You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@iceberg.apache.org by huaxin gao <hu...@gmail.com> on 2026/08/13 02:07:17 UTC
[VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Hi all,
Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
Iceberg V4" [1], I'd like to call a vote on the following proposal for the
V4 table spec.
Proposal
--------
1. Writing new equality deletes is forbidden for V4 tables: the V4 metadata
will not define equality deletes as an allowed entry type.
2. Reading equality deletes remains supported in the reference
implementation for backward compatibility, both for existing V2/V3
tables and for equality deletes carried over into upgraded V4 tables.
3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
of data or delete files). Existing equality deletes remain in carried-
over V2/V3 delete manifests; converting them to deletion vectors is a
separate, optional maintenance action.
Rationale
---------
Equality deletes impose an asymmetric cost paid on every read, complicate
the format, and block features such as CDC, row lineage, and incremental
index/materialized-view maintenance. Deletion vectors make deletion a flat,
one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
viable replacement path, so we do not need the full replacement completed
before forbidding new equality deletes in V4.
The vote will be open for at least 72 hours.
[ ] +1 Forbid writing equality deletes in V4
[ ] +0
[ ] -1 Do not forbid (please explain)
[1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
Thanks,
Huaxin
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Anurag Mantripragada <an...@apache.org>.
+1 (non-binding)
On Thu, Aug 13, 2026 at 10:49 AM Amogh Jahagirdar <2a...@gmail.com> wrote:
> +1 (binding) Thanks Huaxin!
>
> On Thu, Aug 13, 2026 at 10:46 AM Anoop Johnson <an...@apache.org> wrote:
>
>> +1 (non-binding) to deprecate equality deletes in v4. Thank you, Huaxin!
>>
>> On Thu, Aug 13, 2026 at 10:36 AM Ryan Blue <bl...@apache.org> wrote:
>>
>>> +1 (binding) to deprecate equality deletes. Thanks, Huaxin!
>>>
>>> Ryan
>>>
>>> On Wed, Aug 12, 2026 at 8:16 PM huaxin gao <hu...@gmail.com>
>>> wrote:
>>>
>>>> Hi Manu,
>>>>
>>>>
>>>> No rewrite of the existing manifests is needed. The V4 root manifest
>>>> references the existing V2/V3 manifests as-is (data manifests and the older
>>>> delete manifests), so upgrade stays metadata-only regardless of table size;
>>>> only the new top-level V4 metadata is written. The equality deletes in
>>>> those carried-over delete manifests are still applied on read (with 2-phase
>>>> planning) until they age off or are converted to DVs by a separate
>>>> maintenance action.
>>>>
>>>>
>>>> Thanks,
>>>>
>>>> Huaxin
>>>>
>>>> On Thu, Aug 13, 2026 at 10:18 AM Manu Zhang <ow...@gmail.com>
>>>> wrote:
>>>>
>>>>> Hi Huaxin,
>>>>>
>>>>> To clarify the third point: will upgrading a V2/V3 table to a V4 table
>>>>> require a manifest rewrite? If so, how will that process be handled?
>>>>>
>>>>> Thanks,
>>>>> Manu
>>>>>
>>>>>
>>>>> On Thu, Aug 13, 2026 at 10:07 AM huaxin gao <hu...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes
>>>>>> in
>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal
>>>>>> for the
>>>>>> V4 table spec.
>>>>>>
>>>>>> Proposal
>>>>>> --------
>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>>> metadata
>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>
>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>> tables.
>>>>>>
>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>>> rewrite
>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>> carried-
>>>>>> over V2/V3 delete manifests; converting them to deletion vectors
>>>>>> is a
>>>>>> separate, optional maintenance action.
>>>>>>
>>>>>> Rationale
>>>>>> ---------
>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>> complicate
>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>> incremental
>>>>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>>>>> flat,
>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates
>>>>>> a
>>>>>> viable replacement path, so we do not need the full replacement
>>>>>> completed
>>>>>> before forbidding new equality deletes in V4.
>>>>>>
>>>>>> The vote will be open for at least 72 hours.
>>>>>>
>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>> [ ] +0
>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>
>>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>
>>>>>> Thanks,
>>>>>> Huaxin
>>>>>>
>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Amogh Jahagirdar <2a...@gmail.com>.
+1 (binding) Thanks Huaxin!
On Thu, Aug 13, 2026 at 10:46 AM Anoop Johnson <an...@apache.org> wrote:
> +1 (non-binding) to deprecate equality deletes in v4. Thank you, Huaxin!
>
> On Thu, Aug 13, 2026 at 10:36 AM Ryan Blue <bl...@apache.org> wrote:
>
>> +1 (binding) to deprecate equality deletes. Thanks, Huaxin!
>>
>> Ryan
>>
>> On Wed, Aug 12, 2026 at 8:16 PM huaxin gao <hu...@gmail.com>
>> wrote:
>>
>>> Hi Manu,
>>>
>>>
>>> No rewrite of the existing manifests is needed. The V4 root manifest
>>> references the existing V2/V3 manifests as-is (data manifests and the older
>>> delete manifests), so upgrade stays metadata-only regardless of table size;
>>> only the new top-level V4 metadata is written. The equality deletes in
>>> those carried-over delete manifests are still applied on read (with 2-phase
>>> planning) until they age off or are converted to DVs by a separate
>>> maintenance action.
>>>
>>>
>>> Thanks,
>>>
>>> Huaxin
>>>
>>> On Thu, Aug 13, 2026 at 10:18 AM Manu Zhang <ow...@gmail.com>
>>> wrote:
>>>
>>>> Hi Huaxin,
>>>>
>>>> To clarify the third point: will upgrading a V2/V3 table to a V4 table
>>>> require a manifest rewrite? If so, how will that process be handled?
>>>>
>>>> Thanks,
>>>> Manu
>>>>
>>>>
>>>> On Thu, Aug 13, 2026 at 10:07 AM huaxin gao <hu...@gmail.com>
>>>> wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes
>>>>> in
>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for
>>>>> the
>>>>> V4 table spec.
>>>>>
>>>>> Proposal
>>>>> --------
>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>> metadata
>>>>> will not define equality deletes as an allowed entry type.
>>>>>
>>>>> 2. Reading equality deletes remains supported in the reference
>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>> tables and for equality deletes carried over into upgraded V4
>>>>> tables.
>>>>>
>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>> rewrite
>>>>> of data or delete files). Existing equality deletes remain in
>>>>> carried-
>>>>> over V2/V3 delete manifests; converting them to deletion vectors is
>>>>> a
>>>>> separate, optional maintenance action.
>>>>>
>>>>> Rationale
>>>>> ---------
>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>> complicate
>>>>> the format, and block features such as CDC, row lineage, and
>>>>> incremental
>>>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>>>> flat,
>>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>>>> viable replacement path, so we do not need the full replacement
>>>>> completed
>>>>> before forbidding new equality deletes in V4.
>>>>>
>>>>> The vote will be open for at least 72 hours.
>>>>>
>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>> [ ] +0
>>>>> [ ] -1 Do not forbid (please explain)
>>>>>
>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>
>>>>> Thanks,
>>>>> Huaxin
>>>>>
>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Anoop Johnson <an...@apache.org>.
+1 (non-binding) to deprecate equality deletes in v4. Thank you, Huaxin!
On Thu, Aug 13, 2026 at 10:36 AM Ryan Blue <bl...@apache.org> wrote:
> +1 (binding) to deprecate equality deletes. Thanks, Huaxin!
>
> Ryan
>
> On Wed, Aug 12, 2026 at 8:16 PM huaxin gao <hu...@gmail.com> wrote:
>
>> Hi Manu,
>>
>>
>> No rewrite of the existing manifests is needed. The V4 root manifest
>> references the existing V2/V3 manifests as-is (data manifests and the older
>> delete manifests), so upgrade stays metadata-only regardless of table size;
>> only the new top-level V4 metadata is written. The equality deletes in
>> those carried-over delete manifests are still applied on read (with 2-phase
>> planning) until they age off or are converted to DVs by a separate
>> maintenance action.
>>
>>
>> Thanks,
>>
>> Huaxin
>>
>> On Thu, Aug 13, 2026 at 10:18 AM Manu Zhang <ow...@gmail.com>
>> wrote:
>>
>>> Hi Huaxin,
>>>
>>> To clarify the third point: will upgrading a V2/V3 table to a V4 table
>>> require a manifest rewrite? If so, how will that process be handled?
>>>
>>> Thanks,
>>> Manu
>>>
>>>
>>> On Thu, Aug 13, 2026 at 10:07 AM huaxin gao <hu...@gmail.com>
>>> wrote:
>>>
>>>> Hi all,
>>>>
>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for
>>>> the
>>>> V4 table spec.
>>>>
>>>> Proposal
>>>> --------
>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>> metadata
>>>> will not define equality deletes as an allowed entry type.
>>>>
>>>> 2. Reading equality deletes remains supported in the reference
>>>> implementation for backward compatibility, both for existing V2/V3
>>>> tables and for equality deletes carried over into upgraded V4 tables.
>>>>
>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>> rewrite
>>>> of data or delete files). Existing equality deletes remain in
>>>> carried-
>>>> over V2/V3 delete manifests; converting them to deletion vectors is a
>>>> separate, optional maintenance action.
>>>>
>>>> Rationale
>>>> ---------
>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>> complicate
>>>> the format, and block features such as CDC, row lineage, and incremental
>>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>>> flat,
>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>>> viable replacement path, so we do not need the full replacement
>>>> completed
>>>> before forbidding new equality deletes in V4.
>>>>
>>>> The vote will be open for at least 72 hours.
>>>>
>>>> [ ] +1 Forbid writing equality deletes in V4
>>>> [ ] +0
>>>> [ ] -1 Do not forbid (please explain)
>>>>
>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>
>>>> Thanks,
>>>> Huaxin
>>>>
>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Ryan Blue <bl...@apache.org>.
+1 (binding) to deprecate equality deletes. Thanks, Huaxin!
Ryan
On Wed, Aug 12, 2026 at 8:16 PM huaxin gao <hu...@gmail.com> wrote:
> Hi Manu,
>
>
> No rewrite of the existing manifests is needed. The V4 root manifest
> references the existing V2/V3 manifests as-is (data manifests and the older
> delete manifests), so upgrade stays metadata-only regardless of table size;
> only the new top-level V4 metadata is written. The equality deletes in
> those carried-over delete manifests are still applied on read (with 2-phase
> planning) until they age off or are converted to DVs by a separate
> maintenance action.
>
>
> Thanks,
>
> Huaxin
>
> On Thu, Aug 13, 2026 at 10:18 AM Manu Zhang <ow...@gmail.com>
> wrote:
>
>> Hi Huaxin,
>>
>> To clarify the third point: will upgrading a V2/V3 table to a V4 table
>> require a manifest rewrite? If so, how will that process be handled?
>>
>> Thanks,
>> Manu
>>
>>
>> On Thu, Aug 13, 2026 at 10:07 AM huaxin gao <hu...@gmail.com>
>> wrote:
>>
>>> Hi all,
>>>
>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for
>>> the
>>> V4 table spec.
>>>
>>> Proposal
>>> --------
>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>> metadata
>>> will not define equality deletes as an allowed entry type.
>>>
>>> 2. Reading equality deletes remains supported in the reference
>>> implementation for backward compatibility, both for existing V2/V3
>>> tables and for equality deletes carried over into upgraded V4 tables.
>>>
>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>>> of data or delete files). Existing equality deletes remain in carried-
>>> over V2/V3 delete manifests; converting them to deletion vectors is a
>>> separate, optional maintenance action.
>>>
>>> Rationale
>>> ---------
>>> Equality deletes impose an asymmetric cost paid on every read, complicate
>>> the format, and block features such as CDC, row lineage, and incremental
>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>> flat,
>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>> viable replacement path, so we do not need the full replacement completed
>>> before forbidding new equality deletes in V4.
>>>
>>> The vote will be open for at least 72 hours.
>>>
>>> [ ] +1 Forbid writing equality deletes in V4
>>> [ ] +0
>>> [ ] -1 Do not forbid (please explain)
>>>
>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>
>>> Thanks,
>>> Huaxin
>>>
>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by huaxin gao <hu...@gmail.com>.
Hi Manu,
No rewrite of the existing manifests is needed. The V4 root manifest
references the existing V2/V3 manifests as-is (data manifests and the older
delete manifests), so upgrade stays metadata-only regardless of table size;
only the new top-level V4 metadata is written. The equality deletes in
those carried-over delete manifests are still applied on read (with 2-phase
planning) until they age off or are converted to DVs by a separate
maintenance action.
Thanks,
Huaxin
On Thu, Aug 13, 2026 at 10:18 AM Manu Zhang <ow...@gmail.com> wrote:
> Hi Huaxin,
>
> To clarify the third point: will upgrading a V2/V3 table to a V4 table
> require a manifest rewrite? If so, how will that process be handled?
