+1

On Thu, Feb 5, 2026, 02:27 Alex Stephen via dev <[email protected]>
wrote:

> +1 (non-binding) to adding this!
>
> The spec is such a great way to understand what the REST Catalog does/not
> support. I always love adding missing things to it!
>
> On Wed, Feb 4, 2026 at 4:27 PM Yufei Gu <[email protected]> wrote:
>
>> +1 on adding the missing header. Thanks for raising this, Alex.
>>
>>
>> Yufei
>>
>>
>> On Wed, Feb 4, 2026 at 10:17 AM Prashant Singh <[email protected]>
>> wrote:
>>
>>> +1 to the adding of this header :)
>>> Thanks Alex !
>>>
>>> Best,
>>> Prashant
>>>
>>> On Wed, Feb 4, 2026 at 7:27 AM Alexandre Dutra <[email protected]>
>>> wrote:
>>>
>>>> Hi all,
>>>>
>>>> Thanks for confirming! Since it seems an oversight, here is a PR to add
>>>> it:
>>>>
>>>> https://github.com/apache/iceberg/pull/15231
>>>>
>>>> Thanks,
>>>> Alex
>>>>
>>>> On Wed, Feb 4, 2026 at 2:37 PM Gábor Kaszab <[email protected]>
>>>> wrote:
>>>> >
>>>> > +1 for adding this header to registerTable
>>>> >
>>>> > Best Regards,
>>>> > Gabor
>>>> >
>>>> > Eduard Tudenhöfner <[email protected]> ezt írta (időpont:
>>>> 2026. febr. 4., Sze, 13:32):
>>>> >>
>>>> >> This seems like an oversight to me, so it would be great if you
>>>> could open a PR to address this. I think we might be missing that Header in
>>>> a few other places as well, since Prashant opened
>>>> https://github.com/apache/iceberg/pull/14781
>>>> >>
>>>> >> On Tue, Feb 3, 2026 at 10:42 PM Alexandre Dutra <[email protected]>
>>>> wrote:
>>>> >>>
>>>> >>> Hi team,
>>>> >>>
>>>> >>> I've observed that the registerTable endpoint appears to be missing
>>>> >>> support for data access delegation [1]. IOW, it does not currently
>>>> >>> recognize the data-access header parameter [2].
>>>> >>>
>>>> >>> This is surprising, because the registerTable endpoint is
>>>> structurally
>>>> >>> similar to the loadTable or createTable endpoints, and all of them
>>>> >>> return the same LoadTableResult object.
>>>> >>>
>>>> >>> The consequence of this omission is that the returned table object
>>>> is
>>>> >>> likely unusable from a client perspective, as it will probably lack
>>>> >>> necessary properties for credential vending (or remote signing).
>>>> >>>
>>>> >>> Could this be an oversight? If so, I would be happy to contribute a
>>>> PR
>>>> >>> to add the required header support. If there's a specific reason for
>>>> >>> this difference, I'd appreciate it if anyone could explain why.
>>>> >>>
>>>> >>> Thanks,
>>>> >>> Alex
>>>> >>>
>>>> >>> [1]:
>>>> https://github.com/apache/iceberg/blob/d0654440a4cd00feaff73c53cd3ef9fb787b92dc/open-api/rest-catalog-open-api.yaml#L883-L898
>>>> >>> [2]:
>>>> https://github.com/apache/iceberg/blob/d0654440a4cd00feaff73c53cd3ef9fb787b92dc/open-api/rest-catalog-open-api.yaml#L1950-L1965
>>>>
>>>

Reply via email to