Suport for OneOf Input Objects

[cabeca]

Question

Hello!

I tried to look for it in the Roadmap, and in the issue tracker but I could not find any information about it. Apologies if I missed an obvious place.
I was wondering if there are any plans to support OneOf Input Objects on apollo-ios, now that it has been added to the GraphQL spec.
Is this addition being considered an if so, do you have any rough timeline for it?

Thanks!

[AnthonyMDev]

Hi @cabeca. We've added this to our roadmap. The RFC hasn't been officially added to the spec yet, but we do anticipate that it will be very soon.

No guarantees on a timeline for this, but once it's finalized and added to the spec, we'll get to it. We anticipate the work to support this to be relatively minimal.

[cabeca]

Thank you for the update!

[github-actions]

Do you have any feedback for the maintainers? Please tell us by taking a one-minute survey. Your responses will help us understand Apollo iOS usage and allow us to serve you better.

[ruudk]

FWIW, it's now officially supported: https://graphql.org/blog/2025-09-04-multioption-inputs-with-oneof/

[calvincestari]

@ruudk - Apollo iOS is ready, @oneOf has been implemented and available since 1.16.0.

[glenngijsberts]

@calvincestari Hey Calvin, there is still an issue that the Apollo iOS codegen is not fetching the isOneOf property in the introspection query and so the types aren't generated properly when downloading the schema. Since this is now officially supported, does this PR makes more sense now? apollographql/apollo-ios-dev#705. I haven't looked at the alternative "hack" that Android was doing, but I guess that's not needed anymore?

[calvincestari]

there is still an issue that the Apollo iOS codegen is not fetching the isOneOf property in the introspection query and so the types aren't generated properly when downloading the schema.
Ah yes I remember now!

I haven't looked at the alternative "hack" that Android was doing, but I guess that's not needed anymore?

We can't break compatibility with existing servers which are slow to implement the new spec. I think we will need to look into the approach that @BoD shared. I haven't looked at the full spec update so I'm not sure what else may need to be wrapped up into this too.