>
> Thanks,
> Manu
>
>
> On Thu, Aug 13, 2026 at 10:07 AM huaxin gao <hu...@gmail.com>
> wrote:
>
>> Hi all,
>>
>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
>> V4 table spec.
>>
>> Proposal
>> --------
>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>> metadata
>> will not define equality deletes as an allowed entry type.
>>
>> 2. Reading equality deletes remains supported in the reference
>> implementation for backward compatibility, both for existing V2/V3
>> tables and for equality deletes carried over into upgraded V4 tables.
>>
>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>> of data or delete files). Existing equality deletes remain in carried-
>> over V2/V3 delete manifests; converting them to deletion vectors is a
>> separate, optional maintenance action.
>>
>> Rationale
>> ---------
>> Equality deletes impose an asymmetric cost paid on every read, complicate
>> the format, and block features such as CDC, row lineage, and incremental
>> index/materialized-view maintenance. Deletion vectors make deletion a
>> flat,
>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>> viable replacement path, so we do not need the full replacement completed
>> before forbidding new equality deletes in V4.
>>
>> The vote will be open for at least 72 hours.
>>
>> [ ] +1 Forbid writing equality deletes in V4
>> [ ] +0
>> [ ] -1 Do not forbid (please explain)
>>
>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>
>> Thanks,
>> Huaxin
>>
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Manu Zhang <ow...@gmail.com>.
Hi Huaxin,
To clarify the third point: will upgrading a V2/V3 table to a V4 table
require a manifest rewrite? If so, how will that process be handled?
Thanks,
Manu
On Thu, Aug 13, 2026 at 10:07 AM huaxin gao <hu...@gmail.com> wrote:
> Hi all,
>
> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
> V4 table spec.
>
> Proposal
> --------
> 1. Writing new equality deletes is forbidden for V4 tables: the V4 metadata
> will not define equality deletes as an allowed entry type.
>
> 2. Reading equality deletes remains supported in the reference
> implementation for backward compatibility, both for existing V2/V3
> tables and for equality deletes carried over into upgraded V4 tables.
>
> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
> of data or delete files). Existing equality deletes remain in carried-
> over V2/V3 delete manifests; converting them to deletion vectors is a
> separate, optional maintenance action.
>
> Rationale
> ---------
> Equality deletes impose an asymmetric cost paid on every read, complicate
> the format, and block features such as CDC, row lineage, and incremental
> index/materialized-view maintenance. Deletion vectors make deletion a flat,
> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
> viable replacement path, so we do not need the full replacement completed
> before forbidding new equality deletes in V4.
>
> The vote will be open for at least 72 hours.
>
> [ ] +1 Forbid writing equality deletes in V4
> [ ] +0
> [ ] -1 Do not forbid (please explain)
>
> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>
> Thanks,
> Huaxin
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Shawn Chang <ye...@gmail.com>.
+1 (non-binding)
Thanks Huaxin!
On Thu, Aug 13, 2026 at 10:58 AM Neelesh Salian <ne...@gmail.com>
wrote:
> +1 (non binding) Thank you Huaxin.
>
> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com> wrote:
>
>> Hi all,
>>
>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
>> V4 table spec.
>>
>> Proposal
>> --------
>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>> metadata
>> will not define equality deletes as an allowed entry type.
>>
>> 2. Reading equality deletes remains supported in the reference
>> implementation for backward compatibility, both for existing V2/V3
>> tables and for equality deletes carried over into upgraded V4 tables.
>>
>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>> of data or delete files). Existing equality deletes remain in carried-
>> over V2/V3 delete manifests; converting them to deletion vectors is a
>> separate, optional maintenance action.
>>
>> Rationale
>> ---------
>> Equality deletes impose an asymmetric cost paid on every read, complicate
>> the format, and block features such as CDC, row lineage, and incremental
>> index/materialized-view maintenance. Deletion vectors make deletion a
>> flat,
>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>> viable replacement path, so we do not need the full replacement completed
>> before forbidding new equality deletes in V4.
>>
>> The vote will be open for at least 72 hours.
>>
>> [ ] +1 Forbid writing equality deletes in V4
>> [ ] +0
>> [ ] -1 Do not forbid (please explain)
>>
>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>
>> Thanks,
>> Huaxin
>>
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Renjie Liu <li...@gmail.com>.
+1 binding, thanks Huaxin
On Fri, Aug 21, 2026 at 10:11 AM huaxin gao <hu...@gmail.com> wrote:
> I will submit a PR soon for the spec change.
>
> Thanks,
> Huaxin
>
> On Fri, Aug 21, 2026 at 7:12 AM Xiening Dai <xn...@apache.org> wrote:
>
>> Thanks for organizing the vote. Since we have a result now, do we plan to
>> update the spec?
>>
>> On 2026/08/18 06:37:25 huaxin gao wrote:
>> > The vote passes with 7 binding +1's and 17 non-binding +1's, and no 0 or
>> > -1's.
>> >
>> > Thanks everyone for the voting and discussion!
>> >
>> > Best,
>> > Huaxin
>> >
>> > On Tue, Aug 18, 2026 at 4:24 AM Daniel Weeks <dw...@apache.org> wrote:
>> >
>> > > +1 (binding)
>> > >
>> > > On Sat, Aug 15, 2026 at 9:06 AM Junwang Zhao <zh...@gmail.com>
>> wrote:
>> > >
>> > >> +1 (non-binding)
>> > >>
>> > >> On Sat, Aug 15, 2026 at 11:32 PM Maximilian Michels <mx...@apache.org>
>> > >> wrote:
>> > >> >
>> > >> > +1 (non-binding)
>> > >> >
>> > >> > Thanks,
>> > >> > Max
>> > >> >
>> > >> > Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026
>> um
>> > >> 22:07:
>> > >> >>
>> > >> >> +1
>> > >> >>
>> > >> >> Thanks
>> > >> >> Szehon
>> > >> >>
>> > >> >> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <
>> mbutrovich@apache.org>
>> > >> wrote:
>> > >> >>>
>> > >> >>> +1 (non-binding)
>> > >> >>>
>> > >> >>> Thanks Huaxin!
>> > >> >>>
>> > >> >>> -Matt
>> > >> >>>
>> > >> >>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
>> > >> dev@iceberg.apache.org> wrote:
>> > >> >>>>
>> > >> >>>> +1 non-binding
>> > >> >>>>
>> > >> >>>>
>> > >> >>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
>> > >> dev@iceberg.apache.org> wrote:
>> > >> >>>>>
>> > >> >>>>> +1 non-binding
>> > >> >>>>>
>> > >> >>>>> Thanks for driving this!
>> > >> >>>>>
>> > >> >>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com>
>> wrote:
>> > >> >>>>>>
>> > >> >>>>>> +1 (non-binding)
>> > >> >>>>>>
>> > >> >>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <
>> > >> owenzhang1990@gmail.com> wrote:
>> > >> >>>>>>>
>> > >> >>>>>>> +1 (non-binding)
>> > >> >>>>>>>
>> > >> >>>>>>> Thanks Huaxin!
>> > >> >>>>>>>
>> > >> >>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <
>> bryanck@gmail.com>
>> > >> wrote:
>> > >> >>>>>>>>
>> > >> >>>>>>>> +1 non-binding
>> > >> >>>>>>>>
>> > >> >>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>> > >> dev@iceberg.apache.org> wrote:
>> > >> >>>>>>>>
>> > >> >>>>>>>> +1 (non-binding)
>> > >> >>>>>>>>
>> > >> >>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <
>> flyrain000@gmail.com>
>> > >> wrote:
>> > >> >>>>>>>>>
>> > >> >>>>>>>>> +1(binding)
>> > >> >>>>>>>>> Yufei
>> > >> >>>>>>>>>
>> > >> >>>>>>>>>
>> > >> >>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <
>> > >> vaquar.khan@gmail.com> wrote:
>> > >> >>>>>>>>>>
>> > >> >>>>>>>>>> +1
>> > >> >>>>>>>>>>
>> > >> >>>>>>>>>> Regards,
>> > >> >>>>>>>>>> Viquar Khan
>> > >> >>>>>>>>>>
>> > >> >>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <
>> stevenz3wu@gmail.com>
>> > >> wrote:
>> > >> >>>>>>>>>>>
>> > >> >>>>>>>>>>> +1 (binding)
>> > >> >>>>>>>>>>>
>> > >> >>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks
>> everyone
>> > >> for contributing.
>> > >> >>>>>>>>>>>
>> > >> >>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>> > >> dev@iceberg.apache.org> wrote:
>> > >> >>>>>>>>>>>>
>> > >> >>>>>>>>>>>> +1 (non-binding)
>> > >> >>>>>>>>>>>>
>> > >> >>>>>>>>>>>> Thanks
>> > >> >>>>>>>>>>>>
>> > >> >>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>> > >> russell.spitzer@gmail.com> wrote:
>> > >> >>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>> +1 (binding)
>> > >> >>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>> > >> hongyue.apache@gmail.com> wrote:
>> > >> >>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>> +1 (non-binding)
>> > >> >>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>> Thanks Huaxin!
>> > >> >>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>> > >> graziadei.gianluca@gmail.com> wrote:
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> +1 (non-binding )
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor
>> for
>> > >> me.
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> An equality delete allows the writer to offload work
>> onto
>> > >> every subsequent read, and this deferred cost is neither bounded nor
>> > >> visible to the party that generated it. Deletion vectors make that
>> cost
>> > >> explicit and ensure it is paid only once.
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> I would also like to highlight point 3, which I
>> think is
>> > >> underrated: making the upgrade metadata-only decouples V4 adoption
>> from
>> > >> delete file migration. Users can move to V4 on their own timeline and
>> > >> convert equality deletes as a background maintenance task. Coupling
>> the two
>> > >> would have turned V4 adoption into a weeks-long operational project
>> for
>> > >> large tables.
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> Cheers,
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> Gianluca
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh
>> Salian <
>> > >> neeleshssalian@gmail.com> ha scritto:
>> > >> >>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>> > >> >>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
>> > >> huaxin.gao11@gmail.com> wrote:
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> Hi all,
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS]
>> Deprecate
>> > >> Equality Deletes in
>> > >> >>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the
>> > >> following proposal for the
>> > >> >>>>>>>>>>>>>>>>> V4 table spec.
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> Proposal
>> > >> >>>>>>>>>>>>>>>>> --------
>> > >> >>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4
>> > >> tables: the V4 metadata
>> > >> >>>>>>>>>>>>>>>>> will not define equality deletes as an allowed
>> entry
>> > >> type.
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in
>> the
>> > >> reference
>> > >> >>>>>>>>>>>>>>>>> implementation for backward compatibility, both
>> for
>> > >> existing V2/V3
>> > >> >>>>>>>>>>>>>>>>> tables and for equality deletes carried over
>> into
>> > >> upgraded V4 tables.
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only
>> (no
>> > >> synchronous rewrite
>> > >> >>>>>>>>>>>>>>>>> of data or delete files). Existing equality
>> deletes
>> > >> remain in carried-
>> > >> >>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to
>> > >> deletion vectors is a
>> > >> >>>>>>>>>>>>>>>>> separate, optional maintenance action.
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> Rationale
>> > >> >>>>>>>>>>>>>>>>> ---------
>> > >> >>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on
>> > >> every read, complicate
>> > >> >>>>>>>>>>>>>>>>> the format, and block features such as CDC, row
>> > >> lineage, and incremental
>> > >> >>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion
>> vectors
>> > >> make deletion a flat,
>> > >> >>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes
>> > >> work demonstrates a
>> > >> >>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
>> > >> replacement completed
>> > >> >>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>> > >> >>>>>>>>>>>>>>>>> [ ] +0
>> > >> >>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> [1]
>> > >> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>> > >> >>>>>>>>>>>>>>>>>
>> > >> >>>>>>>>>>>>>>>>> Thanks,
>> > >> >>>>>>>>>>>>>>>>> Huaxin
>> > >> >>>>>>>>
>> > >> >>>>>>>>
>> > >>
>> > >>
>> > >> --
>> > >> Regards
>> > >> Junwang Zhao
>> > >>
>> > >
>> >
>>
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by huaxin gao <hu...@gmail.com>.
I will submit a PR soon for the spec change.
Thanks,
Huaxin
On Fri, Aug 21, 2026 at 7:12 AM Xiening Dai <xn...@apache.org> wrote:
> Thanks for organizing the vote. Since we have a result now, do we plan to
> update the spec?
>
> On 2026/08/18 06:37:25 huaxin gao wrote:
> > The vote passes with 7 binding +1's and 17 non-binding +1's, and no 0 or
> > -1's.
> >
> > Thanks everyone for the voting and discussion!
