Hi Knut,

It wouldn't be a real sequence (incrementing integer), but one easy way to
generate an ID for your article data is through the Add a checksum
transform [1].

[1]
https://hop.apache.org//manual/latest/pipeline/transforms/addchecksum.html

Regards,
Bart


On Mon, May 27, 2024 at 7:33 PM Knut Meyer <[email protected]> wrote:

> Hi,
>
> to narrow down my question: How can I define a workflow-wide
> counter/variable that I can safely use from parallel pipelines and thus
> threads?
>
> Is the database sequence the only option, or can I implement a Java-based
> solution?
>
> Thanks a lot!
>
> Best regards,
> Knut
>
>
> On 25/05/2024 19:39, Knut Meyer wrote:
>
> Hi together,
>
> I am migrating a project from Talend Open Studio to Apache Hop and need
> some technical advice.
>
> I am processing article data in a pipeline that runs multiple times in
> parallel using ‘Specify copies’. I only want to assign a new ID to article
> data that does not yet have an ID. For this reason, I cannot use the
> existing sequence transformations.
>
> My questions
> - How can I define a comprehensive sequence that I can use in parallel
> pipelines? Is this possible with a ‘User defined Java Class’ or should I
> map the sequence in a database?
> - Which transformations would you use to map the logic ‘do not assign an
> ID yet, then get the next sequence and set it in the field of the line’?
>
> Many thanks for the work on Apache Hop and your support!
>
> Best regards,
> Knut Meyer
>
>
>

Reply via email to