You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@fluss.apache.org by Hongshun Wang <lo...@gmail.com> on 2025/12/23 02:05:25 UTC
[Discuss] FIP-23: Fluss Support Add Column At Last
Hi devs,
I'd like to start a discussion about Fluss Support Add Column At Last[1].
In production, Add Column At Last is required. Combined with the lenient
mode of Flink CDC YAML, if data is pre-created in the widest type, it can
generally meet most production requirements.
The lenient mode of Flink CDC behaves as follows:
1. Adding nullable columns: Automatically appends the corresponding column
to the end of the result table schema and synchronizes the new column's
data.
2. Dropping nullable columns: Does not directly remove the column from the
result table but automatically fills the column's data with NULL values.
3. Adding non-nullable columns: Automatically appends the corresponding
column to the end of the result table schema and synchronizes the new
column's data. The newly added column is set to nullable by default, and
data prior to the column addition is automatically set to NULL.
4. Renaming columns: Treated as a combination of adding and dropping
columns. Directly appends the renamed column to the end of the result table
and automatically fills the original column's data with NULL values. For
example, if col_a is renamed to col_b, col_b is appended to the end of the
result table, and the original col_a data is automatically set to NULL.
Any feedback and suggestions are welcome!
[1]
https://cwiki.apache.org/confluence/display/FLUSS/FIP-23+Fluss+Support+Add+Column+At+Last
Re: [Discuss] FIP-23: Fluss Support Add Column At Last
Posted by Jark Wu <im...@gmail.com>.
Thank you, Hongshun, for this long-requested feature! It’s a critical
step toward making Fluss mature and production-ready.
The FIP looks solid to me. I’m +1 to start the vote.
Best regards,
Jark
On Tue, 23 Dec 2025 at 10:06, Hongshun Wang <lo...@gmail.com> wrote:
>
> Hi devs,
> I'd like to start a discussion about Fluss Support Add Column At Last[1].
>
> In production, Add Column At Last is required. Combined with the lenient
> mode of Flink CDC YAML, if data is pre-created in the widest type, it can
> generally meet most production requirements.
>
> The lenient mode of Flink CDC behaves as follows:
> 1. Adding nullable columns: Automatically appends the corresponding column
> to the end of the result table schema and synchronizes the new column's
> data.
> 2. Dropping nullable columns: Does not directly remove the column from the
> result table but automatically fills the column's data with NULL values.
> 3. Adding non-nullable columns: Automatically appends the corresponding
> column to the end of the result table schema and synchronizes the new
> column's data. The newly added column is set to nullable by default, and
> data prior to the column addition is automatically set to NULL.
> 4. Renaming columns: Treated as a combination of adding and dropping
> columns. Directly appends the renamed column to the end of the result table
> and automatically fills the original column's data with NULL values. For
> example, if col_a is renamed to col_b, col_b is appended to the end of the
> result table, and the original col_a data is automatically set to NULL.
>
> Any feedback and suggestions are welcome!
>
> [1]
> https://cwiki.apache.org/confluence/display/FLUSS/FIP-23+Fluss+Support+Add+Column+At+Last
Re: [Discuss] FIP-23: Fluss Support Add Column At Last
Posted by Mehul Batra <me...@gmail.com>.
Thank you, Hongshun, for driving this!
Overall solid FIP +1
Best regards,
Mehul Batra
On Thu, Dec 25, 2025 at 9:51 PM Keith Lee <le...@gmail.com>
wrote:
> Speaking as a new contributor, thank you for the easy to read/understand
> FIP!
>
> Best regards
> Keith Lee
>
>
> On Thu, 25 Dec 2025 at 09:53, Leonard Xu <xb...@gmail.com> wrote:
>
> > Thanks for driving this FIP, Hongshun.
> >
> > It will not only help Fluss has the ability to support schema evolution
> > feature within Flink CDC, but also enrich the DML ability of Fluss
> itself.
> >
> > Best,
> > Leonard
> >
> >
> >
> > > 2025 12月 23 10:05，Hongshun Wang <lo...@gmail.com> 写道：
> > >
> > > Hi devs,
> > > I'd like to start a discussion about Fluss Support Add Column At
> Last[1].
> > >
> > > In production, Add Column At Last is required. Combined with the
> lenient
> > > mode of Flink CDC YAML, if data is pre-created in the widest type, it
> can
> > > generally meet most production requirements.
> > >
> > > The lenient mode of Flink CDC behaves as follows:
> > > 1. Adding nullable columns: Automatically appends the corresponding
> > column
> > > to the end of the result table schema and synchronizes the new column's
> > > data.
> > > 2. Dropping nullable columns: Does not directly remove the column from
> > the
> > > result table but automatically fills the column's data with NULL
> values.