> >
> > Best,
> > Huaxin
> >
> > On Tue, Aug 18, 2026 at 4:24 AM Daniel Weeks <dw...@apache.org> wrote:
> >
> > > +1 (binding)
> > >
> > > On Sat, Aug 15, 2026 at 9:06 AM Junwang Zhao <zh...@gmail.com>
> wrote:
> > >
> > >> +1 (non-binding)
> > >>
> > >> On Sat, Aug 15, 2026 at 11:32 PM Maximilian Michels <mx...@apache.org>
> > >> wrote:
> > >> >
> > >> > +1 (non-binding)
> > >> >
> > >> > Thanks,
> > >> > Max
> > >> >
> > >> > Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026 um
> > >> 22:07:
> > >> >>
> > >> >> +1
> > >> >>
> > >> >> Thanks
> > >> >> Szehon
> > >> >>
> > >> >> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <
> mbutrovich@apache.org>
> > >> wrote:
> > >> >>>
> > >> >>> +1 (non-binding)
> > >> >>>
> > >> >>> Thanks Huaxin!
> > >> >>>
> > >> >>> -Matt
> > >> >>>
> > >> >>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
> > >> dev@iceberg.apache.org> wrote:
> > >> >>>>
> > >> >>>> +1 non-binding
> > >> >>>>
> > >> >>>>
> > >> >>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
> > >> dev@iceberg.apache.org> wrote:
> > >> >>>>>
> > >> >>>>> +1 non-binding
> > >> >>>>>
> > >> >>>>> Thanks for driving this!
> > >> >>>>>
> > >> >>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com>
> wrote:
> > >> >>>>>>
> > >> >>>>>> +1 (non-binding)
> > >> >>>>>>
> > >> >>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <
> > >> owenzhang1990@gmail.com> wrote:
> > >> >>>>>>>
> > >> >>>>>>> +1 (non-binding)
> > >> >>>>>>>
> > >> >>>>>>> Thanks Huaxin!
> > >> >>>>>>>
> > >> >>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <
> bryanck@gmail.com>
> > >> wrote:
> > >> >>>>>>>>
> > >> >>>>>>>> +1 non-binding
> > >> >>>>>>>>
> > >> >>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
> > >> dev@iceberg.apache.org> wrote:
> > >> >>>>>>>>
> > >> >>>>>>>> +1 (non-binding)
> > >> >>>>>>>>
> > >> >>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <
> flyrain000@gmail.com>
> > >> wrote:
> > >> >>>>>>>>>
> > >> >>>>>>>>> +1(binding)
> > >> >>>>>>>>> Yufei
> > >> >>>>>>>>>
> > >> >>>>>>>>>
> > >> >>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <
> > >> vaquar.khan@gmail.com> wrote:
> > >> >>>>>>>>>>
> > >> >>>>>>>>>> +1
> > >> >>>>>>>>>>
> > >> >>>>>>>>>> Regards,
> > >> >>>>>>>>>> Viquar Khan
> > >> >>>>>>>>>>
> > >> >>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <
> stevenz3wu@gmail.com>
> > >> wrote:
> > >> >>>>>>>>>>>
> > >> >>>>>>>>>>> +1 (binding)
> > >> >>>>>>>>>>>
> > >> >>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks
> everyone
> > >> for contributing.
> > >> >>>>>>>>>>>
> > >> >>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
> > >> dev@iceberg.apache.org> wrote:
> > >> >>>>>>>>>>>>
> > >> >>>>>>>>>>>> +1 (non-binding)
> > >> >>>>>>>>>>>>
> > >> >>>>>>>>>>>> Thanks
> > >> >>>>>>>>>>>>
> > >> >>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
> > >> russell.spitzer@gmail.com> wrote:
> > >> >>>>>>>>>>>>>
> > >> >>>>>>>>>>>>> +1 (binding)
> > >> >>>>>>>>>>>>>
> > >> >>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
> > >> hongyue.apache@gmail.com> wrote:
> > >> >>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>> +1 (non-binding)
> > >> >>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>> Thanks Huaxin!
> > >> >>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
> > >> graziadei.gianluca@gmail.com> wrote:
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> +1 (non-binding )
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor
> for
> > >> me.
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> An equality delete allows the writer to offload work
> onto
> > >> every subsequent read, and this deferred cost is neither bounded nor
> > >> visible to the party that generated it. Deletion vectors make that
> cost
> > >> explicit and ensure it is paid only once.
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> I would also like to highlight point 3, which I think
> is
> > >> underrated: making the upgrade metadata-only decouples V4 adoption
> from
> > >> delete file migration. Users can move to V4 on their own timeline and
> > >> convert equality deletes as a background maintenance task. Coupling
> the two
> > >> would have turned V4 adoption into a weeks-long operational project
> for
> > >> large tables.
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> Cheers,
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> Gianluca
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh
> Salian <
> > >> neeleshssalian@gmail.com> ha scritto:
> > >> >>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
> > >> >>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
> > >> huaxin.gao11@gmail.com> wrote:
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> Hi all,
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate
> > >> Equality Deletes in
> > >> >>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the
> > >> following proposal for the
> > >> >>>>>>>>>>>>>>>>> V4 table spec.
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> Proposal
> > >> >>>>>>>>>>>>>>>>> --------
> > >> >>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4
> > >> tables: the V4 metadata
> > >> >>>>>>>>>>>>>>>>> will not define equality deletes as an allowed
> entry
> > >> type.
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the
> > >> reference
> > >> >>>>>>>>>>>>>>>>> implementation for backward compatibility, both
> for
> > >> existing V2/V3
> > >> >>>>>>>>>>>>>>>>> tables and for equality deletes carried over into
> > >> upgraded V4 tables.
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only
> (no
> > >> synchronous rewrite
> > >> >>>>>>>>>>>>>>>>> of data or delete files). Existing equality
> deletes
> > >> remain in carried-
> > >> >>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to
> > >> deletion vectors is a
> > >> >>>>>>>>>>>>>>>>> separate, optional maintenance action.
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> Rationale
> > >> >>>>>>>>>>>>>>>>> ---------
> > >> >>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on
> > >> every read, complicate
> > >> >>>>>>>>>>>>>>>>> the format, and block features such as CDC, row
> > >> lineage, and incremental
> > >> >>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion
> vectors
> > >> make deletion a flat,
> > >> >>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes
> > >> work demonstrates a
> > >> >>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
> > >> replacement completed
> > >> >>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
> > >> >>>>>>>>>>>>>>>>> [ ] +0
> > >> >>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> [1]
> > >> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
> > >> >>>>>>>>>>>>>>>>>
> > >> >>>>>>>>>>>>>>>>> Thanks,
> > >> >>>>>>>>>>>>>>>>> Huaxin
> > >> >>>>>>>>
> > >> >>>>>>>>
> > >>
> > >>
> > >> --
> > >> Regards
> > >> Junwang Zhao
> > >>
> > >
> >
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Xiening Dai <xn...@apache.org>.
Thanks for organizing the vote. Since we have a result now, do we plan to update the spec?
On 2026/08/18 06:37:25 huaxin gao wrote:
> The vote passes with 7 binding +1's and 17 non-binding +1's, and no 0 or
> -1's.
>
> Thanks everyone for the voting and discussion!
>
> Best,
> Huaxin
>
> On Tue, Aug 18, 2026 at 4:24 AM Daniel Weeks <dw...@apache.org> wrote:
>
> > +1 (binding)
> >
> > On Sat, Aug 15, 2026 at 9:06 AM Junwang Zhao <zh...@gmail.com> wrote:
> >
> >> +1 (non-binding)
> >>
> >> On Sat, Aug 15, 2026 at 11:32 PM Maximilian Michels <mx...@apache.org>
> >> wrote:
> >> >
> >> > +1 (non-binding)
> >> >
> >> > Thanks,
> >> > Max
> >> >
> >> > Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026 um
> >> 22:07:
> >> >>
> >> >> +1
> >> >>
> >> >> Thanks
> >> >> Szehon
> >> >>
> >> >> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <mb...@apache.org>
> >> wrote:
> >> >>>
> >> >>> +1 (non-binding)
> >> >>>
> >> >>> Thanks Huaxin!
> >> >>>
> >> >>> -Matt
> >> >>>
> >> >>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
> >> dev@iceberg.apache.org> wrote:
> >> >>>>
> >> >>>> +1 non-binding
> >> >>>>
> >> >>>>
> >> >>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
> >> dev@iceberg.apache.org> wrote:
> >> >>>>>
> >> >>>>> +1 non-binding
> >> >>>>>
> >> >>>>> Thanks for driving this!
> >> >>>>>
> >> >>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
> >> >>>>>>
> >> >>>>>> +1 (non-binding)
> >> >>>>>>
> >> >>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <
> >> owenzhang1990@gmail.com> wrote:
> >> >>>>>>>
> >> >>>>>>> +1 (non-binding)
> >> >>>>>>>
> >> >>>>>>> Thanks Huaxin!
> >> >>>>>>>
> >> >>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com>
> >> wrote:
> >> >>>>>>>>
> >> >>>>>>>> +1 non-binding
> >> >>>>>>>>
> >> >>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
> >> dev@iceberg.apache.org> wrote:
> >> >>>>>>>>
> >> >>>>>>>> +1 (non-binding)
> >> >>>>>>>>
> >> >>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com>
> >> wrote:
> >> >>>>>>>>>
> >> >>>>>>>>> +1(binding)
> >> >>>>>>>>> Yufei
> >> >>>>>>>>>
> >> >>>>>>>>>
> >> >>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <
> >> vaquar.khan@gmail.com> wrote:
> >> >>>>>>>>>>
> >> >>>>>>>>>> +1
> >> >>>>>>>>>>
> >> >>>>>>>>>> Regards,
> >> >>>>>>>>>> Viquar Khan
> >> >>>>>>>>>>
> >> >>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com>
> >> wrote:
> >> >>>>>>>>>>>
> >> >>>>>>>>>>> +1 (binding)
> >> >>>>>>>>>>>
> >> >>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone
> >> for contributing.
> >> >>>>>>>>>>>
> >> >>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
> >> dev@iceberg.apache.org> wrote:
> >> >>>>>>>>>>>>
> >> >>>>>>>>>>>> +1 (non-binding)
> >> >>>>>>>>>>>>
> >> >>>>>>>>>>>> Thanks
> >> >>>>>>>>>>>>
> >> >>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
> >> russell.spitzer@gmail.com> wrote:
> >> >>>>>>>>>>>>>
> >> >>>>>>>>>>>>> +1 (binding)
> >> >>>>>>>>>>>>>
> >> >>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
> >> hongyue.apache@gmail.com> wrote:
> >> >>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>> +1 (non-binding)
> >> >>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>> Thanks Huaxin!
> >> >>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
> >> graziadei.gianluca@gmail.com> wrote:
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> +1 (non-binding )
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for
> >> me.
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> An equality delete allows the writer to offload work onto
> >> every subsequent read, and this deferred cost is neither bounded nor
> >> visible to the party that generated it. Deletion vectors make that cost
> >> explicit and ensure it is paid only once.
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> I would also like to highlight point 3, which I think is
> >> underrated: making the upgrade metadata-only decouples V4 adoption from
> >> delete file migration. Users can move to V4 on their own timeline and
> >> convert equality deletes as a background maintenance task. Coupling the two
> >> would have turned V4 adoption into a weeks-long operational project for
> >> large tables.
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> Cheers,
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> Gianluca
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
> >> neeleshssalian@gmail.com> ha scritto:
> >> >>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
> >> >>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
> >> huaxin.gao11@gmail.com> wrote:
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> Hi all,
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate
> >> Equality Deletes in
> >> >>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the
> >> following proposal for the
> >> >>>>>>>>>>>>>>>>> V4 table spec.
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> Proposal
> >> >>>>>>>>>>>>>>>>> --------
> >> >>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4
> >> tables: the V4 metadata
> >> >>>>>>>>>>>>>>>>> will not define equality deletes as an allowed entry
> >> type.
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the
> >> reference
> >> >>>>>>>>>>>>>>>>> implementation for backward compatibility, both for
> >> existing V2/V3
> >> >>>>>>>>>>>>>>>>> tables and for equality deletes carried over into
> >> upgraded V4 tables.
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
> >> synchronous rewrite
> >> >>>>>>>>>>>>>>>>> of data or delete files). Existing equality deletes
> >> remain in carried-
> >> >>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to
> >> deletion vectors is a
> >> >>>>>>>>>>>>>>>>> separate, optional maintenance action.
