+user <user@beam.apache.org>

---------- Forwarded message ---------
From: Chamikara Jayalath <chamik...@google.com>
Date: Wed, Oct 12, 2022 at 9:38 AM
Subject: Re: Cross Language
To: phani geeth <bvpgeeth0...@gmail.com>




On Wed, Oct 12, 2022 at 12:25 AM phani geeth <bvpgeeth0...@gmail.com> wrote:

> Hi Chamikara,
>
> Thanks for replying.
>
> My use case is to write a beam pipeline in Java for kafkaIO and call a
> python transform as expansion service that should be run on dataflow.
>
>
Yeah, using Python transform from a Java pipeline is supported for
Dataflow. Please see the following for a quickstart guide on this.

https://beam.apache.org/documentation/sdks/java-multi-language-pipelines/

Thank,
Cham


> Will the above scenario work if it does how can implement python transform
> in Java.
>
> Regards,
> Phani Geeth
>
> On Tue, 11 Oct, 2022, 7:19 pm Chamikara Jayalath via user, <
> user@beam.apache.org> wrote:
>
>> Is there a specific I/O connector you are hoping to use ?
>>
>> Thanks,
>> Cham
>>
>> On Tue, Oct 11, 2022 at 4:31 AM Alexey Romanenko <
>> aromanenko....@gmail.com> wrote:
>>
>>> Yes, it’s possible though Java IO connector should support being used
>>> via X-language.
>>>
>>> For more details regarding which connector supports this, you may want
>>> to take a look on IO connectors table on this page [1] and check if the
>>> required connector is supported "via X-language" for Python SDK.
>>>
>>> [1] https://beam.apache.org/documentation/io/connectors/
>>>
>>> —
>>> Alexey
>>>
>>> > On 11 Oct 2022, at 11:06, phani geeth <bvpgeeth0...@gmail.com> wrote:
>>> >
>>> > Hi,
>>> >
>>> > Does the present cross Language functionality support creating custom
>>> Java transforms and calling from python in Dataflow runner.
>>> >
>>> > Use case: use existing Java IO as cross Language transform and call in
>>> python pipeline in Dataflow runner.
>>> >
>>> >
>>> > Regards
>>> > Phani Geeth
>>>
>>>

Reply via email to