Best practices for external ids?

[vprivat-ads]

Hi!
I wonder what are the best practices to deal with external ids.
In our project we need to keep different kind of ids from external providers, and couldn't find a better way than to use properties like:

• provider1:id
• provider2:id

With dedicated STAC extensions that define these properties. It's a bit cumbersome when it results in creating an extension that only adds a single external id. I would like to know if there is a more generic approach.

@m-mohr I found this comment that looks related, did you create the extension you had in mind at the time?

[m-mohr]

@vprivat-ads There's a property called "externalIds" in Records and we could use that and just make a very simple extension out of it.
The issue that we faced in other projects and didn't make it into records though, is that some providers want to not just list random IDs in a list but keep track of there the ID came from (e.g. provider1 oder provider2 in your case).

As externalIds is just an array of strings, you'd either need to do this manually (e.g. prepending the id with provider1 and provider 2, so that the ID 1234-567-890 becomes provider1:1234-567-890 or so) or we'd need to define something that's not aligned with Records. That's exactly the issue that was discussed in opengeospatial/ogcapi-records#350. An alternative could be themes, but I found themes as a construct a bit too complex for search. Edit: See comment

I did not create anything at that time, ESA decided to just add those IDs to their EOPF extension, but I think it's time for a solution within STAC.

[m-mohr]

It seems Records made some last minute changes before publishing it and since I last looked at it.

I see that externalIds is now an array of objects instead of an array of strings, see https://github.com/opengeospatial/ogcapi-records/blob/master/core/openapi/schemas/recordCommonProperties.yaml#L62

So this should solve the issue. We should just create an extension based on that property in STAC and use it for your use case as well.

So you can now do something like this:

"externalIds": [
  { "scheme": "provider1", "value": "123" },
  { "scheme": "provider2", "value": "321" },
]

Just ensure that you can search for it easily with queryables...

[vprivat-ads]

nice, thank you! I'll give it a try in the coming weeks 👍

[vprivat-ads]

Hi @m-mohr, thanks a lot for creating the https://github.com/stac-extensions/external-ids! I will rework https://github.com/RS-PYTHON/cadip-stac-extension/ to use it. I wonder if you have already an example where you use the new extension?

[m-mohr]

Oh yeah, I forgot about this discussion, no problem. We have a couple of potential usecases where we discuss it as an option, but haven't implemented it yet.

[vprivat-ads]

If someone is looking for an example: here's our implementation of the extension: https://github.com/RS-PYTHON/cadip-stac-extension/pull/7/files