How to handle large data volumes caused by data expansion?

[QcoderWho]

I am an SOA toolchain development engineer currently using syson for model design. I've encountered an issue where converting a Java bean to a syson model results in a significant expansion of the number of model nodes. For instance, I am designing a model to store a large amount of frame and signal data. There are reference relationships between frame and signals, and each frame and signal stores several attributes. The model is designed as follows: ItemUsage -> AttributeUsage -> LiteralString. After testing the addition of 10,000 frames and signals, the Resource size in the database reached several hundred megabytes, causing both the saving and loading of IEditingContext to become very slow. I want to split the Resource into multiple Resource files based on business logic. However, after splitting, the parent-child relationships between model elements are lost because the eContainer property of the model elements becomes null after they are moved to new Resource files. Are there any effective solutions to this problem?

[AxelRICHARD]

Hello,

Thank you for using SysON.

Based on your description, I assumed that you are manipulating EMF models.
So it is an EMF matter.
If you use the appropriate EMF APIs inside a ResourceSet containing all your split resources, you should then be able to:

• create your elements directly inside the split resources
• or create your elements inside the first resource and then move them into the split resources.

But I'm not sure it will change anything to split the resources, because they still will be part of the same project and the same editing context when loaded.

Furthermore, if you need professional support, you can also visit this page: https://doc.mbse-syson.org/syson/v2025.12.0/user-manual/support.html

Regards,

[QcoderWho]

Hello,

Thank you for using SysON.

Based on your description, I assumed that you are manipulating EMF models. So it is an EMF matter. If you use the appropriate EMF APIs inside a ResourceSet containing all your split resources, you should then be able to:

• create your elements directly inside the split resources
• or create your elements inside the first resource and then move them into the split resources.

But I'm not sure it will change anything to split the resources, because they still will be part of the same project and the same editing context when loaded.

Thank you for your response. Yes, this is an EMF-related issue. I have tried moving elements into the separated resources, but this approach disrupts the original tree structure. I will continue to explore using custom proxy objects and rewriting part of the source code to achieve seamless migration. The cause of the data volume explosion is the large number of Relationship nodes and the overly complex modeling of record-type data. I will also attempt to address this issue by extending the Syson meta-model, designing a dedicated meta-model for record-type data.