you can have multiple pipeline run configurations in a single project, and
only use your transactional run configuration for the pipelines you want to
run transactional.

Regards,
Bart


On Tue, Aug 27, 2024 at 11:53 AM Davide Cisco <[email protected]> wrote:

> Hi Bart,
>
> if I get correctly, the safest way to do that would be:
> - create a new Workflow Run Configuration (WRC) with the "Make this
> pipeline transactional" option enabled
> - set this new WRC to the "critical" pipeline that loads/copies data
>
> If I modified the setting directly in the default "local" WRC, I would
> have transactions enabled for all the pipelines in the Hop workspace (even
> when those are not necessary), correct?
>
> Thanks for your support
>
> DC
>
>
> Il giorno mar 27 ago 2024 alle ore 11:29 Bart Maertens <
> [email protected]> ha scritto:
>
>> Hi Davide,
>>
>> The local native pipeline run configuration [1] has a "Make this pipeline
>> transactional" option that allows you to do what you're looking for.
>>
>> [1]
>> https://hop.apache.org/manual/latest/pipeline/pipeline-run-configurations/native-local-pipeline-engine.html
>>
>> Regards,
>> Bart
>>
>> On Tue, Aug 27, 2024 at 11:24 AM Davide Cisco <[email protected]>
>> wrote:
>>
>>> Hello,
>>>
>>> I would like to write a pipeline/workflow that allows me to control a
>>> transaction when writing in a database, but I can't find the proper
>>> transforms/actions.
>>>
>>> More specifically, what I would like to do is the following:
>>> - connect to a database and begin a transaction
>>> - truncate and store data in various tables (coming from various
>>> sources: files, web services, another DB)
>>> - end the transaction and commit if everything went fine, otherwise
>>> rollback (or abort without commit)
>>>
>>> By using just the Table output transform the tables are correctly
>>> updated if everything went fine, but in case of errors the default options
>>> leave the database in an inconsistent state: tables could have been partly
>>> populated, or some or those may have been updated and some others not.
>>>
>>> Is there any way to control transactions in Hop (excluding the - partial
>>> - workaround of copying data in a temporary/secondary table and then move
>>> them in the corresponding primary table)?
>>>
>>> Thanks for your help
>>>
>>> DC
>>>
>>

Reply via email to