a copy of this item in Content.  I would like the option to disable this or at least hide these copies.  (is there a step or setting I am not seeing?)  I understand this may be the dropdown that sometimes shows 'latest' vs 'cached', perhaps creating this as a result.  If it is the way the backend works though, I would want to hide it from our Orgs so we never see these copies.  Content will end up with hundreds or thousands of copies datasets taking up space and creating confusion for Org users.

Status changed to: Under Consideration


Thanks for your feedback. For private content, Hub uses the "Export" function to create a downloadable file, just like you'd use if you were on ArcGIS Online. As a result, we need to create an exported item (under "item-exports"). We have longer-term plans to move away from the process of caching files. We're still evaluating if in the mid-term we can technically implement an automated deletion of these files. I do not believe that there is a way to hide these files from Content > Search in ArcGIS Online, so for now you'll need to either restrict your search to certain groups, or delete the folder.


Perfect. Implementing auto-deleting in the near term would be great. And in the long-term, making this concept invisible to the user / Org. Perhaps another short-term approach could be a notebook Solutions could provide that would do this auto-deleting.  It appears this is only an issue for Private items, because as you outlined, it is a different API (Portal). I had been concerned that for the public OpenData site, hundreds of thousands of residents would overflow the content with cached copies. Nevertheless, I would also want to hide/remove these items for Org users, as well. There's no use case for keeping them. Or at least for exposing them visibly. (I understand they are used for the caching for the backend)


To summarize if anyone comes across this thread: this is only an issue when data is not shared public to Everyone.  Private data uses Portal's API and public data uses AGOL. (I would suggest in the long term to go from two APIs to one on the backend)  However I would like to see private data behave this way also in a future update, where it does not create copies when downloaded; and is searchable by name.


Thank you and the team again for implementing CDN caching for Enterprise items. I saw the potential on that right away and I think is a huge development for Open Data and Hub.