Options for choosing what download format is available

[ArtRod_ODF]

We use the Hub open data platform to provide access to several datasets that are impacted when downloaded as a shapefile. Impacts include truncated field names, coded domain descriptions not being transferred, and more. It would be great to provide this data as open for download, but remove the option to download as a shapefile.

Thanks!

-arthur

[BruceSmith]

This would make a great enhancement to the HUB. Additionally, it would be beneficial to enable administrators to choose preferred download options for datasets.

[KatieMorgan]

Yes please we would love this, we are also impacted by the Shapefile issues. I'd like this to be available as a site-wide option for all our datasets, but also individually per dataset. 

Ideally it would be good to extend this to all formats! Occasionally we have datasets we know just won't work as well in a specific format for whatever reason and having this control would provide a better experience for our site users, and remove the need to include (probably complex for users) instructions in the dataset summary if there are any particular format considerations. 

[ThomasHervey1]

Thank you all for your comments. We are in the planning phase to offer the ability to hide/show/rearrange certain download formats on an individual item basis.

@ArtRod_ODF @BruceSmith @KatieMorgan to help with prioritization, I'm curious to know more about your needs for disabling download formats and I have a few questions.

• Beyond problematic shapefiles, are there other formats that give you problems?
• Can you elaborate on why you would like certain users to hide/show/rearrange formats?
• When compared to the download buttons, how often do your users use the download API or API explorer to re-use content?
• If an item is shared across multiple Hub sites, should the formats display the same way across all of them?

[KatieMorgan]

@ThomasHervey1 that's great news, really pleased this is being looked at. 

You mentioned shapefile which is the main problematic one due to the performance issues and truncated field name issues etc. already mentioned. 

CSV is also problematic because it automatically re-formats date fields to the US style mm/dd/yyyy but doesn't 'understand' that the date is encoded in this way so then when we open the files the dates become invalid. This is complicated to explain but for example in our feature service table and in all other download formats today's date would appear as a date type field with a value of 20/06/2024 or 20240620 depending on the format, and a date picker can parse that as the correct date. But the CSV format would reformat it as 06/20/2024, but it 'thinks' that the format is dd/mm/yyyy creating an invalid date. I have a call open with Esri UK about this separately but I've noticed it on a number of providers' Hub sites. 

That said, we would not hide CSV because it's widely used, but for some key public datasets that are generally downloaded by non-technical users and opened in Excel I'd like to prioritise Excel files at the top of the downloads list. 

Essentially we know certain datasets will be used more as spreadsheet downloads and users of those may find the wide array of formats a bit confusing so we'd want to prioritise the more 'basic' formats a the top of the list. And then some will be used more as GIS downloads, and we'd want to prioritise those differently. From our user communities I imagine some formats like SQL-lite will be used less. 

Our Hub site has only been live for a short time so I don't have data yet on whether users are using the downloads or APIs more, but I'd imagine it will be strongly weighted towards downloads as they are often used to share data with partners/contractors, or to answer requests for information from the public. 

We are only using this in one Hub site currently but I imagine if we created multiple we'd want to have the ability to configure the data downloads differently on each site. 

Hope that helps but @ me if you need anything else!

[ArtRod_ODF]

@ThomasHervey1 Thank you for reaching out! Happy to provide more details about our situation.

The only real problematic format for our example is the Shapefile. However, this dataset experiences frequent geometry updates as well as attributes, so the CSV and Excel formats are not really useful, and not recommended. (The problems with CSV and dates as @KatieMorgan mentioned are really good points.)

This would not be user dependent, instead it is more dataset dependent. The idea would be to turn off the Shapefile option from at least the Open Data download page for this dataset. All users would experience the same thing.

We don't have exact numbers on downloads vs service use. We are promoting the Service use as much as possible. We do have some users in remote locations with limited bandwidth, or even taking the data offline, which may explain some of the download needs. We also have some users accessing data from other GIS software platforms. Is there a way to check number of downloads and/or compare against hits on the Service?

Having the format maintained across all Hub sites would work well.