Hi Hans,

Thanks for your response.

Just to clarify—my question is not about the merging itself, but rather
what happens *after* the merge. The issue arises from how duplicated field
names are handled in the downstream transforms.

Specifically, the merged output shows duplicated fields with a _1 suffix
(as expected) when you use *"Show output fields"* or *"Get fields"* in the
downstream transform. However, when the pipeline runs, it throws an error
saying the field with the _1 suffix cannot be found.

This suggests a mismatch between the metadata inferred during pipeline
design and what is actually available during execution.

Thanks,
Maria

On Thu, Jun 26, 2025 at 5:23 PM Hans Van Akelyen <[email protected]>
wrote:

> Hi Maria,
>
> It’s actually something that should not have worked in PDI either. It’s by
> accident that it does. You get giant warnings when merging 2 streams that
> do not line up. So this will be an opportunity to clean up those errors.
>
> But you have me a bit confused about the duplicate field name, the error
> in the ticket might be valid that there is something wrong in a newly
> created field resulting in _1. But that is not related to merging 2 streams.
>
>
> Kr,
> Hans
>
> On Thu, 26 Jun 2025 at 09:01, Maria Lim <[email protected]> wrote:
>
>> Thanks Hans.
>>
>> Do you mean, a redesign of the pipeline is inevitable for this case? Do
>> you know if this is a known issue (duplicate field name not correctly
>> passed) for migrated PDI job only, or for all HOP pipelines?
>>
>> Kind regards,
>> Maria
>>
>> On Thu, Jun 26, 2025 at 4:27 PM Hans Van Akelyen <
>> [email protected]> wrote:
>>
>>> Hi Maria,
>>>
>>> By default when merging streams the layout should be the same on both
>>> sides. Alternatively when merging 2 streams with different lay-outs you can
>>> use the stream schema merge transform
>>>
>>> https://hop.apache.org//manual/latest/pipeline/transforms/streamschemamerge.html
>>>
>>> This transform will make sure the result is valid.
>>>
>>> Kr,
>>> Hans
>>>
>>>
>>> On Thu, 26 Jun 2025 at 07:39, Maria Lim <[email protected]> wrote:
>>>
>>>> Hi Apache HOP Support Team,
>>>>
>>>> I’m currently exploring the migration of ETLs from PDI to Apache HOP.
>>>>
>>>> While testing a pipeline, I encountered a failure at the "Select
>>>> values" transform following the merging of two streams. The error
>>>> indicates that a field with the suffix "_1" could not be found—this
>>>> corresponds to a duplicated fieldname resulting from the merge.
>>>>
>>>> The issue appears to align with the one reported on GitHub last year:
>>>> https://github.com/apache/hop/issues/3849
>>>>
>>>> After reviewing that discussion, I attempted to rebuild the pipeline
>>>> entirely in HOP, hoping for a different result. Unfortunately, the
>>>> issue persists.
>>>>
>>>> Aside from known workarounds like renaming fields prior to the merge
>>>> to avoid duplication, is there a more direct or quicker fix available
>>>> for this behaviour?
>>>>
>>>> Appreciate your insights and support.
>>>>
>>>> Kind regards,
>>>> Maria
>>>>
>>>

Reply via email to