> > > 3. Adding non-nullable columns: Automatically appends the corresponding
> > > column to the end of the result table schema and synchronizes the new
> > > column's data. The newly added column is set to nullable by default,
> and
> > > data prior to the column addition is automatically set to NULL.
> > > 4. Renaming columns: Treated as a combination of adding and dropping
> > > columns. Directly appends the renamed column to the end of the result
> > table
> > > and automatically fills the original column's data with NULL values.
> For
> > > example, if col_a is renamed to col_b, col_b is appended to the end of
> > the
> > > result table, and the original col_a data is automatically set to NULL.
> > >
> > > Any feedback and suggestions are welcome!
> > >
> > > [1]
> > >
> >
> https://cwiki.apache.org/confluence/display/FLUSS/FIP-23+Fluss+Support+Add+Column+At+Last
> >
> >
>
Re: [Discuss] FIP-23: Fluss Support Add Column At Last
Posted by Keith Lee <le...@gmail.com>.
Speaking as a new contributor, thank you for the easy to read/understand
FIP!
Best regards
Keith Lee
On Thu, 25 Dec 2025 at 09:53, Leonard Xu <xb...@gmail.com> wrote:
> Thanks for driving this FIP, Hongshun.
>
> It will not only help Fluss has the ability to support schema evolution
> feature within Flink CDC, but also enrich the DML ability of Fluss itself.
>
> Best,
> Leonard
>
>
>
> > 2025 12月 23 10:05，Hongshun Wang <lo...@gmail.com> 写道：
> >
> > Hi devs,
> > I'd like to start a discussion about Fluss Support Add Column At Last[1].
> >
> > In production, Add Column At Last is required. Combined with the lenient
> > mode of Flink CDC YAML, if data is pre-created in the widest type, it can
> > generally meet most production requirements.
> >
> > The lenient mode of Flink CDC behaves as follows:
> > 1. Adding nullable columns: Automatically appends the corresponding
> column
> > to the end of the result table schema and synchronizes the new column's
> > data.
> > 2. Dropping nullable columns: Does not directly remove the column from
> the
> > result table but automatically fills the column's data with NULL values.
> > 3. Adding non-nullable columns: Automatically appends the corresponding
> > column to the end of the result table schema and synchronizes the new
> > column's data. The newly added column is set to nullable by default, and
> > data prior to the column addition is automatically set to NULL.
> > 4. Renaming columns: Treated as a combination of adding and dropping
> > columns. Directly appends the renamed column to the end of the result
> table
> > and automatically fills the original column's data with NULL values. For
> > example, if col_a is renamed to col_b, col_b is appended to the end of
> the
> > result table, and the original col_a data is automatically set to NULL.
> >
> > Any feedback and suggestions are welcome!
> >
> > [1]
> >
> https://cwiki.apache.org/confluence/display/FLUSS/FIP-23+Fluss+Support+Add+Column+At+Last
>
>
Re: [Discuss] FIP-23: Fluss Support Add Column At Last
Posted by Leonard Xu <xb...@gmail.com>.
Thanks for driving this FIP, Hongshun.
It will not only help Fluss has the ability to support schema evolution feature within Flink CDC, but also enrich the DML ability of Fluss itself.
Best,
Leonard
> 2025 12月 23 10:05，Hongshun Wang <lo...@gmail.com> 写道：
>
> Hi devs,
> I'd like to start a discussion about Fluss Support Add Column At Last[1].
>
> In production, Add Column At Last is required. Combined with the lenient
> mode of Flink CDC YAML, if data is pre-created in the widest type, it can
> generally meet most production requirements.
>
> The lenient mode of Flink CDC behaves as follows:
> 1. Adding nullable columns: Automatically appends the corresponding column
> to the end of the result table schema and synchronizes the new column's
> data.
> 2. Dropping nullable columns: Does not directly remove the column from the
> result table but automatically fills the column's data with NULL values.
> 3. Adding non-nullable columns: Automatically appends the corresponding
> column to the end of the result table schema and synchronizes the new
> column's data. The newly added column is set to nullable by default, and
> data prior to the column addition is automatically set to NULL.
> 4. Renaming columns: Treated as a combination of adding and dropping
> columns. Directly appends the renamed column to the end of the result table
> and automatically fills the original column's data with NULL values. For
> example, if col_a is renamed to col_b, col_b is appended to the end of the
> result table, and the original col_a data is automatically set to NULL.
>
> Any feedback and suggestions are welcome!
>
> [1]
> https://cwiki.apache.org/confluence/display/FLUSS/FIP-23+Fluss+Support+Add+Column+At+Last