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> Rationale
> >> >>>>>>>>>>>>>>>>> ---------
> >> >>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on
> >> every read, complicate
> >> >>>>>>>>>>>>>>>>> the format, and block features such as CDC, row
> >> lineage, and incremental
> >> >>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors
> >> make deletion a flat,
> >> >>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes
> >> work demonstrates a
> >> >>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
> >> replacement completed
> >> >>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
> >> >>>>>>>>>>>>>>>>> [ ] +0
> >> >>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> [1]
> >> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
> >> >>>>>>>>>>>>>>>>>
> >> >>>>>>>>>>>>>>>>> Thanks,
> >> >>>>>>>>>>>>>>>>> Huaxin
> >> >>>>>>>>
> >> >>>>>>>>
> >>
> >>
> >> --
> >> Regards
> >> Junwang Zhao
> >>
> >
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by huaxin gao <hu...@gmail.com>.
The vote passes with 7 binding +1's and 17 non-binding +1's, and no 0 or
-1's.
Thanks everyone for the voting and discussion!
Best,
Huaxin
On Tue, Aug 18, 2026 at 4:24 AM Daniel Weeks <dw...@apache.org> wrote:
> +1 (binding)
>
> On Sat, Aug 15, 2026 at 9:06 AM Junwang Zhao <zh...@gmail.com> wrote:
>
>> +1 (non-binding)
>>
>> On Sat, Aug 15, 2026 at 11:32 PM Maximilian Michels <mx...@apache.org>
>> wrote:
>> >
>> > +1 (non-binding)
>> >
>> > Thanks,
>> > Max
>> >
>> > Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026 um
>> 22:07:
>> >>
>> >> +1
>> >>
>> >> Thanks
>> >> Szehon
>> >>
>> >> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <mb...@apache.org>
>> wrote:
>> >>>
>> >>> +1 (non-binding)
>> >>>
>> >>> Thanks Huaxin!
>> >>>
>> >>> -Matt
>> >>>
>> >>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
>> dev@iceberg.apache.org> wrote:
>> >>>>
>> >>>> +1 non-binding
>> >>>>
>> >>>>
>> >>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
>> dev@iceberg.apache.org> wrote:
>> >>>>>
>> >>>>> +1 non-binding
>> >>>>>
>> >>>>> Thanks for driving this!
>> >>>>>
>> >>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
>> >>>>>>
>> >>>>>> +1 (non-binding)
>> >>>>>>
>> >>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <
>> owenzhang1990@gmail.com> wrote:
>> >>>>>>>
>> >>>>>>> +1 (non-binding)
>> >>>>>>>
>> >>>>>>> Thanks Huaxin!
>> >>>>>>>
>> >>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com>
>> wrote:
>> >>>>>>>>
>> >>>>>>>> +1 non-binding
>> >>>>>>>>
>> >>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>> dev@iceberg.apache.org> wrote:
>> >>>>>>>>
>> >>>>>>>> +1 (non-binding)
>> >>>>>>>>
>> >>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com>
>> wrote:
>> >>>>>>>>>
>> >>>>>>>>> +1(binding)
>> >>>>>>>>> Yufei
>> >>>>>>>>>
>> >>>>>>>>>
>> >>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <
>> vaquar.khan@gmail.com> wrote:
>> >>>>>>>>>>
>> >>>>>>>>>> +1
>> >>>>>>>>>>
>> >>>>>>>>>> Regards,
>> >>>>>>>>>> Viquar Khan
>> >>>>>>>>>>
>> >>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com>
>> wrote:
>> >>>>>>>>>>>
>> >>>>>>>>>>> +1 (binding)
>> >>>>>>>>>>>
>> >>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone
>> for contributing.
>> >>>>>>>>>>>
>> >>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>> dev@iceberg.apache.org> wrote:
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> +1 (non-binding)
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> Thanks
>> >>>>>>>>>>>>
>> >>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>> russell.spitzer@gmail.com> wrote:
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> +1 (binding)
>> >>>>>>>>>>>>>
>> >>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>> hongyue.apache@gmail.com> wrote:
>> >>>>>>>>>>>>>>
>> >>>>>>>>>>>>>> +1 (non-binding)
>> >>>>>>>>>>>>>>
>> >>>>>>>>>>>>>> Thanks Huaxin!
>> >>>>>>>>>>>>>>
>> >>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>> graziadei.gianluca@gmail.com> wrote:
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> +1 (non-binding )
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for
>> me.
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> An equality delete allows the writer to offload work onto
>> every subsequent read, and this deferred cost is neither bounded nor
>> visible to the party that generated it. Deletion vectors make that cost
>> explicit and ensure it is paid only once.
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> I would also like to highlight point 3, which I think is
>> underrated: making the upgrade metadata-only decouples V4 adoption from
>> delete file migration. Users can move to V4 on their own timeline and
>> convert equality deletes as a background maintenance task. Coupling the two
>> would have turned V4 adoption into a weeks-long operational project for
>> large tables.
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> Cheers,
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> Gianluca
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>> neeleshssalian@gmail.com> ha scritto:
>> >>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>> >>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
>> huaxin.gao11@gmail.com> wrote:
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> Hi all,
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate
>> Equality Deletes in
>> >>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the
>> following proposal for the
>> >>>>>>>>>>>>>>>>> V4 table spec.
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> Proposal
>> >>>>>>>>>>>>>>>>> --------
>> >>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4
>> tables: the V4 metadata
>> >>>>>>>>>>>>>>>>> will not define equality deletes as an allowed entry
>> type.
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the
>> reference
>> >>>>>>>>>>>>>>>>> implementation for backward compatibility, both for
>> existing V2/V3
>> >>>>>>>>>>>>>>>>> tables and for equality deletes carried over into
>> upgraded V4 tables.
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>> synchronous rewrite
>> >>>>>>>>>>>>>>>>> of data or delete files). Existing equality deletes
>> remain in carried-
>> >>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to
>> deletion vectors is a
>> >>>>>>>>>>>>>>>>> separate, optional maintenance action.
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> Rationale
>> >>>>>>>>>>>>>>>>> ---------
>> >>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on
>> every read, complicate
>> >>>>>>>>>>>>>>>>> the format, and block features such as CDC, row
>> lineage, and incremental
>> >>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors
>> make deletion a flat,
>> >>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes
>> work demonstrates a
>> >>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
>> replacement completed
>> >>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>> >>>>>>>>>>>>>>>>> [ ] +0
>> >>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> [1]
>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>> >>>>>>>>>>>>>>>>>
>> >>>>>>>>>>>>>>>>> Thanks,
>> >>>>>>>>>>>>>>>>> Huaxin
>> >>>>>>>>
>> >>>>>>>>
>>
>>
>> --
>> Regards
>> Junwang Zhao
>>
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Daniel Weeks <dw...@apache.org>.
+1 (binding)
On Sat, Aug 15, 2026 at 9:06 AM Junwang Zhao <zh...@gmail.com> wrote:
> +1 (non-binding)
>
> On Sat, Aug 15, 2026 at 11:32 PM Maximilian Michels <mx...@apache.org>
> wrote:
> >
> > +1 (non-binding)
> >
> > Thanks,
> > Max
> >
> > Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026 um
> 22:07:
> >>
> >> +1
> >>
> >> Thanks
> >> Szehon
> >>
> >> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <mb...@apache.org>
> wrote:
> >>>
> >>> +1 (non-binding)
> >>>
> >>> Thanks Huaxin!
> >>>
> >>> -Matt
> >>>
> >>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
> dev@iceberg.apache.org> wrote:
> >>>>
> >>>> +1 non-binding
> >>>>
> >>>>
> >>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
> dev@iceberg.apache.org> wrote:
> >>>>>
> >>>>> +1 non-binding
> >>>>>
> >>>>> Thanks for driving this!
> >>>>>
> >>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
> >>>>>>
> >>>>>> +1 (non-binding)
> >>>>>>
> >>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com>
> wrote:
> >>>>>>>
> >>>>>>> +1 (non-binding)
> >>>>>>>
> >>>>>>> Thanks Huaxin!
> >>>>>>>
> >>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com>
> wrote:
> >>>>>>>>
> >>>>>>>> +1 non-binding
> >>>>>>>>
> >>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
> dev@iceberg.apache.org> wrote:
> >>>>>>>>
> >>>>>>>> +1 (non-binding)
> >>>>>>>>
> >>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com>
> wrote:
> >>>>>>>>>
> >>>>>>>>> +1(binding)
> >>>>>>>>> Yufei
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <
> vaquar.khan@gmail.com> wrote:
> >>>>>>>>>>
> >>>>>>>>>> +1
> >>>>>>>>>>
> >>>>>>>>>> Regards,
> >>>>>>>>>> Viquar Khan
> >>>>>>>>>>
> >>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com>
> wrote:
> >>>>>>>>>>>
> >>>>>>>>>>> +1 (binding)
> >>>>>>>>>>>
> >>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone
> for contributing.
> >>>>>>>>>>>
> >>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
> dev@iceberg.apache.org> wrote:
> >>>>>>>>>>>>
> >>>>>>>>>>>> +1 (non-binding)
> >>>>>>>>>>>>
> >>>>>>>>>>>> Thanks
> >>>>>>>>>>>>
> >>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
> russell.spitzer@gmail.com> wrote:
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> +1 (binding)
> >>>>>>>>>>>>>
> >>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
> hongyue.apache@gmail.com> wrote:
> >>>>>>>>>>>>>>
> >>>>>>>>>>>>>> +1 (non-binding)
> >>>>>>>>>>>>>>
> >>>>>>>>>>>>>> Thanks Huaxin!
> >>>>>>>>>>>>>>
> >>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
> graziadei.gianluca@gmail.com> wrote:
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> +1 (non-binding )
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> An equality delete allows the writer to offload work onto
> every subsequent read, and this deferred cost is neither bounded nor
> visible to the party that generated it. Deletion vectors make that cost
> explicit and ensure it is paid only once.
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> I would also like to highlight point 3, which I think is
> underrated: making the upgrade metadata-only decouples V4 adoption from
> delete file migration. Users can move to V4 on their own timeline and
> convert equality deletes as a background maintenance task. Coupling the two
> would have turned V4 adoption into a weeks-long operational project for
> large tables.
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> Cheers,
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> Gianluca
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
> neeleshssalian@gmail.com> ha scritto:
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
> >>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
> huaxin.gao11@gmail.com> wrote:
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Hi all,
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate
> Equality Deletes in
> >>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the
> following proposal for the
> >>>>>>>>>>>>>>>>> V4 table spec.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Proposal
> >>>>>>>>>>>>>>>>> --------
> >>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4
> tables: the V4 metadata
> >>>>>>>>>>>>>>>>> will not define equality deletes as an allowed entry
> type.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the
> reference
> >>>>>>>>>>>>>>>>> implementation for backward compatibility, both for
> existing V2/V3
> >>>>>>>>>>>>>>>>> tables and for equality deletes carried over into
> upgraded V4 tables.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
> synchronous rewrite
> >>>>>>>>>>>>>>>>> of data or delete files). Existing equality deletes
> remain in carried-
> >>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to
> deletion vectors is a
> >>>>>>>>>>>>>>>>> separate, optional maintenance action.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Rationale
> >>>>>>>>>>>>>>>>> ---------
> >>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every
> read, complicate
> >>>>>>>>>>>>>>>>> the format, and block features such as CDC, row lineage,
> and incremental
> >>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors
> make deletion a flat,
> >>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
> demonstrates a
> >>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
> replacement completed
> >>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
> >>>>>>>>>>>>>>>>> [ ] +0
> >>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> [1]
> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
> >>>>>>>>>>>>>>>>>
> >>>>>>>>>>>>>>>>> Thanks,
> >>>>>>>>>>>>>>>>> Huaxin
> >>>>>>>>
> >>>>>>>>
>
>
> --
> Regards
> Junwang Zhao
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Junwang Zhao <zh...@gmail.com>.
+1 (non-binding)
On Sat, Aug 15, 2026 at 11:32 PM Maximilian Michels <mx...@apache.org> wrote:
>
> +1 (non-binding)
>
> Thanks,
> Max
>
> Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026 um 22:07:
>>
>> +1
>>
>> Thanks
>> Szehon
>>
>> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <mb...@apache.org> wrote:
>>>
>>> +1 (non-binding)
>>>
>>> Thanks Huaxin!
>>>
>>> -Matt
>>>
>>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <de...@iceberg.apache.org> wrote:
>>>>
>>>> +1 non-binding
>>>>
>>>>
>>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <de...@iceberg.apache.org> wrote:
>>>>>
>>>>> +1 non-binding
>>>>>
>>>>> Thanks for driving this!
>>>>>
>>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
>>>>>>
>>>>>> +1 (non-binding)
>>>>>>
>>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com> wrote:
>>>>>>>
>>>>>>> +1 (non-binding)
>>>>>>>
>>>>>>> Thanks Huaxin!
