Very unstable version v0.20

[Pelayo-Reguera]

Honestly, I don't know what this error means, but it doens't allow me to finish the pipeline (see details). Apart from that, it's quite unstable the application, because now de GUI gets freeze like 5 or 10 times a day. I will be posting the errors if the console says something, but for the moment, it would be nice to not freeze the screen while trying to load execution information of previous pipelines or workflows executions. For example, if I change to the execution perspective and needs to take a time until the download completes all the executions, it would be nice to have a progress bar and during the download be able to keep working.

Details

S NULL.0�[0m
�[31m2026/10/08 08:40:35 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/inquilino IS NULL.0�[0m
�[31m2026/10/08 08:40:35 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/grupo IS NULL 2.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/inquilino IS NOT NULL AND nombre NOT [BAJA] 2.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/inquilino IS NOT NULL AND nombre NOT [BAJA].0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/Filter rows.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/ID inquilino IS NULL 2.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/grupo IS NULL.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/inquilino IS NULL 2.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/Referencia NOT NULL.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/Dummy (do nothing) 5.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/Dummy (do nothing) 3.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/ID inquilino IS NULL.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/inquilino IS NULL.0�[0m
�[31m2026/10/08 08:40:46 - General - ERROR: Skipping execution data set without row metadata: FirstOutput/grupo IS NULL 2.0�[0m

[bamaer]

By "very unstable v0.20", I assume you mean you found a very specific issue in the Neo4j execution information location in Apache Hop 2.20.0-SNAPSHOT.
If that's the case, can you share a screenshot of your pipeline and let us know if it (at least partially) runs with 0 rows?

[hansva]

can you provide a reproducible sample and tell us how your execution information location is configured, a lot will depend on that.

[bamaer]

It took a closer look, this was probably caused by PR #8625. Shouldn't be too hard to fix. Create a ticket and we'll look into it.

[Pelayo-Reguera]

can you provide a reproducible sample and tell us how your execution information location is configured, a lot will depend on that.

[mattcasters]

Issue #8625 is unlikely to hit Neo4j as that one doesn't use the caching base code (essentially a JSON serializer).
My money is on #8865 which fixed an inefficiency in storing/finding parent/child relationships between executions. Some LLM (Gemini 3.8 Flash) was supposed to fix the storing of the parenting but grew a brain and I overlooked it. My advice would be to try a recent version and see if this goes away.
Thanks for telling us about it!

[Pelayo-Reguera]

Yes! I compiled a recent version of it and now it works. At least the pipelines finish and don't show repeatedly the messages I mentioned before.

However, I still have the problem of memory overflow when doing the commit of the execution information against the Neo4J database, although that's discussed in other conversation #8204 suppposedly addressed in #8282. I have configured the maximum length of strings and JSON but it seems that nothing changes.

[mattcasters]

It all depends on what you're doing of-course. Try not saving individual rows into Neo4j. If that solves the out-of-memory issues, it's really just that you don't have enough memory. You can give Hop more memory. The default of 2GB might not be enough for the use-case.

[Pelayo-Reguera]

And how do I not store individual rows?

[mattcasters]

Don't use a data profile in the pipeline run configuration. Or you can store fewer rows, only the Last 10 for example.