Latest Schema method does not return "Schema Id" about avro

[nrwiersma]

The fix can be found in v1.2.0, with thanks to @mcgrawia for providing it.

[nrwiersma]

The subject latest endpoint does not return the schema id, just the version and the schema as seen here:

https://docs.confluent.io/current/schema-registry/develop/api.html

{
  "name": "test",
  "version": 1,
  "schema": "{\"type\": \"string\"}"
}

[mcgrawia]

Hi @nrwiersma I was trying to figure this out as well. Looking at the doc you linked above, it looks like the GET /subjects/(string: subject)/versions/(versionId: version) endpoint does return the schema ID: https://docs.confluent.io/current/schema-registry/develop/api.html#get--subjects-(string-%20subject)-versions-(versionId-%20version)

It isn't in the example response but it is in the response description. I was also able to confirm running the registry locally with the confluentinc/cp-schema-registry:latest docker image.

Response JSON Object:
--
subject (string) – Name of the subject that this schema is registered under
id (int) – Globally unique identifier of the schema
version (int) – Version of the returned schema
schema (string) – The schema string

Sample response:

{
    "subject": "test",
    "version": 1,
    "id": 2,
    "schema": "..."
}

thanks!

[nrwiersma]

No, there is a mismatch between the reported returns and the example. Interestingly I checked https://github.com/salsify/avro-schema-registry which I have for a long time used as my registry, and it does not return the id in this command. I assume it was added since the client was added.

I will gladly add an additional function to get it, to not break backward compatibility, but it should be noted for the kafka case, this should not be needed. We instead used fingerprint checking that is provided by https://github.com/salsify/avro-schema-registry which tells you if a schema is registered and what its id is. This allows the publication of multiple versions of the schema at the same time by multiple versions of your application.

[mcgrawia]

That would be fantastic, thank you. Let me know if a PR would be helpful. I'm relatively new to go but happy to try my hand at one.

One of the reason my use cases needs this ability is because our producing application consumes unique messages from customers. The application looks up the latest schema for a customer by subject, then writes the data to kafka. I don't think we would be able to use the pattern you described because our application does not have the schema in the source code?

[nrwiersma]

Fair enough.

If you would like to try your hand at it, that would be great, I am always glad to get PRs. We cannot add it to the current function at this time, as it would be backward breaking. I would suggest you add function GetLatestSchemaId with its own payload. You can follow the tests for GetLatestSchema as coveralls will not let it pass without them.

Have fun 😄

[mcgrawia]

@nrwiersma sounds good. I have a PR ready but I'm getting a "access denied" error when trying to push my new branch to the repo. Could you give me write access or is there another way to create a PR?

Thanks

[nrwiersma]

Hi. You need to make a fork of it and then make your branch. You can then create the PR against master on the main repo.

[mcgrawia]

thanks that did the trick

[nrwiersma]

Fixed by #56