>>>>>>>
>>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com> wrote:
>>>>>>>>
>>>>>>>> +1 non-binding
>>>>>>>>
>>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <de...@iceberg.apache.org> wrote:
>>>>>>>>
>>>>>>>> +1 (non-binding)
>>>>>>>>
>>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com> wrote:
>>>>>>>>>
>>>>>>>>> +1(binding)
>>>>>>>>> Yufei
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>> +1
>>>>>>>>>>
>>>>>>>>>> Regards,
>>>>>>>>>> Viquar Khan
>>>>>>>>>>
>>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>>>>>>>>>>>
>>>>>>>>>>> +1 (binding)
>>>>>>>>>>>
>>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone for contributing.
>>>>>>>>>>>
>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <de...@iceberg.apache.org> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks
>>>>>>>>>>>>
>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <ru...@gmail.com> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> +1 (binding)
>>>>>>>>>>>>>
>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <ho...@gmail.com> wrote:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks Huaxin!
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <gr...@gmail.com> wrote:
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> +1 (non-binding )
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> An equality delete allows the writer to offload work onto every subsequent read, and this deferred cost is neither bounded nor visible to the party that generated it. Deletion vectors make that cost explicit and ensure it is paid only once.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I would also like to highlight point 3, which I think is underrated: making the upgrade metadata-only decouples V4 adoption from delete file migration. Users can move to V4 on their own timeline and convert equality deletes as a background maintenance task. Coupling the two would have turned V4 adoption into a weeks-long operational project for large tables.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Cheers,
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Gianluca
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <ne...@gmail.com> ha scritto:
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com> wrote:
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
>>>>>>>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Proposal
>>>>>>>>>>>>>>>>> --------
>>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4 metadata
>>>>>>>>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>>>>>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>>>>>>>>>>>>>> tables and for equality deletes carried over into upgraded V4 tables.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>>>>>>>>>>>>>>>>> of data or delete files). Existing equality deletes remain in carried-
>>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion vectors is a
>>>>>>>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Rationale
>>>>>>>>>>>>>>>>> ---------
>>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read, complicate
>>>>>>>>>>>>>>>>> the format, and block features such as CDC, row lineage, and incremental
>>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make deletion a flat,
>>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full replacement completed
>>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>>>>>>>> [ ] +0
>>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>>>> Huaxin
>>>>>>>>
>>>>>>>>
--
Regards
Junwang Zhao
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Maximilian Michels <mx...@apache.org>.
+1 (non-binding)
Thanks,
Max
Szehon Ho <sz...@gmail.com> schrieb am Fr. 14. Aug. 2026 um 22:07:
> +1
>
> Thanks
> Szehon
>
> On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <mb...@apache.org>
> wrote:
>
>> +1 (non-binding)
>>
>> Thanks Huaxin!
>>
>> -Matt
>>
>> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
>> dev@iceberg.apache.org> wrote:
>>
>>> +1 non-binding
>>>
>>>
>>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
>>> dev@iceberg.apache.org> wrote:
>>>
>>>> +1 non-binding
>>>>
>>>> Thanks for driving this!
>>>>
>>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
>>>>
>>>>> +1 (non-binding)
>>>>>
>>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> +1 (non-binding)
>>>>>>
>>>>>> Thanks Huaxin!
>>>>>>
>>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> +1 non-binding
>>>>>>>
>>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>>
>>>>>>> +1 (non-binding)
>>>>>>>
>>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> +1(binding)
>>>>>>>> Yufei
>>>>>>>>
>>>>>>>>
>>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> +1
>>>>>>>>>
>>>>>>>>> Regards,
>>>>>>>>> Viquar Khan
>>>>>>>>>
>>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> +1 (binding)
>>>>>>>>>>
>>>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>>>>>>>> contributing.
>>>>>>>>>>
>>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>>>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>>>>>
>>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>>
>>>>>>>>>>> Thanks
>>>>>>>>>>>
>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>>>>>>>> russell.spitzer@gmail.com> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> +1 (binding)
>>>>>>>>>>>>
>>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>>>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks Huaxin!
>>>>>>>>>>>>>
>>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>>> +1 (non-binding )
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> An equality delete allows the writer to offload work onto
>>>>>>>>>>>>>> every subsequent read, and this deferred cost is neither bounded nor
>>>>>>>>>>>>>> visible to the party that generated it. Deletion vectors make that cost
>>>>>>>>>>>>>> explicit and ensure it is paid only once.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>>>>>>>> large tables.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Cheers,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Gianluca
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
>>>>>>>>>>>>>>> huaxin.gao11@gmail.com> wrote:
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate
>>>>>>>>>>>>>>>> Equality Deletes in
>>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>>>>>>>> proposal for the
>>>>>>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Proposal
>>>>>>>>>>>>>>>> --------
>>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables:
>>>>>>>>>>>>>>>> the V4 metadata
>>>>>>>>>>>>>>>> will not define equality deletes as an allowed entry
>>>>>>>>>>>>>>>> type.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the
>>>>>>>>>>>>>>>> reference
>>>>>>>>>>>>>>>> implementation for backward compatibility, both for
>>>>>>>>>>>>>>>> existing V2/V3
>>>>>>>>>>>>>>>> tables and for equality deletes carried over into
>>>>>>>>>>>>>>>> upgraded V4 tables.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>>>>>>>>>>>>>>>> synchronous rewrite
>>>>>>>>>>>>>>>> of data or delete files). Existing equality deletes
>>>>>>>>>>>>>>>> remain in carried-
>>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>>>>>>>> vectors is a
>>>>>>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Rationale
>>>>>>>>>>>>>>>> ---------
>>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every
>>>>>>>>>>>>>>>> read, complicate
>>>>>>>>>>>>>>>> the format, and block features such as CDC, row lineage,
>>>>>>>>>>>>>>>> and incremental
>>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>>>>>>>> deletion a flat,
>>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>>>>>>>> demonstrates a
>>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
>>>>>>>>>>>>>>>> replacement completed
>>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>>>>>>> [ ] +0
>>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> [1]
>>>>>>>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>>> Huaxin
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>
>>>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Szehon Ho <sz...@gmail.com>.
+1
Thanks
Szehon
On Fri, Aug 14, 2026 at 10:56 AM Matt Butrovich <mb...@apache.org>
wrote:
> +1 (non-binding)
>
> Thanks Huaxin!
>
> -Matt
>
> On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <
> dev@iceberg.apache.org> wrote:
>
>> +1 non-binding
>>
>>
>> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
>> dev@iceberg.apache.org> wrote:
>>
>>> +1 non-binding
>>>
>>> Thanks for driving this!
>>>
>>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
>>>
>>>> +1 (non-binding)
>>>>
>>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com>
>>>> wrote:
>>>>
>>>>> +1 (non-binding)
>>>>>
>>>>> Thanks Huaxin!
>>>>>
>>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> +1 non-binding
>>>>>>
>>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>
>>>>>> +1 (non-binding)
>>>>>>
>>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> +1(binding)
>>>>>>> Yufei
>>>>>>>
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> +1
>>>>>>>>
>>>>>>>> Regards,
>>>>>>>> Viquar Khan
>>>>>>>>
>>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> +1 (binding)
>>>>>>>>>
>>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>>>>>>> contributing.
>>>>>>>>>
>>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>>>>
>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>
>>>>>>>>>> Thanks
>>>>>>>>>>
>>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>>>>>>> russell.spitzer@gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>>> +1 (binding)
>>>>>>>>>>>
>>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks Huaxin!
>>>>>>>>>>>>
>>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> +1 (non-binding )
>>>>>>>>>>>>>
>>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>>>>>
>>>>>>>>>>>>> An equality delete allows the writer to offload work onto
>>>>>>>>>>>>> every subsequent read, and this deferred cost is neither bounded nor
>>>>>>>>>>>>> visible to the party that generated it. Deletion vectors make that cost
>>>>>>>>>>>>> explicit and ensure it is paid only once.
>>>>>>>>>>>>>
>>>>>>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>>>>>>> large tables.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Cheers,
>>>>>>>>>>>>>
>>>>>>>>>>>>> Gianluca
>>>>>>>>>>>>>
>>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>>>>>>
>>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
>>>>>>>>>>>>>> huaxin.gao11@gmail.com> wrote:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate
>>>>>>>>>>>>>>> Equality Deletes in
>>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>>>>>>> proposal for the
>>>>>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Proposal
>>>>>>>>>>>>>>> --------
>>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables:
>>>>>>>>>>>>>>> the V4 metadata
>>>>>>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the
>>>>>>>>>>>>>>> reference
>>>>>>>>>>>>>>> implementation for backward compatibility, both for
>>>>>>>>>>>>>>> existing V2/V3
>>>>>>>>>>>>>>> tables and for equality deletes carried over into
>>>>>>>>>>>>>>> upgraded V4 tables.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>>>>>>>>>>>>>>> synchronous rewrite
>>>>>>>>>>>>>>> of data or delete files). Existing equality deletes
>>>>>>>>>>>>>>> remain in carried-
>>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>>>>>>> vectors is a
>>>>>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Rationale
>>>>>>>>>>>>>>> ---------
>>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every
>>>>>>>>>>>>>>> read, complicate
>>>>>>>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>>>>>>>> incremental
>>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>>>>>>> deletion a flat,
>>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>>>>>>> demonstrates a
>>>>>>>>>>>>>>> viable replacement path, so we do not need the full
>>>>>>>>>>>>>>> replacement completed
>>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>>>>>> [ ] +0
>>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> [1]
>>>>>>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>>> Huaxin
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Matt Butrovich <mb...@apache.org>.
+1 (non-binding)
Thanks Huaxin!
-Matt
On Fri, Aug 14, 2026 at 9:34 AM Talat Uyarer via dev <de...@iceberg.apache.org>
wrote:
> +1 non-binding
>
>
> On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <
> dev@iceberg.apache.org> wrote:
>
>> +1 non-binding
>>
>> Thanks for driving this!
>>
>> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
>>
>>> +1 (non-binding)
>>>
>>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com>
>>> wrote:
>>>
>>>> +1 (non-binding)
>>>>
>>>> Thanks Huaxin!
>>>>
>>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com> wrote:
>>>>
>>>>> +1 non-binding
>>>>>
>>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>>>>> dev@iceberg.apache.org> wrote:
>>>>>
>>>>> +1 (non-binding)
>>>>>
>>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> +1(binding)
>>>>>> Yufei
>>>>>>
>>>>>>
>>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> +1
>>>>>>>
>>>>>>> Regards,
>>>>>>> Viquar Khan
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> +1 (binding)
>>>>>>>>
>>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>>>>>> contributing.
>>>>>>>>
>>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>>>
>>>>>>>>> +1 (non-binding)
>>>>>>>>>
>>>>>>>>> Thanks
>>>>>>>>>
>>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>>>>>> russell.spitzer@gmail.com> wrote:
>>>>>>>>>
>>>>>>>>>> +1 (binding)
>>>>>>>>>>
>>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>>
>>>>>>>>>>> Thanks Huaxin!
>>>>>>>>>>>
>>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> +1 (non-binding )
>>>>>>>>>>>>
>>>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>>>>
>>>>>>>>>>>> An equality delete allows the writer to offload work onto every
>>>>>>>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>>>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>>>>>>>> ensure it is paid only once.
>>>>>>>>>>>>
>>>>>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>>>>>> large tables.
>>>>>>>>>>>>
>>>>>>>>>>>> Cheers,
>>>>>>>>>>>>
>>>>>>>>>>>> Gianluca
>>>>>>>>>>>>
>>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>>>>>
>>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>>>>
>>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
>>>>>>>>>>>>> huaxin.gao11@gmail.com> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>>>>>>>> Deletes in
>>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>>>>>> proposal for the
>>>>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Proposal
>>>>>>>>>>>>>> --------
>>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables:
>>>>>>>>>>>>>> the V4 metadata
>>>>>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>>>>>> implementation for backward compatibility, both for
>>>>>>>>>>>>>> existing V2/V3
>>>>>>>>>>>>>> tables and for equality deletes carried over into upgraded
>>>>>>>>>>>>>> V4 tables.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>>>>>>>>>>>>>> synchronous rewrite
>>>>>>>>>>>>>> of data or delete files). Existing equality deletes remain
>>>>>>>>>>>>>> in carried-
>>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>>>>>> vectors is a
>>>>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Rationale
>>>>>>>>>>>>>> ---------
>>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every
>>>>>>>>>>>>>> read, complicate
>>>>>>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>>>>>>> incremental
>>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>>>>>> deletion a flat,
>>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>>>>>> demonstrates a
>>>>>>>>>>>>>> viable replacement path, so we do not need the full
>>>>>>>>>>>>>> replacement completed
>>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>>>>> [ ] +0
>>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [1]
>>>>>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>> Huaxin
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Talat Uyarer via dev <de...@iceberg.apache.org>.
+1 non-binding
On Thu, Aug 13, 2026 at 8:31 PM Alex Stephen via dev <de...@iceberg.apache.org>
wrote:
> +1 non-binding
>
> Thanks for driving this!
>
> On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
>
>> +1 (non-binding)
>>
>> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com>
>> wrote:
>>
>>> +1 (non-binding)
>>>
>>> Thanks Huaxin!
>>>
>>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com> wrote:
>>>
>>>> +1 non-binding
>>>>
>>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>>>> dev@iceberg.apache.org> wrote:
>>>>
>>>> +1 (non-binding)
>>>>
>>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com> wrote:
>>>>
>>>>> +1(binding)
>>>>> Yufei
>>>>>
>>>>>
>>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> +1
>>>>>>
>>>>>> Regards,
>>>>>> Viquar Khan
>>>>>>
>>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>>>>>>
>>>>>>> +1 (binding)
>>>>>>>
>>>>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>>>>> contributing.
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>>
>>>>>>>> +1 (non-binding)
>>>>>>>>
>>>>>>>> Thanks
>>>>>>>>
>>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>>>>> russell.spitzer@gmail.com> wrote:
>>>>>>>>
>>>>>>>>> +1 (binding)
>>>>>>>>>
>>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>>>>
>>>>>>>>>> +1 (non-binding)
>>>>>>>>>>
>>>>>>>>>> Thanks Huaxin!
>>>>>>>>>>
>>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>>> +1 (non-binding )
>>>>>>>>>>>
>>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>>>
>>>>>>>>>>> An equality delete allows the writer to offload work onto every
>>>>>>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>>>>>>> ensure it is paid only once.
>>>>>>>>>>>
>>>>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>>>>> large tables.
>>>>>>>>>>>
>>>>>>>>>>> Cheers,
>>>>>>>>>>>
>>>>>>>>>>> Gianluca
>>>>>>>>>>>
>>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>>>>
>>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>>>
>>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <
>>>>>>>>>>>> huaxin.gao11@gmail.com> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>>
>>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>>>>>>> Deletes in
>>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>>>>> proposal for the
>>>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Proposal
>>>>>>>>>>>>> --------
>>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables:
>>>>>>>>>>>>> the V4 metadata
>>>>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>>>>
>>>>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>>>>> implementation for backward compatibility, both for
>>>>>>>>>>>>> existing V2/V3
>>>>>>>>>>>>> tables and for equality deletes carried over into upgraded
>>>>>>>>>>>>> V4 tables.
>>>>>>>>>>>>>
>>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>>>>>>>>>>>>> synchronous rewrite
>>>>>>>>>>>>> of data or delete files). Existing equality deletes remain
>>>>>>>>>>>>> in carried-
>>>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>>>>> vectors is a
>>>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Rationale
>>>>>>>>>>>>> ---------
>>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>>>>>>>> complicate
>>>>>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>>>>>> incremental
>>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>>>>> deletion a flat,
>>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>>>>> demonstrates a
>>>>>>>>>>>>> viable replacement path, so we do not need the full
>>>>>>>>>>>>> replacement completed
>>>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>>>
>>>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>>>
>>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>>>> [ ] +0
>>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>>>
>>>>>>>>>>>>> [1]
>>>>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>> Huaxin
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Alex Stephen via dev <de...@iceberg.apache.org>.
+1 non-binding
Thanks for driving this!
On Thu, Aug 13, 2026 at 7:17 PM Gang Wu <us...@gmail.com> wrote:
> +1 (non-binding)
>
> On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com>
> wrote:
>
>> +1 (non-binding)
>>
>> Thanks Huaxin!
>>
>> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com> wrote:
>>
>>> +1 non-binding
>>>
>>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>>> dev@iceberg.apache.org> wrote:
>>>
>>> +1 (non-binding)
>>>
>>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com> wrote:
>>>
>>>> +1(binding)
>>>> Yufei
>>>>
>>>>
>>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>>>> wrote:
>>>>
>>>>> +1
>>>>>
>>>>> Regards,
>>>>> Viquar Khan
>>>>>
>>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>>>>>
>>>>>> +1 (binding)
>>>>>>
>>>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>>>> contributing.
>>>>>>
>>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>>>> dev@iceberg.apache.org> wrote:
>>>>>>
>>>>>>> +1 (non-binding)
>>>>>>>
>>>>>>> Thanks
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>>>> russell.spitzer@gmail.com> wrote:
>>>>>>>
>>>>>>>> +1 (binding)
>>>>>>>>
>>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>>>
>>>>>>>>> +1 (non-binding)
>>>>>>>>>
>>>>>>>>> Thanks Huaxin!
>>>>>>>>>
>>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>>>
>>>>>>>>>> +1 (non-binding )
>>>>>>>>>>
>>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>>
>>>>>>>>>> An equality delete allows the writer to offload work onto every
>>>>>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>>>>>> ensure it is paid only once.
>>>>>>>>>>
>>>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>>>> large tables.
>>>>>>>>>>
>>>>>>>>>> Cheers,
>>>>>>>>>>
>>>>>>>>>> Gianluca
>>>>>>>>>>
>>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>>>
>>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>>
>>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>>>>>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> Hi all,
>>>>>>>>>>>>
>>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>>>>>> Deletes in
>>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>>>> proposal for the
>>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>>
>>>>>>>>>>>> Proposal
>>>>>>>>>>>> --------
>>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the
>>>>>>>>>>>> V4 metadata
>>>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>>>
>>>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>>>> implementation for backward compatibility, both for existing
>>>>>>>>>>>> V2/V3
>>>>>>>>>>>> tables and for equality deletes carried over into upgraded
>>>>>>>>>>>> V4 tables.
>>>>>>>>>>>>
>>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>>>>>>>>>>>> synchronous rewrite
>>>>>>>>>>>> of data or delete files). Existing equality deletes remain
>>>>>>>>>>>> in carried-
>>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>>>> vectors is a
>>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>>
>>>>>>>>>>>> Rationale
>>>>>>>>>>>> ---------
>>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>>>>>>> complicate
>>>>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>>>>> incremental
>>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>>>> deletion a flat,
>>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>>>> demonstrates a
>>>>>>>>>>>> viable replacement path, so we do not need the full replacement
>>>>>>>>>>>> completed
>>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>>
>>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>>
>>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>>> [ ] +0
>>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>>
>>>>>>>>>>>> [1]
>>>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>> Huaxin
>>>>>>>>>>>>
>>>>>>>>>>>
>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Gang Wu <us...@gmail.com>.
+1 (non-binding)
On Fri, Aug 14, 2026 at 9:37 AM Manu Zhang <ow...@gmail.com> wrote:
> +1 (non-binding)
>
> Thanks Huaxin!
>
> On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com> wrote:
>
>> +1 non-binding
>>
>> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
>> dev@iceberg.apache.org> wrote:
>>
>> +1 (non-binding)
>>
>> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com> wrote:
>>
>>> +1(binding)
>>> Yufei
>>>
>>>
>>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>>> wrote:
>>>
>>>> +1
>>>>
>>>> Regards,
>>>> Viquar Khan
>>>>
>>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>>>>
>>>>> +1 (binding)
>>>>>
>>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>>> contributing.
>>>>>
>>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>>> dev@iceberg.apache.org> wrote:
>>>>>
>>>>>> +1 (non-binding)
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>>> russell.spitzer@gmail.com> wrote:
>>>>>>
>>>>>>> +1 (binding)
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>>
>>>>>>>> +1 (non-binding)
>>>>>>>>
>>>>>>>> Thanks Huaxin!
>>>>>>>>
>>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>>
>>>>>>>>> +1 (non-binding )
>>>>>>>>>
>>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>>
>>>>>>>>> An equality delete allows the writer to offload work onto every
>>>>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>>>>> ensure it is paid only once.
>>>>>>>>>
>>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>>> large tables.
>>>>>>>>>
>>>>>>>>> Cheers,
>>>>>>>>>
>>>>>>>>> Gianluca
>>>>>>>>>
>>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>>
>>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>>
>>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>>>>>>> wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi all,
>>>>>>>>>>>
>>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>>>>> Deletes in
>>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>>> proposal for the
>>>>>>>>>>> V4 table spec.
>>>>>>>>>>>
>>>>>>>>>>> Proposal
>>>>>>>>>>> --------
>>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the
>>>>>>>>>>> V4 metadata
>>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>>
>>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>>> implementation for backward compatibility, both for existing
>>>>>>>>>>> V2/V3
>>>>>>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>>>>>>> tables.
>>>>>>>>>>>
>>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no
>>>>>>>>>>> synchronous rewrite
>>>>>>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>>>>>>> carried-
>>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>>> vectors is a
>>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>>
>>>>>>>>>>> Rationale
>>>>>>>>>>> ---------
>>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>>>>>> complicate
>>>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>>>> incremental
>>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>>> deletion a flat,
>>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>>> demonstrates a
>>>>>>>>>>> viable replacement path, so we do not need the full replacement
>>>>>>>>>>> completed
>>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>>
>>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>>
>>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>>> [ ] +0
>>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>>
>>>>>>>>>>> [1]
>>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>> Huaxin
>>>>>>>>>>>
>>>>>>>>>>
>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Manu Zhang <ow...@gmail.com>.
+1 (non-binding)
Thanks Huaxin!
On Fri, Aug 14, 2026 at 7:06 AM Bryan Keller <br...@gmail.com> wrote:
> +1 non-binding
>
> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <
> dev@iceberg.apache.org> wrote:
>
> +1 (non-binding)
>
> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com> wrote:
>
>> +1(binding)
>> Yufei
>>
>>
>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com>
>> wrote:
>>
>>> +1
>>>
>>> Regards,
>>> Viquar Khan
>>>
>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>>>
>>>> +1 (binding)
>>>>
>>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>>> contributing.
>>>>
>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>>> dev@iceberg.apache.org> wrote:
>>>>
>>>>> +1 (non-binding)
>>>>>
>>>>> Thanks
>>>>>
>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>>> russell.spitzer@gmail.com> wrote:
>>>>>
>>>>>> +1 (binding)
>>>>>>
>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>>> hongyue.apache@gmail.com> wrote:
>>>>>>
>>>>>>> +1 (non-binding)
>>>>>>>
>>>>>>> Thanks Huaxin!
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>>
>>>>>>>> +1 (non-binding )
>>>>>>>>
>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>
>>>>>>>> An equality delete allows the writer to offload work onto every
>>>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>>>> ensure it is paid only once.
>>>>>>>>
>>>>>>>> I would also like to highlight point 3, which I think is
>>>>>>>> underrated: making the upgrade metadata-only decouples V4 adoption from
>>>>>>>> delete file migration. Users can move to V4 on their own timeline and
>>>>>>>> convert equality deletes as a background maintenance task. Coupling the two
>>>>>>>> would have turned V4 adoption into a weeks-long operational project for
>>>>>>>> large tables.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>>
>>>>>>>> Gianluca
>>>>>>>>
>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>>
>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>
>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>>>>>> wrote:
>>>>>>>>>
>>>>>>>>>> Hi all,
>>>>>>>>>>
>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>>>> Deletes in
>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following
>>>>>>>>>> proposal for the
>>>>>>>>>> V4 table spec.
>>>>>>>>>>
>>>>>>>>>> Proposal
>>>>>>>>>> --------
>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the
>>>>>>>>>> V4 metadata
>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>
>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>> implementation for backward compatibility, both for existing
>>>>>>>>>> V2/V3
>>>>>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>>>>>> tables.
>>>>>>>>>>
>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>>>>>>> rewrite
>>>>>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>>>>>> carried-
>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>>> vectors is a
>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>
>>>>>>>>>> Rationale
>>>>>>>>>> ---------
>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>>>>> complicate
>>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>>> incremental
>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>>> deletion a flat,
>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>>> demonstrates a
>>>>>>>>>> viable replacement path, so we do not need the full replacement
>>>>>>>>>> completed
>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>
>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>
>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>> [ ] +0
>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>
>>>>>>>>>> [1]
>>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Huaxin
>>>>>>>>>>
>>>>>>>>>
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Bryan Keller <br...@gmail.com>.
+1 non-binding
> On Aug 13, 2026, at 3:55 PM, Andrei Tserakhau via dev <de...@iceberg.apache.org> wrote:
>
> +1 (non-binding)
>
> On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <flyrain000@gmail.com <ma...@gmail.com>> wrote:
>> +1(binding)
>> Yufei
>>
>>
>> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <vaquar.khan@gmail.com <ma...@gmail.com>> wrote:
>>> +1
>>>
>>> Regards,
>>> Viquar Khan
>>>
>>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <stevenz3wu@gmail.com <ma...@gmail.com>> wrote:
>>>> +1 (binding)
>>>>
>>>> Thanks Huaxin for driving the discussion and thanks everyone for contributing.
>>>>
>>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <dev@iceberg.apache.org <ma...@iceberg.apache.org>> wrote:
>>>>> +1 (non-binding)
>>>>>
>>>>> Thanks
>>>>>
>>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <russell.spitzer@gmail.com <ma...@gmail.com>> wrote:
>>>>>> +1 (binding)
>>>>>>
>>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <hongyue.apache@gmail.com <ma...@gmail.com>> wrote:
>>>>>>> +1 (non-binding)
>>>>>>>
>>>>>>> Thanks Huaxin!
>>>>>>>
>>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <graziadei.gianluca@gmail.com <ma...@gmail.com>> wrote:
>>>>>>>> +1 (non-binding )
>>>>>>>>
>>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>>
>>>>>>>> An equality delete allows the writer to offload work onto every subsequent read, and this deferred cost is neither bounded nor visible to the party that generated it. Deletion vectors make that cost explicit and ensure it is paid only once.
>>>>>>>>
>>>>>>>> I would also like to highlight point 3, which I think is underrated: making the upgrade metadata-only decouples V4 adoption from delete file migration. Users can move to V4 on their own timeline and convert equality deletes as a background maintenance task. Coupling the two would have turned V4 adoption into a weeks-long operational project for large tables.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>>
>>>>>>>> Gianluca
>>>>>>>>
>>>>>>>>
>>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <neeleshssalian@gmail.com <ma...@gmail.com>> ha scritto:
>>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>>
>>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <huaxin.gao11@gmail.com <ma...@gmail.com>> wrote:
>>>>>>>>>> Hi all,
>>>>>>>>>>
>>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
>>>>>>>>>> V4 table spec.
>>>>>>>>>>
>>>>>>>>>> Proposal
>>>>>>>>>> --------
>>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4 metadata
>>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>>
>>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>>>>>>> tables and for equality deletes carried over into upgraded V4 tables.
>>>>>>>>>>
>>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>>>>>>>>>> of data or delete files). Existing equality deletes remain in carried-
>>>>>>>>>> over V2/V3 delete manifests; converting them to deletion vectors is a
>>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>>
>>>>>>>>>> Rationale
>>>>>>>>>> ---------
>>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read, complicate
>>>>>>>>>> the format, and block features such as CDC, row lineage, and incremental
>>>>>>>>>> index/materialized-view maintenance. Deletion vectors make deletion a flat,
>>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>>>>>>>>> viable replacement path, so we do not need the full replacement completed
>>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>>
>>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>>
>>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>>> [ ] +0
>>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>>
>>>>>>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6 <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> Huaxin
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Andrei Tserakhau via dev <de...@iceberg.apache.org>.
+1 (non-binding)
On Fri, Aug 14, 2026 at 12:16 AM Yufei Gu <fl...@gmail.com> wrote:
> +1(binding)
> Yufei
>
>
> On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com> wrote:
>
>> +1
>>
>> Regards,
>> Viquar Khan
>>
>> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>>
>>> +1 (binding)
>>>
>>> Thanks Huaxin for driving the discussion and thanks everyone for
>>> contributing.
>>>
>>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>>> dev@iceberg.apache.org> wrote:
>>>
>>>> +1 (non-binding)
>>>>
>>>> Thanks
>>>>
>>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>>> russell.spitzer@gmail.com> wrote:
>>>>
>>>>> +1 (binding)
>>>>>
>>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <
>>>>> hongyue.apache@gmail.com> wrote:
>>>>>
>>>>>> +1 (non-binding)
>>>>>>
>>>>>> Thanks Huaxin!
>>>>>>
>>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>>
>>>>>>> +1 (non-binding )
>>>>>>>
>>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>>
>>>>>>> An equality delete allows the writer to offload work onto every
>>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>>> ensure it is paid only once.
>>>>>>>
>>>>>>> I would also like to highlight point 3, which I think is underrated:
>>>>>>> making the upgrade metadata-only decouples V4 adoption from delete file
>>>>>>> migration. Users can move to V4 on their own timeline and convert equality
>>>>>>> deletes as a background maintenance task. Coupling the two would have
>>>>>>> turned V4 adoption into a weeks-long operational project for large tables.
>>>>>>>
>>>>>>> Cheers,
>>>>>>>
>>>>>>> Gianluca
>>>>>>>
>>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>>
>>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>>
>>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>>>>> wrote:
>>>>>>>>
>>>>>>>>> Hi all,
>>>>>>>>>
>>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>>> Deletes in
>>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal
>>>>>>>>> for the
>>>>>>>>> V4 table spec.
>>>>>>>>>
>>>>>>>>> Proposal
>>>>>>>>> --------
>>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>>>>>> metadata
>>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>>
>>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>>> implementation for backward compatibility, both for existing
>>>>>>>>> V2/V3
>>>>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>>>>> tables.
>>>>>>>>>
>>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>>>>>> rewrite
>>>>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>>>>> carried-
>>>>>>>>> over V2/V3 delete manifests; converting them to deletion
>>>>>>>>> vectors is a
>>>>>>>>> separate, optional maintenance action.
>>>>>>>>>
>>>>>>>>> Rationale
>>>>>>>>> ---------
>>>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>>>> complicate
>>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>>> incremental
>>>>>>>>> index/materialized-view maintenance. Deletion vectors make
>>>>>>>>> deletion a flat,
>>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>>> demonstrates a
>>>>>>>>> viable replacement path, so we do not need the full replacement
>>>>>>>>> completed
>>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>>
>>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>>
>>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>>> [ ] +0
>>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>>
>>>>>>>>> [1]
>>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> Huaxin
>>>>>>>>>
>>>>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Yufei Gu <fl...@gmail.com>.
+1(binding)
Yufei
On Thu, Aug 13, 2026 at 1:11 PM vaquar khan <va...@gmail.com> wrote:
> +1
>
> Regards,
> Viquar Khan
>
> On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
>
>> +1 (binding)
>>
>> Thanks Huaxin for driving the discussion and thanks everyone for
>> contributing.
>>
>> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <
>> dev@iceberg.apache.org> wrote:
>>
>>> +1 (non-binding)
>>>
>>> Thanks
>>>
>>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>>> russell.spitzer@gmail.com> wrote:
>>>
>>>> +1 (binding)
>>>>
>>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <ho...@gmail.com>
>>>> wrote:
>>>>
>>>>> +1 (non-binding)
>>>>>
>>>>> Thanks Huaxin!
>>>>>
>>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>>> graziadei.gianluca@gmail.com> wrote:
>>>>>
>>>>>> +1 (non-binding )
>>>>>>
>>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>>
>>>>>> An equality delete allows the writer to offload work onto every
>>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>>> ensure it is paid only once.
>>>>>>
>>>>>> I would also like to highlight point 3, which I think is underrated:
>>>>>> making the upgrade metadata-only decouples V4 adoption from delete file
>>>>>> migration. Users can move to V4 on their own timeline and convert equality
>>>>>> deletes as a background maintenance task. Coupling the two would have
>>>>>> turned V4 adoption into a weeks-long operational project for large tables.
>>>>>>
>>>>>> Cheers,
>>>>>>
>>>>>> Gianluca
>>>>>>
>>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>>
>>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>>
>>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi all,
>>>>>>>>
>>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>>> Deletes in
>>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal
>>>>>>>> for the
>>>>>>>> V4 table spec.
>>>>>>>>
>>>>>>>> Proposal
>>>>>>>> --------
>>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>>>>> metadata
>>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>>
>>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>>> implementation for backward compatibility, both for existing
>>>>>>>> V2/V3
>>>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>>>> tables.
>>>>>>>>
>>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>>>>> rewrite
>>>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>>>> carried-
>>>>>>>> over V2/V3 delete manifests; converting them to deletion vectors
>>>>>>>> is a
>>>>>>>> separate, optional maintenance action.
>>>>>>>>
>>>>>>>> Rationale
>>>>>>>> ---------
>>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>>> complicate
>>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>>> incremental
>>>>>>>> index/materialized-view maintenance. Deletion vectors make deletion
>>>>>>>> a flat,
>>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>>> demonstrates a
>>>>>>>> viable replacement path, so we do not need the full replacement
>>>>>>>> completed
>>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>>
>>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>>
>>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>>> [ ] +0
>>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>>
>>>>>>>> [1]
>>>>>>>> https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>>
>>>>>>>> Thanks,
>>>>>>>> Huaxin
>>>>>>>>
>>>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by vaquar khan <va...@gmail.com>.
+1
Regards,
Viquar Khan
On Thu, Aug 13, 2026, 3:04 PM Steven Wu <st...@gmail.com> wrote:
> +1 (binding)
>
> Thanks Huaxin for driving the discussion and thanks everyone for
> contributing.
>
> On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <de...@iceberg.apache.org>
> wrote:
>
>> +1 (non-binding)
>>
>> Thanks
>>
>> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
>> russell.spitzer@gmail.com> wrote:
>>
>>> +1 (binding)
>>>
>>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <ho...@gmail.com>
>>> wrote:
>>>
>>>> +1 (non-binding)
>>>>
>>>> Thanks Huaxin!
>>>>
>>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>>> graziadei.gianluca@gmail.com> wrote:
>>>>
>>>>> +1 (non-binding )
>>>>>
>>>>> The asymmetric cost argument is the decisive factor for me.
>>>>>
>>>>> An equality delete allows the writer to offload work onto every
>>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>>> ensure it is paid only once.
>>>>>
>>>>> I would also like to highlight point 3, which I think is underrated:
>>>>> making the upgrade metadata-only decouples V4 adoption from delete file
>>>>> migration. Users can move to V4 on their own timeline and convert equality
>>>>> deletes as a background maintenance task. Coupling the two would have
>>>>> turned V4 adoption into a weeks-long operational project for large tables.
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Gianluca
>>>>>
>>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>>> neeleshssalian@gmail.com> ha scritto:
>>>>>
>>>>>> +1 (non binding) Thank you Huaxin.
>>>>>>
>>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Hi all,
>>>>>>>
>>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality
>>>>>>> Deletes in
>>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal
>>>>>>> for the
>>>>>>> V4 table spec.
>>>>>>>
>>>>>>> Proposal
>>>>>>> --------
>>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>>>> metadata
>>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>>
>>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>>> tables.
>>>>>>>
>>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>>>> rewrite
>>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>>> carried-
>>>>>>> over V2/V3 delete manifests; converting them to deletion vectors
>>>>>>> is a
>>>>>>> separate, optional maintenance action.
>>>>>>>
>>>>>>> Rationale
>>>>>>> ---------
>>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>>> complicate
>>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>>> incremental
>>>>>>> index/materialized-view maintenance. Deletion vectors make deletion
>>>>>>> a flat,
>>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work
>>>>>>> demonstrates a
>>>>>>> viable replacement path, so we do not need the full replacement
>>>>>>> completed
>>>>>>> before forbidding new equality deletes in V4.
>>>>>>>
>>>>>>> The vote will be open for at least 72 hours.
>>>>>>>
>>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>>> [ ] +0
>>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>>
>>>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Huaxin
>>>>>>>
>>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Steven Wu <st...@gmail.com>.
+1 (binding)
Thanks Huaxin for driving the discussion and thanks everyone for
contributing.
On Thu, Aug 13, 2026 at 11:18 AM Xin Huang via dev <de...@iceberg.apache.org>
wrote:
> +1 (non-binding)
>
> Thanks
>
> On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <
> russell.spitzer@gmail.com> wrote:
>
>> +1 (binding)
>>
>> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <ho...@gmail.com>
>> wrote:
>>
>>> +1 (non-binding)
>>>
>>> Thanks Huaxin!
>>>
>>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>>> graziadei.gianluca@gmail.com> wrote:
>>>
>>>> +1 (non-binding )
>>>>
>>>> The asymmetric cost argument is the decisive factor for me.
>>>>
>>>> An equality delete allows the writer to offload work onto every
>>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>>> the party that generated it. Deletion vectors make that cost explicit and
>>>> ensure it is paid only once.
>>>>
>>>> I would also like to highlight point 3, which I think is underrated:
>>>> making the upgrade metadata-only decouples V4 adoption from delete file
>>>> migration. Users can move to V4 on their own timeline and convert equality
>>>> deletes as a background maintenance task. Coupling the two would have
>>>> turned V4 adoption into a weeks-long operational project for large tables.
>>>>
>>>> Cheers,
>>>>
>>>> Gianluca
>>>>
>>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>>> neeleshssalian@gmail.com> ha scritto:
>>>>
>>>>> +1 (non binding) Thank you Huaxin.
>>>>>
>>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes
>>>>>> in
>>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal
>>>>>> for the
>>>>>> V4 table spec.
>>>>>>
>>>>>> Proposal
>>>>>> --------
>>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>>> metadata
>>>>>> will not define equality deletes as an allowed entry type.
>>>>>>
>>>>>> 2. Reading equality deletes remains supported in the reference
>>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>>> tables and for equality deletes carried over into upgraded V4
>>>>>> tables.
>>>>>>
>>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>>> rewrite
>>>>>> of data or delete files). Existing equality deletes remain in
>>>>>> carried-
>>>>>> over V2/V3 delete manifests; converting them to deletion vectors
>>>>>> is a
>>>>>> separate, optional maintenance action.
>>>>>>
>>>>>> Rationale
>>>>>> ---------
>>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>>> complicate
>>>>>> the format, and block features such as CDC, row lineage, and
>>>>>> incremental
>>>>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>>>>> flat,
>>>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates
>>>>>> a
>>>>>> viable replacement path, so we do not need the full replacement
>>>>>> completed
>>>>>> before forbidding new equality deletes in V4.
>>>>>>
>>>>>> The vote will be open for at least 72 hours.
>>>>>>
>>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>>> [ ] +0
>>>>>> [ ] -1 Do not forbid (please explain)
>>>>>>
>>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>>
>>>>>> Thanks,
>>>>>> Huaxin
>>>>>>
>>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Xin Huang via dev <de...@iceberg.apache.org>.
+1 (non-binding)
Thanks
On Thu, Aug 13, 2026 at 11:13 AM Russell Spitzer <ru...@gmail.com>
wrote:
> +1 (binding)
>
> On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <ho...@gmail.com>
> wrote:
>
>> +1 (non-binding)
>>
>> Thanks Huaxin!
>>
>> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
>> graziadei.gianluca@gmail.com> wrote:
>>
>>> +1 (non-binding )
>>>
>>> The asymmetric cost argument is the decisive factor for me.
>>>
>>> An equality delete allows the writer to offload work onto every
>>> subsequent read, and this deferred cost is neither bounded nor visible to
>>> the party that generated it. Deletion vectors make that cost explicit and
>>> ensure it is paid only once.
>>>
>>> I would also like to highlight point 3, which I think is underrated:
>>> making the upgrade metadata-only decouples V4 adoption from delete file
>>> migration. Users can move to V4 on their own timeline and convert equality
>>> deletes as a background maintenance task. Coupling the two would have
>>> turned V4 adoption into a weeks-long operational project for large tables.
>>>
>>> Cheers,
>>>
>>> Gianluca
>>>
>>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>>> neeleshssalian@gmail.com> ha scritto:
>>>
>>>> +1 (non binding) Thank you Huaxin.
>>>>
>>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com>
>>>> wrote:
>>>>
>>>>> Hi all,
>>>>>
>>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes
>>>>> in
>>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for
>>>>> the
>>>>> V4 table spec.
>>>>>
>>>>> Proposal
>>>>> --------
>>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>>> metadata
>>>>> will not define equality deletes as an allowed entry type.
>>>>>
>>>>> 2. Reading equality deletes remains supported in the reference
>>>>> implementation for backward compatibility, both for existing V2/V3
>>>>> tables and for equality deletes carried over into upgraded V4
>>>>> tables.
>>>>>
>>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>>> rewrite
>>>>> of data or delete files). Existing equality deletes remain in
>>>>> carried-
>>>>> over V2/V3 delete manifests; converting them to deletion vectors is
>>>>> a
>>>>> separate, optional maintenance action.
>>>>>
>>>>> Rationale
>>>>> ---------
>>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>>> complicate
>>>>> the format, and block features such as CDC, row lineage, and
>>>>> incremental
>>>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>>>> flat,
>>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>>>> viable replacement path, so we do not need the full replacement
>>>>> completed
>>>>> before forbidding new equality deletes in V4.
>>>>>
>>>>> The vote will be open for at least 72 hours.
>>>>>
>>>>> [ ] +1 Forbid writing equality deletes in V4
>>>>> [ ] +0
>>>>> [ ] -1 Do not forbid (please explain)
>>>>>
>>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>> <https://urldefense.com/v3/__https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6__;!!LIr3w8kk_Xxm!vsYT4Z_ihSMhUpgRXcXgEUbArKyPoG3WKwSCXmmMtC_0Tg3ICMgaSOoq7WzaKsMaK4-c_vXUCsbXEVhdw7RV0di7e8S-yLUE$>
>>>>>
>>>>> Thanks,
>>>>> Huaxin
>>>>>
>>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Russell Spitzer <ru...@gmail.com>.
+1 (binding)
On Thu, Aug 13, 2026 at 1:07 PM Hongyue Zhang <ho...@gmail.com>
wrote:
> +1 (non-binding)
>
> Thanks Huaxin!
>
> On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
> graziadei.gianluca@gmail.com> wrote:
>
>> +1 (non-binding )
>>
>> The asymmetric cost argument is the decisive factor for me.
>>
>> An equality delete allows the writer to offload work onto every
>> subsequent read, and this deferred cost is neither bounded nor visible to
>> the party that generated it. Deletion vectors make that cost explicit and
>> ensure it is paid only once.
>>
>> I would also like to highlight point 3, which I think is underrated:
>> making the upgrade metadata-only decouples V4 adoption from delete file
>> migration. Users can move to V4 on their own timeline and convert equality
>> deletes as a background maintenance task. Coupling the two would have
>> turned V4 adoption into a weeks-long operational project for large tables.
>>
>> Cheers,
>>
>> Gianluca
>>
>> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
>> neeleshssalian@gmail.com> ha scritto:
>>
>>> +1 (non binding) Thank you Huaxin.
>>>
>>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com> wrote:
>>>
>>>> Hi all,
>>>>
>>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for
>>>> the
>>>> V4 table spec.
>>>>
>>>> Proposal
>>>> --------
>>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>>> metadata
>>>> will not define equality deletes as an allowed entry type.
>>>>
>>>> 2. Reading equality deletes remains supported in the reference
>>>> implementation for backward compatibility, both for existing V2/V3
>>>> tables and for equality deletes carried over into upgraded V4 tables.
>>>>
>>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous
>>>> rewrite
>>>> of data or delete files). Existing equality deletes remain in
>>>> carried-
>>>> over V2/V3 delete manifests; converting them to deletion vectors is a
>>>> separate, optional maintenance action.
>>>>
>>>> Rationale
>>>> ---------
>>>> Equality deletes impose an asymmetric cost paid on every read,
>>>> complicate
>>>> the format, and block features such as CDC, row lineage, and incremental
>>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>>> flat,
>>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>>> viable replacement path, so we do not need the full replacement
>>>> completed
>>>> before forbidding new equality deletes in V4.
>>>>
>>>> The vote will be open for at least 72 hours.
>>>>
>>>> [ ] +1 Forbid writing equality deletes in V4
>>>> [ ] +0
>>>> [ ] -1 Do not forbid (please explain)
>>>>
>>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>>
>>>> Thanks,
>>>> Huaxin
>>>>
>>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Hongyue Zhang <ho...@gmail.com>.
+1 (non-binding)
Thanks Huaxin!
On Thu, Aug 13, 2026 at 11:03 Gianluca Graziadei <
graziadei.gianluca@gmail.com> wrote:
> +1 (non-binding )
>
> The asymmetric cost argument is the decisive factor for me.
>
> An equality delete allows the writer to offload work onto every subsequent
> read, and this deferred cost is neither bounded nor visible to the party
> that generated it. Deletion vectors make that cost explicit and ensure it
> is paid only once.
>
> I would also like to highlight point 3, which I think is underrated:
> making the upgrade metadata-only decouples V4 adoption from delete file
> migration. Users can move to V4 on their own timeline and convert equality
> deletes as a background maintenance task. Coupling the two would have
> turned V4 adoption into a weeks-long operational project for large tables.
>
> Cheers,
>
> Gianluca
>
> Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
> neeleshssalian@gmail.com> ha scritto:
>
>> +1 (non binding) Thank you Huaxin.
>>
>> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com> wrote:
>>
>>> Hi all,
>>>
>>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>>> Iceberg V4" [1], I'd like to call a vote on the following proposal for
>>> the
>>> V4 table spec.
>>>
>>> Proposal
>>> --------
>>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>>> metadata
>>> will not define equality deletes as an allowed entry type.
>>>
>>> 2. Reading equality deletes remains supported in the reference
>>> implementation for backward compatibility, both for existing V2/V3
>>> tables and for equality deletes carried over into upgraded V4 tables.
>>>
>>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>>> of data or delete files). Existing equality deletes remain in carried-
>>> over V2/V3 delete manifests; converting them to deletion vectors is a
>>> separate, optional maintenance action.
>>>
>>> Rationale
>>> ---------
>>> Equality deletes impose an asymmetric cost paid on every read, complicate
>>> the format, and block features such as CDC, row lineage, and incremental
>>> index/materialized-view maintenance. Deletion vectors make deletion a
>>> flat,
>>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>>> viable replacement path, so we do not need the full replacement completed
>>> before forbidding new equality deletes in V4.
>>>
>>> The vote will be open for at least 72 hours.
>>>
>>> [ ] +1 Forbid writing equality deletes in V4
>>> [ ] +0
>>> [ ] -1 Do not forbid (please explain)
>>>
>>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>>
>>> Thanks,
>>> Huaxin
>>>
>>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Gianluca Graziadei <gr...@gmail.com>.
+1 (non-binding )
The asymmetric cost argument is the decisive factor for me.
An equality delete allows the writer to offload work onto every subsequent
read, and this deferred cost is neither bounded nor visible to the party
that generated it. Deletion vectors make that cost explicit and ensure it
is paid only once.
I would also like to highlight point 3, which I think is underrated: making
the upgrade metadata-only decouples V4 adoption from delete file migration.
Users can move to V4 on their own timeline and convert equality deletes as
a background maintenance task. Coupling the two would have turned V4
adoption into a weeks-long operational project for large tables.
Cheers,
Gianluca
Il giorno gio 13 ago 2026 alle ore 19:57 Neelesh Salian <
neeleshssalian@gmail.com> ha scritto:
> +1 (non binding) Thank you Huaxin.
>
> On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com> wrote:
>
>> Hi all,
>>
>> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
>> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
>> V4 table spec.
>>
>> Proposal
>> --------
>> 1. Writing new equality deletes is forbidden for V4 tables: the V4
>> metadata
>> will not define equality deletes as an allowed entry type.
>>
>> 2. Reading equality deletes remains supported in the reference
>> implementation for backward compatibility, both for existing V2/V3
>> tables and for equality deletes carried over into upgraded V4 tables.
>>
>> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
>> of data or delete files). Existing equality deletes remain in carried-
>> over V2/V3 delete manifests; converting them to deletion vectors is a
>> separate, optional maintenance action.
>>
>> Rationale
>> ---------
>> Equality deletes impose an asymmetric cost paid on every read, complicate
>> the format, and block features such as CDC, row lineage, and incremental
>> index/materialized-view maintenance. Deletion vectors make deletion a
>> flat,
>> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
>> viable replacement path, so we do not need the full replacement completed
>> before forbidding new equality deletes in V4.
>>
>> The vote will be open for at least 72 hours.
>>
>> [ ] +1 Forbid writing equality deletes in V4
>> [ ] +0
>> [ ] -1 Do not forbid (please explain)
>>
>> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>>
>> Thanks,
>> Huaxin
>>
>
Re: [VOTE] Deprecate equality deletes in Iceberg V4 (forbid new writes)
Posted by Neelesh Salian <ne...@gmail.com>.
+1 (non binding) Thank you Huaxin.
On Wed, Aug 12, 2026 at 19:07 huaxin gao <hu...@gmail.com> wrote:
> Hi all,
>
> Following the discussion thread "[DISCUSS] Deprecate Equality Deletes in
> Iceberg V4" [1], I'd like to call a vote on the following proposal for the
> V4 table spec.
>
> Proposal
> --------
> 1. Writing new equality deletes is forbidden for V4 tables: the V4 metadata
> will not define equality deletes as an allowed entry type.
>
> 2. Reading equality deletes remains supported in the reference
> implementation for backward compatibility, both for existing V2/V3
> tables and for equality deletes carried over into upgraded V4 tables.
>
> 3. Upgrading a V2/V3 table to V4 is metadata-only (no synchronous rewrite
> of data or delete files). Existing equality deletes remain in carried-
> over V2/V3 delete manifests; converting them to deletion vectors is a
> separate, optional maintenance action.
>
> Rationale
> ---------
> Equality deletes impose an asymmetric cost paid on every read, complicate
> the format, and block features such as CDC, row lineage, and incremental
> index/materialized-view maintenance. Deletion vectors make deletion a flat,
> one-time cost, and the Flink ConvertEqualityDeletes work demonstrates a
> viable replacement path, so we do not need the full replacement completed
> before forbidding new equality deletes in V4.
>
> The vote will be open for at least 72 hours.
>
> [ ] +1 Forbid writing equality deletes in V4
> [ ] +0
> [ ] -1 Do not forbid (please explain)
>
> [1] https://lists.apache.org/thread/ks01jpv40qjlvz4yop5tlqv4x5oxbwy6
>
> Thanks,
> Huaxin
>